THE ACCURACY AND PRECISION OF

KINESIOLOGY-STYLE MANUAL MUSCLE TESTING:

DESIGNING AND IMPLEMENTING A SERIES OF

DIAGNOSTIC TEST ACCURACY STUDIES

Anne Marie Jensen

Wolfson College

University of Oxford

A thesis submitted for the degree of

Doctor of Philosophy

Initial Submission 19 June 2014

Revision Submitted 24 October 2014

 ii

ABSTRACT

Introduction: Kinesiology-style manual muscle testing (kMMT) is a non-invasive

assessment method used by various types of practitioners to detect a wide range of target

conditions. It is distinctly different from the muscle testing performed in

orthopaedic/neurological settings and from Applied Kinesiology. Despite being estimated

to be used by over 1 million people worldwide, the usefulness of kMMT has not yet been

established. The aim of this thesis was to assess the validity of kMMT by examining its

accuracy and precision.

Methods: A series of 5 diagnostic test accuracy studies were undertaken. In the first

study, the index test was kMMT, and the target condition was deceit in verbal statements

spoken by Test Patients (TPs). The comparator reference standard was a true gold

standard: the actual verity of the spoken statement. The outcomes of the muscle tests

were interpreted consistently: a weak result indicated a Lie and a strong result indicated a

Truth. A secondary index test was included as a comparator: Intuition, where

Practitioners used intuition (without using kMMT) to ascertain if a Lie or Truth was

spoken. Forty-eight Practitioners were recruited and paired with 48 unique kMMT-naïve

TPs. Each Pair performed 60 kMMTs broken up into 6 blocks of 10, which alternated

with blocks of 10 Intuitions. For each Pair, an overall percent correct was calculated for

both kMMT and Intuition, and their means were compared. Also calculated for both tests

were sensitivity, specificity, positive predictive value and negative predictive value.

The second study was a replication of the first, using a sample size of 20 Pairs and a less

complex procedure. In the third study, grip strength dynamometry replaced kMMT as the

primary index test. In the fourth study, the reproducibility and repeatability of kMMT
 iii

were examined. In the final study, TPs were presented with emotionally-arousing stimuli

in addition to the affect-neutral stimuli used in previous studies, to assess if stimuli

valence impacted kMMT accuracy.

Results: Throughout this series of studies, mean kMMT accuracies (95% Confidence

Intervals; CIs) ranged from 0.594 (0.541 – 0.647) to 0.659 (0.623 - 0.695) and mean

Intuition accuracies, from 0.481 (0.456 - 0.506) to 0.526 (0.488 - 0.564). In all studies,

mean kMMT accuracies were found to be significantly different from mean Intuition

accuracies (p ≤ 0.01), and from Chance (p < 0.01). On the other hand, no difference was

found between grip strength following False statements compared to grip strength

following True statements (p = 0.61). In addition, the Practitioner-TP complex accounted

for 57% of the variation in kMMT accuracy, with 43% unaccounted for. Also, there was

no difference in the mean kMMT accuracy when using emotionally-arousing stimuli

compared to when using affect-neutral stimuli (p = 0.35). Mean sensitivities (95% CI)

ranged from 0.503 (0.421 - 0.584) to 0.659 (0.612 - 0.706) while mean specificities (95%

CI) ranged from 0.638 (0.430 - 0.486) to 0.685 (0.616 - 0.754). Finally, while a number

of participant characteristic seemed to influence kMMT accuracy during one study or

another, no one specific characteristic was found to influence kMMT accuracy

consistently (i.e. across the series of studies).

Discussion: This series of studies has shown that kMMT can be investigated using

rigorous evidence-based health care methods. Furthermore, for distinguishing lies from

truths, kMMT has repeatedly been found to be significantly more accurate than both

Intuition and Chance. Practitioners appear to be an integral part of the kMMT dynamic

because when replaced by a mechanical device (i.e. a grip strength dynamometer),

distinguishing Lies from Truth was not possible. In addition, since specificities seemed to
 iv

be greater than sensitivities, Truths may have been easier to detect than Lies. A limitation

of this series of studies is that I have a potential conflict of interest, in that I am a

practitioner of kMMT who gets paid to perform kMMT. Another limitation is these

results are not generalisable to other applications of kMMT, such as its use in other

paradigms or using muscles other than the deltoid. Also, these results suggest that kMMT

may be about 60% accurate, which is statistically different from Intuition and Chance;

however it has not been established if 60% correct is “good enough” in a clinical

context. As such, further research is needed to assess its clinical utility, such as

randomised controlled trials investigating the effectiveness of whole kMMT technique

systems. Also, future investigators may want to explore what factors, such as specific

Practitioner and TP characteristics, influence kMMT accuracy, and to investigate the

validity of using kMMT to detect other target conditions, using other reference standards

and muscles other than the deltoid.

Summary: This series of diagnostic test accuracy studies has found that kMMT can be

investigated using rigorous methods, and that kMMT used to distinguish Lies from

Truths is significantly more accurate that both Intuition and Chance. Further research is

needed to assess kMMT’s clinical utility.

 v

DEDICATION

This thesis is dedicated to

my father....

Thank you for teaching me perseverance.

“I am thankful to all those who said, ‘No.’

Because of them I did it myself.”

- Albert Einstein
 vi

GRATITUDE
(ACKNOWLEDGEMENTS)

“At times our own light goes out and is rekindled by a spark from
another person. Each of us has cause to think with deep gratitude of
those who have lighted the flame within us.”

— Albert Schweitzer

There were innumerable times over the last 7 years that my flame sputtered and waned,

and were it not for its rekindling at just the right times and places, I would have never

reached this point. There are many of you out there who acted in this capacity – by giving

advice or instruction, by giving moral or financial support, by giving a kind word or

warm hearth, or by kicking me out the nest so I could experience flying myself. To each

and every one of these rekindlers, I am humbly grateful...

 Amanda Burls, DPhil Supervisor – thank you for taking me on when no one else dared

 Richard Stevens, DPhil Supervisor – thank you for your interminable patience

 To anyone who acted as an Advisor / Consultant:

 Bruce Arroll  Mike Clark  Beth Shinkins

 Rafael Perera  Tom Fanshaw

 Tim Kenealy
 Susan Mallett  Adai Ramasamy
 Joanna Stewart
 Sharon Mickan  Jay Triano
 Paul Glasziou
 Jeremy Howick  Charlotte LeBoeuf-Yde
 Carl Heneghan
 Alison Ward  Adrian Stokes
 Dan Lasserson
 Jason Oke  Ben Feakins
 Clare Bankhead
Thank you for your kind guidance and great ideas

 Oxford’s Department of Primary Care Health Sciences and the Department of

Continuing Education – thanks to everyone in these department for their help and
support – especially Jane McCaffrey.
 vii

 Joseph LeDoux – thank you for not “believing” in muscle testing... and starting this
whole ball rolling.

 To other sources of academic inspiration or support:

 Rolf Peters  Bruce Lipton
 Patrick Bossuyt
 George Lewith

 Geoff Miller – thank you for voluntarily reading this from cover to cover.

 My practitioner-colleague-friends:
 Howard Cohen  Leslie Oldershaw  Katherine Moyer

 John Campise  Trish Anton

 Erika Barrantes
 Victoria Moore  Mary Lowther
 Khelly Webb
 Linda Christian  Linda Li
 Wil Bos
 Kim Makoi  Scott Cuthbert
 Kit Macy

Thank you for your help and for trusting me

 Thank you very much for the financial support:

 Virginia Hernly
 Wolfson College
 Santander
 Ellen Blasi

 Chris Guile – for his computer wizardry and his just being there... much of the time

 A HUGE THANK YOU to all the study participants – Practitioners and Test Patients
alike – from all over the world! Thank you SO much for generously sharing with me
your time, space and thoughts

 My friends and fellow cohort in the EBHC DPhil programme (especially Antoine) –
thank you for your inspiration and support, and all the best completing!

 Eleanor and Erland Jensen, my parents – thank you for your giving me in death what
you could not in life

 Ellen Jensen, my sister – thank you for sticking by me

 Tom Jensen, my brother – thanks for teaching me about “wiggle room” – it has come
in handy many times!

 Aunt Christiane Stolte – thank you for making sense

 viii

 Matthew, Connie and Diana Patane, my Australian family – thank you for always
being there

 My friends all over the world:

 Kerri Elston Doherty  Michelle Hogan  Mike Hall

 Suzy Dormer  Judith & Mike Hotek

 Maryellen Stephens
 Lorna Corgat  Francis Murphy
 Finn Jenk
 Lindsay Collins  Louis D'Amico
 Konstans Foskolos
 Curtis Rigney  Caroline Omo
 Greg Hayman
 Steph Essex  Marty Hall
 Roz Gibbs
 Darryl Cole  Martin Enderlin
 Naretha & Japie Nel
 Nick Kovacs (& Derrick)  Jackie Fairbourn
 Merlijn Wolsink
 Mark Postles  Catherine & Eric Four
 Wanyo Jennings

 Jeff Schelling & family  Marcella MacArthur

Thank you for your help, love and support, for encouragement and smiles, for BBQs and
cuppas, for normality and sanity, for insanity and giggles, for fresh air and fresh views.
Each one of you mean so much.

 (The late) Bill Harris – for your unwavering faith and enduring legacy of wisdom:
“Oh and one more thing... remember...”

 BJ Galvin, my star sista – thank you for reminding me time and again how spacious I
actually am

 Maryellen Stephens, Lisa Naera, Frank So & Marie McElhinney – thank you for your
confidence and many referrals

 My Dallas neighbours, Paulette, Brenda, Paul, Brad, Chris, and Karen – thank you for
margaritas and midnight swims, and being there when the roof(s) caved in

 Thanks also to all the body workers, mind workers, mindbody workers and other
healers who looked after me over the years – you all ROCK!

 Wolfson College – thank you for being my “home” for the better part of the last 9
years. Special thanks to my College advisor, Professor Andrew Neil; former Senior
Tutor, Dr. Martin Francis; President, Professor Dame Hermione Lee, Librarian Fiona
Wilkes, and also Rose Truby, John Kirby, Melvin Curtin, Di Wheeler, Cheery
Johnson, and all the maintenance staff, cleaning staff and ground staff.
 ix

 Parker University and Parker Research Institute – thank you for your support during
the initial phase of this journey

 Thank you to all organisations that support kMMT, especially the ICAK, NET, Inc;
TBM, Inc; PSYCH-K; and Health Kinesiology

 Empirisoft Corporation – thank you for your support and technical assistance with
your DirectRT software

 Oxford’s IT Services (formerly OUCS) – thanks for saving the day more than once

 University of Oxford – thank you for an amazing trip!

 Finally, thank you to Dr. George Goodheart, for founding Applied Kinesiology, to
Major Dejarnette, who paved the way, and to John Thie, who followed, for
contributing to the widespread use of kMMT
 TABLE OF CONTENTS

Abstract ..............................................................................................................................ii

Dedication .......................................................................................................................... v

Gratitude (Acknowledgements) ...................................................................................... vi

Table of Contents .............................................................................................................. x

Abbreviations.................................................................................................................. xvi

Glossary ........................................................................................................................... xix

CHAPTER 1 : Introduction ............................................................................................. 1

1.1 The Evolution of MMT ........................................................................................ 2
1.2 The kinesiology-style Manual Muscle Test ......................................................... 5
1.3 Applications of kMMT ......................................................................................... 8
1.4 Interpreting the outcome of the kMMT ................................................................ 9
1.5 The Current Status of the Evidence .................................................................... 10
1.6 kMMT as a “Diagnostic Test?” .......................................................................... 12
1.7 Diagnostic Accuracy of kMMT.......................................................................... 14
1.8 Diagnostic Precision of kMMT .......................................................................... 15
1.9 Choice of Research Topic .................................................................................. 18
1.10 Research Question and Paradigm to Investigate ................................................ 18
1.11 Choice of Populations......................................................................................... 21
1.12 Main Study Aims ................................................................................................ 22
1.13 Chapter 1 – List of Tables and Figures .............................................................. 23

CHAPTER 2 : Study 1 – Estimating the Accuracy of kMMT .................................... 46

2.1 Abstract............................................................................................................... 47
2.2 Introduction ........................................................................................................ 49
2.3 Methods .............................................................................................................. 51
2.3.1 Participants and Setting ............................................................................. 52
2.3.2 Practice Phase ............................................................................................ 53
2.3.3 Test Methods.............................................................................................. 55
 xi

2.3.4 Recording of Results .................................................................................. 66

2.3.5 Blinding ..................................................................................................... 67
2.3.6 Pre- and Post-testing Questionnaires ......................................................... 68
2.3.7 Pilot study .................................................................................................. 69
2.3.8 Statistical Methods ..................................................................................... 70
2.4 Results ................................................................................................................ 72
2.4.1 Participants................................................................................................. 72
2.4.3 Test Results ................................................................................................ 75
2.4.4 Post Hoc Analyses ..................................................................................... 89
2.4.5 Adverse Events .......................................................................................... 92
2.5 Discussion........................................................................................................... 92
2.5.1 Statement of Principal Findings ................................................................. 92
2.5.2 Strengths and Limitations .......................................................................... 97
2.5.3 Comparisons to Other Studies ................................................................. 100
2.5.4 Possible Explanations of Results ............................................................. 101
2.5.5 Implications for clinical practice ............................................................. 103
2.5.6 Unanswered questions and future research .............................................. 104
2.6 Summary........................................................................................................... 106
2.7 Chapter 2 – List of Tables and Figures ............................................................ 107

CHAPTER 3 : Study 2 – Replication of Study 1 ........................................................ 110

3.1 Abstract............................................................................................................. 111
3.2 Introduction ...................................................................................................... 113
3.3 Methods ............................................................................................................ 115
3.3.1 Participants and Setting ........................................................................... 116
3.3.2 Test Methods............................................................................................ 117
3.3.3 Statistical Methods ................................................................................... 119
3.4 Results .............................................................................................................. 120
3.4.1 Participants............................................................................................... 120
3.4.2 Test Results .............................................................................................. 122
3.5 Discussion......................................................................................................... 135
3.5.1 Statement of Principal Findings ............................................................... 135

© drannejensen 2014
 xii

3.5.2 Strengths and Limitations ........................................................................ 137

3.5.3 Possible Explanations of Results ............................................................. 138
3.5.4 Implications for clinical practice ............................................................. 139
3.5.5 Unanswered questions and future research .............................................. 139
3.6 Summary........................................................................................................... 141
3.7 Chapter 3 – List of Tables and Figures ............................................................ 143

CHAPTER 4 : Study 3 – Grip Strength Dynamometry for Lie Detection .............. 145
4.1 Abstract............................................................................................................. 146
4.2 Introduction ...................................................................................................... 148
4.3 Methods ............................................................................................................ 149
4.3.1 Participants and Setting ........................................................................... 149
4.3.2 Test Methods............................................................................................ 150
4.3.3 Statistical Methods ................................................................................... 153
4.4 Results .............................................................................................................. 154
4.4.1 Participants............................................................................................... 154
4.4.2 Test Results .............................................................................................. 155
4.5 Discussion......................................................................................................... 161
4.5.1 Statement of Principal Findings ............................................................... 161
4.5.2 Possible Explanations of Results ............................................................. 162
4.5.3 Strengths and Limitations ........................................................................ 163
4.5.4 Implications for Clinical Practice ............................................................ 164
4.5.5 Unanswered Questions and Future Research........................................... 164
4.6 Summary........................................................................................................... 164
4.7 Chapter 4 – List of Tables and Figures ............................................................ 165

CHAPTER 5 : Study 4 – Exploring the Variation in kMMT Accuracy .................. 166

5.1 Abstract............................................................................................................. 167
5.2 Introduction ...................................................................................................... 169
5.3 Methods ............................................................................................................ 170
5.3.1 Participants and Setting ........................................................................... 171
5.3.2 Test Methods............................................................................................ 173
5.3.3 Statistical Methods ................................................................................... 176

© drannejensen 2014
 xiii

5.4 Results .............................................................................................................. 178

5.4.1 Participants............................................................................................... 178
5.4.2 Test Results .............................................................................................. 181
5.5 Discussion......................................................................................................... 194
5.5.1 Statement of Principal Findings ............................................................... 194
5.5.2 Comparisons to Other Studies ................................................................. 194
5.5.3 Strengths and Limitations ........................................................................ 195
5.5.4 Implications for Clinical Practice ............................................................ 195
5.5.5 Unanswered Questions and Future Research........................................... 196
5.6 Summary........................................................................................................... 196
5.7 Chapter 5 – List of Tables and Figures ............................................................ 197

CHAPTER 6 : Study 5 – Using Emotionally-Arousing Stimuli ................................ 199

6.1 Abstract............................................................................................................. 200
6.2 Introduction ...................................................................................................... 202
6.3 Methods ............................................................................................................ 203
6.3.1 Participants and Setting ........................................................................... 204
6.3.2 The Stimuli .............................................................................................. 204
6.3.3 The Testing Scenario ............................................................................... 206
6.3.4 Statistical Methods ................................................................................... 207
6.4 Results .............................................................................................................. 208
6.4.1 Participants............................................................................................... 208
6.4.2 The Stimuli .............................................................................................. 208
6.4.3 Test Results .............................................................................................. 209
6.5 Discussion......................................................................................................... 222
6.5.1 Statement of Principal Findings ............................................................... 222
6.5.2 Strengths and Limitations ........................................................................ 224
6.5.3 Possible Explanations of Results ............................................................. 225
6.5.4 Implications for clinical practice ............................................................. 227
6.5.5 Unanswered questions and future research .............................................. 227
6.6 Summary........................................................................................................... 228
6.7 Chapter 6 – Study 5 – List of Tables and Figures ............................................ 230

© drannejensen 2014
 xiv

CHAPTER 7 : Discussion ............................................................................................. 232

7.1 Statement of principal findings ........................................................................ 233
7.2 Strengths and limitations .................................................................................. 234
7.3 Comparison of results to other studies ............................................................. 234
7.4 Implications for clinical practice ...................................................................... 237
7.5 Unanswered questions and future research ...................................................... 238
7.6 Summary........................................................................................................... 240
7.7 Chapter 7 – List of Tables and Figures ............................................................ 242

APPENDIX A : Participant Forms ............................................................................. 244

Participant Information Sheet (PIS) / Informed Consent : Studies 1,2,4,5 ................. 245
Participant Information Sheet (PIS) / Informed Consent : Study 3 ............................. 247
Practitioner Instruction Sheet : Study 1 ....................................................................... 249
Participant Instruction Sheet – Test Patient : Study 1 ................................................. 251
Participant Instruction Sheet - Practitioner : Studies 2,5 ............................................. 253
Participant Instruction Sheet – Test Patient : Studies 2,5 ............................................ 255
Participant Instruction Sheet : Study 3 – Grip Strength .............................................. 257
Participant Instruction Sheet – Practitioner : Study 4.................................................. 258
Participant Instruction Sheet – Test Patient : Study 4 ................................................. 260
Pre-Testing Questionnaire – Practitioner : Studies 1,2,5............................................. 262
Pre-Testing Questionnaire – Test Patient : Studies 1,2,5 ............................................ 263
Pre-Testing Questionnaire – Test Patient : Study 3 ..................................................... 264
Pre-Testing Questionnaire – Practitioner : Study 4 ..................................................... 265
Pre-Testing Questionnaire – Test Patient : Study 4 ..................................................... 266
Post-Testing Questionnaire – Practitioner : All Studies .............................................. 267
Post-Testing Questionnaire – Test Patient : All Studies.............................................. 268
Practitioner Score Sheets: Study 4 – Reproducibility & Repeatability ....................... 269
Test Patient Score Sheets: Study 4 – Reproducibility & Repeatability ...................... 271

APPENDIX B : Extra Figures & Tables ..................................................................... 273

List of Extra Figures & Tables .................................................................................... 274
CHAPTER 1 ................................................................................................................ 277
CHAPTER 2 ................................................................................................................ 278

© drannejensen 2014
 xv

CHAPTER 3 ................................................................................................................ 315

CHAPTER 4 ................................................................................................................ 324
CHAPTER 5 ................................................................................................................ 326
CHAPTER 6 ................................................................................................................ 339
CHAPTER 7 ................................................................................................................ 358

APPENDIX C : The Prevalence of Use of kMMT ..................................................... 359

Abstract ........................................................................................................................ 360
Introduction ................................................................................................................. 361
Methods ....................................................................................................................... 362
Results ......................................................................................................................... 362
Discussion .................................................................................................................... 366
Summary ...................................................................................................................... 368

APPENDIX D : STARD Checklists ............................................................................. 369

STARD Checklist – Study 1 – Estimating the Accuracy of kMMT (Chapter 2) ........ 370
STARD Checklist – Study 2 – Replication of Study 1 (Chapter 3) ............................ 372
STARD Checklist – Study 3 – Grip Strength Study (Chapter 4) ................................ 374
STARD Checklist – Study 4 – The Variation of kMMTAccuracy (Chapter 5) .......... 376
STARD Checklist – Study 5 – Emotional Stimuli Study (Chapter 6)......................... 378

Postface (Final Thoughts) ............................................................................................. 380

References ...................................................................................................................... 383

© drannejensen 2014
 xvi

ABBREVIATIONS

AK Applied Kinesiology (technique)

AK-MMT Applied-Kinesiology-style manual muscle testing

ANEW Affective Norms for English Words

ANOVA analysis of variance

ANZCTR Australian New Zealand Clinical Trials Registry

BEST Bio Energetic Synchronization Technique

BSFF Be Set Free Fast (technique)

C congruent

CI confidence interval

CK Clinical Kinesiology

cm centimetre

CRA Contact Reflex Analysis (technique)

DMT dynamometric muscle testing

DNFT Directional Non-Force Technique

EBM evidence-based medicine

ECU European Chiropractors' Union

EEM Eden Energy Medicine (technique)

EFT Emotional Freedom Technique

F false or female (as defined in text)

FN false negative

FP false positive

HK Health Kinesiology

I incongruent

IAPS International Affective Picture System

kg kilogram

© drannejensen 2014
 xvii

kMMT kinesiology-style manual muscle testing

KST Koren Specific Technique

m slope

MMT manual muscle testing

MRT muscle response testing

NAET Nambudripad's Allergy Elimination Techniques

NET Neuro Emotional Technique

NIS Neurological Integration System (technique)

NMT Neuro Modulation Technique

NOT Neural Organization Technique

NPV negative predictive value

NRT Nutritional Response Testing (technique)

OA osteoarthritis

OxTREC Oxford Tropical Research Ethics Committee

PI principal investigator

PIS participant information sheet

PPV positive predictive value

QRA Quantum Reflex Analysis (technique)

ROC receiver operating characteristic

S strong

SD standard deviation

SOT Sacro Occipital Technique

STAI State-Trait Anxiety Inventory

STARD Standards for the Reporting of Diagnostic Accuracy

STO Soft Tissue Orthopedics (technique)

T true

TBM Total Body Modification (technique)

© drannejensen 2014
 xviii

TFT Thought Field Therapy (technique)

TP test patient

UK United Kingdom

US United States

VAS visual analogue scale

W weak

© drannejensen 2014
 xix

GLOSSARY

In this thesis, the following terms are used as defined:

accuracy the amount of agreement between the results from the index test
and those from the reference standard; specifically, in these studies,
the overall fraction correct; see Figure 1.5.

affect the feeling experienced in connection with an emotion.

analytical validity a test's ability to accurately and reliably measure the entity of
interest; answers the question: "Is it true?"

Applied Kinesiology a specific intervention system which uses MMT not to evaluate
muscular strength or power per se, but to evaluate the neural
control of muscle function; its premise is that when there is some
“aberrant nervous system input to a muscle,” it is less likely to be
able to resist an externally applied force, indicating some type of
neurologic dysfunction, which then may be related to some altered
physiological function, such as organ, endocrine or immune
dysfunction; developed by Dr. George Goodheart in the 1960’s.

arm response testing see kinesiology-style manual muscle testing.

Chance the hypothetical situation where either outcome was equally likely:
50-50.

clinical utility a test's clinical value in relation to costs and patient outcomes;
answers the question: "Is it useful?"

clinical validity a test's meaningfulness to other related clinical data; answers the
question: "Is it meaningful?"

congruent in agreement with; one is “congruent” with a concept one believes

to be true.

conscious / consciousness the state of being aware; used interchangeably with aware: If one
is aware of something, one is conscious of it, and vice versa.

deceit concealment or distortion of the truth for the purpose of misleading.

diagnostic test any process that yields information used to inform patient
management; or alternatively, any method for obtaining additional
information on a patient’s health status.

© drannejensen 2014
 xx

dynamometric muscle a test which quantifies MMT by recording the peak force generated
testing by a muscle or a group of muscles when loaded in tension or
compression; specifically grip strength dynamometric muscle
testing.

eccentric contraction a muscular contraction in which the muscle elongates usually while
under load, where the muscle acts to decelerate the joint at the end
of a movement.

efferent carrying nerve impulses from the central nervous system to an

effector, such as a muscle or organ.

facilitated muscle a strong muscle; see strong.

false not in accordance with what is generally accepted as true or factual.

glenohumeral joint the ball-and-socket joint which is part of the shoulder joint
complex.

Hawthorne Effect the alteration of behavior by the participants of a study due to their
awareness of being observed.

incongruent not in agreement with; one is “incongruent” with a concept that one
believes to be untrue.

indicator muscle in AK and other kMMT techniques, the muscle used for testing;
commonly the deltoid, psoas or pectoralis muscles.

inhibited muscle a weak muscle; see weak.

Intuition a Practitioner’s ability to “read” a TP, using only the senses (i.e.
vision, hearing, touch) without the use of kMMT.

to intuit to "read" a TP, using only the senses (i.e. vision, hearing, touch)
without the use of kMMT.

isometric contraction a muscular contraction characterised by increase in tension without

change in muscle length or joint angle.

kinesiology see kinesiology-style manual muscle testing.

© drannejensen 2014
 xxi

kinesiology-style manual a non-invasive assessment method used to obtain additional

muscle testing information about a patient’s health status by applying an external
force to an “indicator” muscle; while the patient holds a specific
joint in a fixed position (usually in partial flexion), the practitioner-
applied pressure first causes an isometric and then an eccentric
contraction; the test is binary with outcomes labelled as “weak” or
“strong” based on its ability to resist the practitioner-applied force;
see indicator muscle, weak and strong.

Likert Scale an ordinal psychometric measurement used to quantify attitudes,

beliefs or opinions;

lying a false statement made with deliberate intent to deceive; an

intentional untruth.

manual muscle testing the general term for a non-invasive assessment method used to
evaluate neuromusculoskeletal integrity; a fundamental component
of physical examinations performed by physiotherapists,
chiropractors, osteopaths and some medical specialists; if not
otherwise specified, refers to Kendall-and-Kendall / orthopaedic
muscle testing only, and does not refer to kMMT.

meridian in acupuncture and Chinese medicine, each of a set of pathways in

the body along which vital energy ("chi") is said to flow; it is said
that there are twelve such pathways associated with specific organs.

muscle checking see kinesiology-style manual muscle testing.

muscle monitoring see kinesiology-style manual muscle testing.

muscle power a muscle’s ability to generate as much force as fast as possible;

Power = Work/Time.

muscle response testing see kinesiology-style manual muscle testing.

muscle strength the ability of a muscle or muscle group to exert force to overcome a
resistance with no concern for time; Strength = Mass * Distance; it
is the result of 3 factors: (1) physiological strength (muscle size,
cross sectional area, available cross bridging, responses to training),
(2) neurological strength (how strong or weak is the signal that tells
the muscle to contract), and (3) mechanical strength (muscle's force
angle on the lever, moment arm length, joint capabilities).

muscle testing see kinesiology-style manual muscle testing.

© drannejensen 2014
 xxii

naïve (or kMMT-naïve) having no prior experience with kMMT.

negative predictive value the chance that if kMMT stayed strong that the statement was True.

nonconscious all that which is not conscious; and inherently different to what has
been described as “subconscious” or “unconscious” in other
contexts; see conscious / consciousness.

Pair refers to the Practitioner-Test Patient dyad.

positive predictive value the chance that if kMMT went weak that the statement was a Lie.

Practitioner the participant who is performing the kMMT.

precision the degree to which repeated measurements under unchanged

conditions show the same results; see Figure 1.5.

repeatability the closeness of the agreement between independent results

obtained with the same method on the identical subject(s), under
the same conditions; or the variability of the measurements
obtained by one person while measuring the same item repeatedly
(intraobserver variability).

reproducibility the closeness of the agreement between independent results

obtained with the same method on the identical subject(s) but under
different conditions; or the variability of the average values
obtained by several observers while measuring the same item
(interobserver variability).

semantic stimuli stimuli which consist of language, such as words and phrases; such
as a spoken statement.

sensitivity the proportion of the Lies that were detected.

specificity the proportion of the Truths that were detected

STARD guidelines developed to improve the quality of reporting of studies

of diagnostic test accuracy, consisting of the statement, a checklist,
flowchart and an explanation/elaboration document.

strong a muscle which is able to resist a practitioner’s downward pressure.

surrogate in the case that a patient’s muscle/s cannot be tested (e.g. infants),
tested instead are the muscles of a substitute, who is often in close
proximity to and / or touching the patient, but not always.

© drannejensen 2014
 xxiii

Test Patient the participant upon whom the kMMT was being performed.

true / truth that which is generally accepted as fact or reality.

valence (used in psychology, especially in discussing emotions) the

intrinsic attractiveness (positive valence) or aversiveness (negative
valence) of an event, object, or situation.

weak a muscle which is unable to resist a practitioner’s downward

pressure.

© drannejensen 2014
 CHAPTER 1
Introduction

“Truth will ultimately prevail where there is pains to bring it to light.”

George Washington
 2

CHAPTER 1 : INTRODUCTION

Manual muscle testing (MMT) is a non-invasive assessment method used to evaluate

neuromusculoskeletal integrity,1 and is a fundamental component of physical

examinations performed by physiotherapists, chiropractors, osteopaths and some medical

specialists.2 Different health professionals use MMT for different purposes, and as a

result, there exists some confusion surrounding the term itself, and how the tests are

performed and interpreted. Consequently, research efforts to assess the validity and

clinical utility of MMT have been difficult to design, to conduct and even to understand;

and as a result, its usefulness as an assessment method has been called into question.3-6

1.1 The Evolution of MMT

First described in the literature in 1915 by Lovett and Martin, MMT was originally used

to assess muscular weakness in polio patients.7, 8 The tests were crude and unspecific, and

little was known about their validity.

In 1949, in their benchmark textbook, Muscles: Testing and Function, Kendall and

Kendall outlined specific methodologies to isolate and test individual muscles or muscle

groups.1, 9, 10 Currently, it is this type of MMT that is used in orthopaedic and neurology

settings to assess neuromusculoskeletal integrity. This form of MMT usually tests for

muscular strength or power, and outcomes are typically graded from 0 to 5, and

interpreted as 5 being normal.7, 10

In the 1960’s, a different use for MMT was developed by a chiropractor, Dr. George

Goodheart.11 In Goodheart’s technique, called Applied Kinesiology (AK), which is

practiced by approximately 40% of American chiropractors,12, 13 specific muscles are

tested (similar to Kendall and Kendall), not to evaluate muscular strength or power per

© drannejensen 2014
 3

se, but to evaluate the neural control of muscle function.11 The basic premise of AK is

that when there is some “aberrant nervous system input to a muscle,” it is less likely to be

able to resist an externally applied force.11 Therefore, target conditions of AK-style MMT

(AK-MMT) include various types of neurologic dysfunction, such as a sciatic neuropathy

or Bell’s Palsy,14 which then may be related to some altered physiological function, such

as organ, endocrine or immune dysfunction.11, 15-18 However, both the origin(s) and the

cause(s) of this irregular neurological input are yet unclear and fervently debated.19-22

One other notable difference between AK-MMT and the Kendall-style MMT is that in

the AK-style, the outcome is binary, and usually labeled “strong” (or “facilitated”) or

“weak” (or “inhibited”).11 So with this divergence in the 1960’s, various approaches

toward MMT began to emerge. While the tests may be similar in appearance, both the

purpose of performing the tests and the interpretation of the test results differ

significantly.

FIGURE 1.1 – The evolution of manual muscle testing (MMT).

© drannejensen 2014
 4

Following on from Goodheart’s work, a third type of MMT emerged (see Figure 1.1).

This third form, which I will refer to as “kinesiology-style MMT” (kMMT), is estimated

to be used in over 70 different therapeutic technique systems (see Appendix C, page

360), and by over 1 million practitioners worldwide (see Appendix C, page 360). While it

is often referred to colloquially as simply “muscle testing,” it has also been referred to by

other names, such as “kinesiologyi,” “muscle response testing,” “arm response testing,”

“arm testing,” “the arm push down test,” “muscle checking,” “muscle monitoring,” and

others. Examples of technique systems that use kMMT include, but are not limited to:

BodyTalk, Contact Reflex Analysis (CRA), Neuro Emotional Technique (NET),

PSYCH-K, and Total Body Modification (TBM). For further clarification of the

different types of MMT, see Figures 1.2 and 1.3.

FIGURE 1.2 – Performance of the MMT : This Venn Diagram describes how the 3 kinds of
MMT are performed. There is little difference between Kendall-style and AK-MMT, but how
kMMT is performed primarily deviates from both.

i
It may be useful here to note that there are now two other disciplines that use of the term “kinesiology:”
(1) “Kinesiology” as in the study of human movement [Twietmeyer G. What is kinesiology? Historical and
philosophical insights. Quest 2012; 64(1): 4-23.], and (2) “Kinesiology Taping” in the field of
Physiotherapy / Physical Therapy [Kahanov L. Kinesio taping®, part 1: An overview of its use in athletes.
Athletic Therapy Today 2007; 12(3): 17-8.] Both are from different fields altogether, and not related to
kMMT.

© drannejensen 2014
 5

FIGURE 1.3 – Target conditions of MMT: This Venn Diagram describes

common target conditions (i.e. dysfunctions) of the 3 kinds of MMT. The Kendall-
style MMT only tests for neuromusculoskeletal dysfunction, while AK-style and
kMMT are used to test many more conditions.

To be clear, it is the third generation of MMT, kMMT, which is of interest to this

research student and the subject under investigation in this series of diagnostic test

accuracy studies. Therefore, I relinquish further description of the other forms of MMT,

and will now focus exclusively on kMMT.

1.2 The kinesiology-style Manual Muscle Test

A kinesiology-style manual muscle test is distinctly different in a number of ways from

its predecessors:

(1) kMMT is not as specific as either the Kendall-style or the AK-style of MMT;

(2) its applications and interpretations of results are not standardised;

(3) normally only one muscle, commonly called “the indicator muscle,” is used for

testing for various target conditions;

© drannejensen 2014
 6

(4) the amount of force applied to the indicator muscle is also not standardised, with

variations ranging from a great deal of pressure to an amount barely perceivable;

(5) the indicator muscle is tested repeatedly as the target conditions change; and

finally,

(6) which muscle is used as the indicator muscle is of little significance.

This last point means that it is not the specific muscle that is of importance, but what the

practitioner is testing for (i.e. the target condition) that is `fundamental. In other words,

once the practitioner decides on the target condition and the interpretation of the

outcomes, he can use any indicator muscle to conduct the test. Which indicator muscle

used may vary with kMMT technique system and practitioner preference, however, the

anterior or lateral deltoids, the hamstring, the latissimus dorsi or the pectoralis major

muscles are often used.

Nevertheless, kMMT does have some similarities to the other forms of MMT as well. For

instance, its basic premise is comparable in that users contend that alterations in efferent

nervous stimulation into a muscle, will cause the muscle to weaken.23, 24 And again, the

cause(s) and source(s) of these alterations are yet unclear. Another similarity is that

patients are asked to resist the practitioner’s applied pressure in an analogous way.

During a kMMT, an external force is likewise applied to a muscle. At first, this

practitioner-applied pressure causes an isometricii then an eccentricii contraction. More

explicitly, during a kMMT, the patient holds a specific joint in a fixed position, usually in

partial flexion. The practitioner then applies pressure, usually into extension, as the

ii
Isometric contraction: muscular contraction not accompanied by movement of the joint; Eccentric
contraction: a muscular contraction in which the muscle elongates usually while under load, where the
muscle acts to decelerate the joint at the end of a movement. (Mosby's Medical Dictionary, 8th edition.
Oxford, UK: Elsevier, 2009.)

© drannejensen 2014
 7

patient resists this pressure using an isometric contraction. For example, the practitioner

may ask the patient to hold his shoulder (i.e. the glenohumeral jointiii) in 90o flexion,

palm facing down, while he tests the anterior deltoid (see Figure 1.4A). Where the

practitioner places his own hand for the application of the force into extension is often a

matter of contention,9 but the location is routinely on the distal forearm of the patient,

just proximal to the wrist joint (see Figure 1.4.B), with the elbow held in full and locked

extension. Some muscle testing practitioners disagree with this placement, as it

contradicts Kendall’s convention of testing one joint at a time,1 since pressure is being

applied to both the shoulder and elbow joints simultaneously. The degree of shoulder

flexion and abduction and elbow flexion may vary, as do the placement of the

practitioner’s testing hand and his non-testing hand. Finally, while the degree of pressure

that the practitioner applies may differ widely, a steady and constant pressure is thought

to minimise bias, whereas abrupt and inconsistent pressure is thought to introduce bias

into the test.

FIGURE 1.4 – Kinesiology-style Manual Muscle Testing (kMMT): (A) An example of

one style of kMMT, (B) An example of where a practitioner might place his or her hand on a
patient’s wrist.

iii
The Glenohumeral Joint is part of the shoulder joint complex.

© drannejensen 2014
 8

Similar to AK-MMT, the test result of kMMT is binary, with the muscle being tested also

customarily labelled “weak” or “strong” based on its ability to resist the practitioner-

applied force.25 Previous research has established that there is a significant difference

between “strong” muscles and “weak” muscles during a muscle test.3, 25-29 Therefore, in

this series of studies, I have not re-examined if there is a difference, but instead, I have

focused on the gap in the literature: if this difference can be used in an accurate and

clinically meaningful way.

1.3 Applications of kMMT

In this section, I will briefly outline a number of applications of kMMT, or more

explicitly, some target conditions which muscle testing practitioners regularly test for

using kMMT. Because kMMT is used to assess for various target conditions, its accuracy

must be considered, and therefore, for clarity, I define accuracy as, “the amount of

agreement between the results from the index test and those from a reference standard.”

Within the 70+ technique systems that use kMMT, there exists literally hundreds of

potential target conditions, ranging from physiological dysfunction to meridian

imbalance to a patient’s level of stress, and others. See Figure 1.3.

For example, in a review of the literature, kMMT was found to accurately predict low

back pain30, simple phobia,31 and food allergies.32 On the other hand, other studies found

that MMT was unable to accurately predict nutritional needs,33-35 nutritional

intolerance,33, 36 thyroid dysfunction,37 exposure to a practitioner-defined noxious

stimulus,33, 38 and chiropractic subluxation detection and correction.39 Irrespective of

these studies failing to demonstrate sufficient accuracy, practitioners still routinely use

kMMT to attempt to detect these conditions.24, 40

© drannejensen 2014
 9

There are many other examples of target conditions regularly tested for using kMMT that

are not yet supported by clinical research. For instance, in the first course of one popular

kMMT technique system called Total Body Modification (TBM), practitioners are taught

to use kMMT to identify depression, anxiety, organ‐centered problems, blood sugar

problems, autonomic nervous system dysregulation, and overall health status.41 In

addition, another kMMT technique, Neuro Emotional Technique® (NET), also teaches

protocols that use kMMT to assess for emotional stress, blood sugar irregularities, and

meridian imbalance.42 Like NET, another widely-practiced technique called Touch for

Health, uses kMMT to assess for emotional stress and meridian imbalance, and also for

food allergies and the need for nutritional supplementation.43, 44

The wide range of applications and heterogeneity of protocols of use of kMMT

contribute to the difficulties of undertaking rigorous trials on the clinical utility of

kMMT. Plus, the varying interpretations of its outcomes have caused further confusion,

which also must be addressed.9

1.4 Interpreting the outcome of the kMMT

The interpretation of the outcome of a muscle test is not consistent throughout kMMT

techniques. In most kMMT techniques, the practitioner decides the outcome of the

kMMT,22-24 however, in PSYCH-K45 and some others, the client decides. The outcome is

usually labeled “strong” if the muscle is able to resist the practitioner’s downward

pressure and “weak” if it is unable to resist the pressure. Consequently, the outcome of

the kMMT is binary (“strong” or “weak”) and is interpreted to denote the presence or

absence of a target condition. However, in some applications of kMMT, a strong result

could indicate the presence of the target condition, and in other applications of kMMT, a

© drannejensen 2014
 10

strong result could indicate the absence of the very same target condition. This can easily

lead to confusion and misinterpretation both in practice and research. Therefore, it

becomes imperative that those doing the testing be clear about the target condition for

which they are testing, and also about the meanings of the potential outcomes of the test.

Consequently, the target condition and the interpretation of the test outcome must be

specifically identified prior to the initiation of the kMMT.

For these reasons, and because the target condition can literally change from one kMMT

to the next, designing diagnostic test accuracy studies for kMMT can be challenging.

Therefore, for kMMT studies to be meaningful, careful consideration must be given to

the choice of target condition, to the interpretation of the test outcome, and to the choice

of reference standard.

1.5 The Current Status of the Evidence

Until the development of the Standards for the Reporting of Diagnostic Accuracy

(STARD) guidelines in 2003, the evaluation of diagnostic techniques lagged behind that

of interventions, and had been notoriously fraught with inconsistencies and bias.46, 47 This

is especially true about the inconsistent use of terminology used to describe the

usefulness of a diagnostic test: Various terms (e.g. accuracy and precision) are confused

in colloquial English, and at times in the scientific literature as well.48 In assessing the

current status of the kMMT literature, this difficulty is further amplified by the confusion

about the terms “muscle testing” and “kinesiology” – which, as previously described, can

have different meanings in different contexts.

With this in mind, using the electronic databases MEDLINE, MANTIS, PsycINFO® and

CINAHL, a literature search was conducted, firstly narrowly, and then broadening. At

© drannejensen 2014
 11

first, I searched for papers that only used kMMT to detect lies (or truths), then I widened

the search to include papers that used kMMT to detect anything, and then I widened it

further to include papers that also used AK-MMT to detect anything. Since Kendall &

Kendall MMT is just used to detect neuromusculoskeletal dysfunction, I ruled out

searching for papers that used this type of muscle testing. The outcome of my search was

25 papers from peer-reviewed journals that used either kMMT or AK-MMT to detect a

specified target condition. Since research on MMT described in conference proceedings

was found to lack or to fail to report rigorous methods,49 only papers published in peer-

reviewed journals were considered. The reference lists of the included papers were also

checked for relevant research, which resulted in no additions. For a complete description

of search strategy, see Appendix Table B.1.1.

Few rigorous studies have attempted to estimate the accuracy of kMMT or AK-MMT.

On the other hand, there are numerous studies that have looked as other characteristics of

MMT, such as reliability,50-56 validity,50, 54, 57 inter-examiner agreement,6, 55 intra-

examiner agreement,56, 58 predictability,30, 59 internal consistency,60 and diagnosis in

general.34, 36, 37, 61-64 The appropriateness of the use of some of these analyses in regards to

kMMT or AK-MMT is questionable, and the sheer number of terms used to describe its

usefulness is frankly confusing.

However, there are some published studies that do report accuracy estimations. Using the

AK-style of MMT, Caruso and Leisman13 report experienced practitioners (5 year

experience) predicted muscle strength more accurately compared to inexperienced

practitioners (<5 year experience), with accuracies of 98% and 64%, respectively. In

other studies it was found that kMMT was used to accurately predict low back pain27 and

simple phobia,28 and AK-MMT accurately predicted food allergies.29 On the other hand,

© drannejensen 2014
 12

further studies found that AK-MMT was unable to accurately predict nutritional needs,30-
32
nutritional intolerance,30,33 and thyroid dysfunction.34

Nevertheless, one study used kMMT to differentiate between true and false statements.

Monti et al.25 found that when a muscle is tested following a true spoken statement, it

yields significantly different results compared to kMMT following false spoken

statements. Their study found that the indicator muscle stays “strong” after a patient

speaks true statements and goes “weak” after a patient speaks false statements. Their

statements were self-referential statements, which used the speaker’s name, as in, “My

name is (insert one’s name or another name)”.25 One problem with using self-referential

statements is that, in all likelihood, both the muscle tester and the test patient were aware

of the verity of the statement, and therefore, both were not blind.iv Even though examiner

bias was controlled for in Monti’s study,25 there is a chance that unblinded testing may

introduce other biases and thus influence the test’s outcome. While it is generally

accepted among those who use various types of MMT that some bias can exist, little is

currently known about the degree of this bias.

1.6 kMMT as a “Diagnostic Test?”

Some may question the appropriateness of using the term “diagnostic test” in regard to

kMMT. A diagnostic test can be described as “any process that yields information used

to inform patient management,”65 or alternatively “any method for obtaining additional

information on a patient’s health status.”66 Since kMMT is used both to obtain

iv
Blinding of the muscle testers was not specifically reported in Monti’s paper. [Monti, D.A., et al., Muscle
test comparisons of congruent and incongruent self-referential statements. Perceptual and Motor Skills,
1999. 88(3): p. 1019-1028.]

© drannejensen 2014
 13

information about a patient and to guide treatment, kMMT would indeed qualify as a

diagnostic test.

Regarding diagnostic tests, two different approaches have emerged. On one hand, in

medicine, a diagnostic test is used to differentiate people with a specific disease from

those without that specific disease,67 and the result of the test is usually binaryv.68 That is,

the condition or disease is either present or absent.69 On the other hand, in psychology

and education, test outcomes are typically graded according to how much of a specific

attribute a patient experiences,68 such as mild, moderate or severe anxiety. Therefore,

variables are not normally binary, but instead fall along continua.68, 69

In either case, in order for the test results to be useful, the assessment method must have

certain characteristics: It must measure what it claims to measure, and it must do this

consistently. With psychometrics, the terms “valid” and “reliable” are used in regard to

the former and latter, respectively.70 In contrast, medical diagnostic tests are referred to

as being “accurate,” and “precise” or “repeatable.”71

Finally, used in conjunction with a thorough history and clinical examination, diagnostic

tests can help lead to a diagnosis and therefore, guide the course of treatment.71, 72

Nonetheless, it is acknowledged that no medical “diagnosis” is made from the results of a

muscle test; rather, practitioners use kMMT to formulate a clinical impression of a

patient’s condition. Furthermore, in clinical practice, multiple kMMTs are performed in

series before any clinical impression is even considered.

v
It is noted that many medical tests are reported as continua initially, yet when diagnosing, a “cut off”
mark is established, whereupon a patient either has or does not have a certain condition; therefore,
ultimately the test becomes binary.

© drannejensen 2014
 14

In light of these considerations, it would be more appropriate to use the medical

terminology, rather than the psychometric terminology. It would also be appropriate to

consider kMMT a diagnostic test and use the methods of diagnostic accuracy studies and

the corresponding terminology. Furthermore, since the evaluation of kMMT as a

diagnostic tool is in its early stages, it is appropriate to start with estimating its accuracy

(overall fraction correct), sensitivity, specificity and precision (reproducibility and

repeatability) in a controlled setting.73 Subsequent studies will be needed to focus on its

clinical utility, or if kMMT is useful in improving or maintaining the health of patients.73

Accordingly, this series of studies have estimated the accuracy and precision of kMMT
46
, and the STARD Statement have been used to report the results (see Appendix D, page

370).74

1.7 Diagnostic Accuracy of kMMT

Accuracy refers to the overall quality of the test.75 A perfectly accurate test (100%

accuracy) would correctly identify all persons in a sample who have the target condition,

and would not wrongly identify anyone in the sample as having the condition when they

did not.

While there are many ways to quantify the accuracy of a diagnostic test,73 I chose as the

primary outcome of many of the studies in this series, the overall fraction correct of using

kMMT to detect lies. Experts in the field of diagnostic accuracy have questioned the

usefulness of overall fraction correct, or simple “accuracy.”74-76 This is because with

most medical conditions, the difference in importance between a false positive and a false

negative test result could be critical.

© drannejensen 2014
 15

In this series of studies assessing the diagnostic accuracy of kMMT, not only am I

asserting that true positives are equal in importance to true negatives, I am also asserting

that the significance of false positives is equivalent to that of false negatives. However,

future research may prove this not to be the case.

In medical testing, there are more common ways of quantifying the accuracy of a

diagnostic test other than by reporting overall fraction correct, such as sensitivity,

specificity, positive predictive value (PPV) and negative predictive value (NPV).73, 75 For

completeness, in many cases I will be reporting these estimations as well, however, their

meanings might need clarification within the context of kMMT:

Sensitivity = The proportion of the Lies that were detected

Specificity = The proportion of the Truths that were detected

PPV = The chance that if kMMT went weak that the statement was a Lie

NPV= The chance that if kMMT stayed strong that the statement was True

In theory, kMMT would be considered perfectly accurate if it could be used to identify

all false statements as being false and all true statements as being true. While it is

generally accepted that no practitioner-interpreted assessment tool is 100% accurate and

totally bias-free, it is clear that the clinical validity of kMMT has not yet been firmly

established. It is hoped that through these studies, the usefulness of kMMT will become

more evident.

1.8 Diagnostic Precision of kMMT

A diagnostic test is considered valid if it is both accurate and precise; therefore, an in

depth look at the precision of kMMT is in order as well.

© drannejensen 2014
 16

Colloquially, the terms accuracy and precision are often used interchangeably. Where

accuracy is defined as “the amount of agreement between the results from the index test

and those from the reference standard,”66 precision can be defined as “the degree to

which repeated measurements under unchanged conditions show the same results.”77 For

clarification of the difference between accuracy and precision, see Figure 1.5. Some

medical statisticians look at a test’s confidence intervals when determining its

precision.78 Other medical researchers think of precision in terms of reproducibility or

repeatability, two terms which are also frequently confused. For clarity, each is defined

as:

Reproducibility, the closeness of the agreement between independent results

obtained with the same method on the identical subject(s) but under different

conditions… or it is the variability of the average values obtained by several

observers while measuring the same item (interobserver variability).48

Repeatability, the closeness of the agreement between independent results

obtained with the same method on the identical subject(s), under the same

conditions… or it is the variability of the measurements obtained by one

person while measuring the same item repeatedly (intraobserver variability).48

The inconsistency of terminologies is not uncommon in scientific literature.48 The use of

imprecise terms undermines the rigorousness of methods, hinders the interpretation of

results, and hence, weakens the credibility of conclusions.11, 48, 75 Since a good deal of

ambiguity already surrounds kMMT, special care was taken to define terms explicitly.

Applying these terms to kMMT, I designed a study that could be used to estimate its

precision both in terms of reproducibility and repeatability. In designing this study, I was

© drannejensen 2014
 17

curious if a Practitioner’s kMMT accuracy was consistent over many Test Patients (TPs),

or if it was TP- or pair-specific; and likewise, if the kMMT accuracy obtained with one

TP was consistent over many Practitioners, or if it was Practitioner- or pair-specific. For

further details about this study, see Chapter 5.

FIGURE 1.5 – Accuracy & precision in diagnostic tests. (A) Accuracy vs. precision
using a graph; (B-E) Comparing accuracy vs. precision using a target metaphor.

(B)

© drannejensen 2014
 18

1.9 Choice of Research Topic

When I began my PDhil, I initially wished to assess the effectiveness of a new stress

reduction technique for minor depressive disorder (a mood disorder).79 However, this

technique uses kMMT to distinguish truth from lies, which in turn guides the therapy.

Since the validity of kMMT for distinguishing truth from lies had not been established, I

was encouraged first to evaluate the accuracy of kMMT used in this manner. I, therefore,

launched a thorough investigation into studies of diagnostic test accuracy and how they

might be applied to kMMT.

1.10 Research Question and Paradigm to Investigate

Since I was interested in exploring if kMMT can accurately distinguish truth from lies, I

chose as my target condition deceit. Therefore, my research question became: Can kMMT

be used to accurately detect lies?

I could have chosen truth as my target condition; indeed, it is my perception that

clinically, kMMT is used just as often to detect truths as lies. In fact, for a period of time,

I was convinced that I should be targeting truth, since I am usually in pursuit of the truth.

However, after a number of vigorous debates with my supervisors, I was persuaded that

lie detection was a more useful construct in the context of the intervention mentioned

above. So, deceit is the target condition in this series of diagnostic test accuracy studies.

Since test accuracy is largely dependent upon how the target condition and reference

standard are defined,80 great care also was taken in defining the terms truth, and

subsequently deceit, as they were used in these studies. Had the definitions been

ambiguous, the test results would not have been able to be interpreted meaningfully,

limiting the ability to use and interpret the test in a clinical context.

© drannejensen 2014
 19

In the following series of studies, deceit is the target condition and the studies test

whether kMMT was useful in distinguishing lying from truth. Here, “truth” refers to the

colloquial usage of the term: that which is generally accepted as fact or reality.81 (This is

as opposed to abstract concepts of “truth,” such as “the Universal Truth” or “the Higher

Truth.”22) In contrast, “lying” refers to the opposite of “truth,” or more specifically: a

false statement made with deliberate intent to deceive; an intentional untruth; a

falsehood.82 It may be useful to emphasise here that lying implies an intent to deceive,

and in the context of the five studies presented within, it is asserted that test patients are

consciously aware when they are indeed lying and when they are telling the truth. This

therefore may not represent all situations in clinical practice, for example, where patients

are not consciously aware of the truth or falsity of statements. Since testing for

nonconsciousvi beliefs occurs in clinical practice, 24, 42, 83 it is, therefore, a legitimate

avenue of research.84 However, the utility of nonconscious beliefs is not explored in this

series of studies, as I thought it important to establish the utility of kMMT for detecting

conscious untruths before introducing an additional level of complexity into the research.

Throughout this series of diagnostic test accuracy studies, therefore, the target condition

is conscious deceit by the test patient.

It is widely accepted that lying is a stress which may cause observable changes in

physiology, some clearly discernable, such as blushing, others less apparent.85, 86 A

number of different technique systems use kMMT to explore the body’s physiological

response to semantic stimuli,25 which may include both cognitive and emotional

components. The semantic stimuli can be spoken statements,23, 24 questions,22 or

vi
The term “nonconscious” can be described as all that which is not conscious, and is inherently different
to what has been described as “subconscious” or “unconscious” in other contexts. [LeDoux, J.E., The
emotional brain: The mysterious underpinnings of emotional life. 1996, New York: Touchstone.]

© drannejensen 2014
 20

concepts.24 I chose to test a commonly used paradigm among muscle testing practitioners

in this series of studies which is: when a patient believes a spoken statement to be true,

then in response to a kMMT, the muscle will stay “strong;” and when a patient believes

the statement to be false, the muscle will become “weak.” [In this paradigm the patient is

the one speaking the statements, and the practitioner performs the muscle test

immediately afterwards.] (See Table 1.1 for a summary of the paradigm.)

TABLE 1.1 – Summary of kMMT paradigm under

investigation in this series of studies

Test Patient’s Spoken Expected Result of
Statement kMMT
TRUE STRONG

FALSE WEAK
kMMT, kinesiology-style Manual Muscle Testing

This paradigm was selected from hundreds of potential paradigms presently used in

clinical kMMT practice for a number of reasons. Firstly, I regularly use this paradigm in

practice myself, as do tens of thousands of other muscle-testing practitioners around the

world. Secondly, on a logical level, distinguishing lies from truths is relatively

straightforward: something is either true or it is not. Therefore, the valence is clear. This

means that a clear reference standard, indeed a gold standard, is possible.

As the legitimacy of kMMT is often called into question,9, 21, 33, 34, 87-95 a rigorously-

designed series of diagnostic test accuracy studies using a gold standard reference test

may resolve the debate, one way or the other. For all these reasons, this paradigm was the

obvious choice as a starting point for this line of research.

© drannejensen 2014
 21

1.11 Choice of Populations

The choice of populations for this series of studies was also given careful consideration.

The process of kMMT takes two individuals (usuallyvii), the Practitioner and the TP,

dynamically enmeshed in a close therapeutic relationship. Because of this, it was

conceived of early in the study design that there would be two distinct populations: (1)

Practitioners, or those performing the kMMT, and (2) Test Patients, upon whom the

kMMT was being performed, and together they would form one unique “pair.”

Despite my initial objections, it was decided that recruitment of TPs would target healthy

individuals with no prior kMMT experience. I initially objected because this does not

represent a true clinical setting, since most patients of kMMT practitioners do have prior

experience with kMMT and present with a specific complaint (i.e. are “unhealthy” in

some way). However, scientific convention that “blind is better” won in the end and

kMMT-naïve TPs were sought.

For Practitioners, health care providers would be the obvious population, but I

deliberated whether to include any practitioner – or only those with kMMT training. In

the pilot study, any practitioner was enrolled, regardless of having had any kMMT-

training, and the results suggested that the kMMT accuracy of non-trained practitioners

was no better than Chance, so in these studies, only kMMT-trained practitioners were

recruited to perform the testing – regardless of technique, training, or profession. In order

to get a true flavor of what was actually happening in clinical practice, a wide net was

cast.

vii
Some kMMT applications purportedly can be done on oneself through self-testing, but a discussion
about self-testing goes beyond the scope of this dissertation.

© drannejensen 2014
 22

1.12 Main Study Aims

Diagnostic tests are used to detect the presence or absence of a target condition,80 and in

studies of diagnostic test accuracy, the results of one test (i.e. The Index Test) are

compared to the results of a Gold or Reference Standard Test, in the same population,

and the amount of their agreement at detecting the target condition is estimated. In these

studies, the target condition is “lying,” kMMT is the primary Index Test, and the actual

verity of the spoken statement can be considered a Gold Standard, since its presence was

definitively known.

The main objective of this series of studies was to estimate the accuracy (overall fraction

correct) and precision of kMMT to distinguish false from true statements under varying

conditions.

Since kMMT styles vary widely from practitioner to practitioner and from technique

system to technique system, a further goal of these studies was to test its accuracy in as

close to a real clinical setting as possible, by using a variety of practitioner types and

allowing each practitioner to choose his own kMMT style, while adhering to the same

study methods. The characteristics of the stimuli presented varied from affect-neutral to

emotionally-charged to subliminal. The target condition remained constant: deceit. The

gold standard test remained constant: actual verity. To these ends, I present a series of

five studies of diagnostic test accuracy using kMMT.

© drannejensen 2014
 23

1.13 Chapter 1 – List of Tables and Figures

1.13.1 Tables

TABLE 1.1 – Summary of kMMT paradigm under investigation in this series of studies.

1.13.2 Figures

FIGURE 1.1 – The evolution of manual muscle testing (MMT).

FIGURE 1.2 – Performance of the MMT - This Venn Diagram describes how the 3 kinds
of MMT are performed. There is little difference between the Kendall-style and the AK-
MMT, but how kMMT is performed deviates from both.

FIGURE 1.3 – Target conditions of MMT - This Venn Diagram describes target
conditions (i.e. dysfunctions) of the 3 kinds of MMT. The Kendall-style MMT only tests
for neuromusculoskeletal dysfunction, while the AK-style and kMMT are used to test for
many more conditions.

FIGURE 1.4 – Kinesiology-style Manual Muscle Testing (kMMT): (A) An example of

one style of kMMT, (B) An example of where a practitioner might place his or her hand
on a patient’s wrist.

FIGURE 1.5 – Accuracy & precision in diagnostic tests. (A) Accuracy vs. precision
using a graph; (B-E) Comparing accuracy vs. precision using a target metaphor.

© drannejensen 2014
 CHAPTER 2
Study 1 – Estimating the Accuracy of kMMT

“All truths are easy to understand once they are discovered;

the point is to discover them.”

Galileo Galilei
 47

CHAPTER 2 : STUDY 1 – ESTIMATING THE ACCURACY OF KMMT

2.1 ABSTRACT

Research Objective: To estimate the accuracy (overall fraction correct) of kinesiology-

style manual muscle testing (kMMT) to distinguish lies from truth in spoken statements,

with varying degrees of blinding.

Methods: A prospective study of diagnostic test accuracy was carried out. Forty-eight

Practitioners who routinely practised kMMT were paired with kMMT-naïve Test Patients

(TPs) and performed 60 kMMTs as TPs spoke True and False statements. Blocks of

kMMT alternated with blocks of Intuition. Other conditions, such as Not-blind and

Practitioner Misled were also introduced. Bias was controlled for using varying degrees

of blinding and randomisation of True and False statements.

Results: kMMT accuracy was found to be 0.659 (95% CI 0.623 - 0.695), while Intuition

accuracy was 0.474 (95% CI 0.449 - 0.500), which were significantly different (p<0.01).

When the mean accuracy of kMMT was compared to the likelihood of Chance (0.500), a

significant difference was also achieved (p<0.01). Testing for various factors that may

have influenced kMMT accuracy failed to detect any correlations.

Summary: kMMT had significant accuracy for distinguishing lies from truths, compared

to both Intuition and Chance. However, despite tracking on a variety of participant

characteristics, no factor was identified that influenced kMMT accuracy. Strengths of this

study include a high degree of blinding, the heterogeneity of the samples, the choice of a

© drannejensen 2014
 48

clear target condition, and the choice of a “gold standard” reference standard, while the

main limitation was its lack of generalisability to other applications of kMMT.

Keywords: sensitivity; specificity; kinesiology; muscle weakness; lie detection;

deception; lying; intuition; arm; upper extremity

© drannejensen 2014
 49

2.2 Introduction

As mentioned in Chapter 1, kinesiology-style Manual Muscle Testing (kMMT) is

estimated to be used by over 1 million practitioners worldwide in hundreds of

applications and for a vast range of target conditions. Yet despite this prevalence, the

validity of kMMT as a diagnostic test has not been clearly established. According to

Bossuyt, early evaluations of a test should focus on answering the question, “Can I trust

the results of this test?”73 This is commonly known as a test’s analytical validity, or its

ability to measure what it is supposed to measure.73 So this chapter describes my first

study in a series of five diagnostic test accuracy studies and is primarily focused on

answering this fundamental question in as general a setting as feasibly and reasonably

practicable. The main purpose was to estimate the accuracy (i.e. overall fraction correct)

of kMMT for detecting lies.

Previous diagnostic studies of kMMT have been conducted with mixed results. One

study using a similar clinical application found that kMMT is able to distinguish true

statements from false statements: They found a muscle is able to resist significantly more

force after speaking true statements compared to after speaking false statements; that is,

true statements result in a “strong” muscle response, while false statements result in a

“weak” muscle response.25 However, the degree of practitioner and/or test patient

blinding in this study is unclear, leaving a gap in the literature. Therefore, another aim of

this first study was to fill this gap by attempting to estimate kMMT accuracy using

clearly differentiated blind and not-blind conditions.

Additionally, it is widely accepted that there are specific physiological events that

spontaneously occur when someone is lying, such as changes in facial expressions, body

language, speech qualities and skin dampness.96, 97 Therefore, it is possible that it is not

© drannejensen 2014
 50

the kMMT itself that allows practitioners to differentiate lies from truth; they may simply

be proficient at detecting these physiological changes in patients. Therefore, a further aim

of this study was to control for this possibility. To this end, a second index test was

enacted, where practitioners were asked to detect the verity of spoken statements without

using kMMT. That is, they were asked to “intuit” the verity of the spoken statement,

using only visual, auditory and kinesthetic clues. Hence, the Intuition condition was

designed to differentiate a practitioner’s kMMT ability from his ability to “read” a

patient.

There is also a further point to be considered. Critics of kMMT have suggested that

practitioners may bias the test toward their preferred outcome, or what they think the

outcome of the test ought to be. As an active kMMT practitioner myself, I acknowledge

that it may be possible to influence the outcome of kMMT. I also acknowledge the

possibility that patients may influence the outcome of kMMT, for example, through

insincerity of effort.98 However, little is currently known about the degree of influence

that practitioners and patients may actually introduce into a muscle test. A final aim of

this study was to investigate if it is possible to sway practitioner bias. Therefore, a

different condition was presented to participants, unbeknownst to them, during the final

part of testing, whereby it was attempted to mislead practitioners to see if their kMMT

accuracies were affected. While this condition did not assess the amount of bias a test

patient may contribute, it helped to ascertain if practitioner bias is a factor that needs

further consideration.

To summarise, the primary aim of this study was to estimate the accuracy of kMMT for

detecting lies in verbal statements spoken by a test patient when the muscle testing

practitioner was blind to the verity of the spoken statement. Secondary aims were (1) to

© drannejensen 2014
 51

detect factors that may influence kMMT accuracy, and (2) to estimate the amount of bias

that may be introduced by the Practitioner and/or Test Patient (see Table 2.1).

TABLE 2.1 – Summary of study aims.

Primary  To estimate the accuracy of kMMT to detect lies

Secondary  To detect factors that may influence kMMT accuracy

 To estimate the amount of bias that may be introduced

by the Practitioner and/or TP
kMMT, kinesiology-style Manual Muscle Testing

Certainly, the primary purpose of this study was clear, which lent itself to a

straightforward study design. However, the secondary aims added complexity to the

methods, which warrant detailed description.

2.3 Methods

This study was a prospective study of diagnostic test accuracy. No participant was

assessed prior to enrolment. This protocol received ethics committee approval in the

United Kingdom (UK) by the Oxford Tropical Research Ethics Committee (OxTREC;

Approval #34-09), and in America by the Parker University Institutional Review Board

for Human Subjects (Approval #R09-09). Also, this study protocol was registered with

two clinical trials registries: the Australian New Zealand Clinical Trials Registry

(ANZCTR; www.anzctr.org.au), and US-based ClinicalTrials.gov. Written informed

consent was obtained from all participants, and all other tenets of the Declaration of

Helsinki were upheld. Finally, this paper was written in accordance with the Standards

for the Reporting of Diagnostic Test Accuracy Studies (STARD) guidelines (see

Appendix D, page 370, for the STARD Checklist).47, 66, 99

© drannejensen 2014
 52

The methods presented here in this section formed the general structure for the

subsequent studies in this series, and as a result, will be outlined in explicit detail. In later

studies, depending upon the study objectives, some features were removed, in others,

elements were changed; however, the fundamental study design remained consistent.

2.3.1 Participants and Setting

Two groups of participants were recruited: (1) Healthcare practitioners (n=48) who

routinely use kMMT in practice (“Practitioners”), and (2) Test Patients (n=48) who were

naïve to kMMT (“TPs”). Each Practitioner was paired with a single TP – and together

they formed a unique testing pair (“Pair”; hence, n=48 unique pairs). Recruitment was by

direct contact (via email or telephone), social media and word of mouth. Any volunteer

was eligible if he or she was aged 18-65 years, had fully functioning and painfree upper

extremities, and was fluent in English. Volunteers were excluded if they were visually-

impaired, deaf or mute. All Practitioners who wished to participate and met the inclusion

criteria were enrolled, regardless of their profession, kMMT technique(s) used, breadth of

kMMT expertise or experience, or number of years in practice. Once a Practitioner was

enrolled, a unique TP (unacquainted with the Practitioner) was then sought who met the

enrolment criteria. See Table 2.2 for a summary of enrolment criteria.

© drannejensen 2014
 53

TABLE 2.2 – Participant enrolment criteria.

Practitioner Test Patient

(n=48) (n=48)
 Aged 18-65 years  Aged 18-65 years
 Fully functioning & painfree arms  Fully functioning & painfree arms
 Fluent in English  Fluent in English
 Not blind, deaf or mute  Not visually-impaired, deaf or mute
 Did not know Test Patient  Did not know Practitioner
 Any type of healthcare professional  No prior experience with kMMT
 Uses kMMT regularly in practice
 Uses any kMMT technique(s)
 Any amount of kMMT experience
 Any amount of kMMT expertise
 Any number of years in practice

To simulate a real diagnostic environment, data were collected in private clinical settings

whenever possible. All recruitment took place in the United Kingdom and in the US, in

the states of Texas, Arizona, New Mexico and California. Once an eligible Practitioner

was enrolled, a TP was sought and recruited from the lay public, and a mutually

convenient time and place for testing was established. After arriving to the testing site,

each participant was given a Participant Information Sheet (PIS) and gave written

informed consent. Then they completed short Pre-Testing Questionnaires which collected

demographic information and their views on kMMT (See Appendix A, page 245). They

then were introduced to the testing equipment and started the Practice Phase, after which

they began the actual testing phase of the study.

2.3.2 Practice Phase

After the initial forms were completed and prior to the commencement of testing both the

Practitioner and the TP were given the opportunity to practise with the test equipment.

Participants removed all watches and bracelets, and were individually instructed how to

operate the equipment. Then they were allowed to practise until they felt they understood

© drannejensen 2014
 54

what was expected of them and felt comfortable with the procedures. Next, the

participants in each Pair were brought together for the first time. Then the Practitioner

arranged the TP in his preferred kMMT position (see Figure 2.1 for examples of kMMT

positions used in this study), and performed up to 5 practice tests in order to gauge the

amount of force that would be required. Once both participants were comfortable with

the equipment, the protocol, their positions and with each other, the testing phase was

commenced.

FIGURE 2.1 – Examples of kMMT testing positions. (A) Both Practitioner and Test Patient
seated, facing each other; (B) The Test Patient seated and the Practitioner standing to one side;
(C) The Test Patient seated and Practitioner standing in front; (D) Both Test Patient and
Practitioner standing, with the Practitioner behind; and (E) & (F) Both the Test Patient and the
Practitioner standing, with the Practitioner in front to one side. Note also the variations of
shoulder position, that in (B) the thumb of the Test Patient is down, and that in (C) the elbow of
the Test Patient is bent.

© drannejensen 2014
 55

2.3.3 Test Methods

In this section, I first describe in detail the target condition, the primary and secondary

index tests, and the reference standard test. See Table 2.3 for a summary of study

methods. Then I meticulously outline the testing procedure, and discuss the choice of

stimuli, how blinding was accomplished, how results were recorded, and the statistical

methods used for analysis.

2.3.3.1 The Target Condition

The target condition for which the Practitioners were testing was if the TPs were lying.i

The presence (i.e. deceit) or absence (i.e. truth) of the target condition was strictly

controlled: TPs were shown pictures on a computer screen, and were instructed (via an

earpiece) to say a specific statement about the picture while viewing the picture on the

computer screen. Sometimes the statement was false (i.e. target condition present) and

sometimes it was true (i.e. target condition absent). For example, if a TP was presented

with a picture of an apple and was instructed to say, “I see an apple,” this statement was

considered to be true. On the other hand, if a TP was presented with a picture of an apple

and was instructed to say, “I see a horse,” this statement was considered to be false,

therefore, a lie. For an example of TP stimuli, click here.ii During the piloting of these

methods, I monitored how explicitly TPs adhered to the instructions and was satisfied

that adherence was excellent. Finally, it was presumed that TPs fully comprehended

when they spoke true and false statements.

i
Note that the words lying and deceit are used interchangeably.
ii
http://youtu.be/itz0FgqWlss

© drannejensen 2014
 56

TABLE 2.3 – Summary of general methods.

Target Condition ……………….. Truthfulness

Primary Index Test ……………… kMMT

Gold Reference Standard ..……… Verity of Spoken Statement

Secondary Index Test …………… Intuition

Secondary Gold Standard ……….. Verity of Spoken Statement

kMMT, kinesiology-style Manual Muscle Testing

2.3.3.2 The Primary Index Test: kMMT

Because of the many applications of kMMT and variations in kMMT techniques, a

protocol for the index test was developed to maintain a degree of uniformity while at the

same time, allowing Practitioners to remain true to their own clinical form: Practitioners

were instructed to only use either the right or left anterior or lateral deltoid for testing

while the TP was either seated or standing. Beyond that, Practitioners were permitted to

individualise their testing methods, as long as the integrity of the study methods was

maintained (e.g. blinding). Also, once a testing position was initiated, this position was

retained for the duration of participation (i.e. they were not permitted to change arms or

positions for the duration of the testing). For examples of various kMMT testing

positions that pairs assumed, see Figure 2.1.

The paradigm used for these studies was: A true statement results in a strong muscle test,

and a false statement results in a weak muscle test. Plus, since the target condition was

deceit, and lying results in a “weak” muscle test, in the context of this study, a “weak”

test was considered a “positive” result and a “strong” test was considered a “negative”

result. While Practitioners were made fully aware of the paradigm to use during testing,

TPs were kept uninformed. In other words, although TPs knew the study involved muscle

© drannejensen 2014
 57

testing, they were not explicitly told: (1) that Practitioners were testing for deceit; (2) that

their arms may go weak when they lied and stay strong when they told the truth; or (3)

that a weak muscle test indicated a lie and a strong muscle test indicated a truth. Nor

were TPs advised of the outcome of the test (“weak” or “strong’) or the Practitioner’s

interpretation of the outcome. However, despite all attempts to keep TPs blind, it was

suspected that some may guess the paradigm; therefore, this was monitored in the Post-

Testing Questionnaire (See Appendix A, page 245).

Keeping in line with common clinical practice, the interpretation of the outcome of the

kMMT was left to the discretion of each Practitioner. After performing a kMMT, the

Practitioner alone decided if the muscle stayed strong or went weak, and recorded the

results himself by entering “S” for “strong” or “W” for “weak” on a keyboard.

2.3.3.3 The Reference Standard: Actual Truth of the Spoken Statement

The reference standard used in this study was the actual truth of the spoken statement,

which was always definitively known. Further, it was presumed that all participants knew

the difference between True and False statements. Also, the true/false polarity of the

statements were randomly presented, with approximately half being true and half being

false, with each pair being presented with a different sequence.

2.3.3.4 The Secondary Index Test: Using Intuition to Distinguish Lies from Truth

As mentioned in the Introduction, there is a chance that it is not the kMMT itself that

allows practitioners to differentiate lies from truths, but his capacity to “read” a patient.

Therefore, to control for this possibility a second index test was enacted whereby the

Practitioner was asked to use intuition (without muscle testing) if spoken statements were

true or false. Taking into consideration the possibility of visual, auditory and kinesthetic

© drannejensen 2014
 58

clues, Practitioners were asked to watch, listen, and touch the skin of their TP’s arm as

they spoke the given statements. Otherwise, the testing scenario remained exactly the

same as during the kMMT, except kMMT was removed and Intuiting was implemented.

2.3.3.5 The Testing Scenario

Both the Practitioner and the TP viewed their own computer screens. The TP was

presented with an affect-neutral picture and was instructed (via an earpiece) to say a

specific statement while viewing the picture. For an example of a Pair performing a

number of kMMTs, click here.iii In this video you can see the TP’s screen and clearly see

when her arm stays strong and goes weak.

In order to randomly blind the Practitioner to the verity of the spoken statement, he also

viewed a computer screen which half the time presented the same picture as the TP’s and

half the time, a blank, black screen. See Figure 2.2 for a layout of the testing scenario.

Moreover, on the participants’ Instructions Sheets (see Appendix A, page 245), it was

explicitly spelled out that the Practitioner’s screen will be displaying either the same

picture as the TP’s or a blank, black screen. In the instance that the Practitioner was

viewing a blank screen, he was blind to the verity of the TP’s statement, and it was these

blind tests only that were used to calculate the primary outcome (i.e. kMMT accuracy as

overall fraction correct). The purpose of presenting to the Practitioner the same picture as

the TP was three-fold: (1) it served to randomly blind the Practitioner, and to randomly

blind the TP to the Practitioner’s blindness; (2) it set up the Misled condition in Blocks 5

and 6 where I attempted to influence Practitioner bias (see below); and (3) it further

served as a quality control, to see if Practitioners could be persuaded to bias the kMMT.

iii
http://youtu.be/13w516uUNqA

© drannejensen 2014
 59

FIGURE 2.2 - Testing scenario layout: (A) Blocks 1-4; (B) Blocks 5-6. The Practitioner (blue)
viewed a monitor (also blue) which the Test Patient could not see and entered his results on a
keyboard. The Test Patient (TP; red) viewed a monitor (also red) which the Practitioner could
see, had an ear piece in his ear through which he received instructions, and used a mouse to
advance his computer to the next picture/statement. Note that while in Blocks 1-4, the
Practitioner was presented with either the same picture as the Test Patient or a blank, black
screen, while in Blocks 5-6, a third possibility is introduced: a picture which was different from
the Test Patient’s picture. Also note that the Principal Investigator (PI) was present in the room
and observing during all assessments.

Blocks 5-6

© drannejensen 2014
 60

Each Practitioner performed 60 kMMTs on the TP, broken up into 6 blocks of 10 tests. In

Blocks 1-4, each single test consisted of this sequence of events:

(1) the TP was presented with a picture on a computer screen and the Practitioner

was presented with either the same picture or a blank, black screen,

(2) the TP was instructed (via an earpiece) to speak a specific statement while

viewing the picture,

(3) the Pair assumed the testing position,

(4) the TP spoke the instructed statement while viewing the picture,

(5) the Practitioner performed the kMMT,

(6) the Practitioner entered the result of the kMMT into the computer (“S” for

“strong” and “W” for “weak”), which advanced his monitor to the next

picture,

(7) the TP pressed his left mouse button, which advanced his monitor to the next

picture / instruction.

Then the sequence was repeated until 10 kMMT repetitions were completed (i.e. 1

kMMT Block). See Figure 2.3 for the Participant Flow Diagram, and Figure 2.4 for

the flow of 1 test repetition.

© drannejensen 2014
 61

FIGURE 2.3 – Participant Flow Diagram : Study 1

kMMT, kinesiology-style Manual Muscle Testing; TP, Test Patient; *Touching wrist & observing

FIGURE 2.4 – Flowchart for one kMMT / Intuition. Participant pairs performed 10
repetitions of this series per Block. kMMT Blocks were alternated with Intuition Blocks,
and each pair performed 6 Blocks of each (kMMT & Intuition).

© drannejensen 2014
 62

Blocks of Intuition alternated with blocks of kMMT. During the piloting of these

methods, it was discovered that participants’ arms fatigued from repeated kMMT;

therefore, the addition of Intuition Blocks afforded a needed rest, as well as serving to

control for Practitioner’s ability to “read” their patients for clues of deceit.

The Intuition Blocks followed the same basic pattern as the kMMT Blocks:

(1) the TP was presented with a picture on a computer screen, and in Blocks 1-4,

the Practitioner was presented with either the same picture or a blank, black

screen,

(2) the TP was instructed (via an earpiece) to speak a specific statement while

viewing the picture,

(3) the TP spoke the instructed statement while viewing the picture, as the

Practitioner watched, listened and touched his arm,

(4) the Practitioner intuited silently to himself if the TP was telling the truth or

lying,

(5) the Practitioner entered his response into the computer (“C” for congruentiv or

truth and “I” for incongruent or lying), which advanced his monitor to the

next picture,

(6) the TP pressed his left mouse button, which advanced his monitor to the next

picture / instruction.

iv
“I” stood for “incongruent” and “C” for “congruent.” In the pilot, I used these terms to mean “lying” and
“truth”, respectively, so in subsequent studies, I kept the same format for convenience. In clinical practice,
these terms are often used in this context: One is “congruent” with a concept one believes to be true, and
“incongruent” with a concept that one believes to be untrue. [Walker, S.W., Neuro Emotional Technique®
Certification Manual. 2004, Encinitas (CA): Neuro Emotional Technique, Inc.]

© drannejensen 2014
 63

Then the sequence was repeated until 10 Intuition repetitions were completed (i.e. 1

Intuition Block). Again, see Figure 2.4 for the flow of 1 test repetition.

Six kMMT Blocks alternated with six Intuition Blocks, and each Pair started with a

kMMT Block and ended with an Intuition Block. Unbeknownst to the participants, in

Blocks 5 & 6, a twist was added to the testing scenario.

As mentioned, in Blocks 1-4 Practitioners were presented half of the time with the same

picture as the TP’s, and half of the time with a blank, black screen. However, in Blocks 5

& 6 for both kMMT and Intuition, 1/3 of the time he was shown the same picture, 1/3 of

the time he was shown a blank, black screen and 1/3 of the time he was shown a picture

different from the TP’s. See Figure 2.2. Since participants were instructed that the

Practitioner’s screen will be showing either the same picture or a blank screen, the

addition of the different picture was theoretically unanticipated, and so served to attempt

to persuade the Practitioner to bias the kMMT.

While designing this study, much consideration went into the choice of 60 for the number

of kMMT repetitions. First, I polled a number of muscle testing colleagues (n=6) and

asked them to count the number of actual muscle tests they performed during one

consultation (i.e. one office visit), and responses ranged from a low of 12 to a high of 80,

with an average of 40 kMMTs/visit. So, I then experimented with the number of

repetitions that could be feasibly done within a 20-minute period, and found that 40-60

kMMTs could be completed comfortably, and at approximately 60 kMMTs, patients

started to show signs of fatigue. Therefore, I used 60 kMMT repetitions as a maximum

per pair.

© drannejensen 2014
 64

2.3.3.6 The Stimuli

The visual stimuli presented were pictures of common items and were selected from the

International Affective Picture System (IAPS; National Institute of Mental Health Center

for Emotion and Attention, University of Florida, Gainesville, FL).100 Pictures with mean

arousal levels between 4 and 7 (neutral to slightly positive valences) were chosen from

the IAPS database,100 and supplemented with additional similarly neutral pictures. The

IAPS is a widely used, standardised testing system consisting of several thousand

pictures of objects and images from everyday life.100-103

Likewise, the pictures were paired with words selected from the Affective Norms for

English Words (ANEW; National Institute of Mental Health Center for Emotion and

Attention, University of Florida, Gainesville, FL), which had a mean arousal valence

between 4.74 and 7.57 (again, neutral to slightly positive valences).104 The ANEW

database is a list of verbal words with normative emotional ratings which complements

the IAPS for a large number of words in the English language.104, 105

For this study, one hundred picture-word pairs were selected and placed into a database.

Of these, 60 were randomly chosen to present to each pair in the kMMT Blocks, and 60

were randomly chosen to present to each pair in the Intuition Blocks. Stimuli were

presented with a unique sequence of stimuli. The order of visual and audio stimuli was

chosen and presented using DirectRT Research Software (Empirisoft Corporation, New

York, NY). Since the choice of stimuli was randomly determined by the DirectRT

Software, it was possible to have stimuli chosen twice – once during kMMT and once

during Intuition; however, it was not possible for stimuli to be presented twice during

kMMT or twice during Intuition. In other words, all 100 stimulus pairs were available to

be chosen without replacement for each of the kMMT and Intuition conditions.

© drannejensen 2014
 65

Furthermore, the valences of the statements (i.e. True or False) were also randomly

selected by DirectRT software, as were the blinding of the Practitioner (i.e. Same Picture

or Blank Screen in Blocks 1-4, or Same Picture or Blank Screen or Different Picture in

Blocks 5 & 6). It is widely understood that the performance of a diagnostic test can

change from one clinical setting to another due to the mix of patients and changes in

prevalence, which may lead to spectrum bias.106, 107 Therefore, since the actual usual

prevalences that occur during a kMMT session are unknown, valence and blinding were

randomly assigned settings,106, 107 serving to mimic a normal spectrum. However, the

participants were not informed of the exact proportions of True and False Statements or

Blind or Not Blind cases, which might have introduced an expectation bias.108 They were

merely told there would be a mixture.

Since both the Practitioner and the TP were presented with visual stimulus on a computer

monitor, participants and equipment were positioned in such a way as to ensure that each

other’s monitor was not visible by the other. Also, the auditory stimulus (i.e. the

instruction what to say) was presented to the TP through a single earpiece, such that it

was inaudible to the Practitioner. Finally, it was presumed that the participants

recognised the picture being presented. For examples of the visual stimuli, see Figure 2.5.

To hear examples of TP’s auditory stimuli, click herevor herevior here.vii To see examples

of the paired stimuli that TPs might have been presented, click here.viii

v
http://www.drannejensen.com/soundbites/apple.wav
vi
http://www.drannejensen.com/soundbites/giraffe.wav
vii
http://www.drannejensen.com/soundbites/bridge.wav
viii
http://www.youtube.com/watch?v=itz0FgqWlss

© drannejensen 2014
 66

2.3.4 Recording of Results

Once the Practitioner performed the kMMT and decided if the muscle stayed “strong” or

went “weak,” he recorded his finding using a computer keyboard. In all cases, the

Practitioner alone decided the outcome of the kMMT (i.e. “strong” or “weak”). He

pressed the “S” key if the muscle stayed “strong” and the “W” key if the muscle went

“weak.” In the case of Intuition, once the Practitioner decided if the TP was lying or

telling the truth, he pressed the “I” key for lying and the “C” key for truth. This action

advanced his computer monitor to the next picture. The outcome of the kMMT was not

divulged to the TP. In addition, during piloting, I assessed if TPs were aware of the

outcome of the kMMT was and if he noticed what the Practitioner was entering on the

keyboard, and I was convinced that in both instances, the TPs were not aware of any test

outcomes.

FIGURE 2.5 – Examples of visual stimuli. (A), (B) and (C) are examples that
could have been presented to either the Test Patient or the Practitioner, while (D) – a
blank, black screen – could have been presented only to the Practitioner.

© drannejensen 2014
 67

2.3.5 Blinding

Since it was an aim of this study to investigate the impact that blinding had on kMMT

accuracy, much thought was given about how to blind participants, both TPs and

Practitioners.

Firstly, since TPs were kMMT-naïve, they were unaware of what kMMT involved and

were unfamiliar with any kMMT paradigms. Also, while it was impossible to blind TPs

to the reaction of their arms, they were not explicitly told when their arm stayed strong or

went weak, and in many instances, the Practitioner’s determination of test outcome

(“strong” or “weak”) was not obvious to me, an observer, during testing. In addition,

because Practitioners were randomly blinded, TPs were effectively blind to the

Practitioner’s blindness. Furthermore TPs were blind to the interpretation of the kMMT

outcome,ix and blind to what the Practitioner entered into the computer (e.g. “S” or “W”).

Furthermore, no findings or results were discussed with the TP during the testing, nor

was the TP’s opinion sought. Finally, TPs were theoretically not blind to the verities of

the statements they spoke. That is, it was presumed that TPs were aware of when they

spoke true statements and when they spoke false statements.

On the other hand, blinding the Practitioner was, in many respects, more straightforward.

To begin with, clearly Practitioners were not blind to the paradigm being used.ix They

were also aware of: (1) the primary aim of the study (i.e. to estimate kMMT accuracy in

detecting deceit); (2) that TPs were naïve to kMMT; (3) that TPs were unaware of the

paradigm being usedix; and (4) TPs were going to be instructed to either lie or tell the

truth. They were, however, randomly intermittently blind to the verity of the spoken

ix
“Weak” was interpreted as “lying” and “Strong” was interpreted as “truth.”

© drannejensen 2014
 68

statement. In addition, in Blocks 5 & 6 when they were misled about the sameness of the

picture they were shown, they were also blind to being blind. Taking all these factors into

consideration, I believe that overall a high level of participant blinding was achieved.

2.3.6 Pre- and Post-testing Questionnaires

Participants were asked to complete two short questionnaires, one before testing started

and one after testing was completed. The actual questionnaires can be found in Appendix

A (page 245). In the Pre-testing Questionnaire, participants were asked questions about

age, gender, handedness, kMMT experience, degrees of confidence, etc. These

confidence questions were included because it is a popular, yet unsubstantiated, belief

that practitioner confidence is associated with muscle testing accuracy. The degrees of

confidence were measured using a 10cm Visual Analogue Scale (VAS) with the left end

marked “None” and the right end marked with “Complete Confidence.” The participants

were asked to use a “|” to mark the VAS, which was subsequently assigned a score out of

10 rounded to the nearest 0.1cm. For example, if the participant drew his mark at 8.1cm

along the line, he was given a score of 8.1 for that item. Lengths of time, such as ages

and years in practice, were kept as continuous variables, while other variables, such as

gender, profession, and kMMT techniques used, were kept as categorical variables.

In the Post-testing Questionnaire, participants were again asked about their degrees of

confidence. In addition, in the Post-testing Questionnaire, TPs were asked to make open-

ended comments about anything they noticed during the kMMT, in order to establish if

they guessed the paradigm under investigation (i.e. hypothesis-guessing), so that

response bias can be measured.109, 110

© drannejensen 2014
 69

2.3.7 Pilot study

The methods described above were developed after considerable informal and then

formal piloting, wherein flaws and difficulties were detected and remedied, sources of

bias reduced and confounders, considered. For instance, I conducted a pilot study (n=12)

with methods similar to this study; however, there were a number of differences. Firstly,

practitioners with and without kMMT training were enrolled, and likewise, included were

TPs who were kMMT-naïve and also those who were not naïve. Secondly, in a number

of instances, the Practitioner and the TP knew each other. Thirdly, the pictures that the

Practitioners were randomly shown could have been: (1) the same picture as the TP, (2) a

blank, black screen, or (3) a picture different to the TP (i.e. the Misled condition, similar

to Blocks 5 & 6). Finally, while the number of kMMTs was the same (60 repetitions), in

the pilot there were no Intuition Blocks: Practitioners performed 60 kMMTs broken up

into 6 Blocks of 10 kMMTs, with simply a 1-minute rest in between each Block. Aside

from these variations, the methodologies were identical.

Twelve Practitioner-TP pairs participated in the pilot. Practitioners trained in kMMT

were found to be 67.7% accurate (95% CI 52.6% to 82.8%), while those untrained were

51.7% accurate (95% CI 46.7% to 56.7%). See Table 2.4 for more pilot results.

© drannejensen 2014
 70

TABLE 2.4 – Results of the pilot study. Comparison of Practitioner characteristics,

trained and untrained in kMMT.
All Practitioners Trained Untrained
(n=12) (n=8) (n=4)
Mean # Years in Practice (SD) 20.4 (10.0) 20.9 (11.2) 19.4 (8.6)
Mean # Years kMMT Experience (SD) 14.9 (7.3)
Range of kMMT Experience (years) 7.6 – 30.0
Self-Ranked kMMT Expertise* (SD) 3.6 (0.5)
Mean Age (years) (SD) 51.3 (6.6) 52.5 (7.1) 48.8 (5.3)
Gender (Male:Female) 10:2 6:2 4:0
Practitioners by Profession:
Chiropractors 9 5 4
Mental Health Professionals 2 2 0
Acupuncturists 1 1 0
Mean kMMT Accuracy† 0.624 0.677 0.517
95% Confidence Interval 0.526 - 0.722 0.526 – 0.828 0.467 – 0.567
kMMT, kinesiology-style Manual Muscle Testing; SD, Standard Deviation; * 0=No Expertise to 4=Expert; †Overall
Fraction Correct

After completing this pilot, I made several distinct changes to the protocol that

significantly strengthened this current study: (1) because the accuracy scores of

Practitioners not trained in kMMT were similar to Chance (0.500), I decided to only

recruit Practitioners explicitly trained in some form of kMMT; (2) to control the

possibility of TP bias, it was decided to only recruit kMMT-naïve TPs; and most notably,

(3) Intuition Blocks were added, alternating with kMMT Blocks, which made it possible

to account for Practitioners perceiving any physiological clues of deceit, while at the

same time allowing the Pairs to rest between kMMT blocks. Finally, performing this

pilot study allowed me to use its results to perform a sample size estimation for this

study.

2.3.8 Statistical Methods

Since the evaluation of the validity of kMMT is in its early stages, and since I am mainly

interested in estimating how well kMMT is at detecting lies, I report error-based

measures of accuracy: overall fraction correct, sensitivity, specificity, positive predictive

© drannejensen 2014
 71

value (PPV) and negative predictive value (NPV) – with 95% confidence intervals (95%

CI). Sensitivity can be described as the nonerror rate of those with the target condition.73,
111
In this study, since the target condition is deceit, sensitivity was the proportion of false

statements correctly identified as false. Specificity can be considered the nonerror rate of

those without the target condition,73, 111 so was calculated by finding the proportion of

true statements correctly identified as true. Furthermore, PPV and NPV were the

probabilities that a weak test result was actually a lie, and a strong test result was actually

a truth, respectively. Since sensitivity and specificity can vary depending upon testing

conditions,73 and since little is known about the impact of false positives (FP) and false

negatives (FN) in kMMT, I am taking FP to be equal in importance to FN, which later

research might find inappropriate. Error-based measures were also reported for the

Intuition condition. See Table 2.5 for a summary of the statistical terms.

Using the pilot data, a sample size for this full-scale study was calculated. I consulted

with a statistician colleague, and after discussion, I decided that I wanted this study to be

powered to 80%. Using the pilot data, I calculated the overall fraction correct for each

pair, and then compared the mean to the proportion which would be expected by Chance

alone (i.e. 0.500). Based on these assumptions and using a 95% confidence interval, I

determined that a study of 48 pairs would have good statistical power to demonstrate

whether trained practitioners can use kMMT to distinguish a lie from a truth. Statistical

advice was sought during the design phase, after piloting, and before data analysis. All

data were analyzed using Stata/IC 12.1 (StataCorp LP, College Station, Texas),

specifically the commands “ttest” and “pwcorr.”

© drannejensen 2014
 72

TABLE 2.5 – Summary of statistical terms of the error-based measures of

accuracy as defined in the context of this study.
Statistical Term Definition within the context of kMMT
The overall percent correct: (TP + TN) / (TP +
accuracy
FP + TN + FN).

sensitivity The proportion of the Lies that were detected.

The proportion of the Truths that were

specificity
detected.
The chance that if kMMT went weak that the
positive predictive value (PPV) statement was a Lie.
The chance that if kMMT stayed strong that the
negative predictive value (NPV) statement was True.

2.4 Results

2.4.1 Participants

Forty-eight unique Practitioner-TP Pairs were enrolled between June 2010 and October

2011. Four volunteer Practitioners were excluded because they did not meet the age

criteria (i.e. they were aged > 65 years), and 2 volunteer TPs were excluded, one lacked

fluency in English and the other was markedly hearing impaired. Of the 48 enrolled

Pairs, there were 32 female and 16 male Practitioners, and 31 female and 17 male TPs.

The mean (Standard Deviation, SD) age for Practitioners was 49.3 (12.0) years, and for

TPs, 40.8 (12.8) years. Of the 48 Practitioners, 20 were chiropractors, 4 mental health

professionals, 2 acupuncturists, 2 naturopaths, 2 massage therapists, 12 other health

professionals, 4 non-health-professionals, and 2 did not respond to this question. Twenty-

six Practitioners were in full-time practice, 13 were in part-time practice, 7 were not

currently practising, and 2 did not respond to this question. The Practitioners’ mean (SD)

number of years in practicex was 14.8 (10.4), the mean (SD) years of kMMT experiencex

x
Two participants did not respond to this question.

© drannejensen 2014
 73

was 12.9 (1.7), and the mean (SD) hours of performing kMMT/dayx was 3.2 (0.6).

Practitioners were also asked to rate their own kMMT expertise using a Likert scale from

0 (None) to 4 (Expert). The mean (SD) self-ranked kMMT Expertisex was found to be

3.1 (0.2), which suggests that the enrolled Practitioners considered themselves

considerably proficient in kMMT. For a summary of Practitioner demographics, see

Table 2.6.

© drannejensen 2014
 74

TABLE 2.6 – Demographics of Practitioners.

© drannejensen 2014
 75

2.4.2 Test Results

Pairs took between 20 minutes and 1 hour to complete their participation. Of the 48 Pairs,

47 completed the muscle testing assessment in full, while 1 Pair completed only Blocks

1-4, due to a technical problem with the testing equipment. In this results section, for

simplicity’s sake and for clarity, the term “kMMT accuracy” means “the overall fraction

correctxi while the Practitioner was blind, in Blocks 1-4 only”, unless otherwise stated.

Histograms of all accuracy scores show that the data are normally distributed (see Figure

2.6 and Appendix Figure B.2.1), so parametric statistics have been applied below.

FIGURE 2.6 – Histogram showing Normal Distributions of accuracies –

indicating use of t-test statistic is appropriate (n=48). (A) kMMT (40
tests), while Practitioner was Blind (Blocks 1-4 only), (B) Intuition (40
Intuits), while Practitioner was Blind (Blocks 1-4 only).

xi
Accuracy = Overall Fraction Correct = (TP+TN) ⁄ (TP+FP+TN+FN) [TP=True Positives; TN= True
Negatives; FP=False Positives; FN=False Negatives]

© drannejensen 2014
 76

FIGURE 2.6 (con’t.)

2.4.2.1 Accuracies: Blocks 1-4

Diagnostic accuracy can be expressing using a number of statistics: Overall fraction

correct, sensitivity, specificity, PPV and NPV. The kMMT 2x2 tables for each Pair are

located in Appendix Table B.2.1. Since the valence of the statements (i.e. True or False)

and the blinding of the Practitioners were both randomly allocated for each Pair, their

prevalences varied from Pair to Pair. This random allocation caused a variation in the

prevalence of the target condition (i.e. Lies, or False Statements). The prevalence of Lies

ranged from 0.25 to 0.67, and was normally distributed (see Appendix Figure B.2.1.A)

with a mean of 0.47, and a 95% Confidence Interval (CI) of 0.45 – 0.50.

© drannejensen 2014
 77

The primary outcome, kMMT overall fraction correct, was estimated by computing the

individual accuracies of the Pairs. Similar computations were done to estimate overall

sensitivity, specificity, PPV and NPV. The mean (95% CI) accuracy for

kMMT while the Practitioner was blind was 0.659 (0.623 - 0.695), and a range of 0.400 -

0.917. The mean (95% CI) accuracy for Intuition was 0.481 (0.456 - 0.506), and with a

range of 0.238 - 0.636. Since these data were normally distributed (see Figure 2.6), a

Student t-test was used to calculate significance. Under these conditions, when the mean

accuracy of kMMT was compared to the mean accuracy of Intuition, a significant

difference was found (p<0.01; see Table 2.7). Likewise, when the mean accuracy of

kMMT was compared to the likelihood of Chancexii (0.500), significance was also

reached (p<0.01), as was the mean accuracy of Intuition compared to Chance (p<0.05).

In addition, visual inspection of a scatterplot of kMMT Accuracy vs. Intuition Accuracy

(see Figure 2.7) suggested that there was no correlation between the two (r = 0.057).

The mean (95% CI) sensitivity for kMMT was 0.568 (0.504 - 0.633) and the mean (95%

CI) specificity for kMMT was 0.734 (0.687 - 0.782), while the mean (95% CI) PPV for

kMMT was 0.663 (0.607 - 0.781) and the mean (95% CI) NPV for kMMT was 0.667

(0.625 - 0.708). See Table 2.7. A receiver operator characteristic (ROC) space (see

Figure 2.8) shows that all but 7 pairs performed better than Chance.112

xii
Chance here refers to the hypothetical situation where either outcome was equally likely: 50-50.

© drannejensen 2014
 78

TABLE 2.7 – Diagnostic Accuracy while Practitioner is Blind (Blocks 1-4): The means
and 95% CIs for Overall Fraction Correct, Sensitivity, Specificity, Positive Predictive Value,
and Negative Predictive Value. (A) For kMMT; (B) For Intuition.

FIGURE 2.7 – Scatterplot of

kMMT Accuracy vs. Intuition
Accuracy – when the Practitioner
is Blind (Block 1-4 only).

FIGURE 2.8 – ROC Spaces and

ROC Curve for kMMT. (A)
Shadow area represents better
performances; (B) Those above the
red diagonal had a tendency toward
finding strong kMMT responses,
while those below the red diagonal
had a tendency toward weak
kMMT responses; (C) ROC Curve.
[In Blocks 1-4 only, when the
Practitioner was Blind (n=48
pairs).]

© drannejensen 2014
 79

FIGURE 2.8 (con’t.)

© drannejensen 2014
 80

The condition in Blocks 1-4 where the Practitioner was shown the same picture as the TP

was introduced primarily for quality control. Therefore, in those instances where the

Practitioner was not blind but actually knew the verity of the TP’s spoken statement, both

the kMMT and Intuition accuracies theoretically should have been a perfect 1.000.

However, this was not the case: The mean (95% CI) accuracy for kMMT was 0.639

(0.585 - 0.693), and for Intuition, 0.631 (0.577 - 0.685). While they were both

significantly different from Chancexii (p<0.01 for both), they were not distinctly different

from each other (p=0.73). Furthermore, when comparing the Not Blind kMMT accuracy

to the Blind kMMT accuracy, no significant difference was found between the scores (p=

0.44), and they were significantly correlated (r = 0.383, p=0.01).

As a check, I also compared the kMMT and Intuition accuracies of False Statements and

True Statements separately, which theoretically should have been the same as sensitivity

and specificity, respectively. Fortunately, they were. See Appendix Table B.2.2. Also,

comparing kMMT Accuracies for True vs. False Statements showed no correlation using

both a visual inspection of its scatterplot (see Appendix Figure B.2.2.Q), and the

calculation of a correlation coefficient (r = -0.1761, p<0.05).

2.4.2.2 Accuracies: Blocks 5 & 6

In the last 2 Blocks, when the Practitioners were intermittently Misled by randomly being

shown pictures that were different from the TP’s, mean (95% CI) accuracies for both

kMMT and Intuition dropped slightly to 0.566 (0.494 - 0.638) and 0.418 (0.351 - 0.484)

respectively, which were still significantly different from each other (p<0.01). See Table

2.8.A.

© drannejensen 2014
 81

TABLE 2.8 – The Impact of Misleading the Practitioner on kMMT & Intuition Accuracies
(for n=48 pairs). kMMT vs. Intuition Accuracies for the Misled Condition;
Misled vs. Not Misled kMMT Accuracies for Blind and Not Blind Conditions.

Also, I compared the kMMT accuracies in the condition where the Practitioners were

Misled (i.e. Blocks 5-6, shown a picture different from the TP) to the 2 conditions where

(1) the Practitioners were Blind and Not Misled (i.e. Blocks 1-4, shown a blank, black

screen), and (2) the Practitioners were Not Blind and Not Misled (i.e. Blocks 1-4, shown

the same picture as the TP). The former comparison showed a significant decrease in

kMMT scores (p<0.01) during the Misled condition, and the latter showed no significant

difference (p= 0.11) between the two conditions. See Table 2.8.B.

2.4.2.3 Pre- & Post-testing Confidence Ratings

Before and after testing, participants were asked to rate their levels of confidence on a

10cm VAS. TPs ranked their levels of confidence they had in kMMT in general, in their

Practitioner and in their Practitioner’s kMMT ability. Similarly, pre- and post-testing,

© drannejensen 2014
 82

Practitioners rated their levels of confidence they had in kMMT in general, in their own

kMMT ability, and in their ability to kMMT their paired TP (post-testing only).

TPs’x Confidence ratings (95% CI) were:

(1) Confidence in kMMT in general (pre-testing): 6.76 (6.16 – 7.36);

(2) Confidence in kMMT in general (post-testing): 7.22 (6.63 – 7.81);

(3) Confidence in Practitioner (pre-testing): 6.95 (6.30 – 7.61);

(4) Confidence in Practitioner (post-testing): 7.63 (7.01 – 8.25);

(5) Confidence in Practitioner’s kMMT ability (pre-testing): 7.00 (6.35 – 7.65); and

(6) Confidence in Practitioner’s kMMT ability (post-testing): 7.76 (7.10 – 8.41).

The increase in TP Confidence in kMMT in general reached significance (p= 0.03), as

did both the increase in TP Confidence in their Paired Practitioner (p= 0.01) and the

increase in TP Confidence in Practitioner’s kMMT ability (p= 0.01).

Practitionersx Confidence ratings (95% CI) were:

(1) Confidence in their own kMMT ability (pre-testing): 8.43 (8.02 – 8.85);

(2) Confidence in their own kMMT ability (post-testing): 8.15 (7.67 – 8.63);

(3) Confidence in kMMT in general (pre-testing): 8.67 (8.22 – 9.12);

(4) Confidence in their ability to kMMT their paired TP (post-testing): 8.42 (7.94 –

8.92); and

(5) Confidence in their own kMMT ability (post-testing): 7.79 (7.12 – 8.46).

Although the Practitioners ratings of Confidence in their own kMMT ability and

Confidence in kMMT in general both dropped, their differences did not reach

significance (p=0.20 and p=0.22, respectively).

© drannejensen 2014
 83

2.4.2.4 Potential Influencers of Accuracies

It is thought that there are many factors that contribute to the accuracy of kMMT. Such

factors include practitioner and/or patient bias, length and extent of practitioner

experience, fatigue and dehydration, to name a few.

Looking to see if and how the TP influences the outcome, I compared the mean

accuracies of those pairs whose TP reported guessing the paradigm (n=21) to those pairs

whose TPs did not report guessing the paradigm (n=27). For those pairs whose TP

reported guessing the paradigm (n=21), the mean accuracy of kMMT was 0.661 (95% CI

0.591-0.730), and for those pairs whose TP did not report guessing the paradigm (n=27),

the mean accuracy of kMMT was 0.649 (95% CI 0.610-0.688). Since I hypothesised that

those pairs whose TPs reported guessing the paradigm, might have a higher accuracy

than the other group, I used a 1-sided t-test for two samples with unequal variances, and

found that they was no significant difference between these two groups (p=0.38). See

Table 2.9.A.1.

Also, because it was suggested to me that Practitioners may “cheat” during testing by

seeing a reflection of the TP’s screen in their eyeglasses, TP’s eye-glass-wearing was

tracked. The mean accuracy of pairs whose TPs wore glasses (n=15) was 0.640 (95% CI

0.592-0.687), and the mean accuracy of pairs whose TPs did not wear glasses (n=31) was

0.661 (95% CI 0.611-0.712). As such, there was no significant difference (p=0.26) in

kMMT accuracies between these two groups. See Table 2.9.A.2.

I also tested to see if Practitioner profession, practising status or self-ranked kMMT

expertise affected kMMT accuracy. The mean accuracy (95% CI) for the 20

chiropractors who participated was 0.670 (0.611 - 0.729), and for the 26 non-

© drannejensen 2014
 84

chiropractorsx, 0.642 (0.593 - 0.691), whose difference did not reach significance

(p=0.45). See Table 2.9.B.1, and the Appendix Table B.2.3. Likewise, the mean accuracy

(95% CI) for those in full-time practice (n=26) was 0.663 (0.612 - 0.715), part-time

practice (n=13), 0.682 (0.618 - 0.746), and not practising (n=7)x 0.569 (0.465 - 0.673).

When the accuracies of all three groups were compared using an Analysis of Variance

(ANOVA), no difference was found between the three groups (p=0.45). See Table

TABLE 2.9.A – The influence of various categorical factors of the Test Patient on
kMMT accuracy. The Test Patient guessing the paradigm, and (2) The Test Patient
wearing glasses during testing.

2.8.B.2. Similarly, the mean kMMT accuracy (95% CI) of those Practitioners who ranked

themselves as “Expert” muscle testers (4/4; n=15) was 0.682 (0.617 - 0.747), of those

who ranked themselves as 3 out of 4 (n=19), 0.666 (0.605 - 0.728), and of those who

ranked themselves as 1 or 2 out of 4 (n=12)xiii, 0.600 (0.528 - 0.672). While the kMMT

accuracies steadily decreased with ranking, no significant differences were found

between these 3 groups (p=0.35). See Table 2.9.B.3.

xiii
Two Practitioners did not respond to this question, and none ranked themselves as a “0”.

© drannejensen 2014
 TABLE 2.9.B – The influence on various categorical characteristics of Practitioner on kMMT Accuracy. (1) Practitioner Profession, (2)
Practitioner’s Practising Status, and (3) Practitioner’s Self-Ranked kMMT Expertise.*

© drannejensen 2014
85
 86

Since females seem to have an advantage over men in empathy and decoding nonverbal

communication,113, 114 participant gender may have had an impact on kMMT

performance as well. Therefore, I compared the kMMT accuracies by the genders of the

Practitioners and TPs, as well as the sameness of the genders of the Pairs. For male

Practitioners (n=16), the mean accuracy (95% CI) was 0.665 (0.608 - 0.722), and for

female Practitioners (n=32), it was 0.656 (0.609 - 0.704), which did not prove to be

significantly different (p=0.81). However, using male TPs (n=17), the mean accuracy

(95% CI) was 0.715 (0.653 - 0.778), and using female TPs (n=31), it was 0.628 (0.586 -

0.671), which was found to be significantly different (p=0.02). For those Practitioner-TP

Pairs of the same gender (Male-Male or Female-Female; n=27 Pairs), the mean accuracy

(95% CI) was 0.658 (0.603 - 0.713), and for different gender Pairs (Male-Female or

Female-Male; n=21 Pairs), it was 0.661 (0.614 - 0.708), which also did not prove to be

significantly different (p=0.92). See Table 2.9.C.

Choice of arms may also have influenced kMMT accuracy. Looking at the Practitioner’s

preferred arm (of their own) with which to perform kMMT, the kMMT accuracy (95%

CI) of those Practitioners that preferred their right arm (n=35) was 0.650 (0.612 - 0.688),

and of those that preferred their left arm (n=16) was 0.655 (0.583 - 0.726), which was not

significantly different (p=0.90). Note that 5 Practitioners indicated that they prefer either

arm, so were included in both groups, and 2 Practitioners did not respond to this question.

Of those Pairs where the TP’s dominant arm was tested (n=18), the mean kMMT

accuracy (95% CI) was 0.678 (0.609 - 0.746), and whose non-dominant arm was tested

(n=28), 0.639 (0.596 - 0.683), which also was not significantly different (p=0.33). See

Table 2.9.D.

© drannejensen 2014
 87

To see if kMMT accuracy differed by Block, I compared the mean kMMT accuracies for

Blocks 1 through 4. Figure 2.9 shows a bar graph of mean kMMT accuracies (with 95%

CIs) by Block, and no trend is obvious. Further exploration revealed no correlation

among kMMT accuracy by Block (see Table 2.10).

TABLE 2.9.C – The influence on various mixed categorical variables on kMMT

Accuracy. (1) Practitioner’s Gender, (2) Test Patient’s Gender, and (3) Sameness of Gender.
kMMT Accuracy
(1) (2) (3)
(C) Practitioner’s Gender Test Patient’s Gender Practitioner & TP Gender
Males Females Males Females Same Different
(n=16) (n=32) (n=17) (n=31) (n=27) (n=21)
Mean 0.665 0.656 0.715 0.628 0.658 0.661
95% CI 0.608-0.722 0.609-0.704 0.653-0.778 0.586-0.671 0.603-0.713 0.614-0.708
Minimum 0.538 0.400 0.500 0.400 0.400 0.500
Maximum 0.875 0.917 0.917 0.846 0.875 0.917
p-value
0.81 0.0216* 0.92
kMMT, kinesiology-style Manual Muscle Testing; CI, Confidence Interval; TP Test Patient; *Significance reached.

TABLE 2.9.D – The influence on various categorical variables on kMMT Accuracy. Choice
of Practitioner’s and Test Patient’s arms.
kMMT Accuracy
Practitioner’s Preferred Choice of Test Patient’s
Own Arm for kMMT†** Arm Used†
(D) Right Left Dominant Non-Dominant
(n=35) (n=16) (n=18) (n=28)
Mean 0.650 0.655 0.678 0.639

95% CI 0.612 - 0.688 0.583 - 0.726 0.609 - 0.746 0.596 - 0.683

Minimum 0.400 0.421 0.421 0.400

Maximum 0.875 0.917 0.917 0.846

p-value 0.90 0.33
†
kMMT, kinesiology-style Manual Muscle Testing; CI, Confidence Interval; Two (2) participants did not respond to
this question; ** Five (5) Practitioners preferred either right or left.

© drannejensen 2014
 88

FIGURE 2.9 – kMMT Accuracy by Block with 95% Confidence Intervals.

TABLE 2.10 – Correlations (r) among kMMT Accuracies by Block. With p-values (n=48).

2.4.2.5 Correlation Testing

In an effort to further understand the relationships between kMMT accuracy and other

participant characteristics, correlation analyses were run, first by visually inspecting

scatterplots (see Appendix Figure B.2.2), and then by creating correlation matrices using

Stata’s “pwcorr” command. Reviewing the scatterplots, no correlations were obvious, so

further statistical analysis was performed. Calculating the correlation matrix using

kMMT accuracy and the Practitioner characteristics of age, years in practice, years

© drannejensen 2014
 89

practising kMMT, and average hours per day using kMMT, I found no relationship

reached significance with kMMT (p<0.05; see Table 2.11). Moreover, with this data I

failed to find any significant correlation for kMMT accuracy with any continuous

variable collected or calculated (see Appendix Table B.2.4).

TABLE 2.11 – Correlations (r) among kMMT Accuracy and Practitioner

demographics. p (2-tailed) <0.05.†

2.4.3 Post Hoc Analyses

Many practitioners that I met during data collection asked me if practitioners from one

technique system were more accurate than from any other system. Curious myself, I

considered doing this analysis. However, due to the fact that this study was powered for a

sample size of n=48 Pairs, the smaller sizes of the sub-samples (n=2 to 22 Pairs; see

Table 2.12 for the breakdown), comparing individual technique systems to each other

using statistical methods may have lacked sufficient power to be meaningful. Its limited

relevance becomes even more questionable when one considers that many Practitioners

(n=17) reported using multiple kMMT technique systems (whereas 29 reported using

only 1, and 2 did not respond). However, appeasing my curiosity, I did conduct an

analysis comparing the kMMT accuracies of the two most frequently reported systems,

© drannejensen 2014
 90

Neuro Emotional Technique (NET) and AK, and all the others combined. The mean

accuracy (95% CI) of the NET practitioners (n=22) was 0.676 (0.622 - 0.731), for AK

practitioners (n=17), 0.658 (0.602 - 0.714), and for all others combined (n=40), 0.652

(0.610 - 0.695). I found no significant difference between these three groups, yet they

were all significantly different from Chancexii (see Table 2.12). For a more detailed

summary of the scores by technique system, see Appendix Table B.2.5.

During data collection, a considerable difference in testing environments was noted, most

remarkably about noise and nearby activity levels. Therefore, a post hoc analysis was

performed to determine if the results from one testing site (“Location X”), which was

particularly loud, differed from the other testing sites. The mean kMMT accuracy (95%

CI) at Location X was 0.627 (0.513 - 0.741), whereas for those Pairs not tested at this

site, the mean accuracy was 0.668 (0.631 - 0.705). Although the mean kMMT accuracy

was less for Location X, the difference between these two groups did not reach

significance (p=0.46). See Table 2.13.

© drannejensen 2014
 TABLE 2.11 - kMMT Accuracy by kMMT Technique System. Neuro Emotional Technique vs. Applied Kinesiology vs. All Others Combined.

kMMT, kinesiology-style Manual Muscle Testing; CI, Confidence Interval; * significance reached; ** Chance here refers to the hypothetical situation where either outcome (strong or weak) was
equally likely: 50-50. † Other kMMT Technique Systems (# Practitioners) included: Total Body Modification (6), BodyTalk (4), Health Kinesiology (4), Touch for Health (4), Contact Reflex Analysis
(3), Psych-K (3), Clinical Kinesiology (2), Kinesionics (2), Nutritional Response Testing (2), and 1 Practitioner each of: Be Set Free Fast (BSFF), Belief System Technique, BioKinesiology, Energy
Kinesiology, GeoTran Integrations, Lifeworks, Nambudripad's Allergy Elimination Techniques (NAET), Neural Organization Technique (NOT), Sacro Occiptial Technique (SOT), Soft Tissue
Orthopedics (STO), Thought Field Therapy (TFT), Wellness Kinesiology, Wholistic Kinesiology.

© drannejensen 2014
91
 92

TABLE 2.13 – Post Hoc Analysis of Testing Locations. kMMT Accuracy

Location X compared to all other locations combined.
kMMT Accuracy

n Mean 95% CI p-value

Location X 10 0.627 0.513 – 0.741

0.4639
NOT Location X 38 0.668 0.631 – 0.705
kMMT, kinesiology-style Manual Muscle Testing; CI, Confidence Interval

2.4.4 Adverse Events

In all testing locations, aside from TP arm fatigue, there were no adverse events reported

from any testing.

2.5 Discussion

It is essential to know the accuracy of any diagnostic test in order for its clinical

usefulness to be established. With the widespread use of kMMT and with its clinical

validity often questioned, estimating its accuracy is the important first step in

determining its usefulness.

2.5.1 Statement of Principal Findings

kMMT used for distinguishing false from true spoken statements was found to be

significantly more accurate than either Chance or Intuition. Furthermore, there seems to

be no correlation between kMMT accuracy and Intuition accuracy: if a Practitioner was

accurate at kMMT it did not mean he would also be accurate at Intuition, or not accurate

at Intuition, or either vice versa. Likewise, no correlation was found between kMMT

accuracy for True vs. False Statements: if a Practitioner scored highly accurate for False

Statements, it did not imply he would score accurately for True Statements, or vice versa.

© drannejensen 2014
 93

In addition, since no significant difference was found between kMMT accuracy while the

Practitioner was Blind (Blocks 1-4) and when the Practitioner was Not Blind (Blocks 1-

4), this seems to indicate that Practitioners honestly reported their results and/or they

could not be persuaded to bias the kMMT – whether Blind or not. Since the kMMT

accuracy for the Misled condition was not 0.000 as theoretically it could have been, this

suggests that Practitioners could not be persuaded to bias the kMMT – even when being

misled. Since the kMMT accuracy for the Misled condition was less than for the Blind

condition (p<0.01), it could have been that the Pairs were simply fatiguing, since the

Misled condition took place last (Blocks 5 & 6). However, since there was no significant

difference between the kMMT accuracy in the Blind condition of Blocks 5 & 6 and the

kMMT accuracy in the Blind condition of Blocks 1-4, fatigue is an unlikely explanation.

Finally, since the drop in kMMT accuracy in the Misled condition reached significance

(p< 0.01) and since Intuition accuracy also dropped (but not significantly; p= 0.11), this

could mean that Practitioners started to doubt themselves while being misled. Regardless

of the reason, it appears from these results that Practitioners honestly reported their

results and/or they could not be persuaded to bias the kMMT.

Diagnostic accuracy can also be expressed in terms of sensitivity and specificity, and

PPV and NPV. Since kMMT sensitivity was calculated to be 0.568 (95% CI 0.504 -

0.633) and specificity, 0.734 (95% CI 0.687 - 0.782), these findings suggest that about

57% of the time Lies were classified correctly, and about 73% of the time, Truths were

classified correctly (which is the identical findings for False and True Statements alone –

see Appendix Table B.2.2). Correspondingly, a PPV of 0.663 (95% CI 0.607 - 0.781)

implies that, in this testing scenario with a mean prevalence of Lies of 0.47 (95% CI 0.45

– 0.50), if the kMMT went “weak”, there is a 66% chance that the statement was a Lie.

© drannejensen 2014
 94

Likewise, a NPV of 0.667 (95% CI 0.625 - 0.708) suggests that if the kMMT stayed

“strong”, there is a 67% chance that the statement was a Truth. Since both PPV and NPV

will vary with prevalence, should the prevalence of Lies change, the PPV and NPV may

change as well. For example, if the prevalence of Lies should increase, it is likely that the

PPV would decrease and the NPV would increase, correspondingly. Translating this

specifically into the context of this study, if the greater the prevalence of Lies, the more

surety one can have that a statement giving a weak kMMT result was is Lie (PPV); and

analogously, the lower the prevalence of Lies, the more certain that a statement following

a strong kMMT result was actually a Truth (NPV).106 Since the actual prevalence of Lies

is not usually known, the predictive values reported here should not be applied

universally.

Also interesting were the differences in the Confidence ratings for both the TP and the

Practitioner. It seems that for kMMT-naïve TPs, merely participating in this study served

to elevate their confidences significantly. Some cynics have noted a similarly heightened

belief in processes they call “ideomotor actions” following a positive personal

experience.115, 116 However, since there was no correlation between kMMT accuracy and

increase in any TP Confidence rating (see Appendix Table B.2.6), I am disinclined to

attribute this to an ideomotor-like experience. (See below for a further discussion of the

Ideomotor Effect.) On the other hand, for Practitioners, participation made no difference

in their Confidence ratings. Plus, their kMMT accuracy scores were not correlated with

their change in Confidence ratings (see Appendix Table B.2.7).

Another notable finding is the lack of detection of factors that influence kMMT accuracy.

Another study on MMT reported finding that those practitioners with at least 5 years’

experience achieved a 98% accuracy compared to those practitioners with less than 5

© drannejensen 2014
 95

years’ experience who achieved a 64% accuracy.27 Conversely, my results failed to find a

correlation between practice experience and kMMT accuracy.

Despite testing for a range of participant characteristics, no significant correlation was

found between kMMT accuracy and any of the following:

 Practitioner’s profession

 Practitioner’s years of practice

 Practitioner’s years practicing kMMT

 Practitioner’s practice status

 Usual hours per day using kMMT

 The kMMT technique system(s) in which the Practitioner was trained

 Practitioner’s self-ranked kMMT expertise

 Practitioner’s or TP’s age

 Practitioner’s gender

 Pair’s sameness of gender

 TP reported guessing the paradigm

 TP wearing glasses

 Choice of Practitioner’s or TP’s arm

 Testing location

 Practitioner’s confidence in his own kMMT ability (before or after testing)

 Practitioner’s confidence in kMMT in general (before or after testing)

 TP’s confidence in kMMT in general (before or after testing)

 TP confidence in paired Practitioner (before or after testing)

© drannejensen 2014
 96

 TP’s confidence in paired Practitioner’s kMMT ability (before or after

testing)

 Practitioner’s confidence in using kMMT on his paired TP (after testing)

While finding no relationships is interesting, and somewhat puzzling, it may simply mean

that the measures or methods employed or my sample size was inadequate to educe the

connection.

Finally, the only significant relationship detected seemed to be that with male TPs, a

higher kMMT accuracy score was achieved compared to female TPs. This seems to

suggest that males may be “easier” to muscle test accurately than females. This in itself is

interesting when considering a number of points. Firstly, females seem to be more

proficient than males at nonverbal sensitivity and social functioning.113, 114 Secondly,

these results that males are easier to test accurately than females is perplexing since it is

the opposite of what I find is widely believed among kMMT practitioners: I have often

heard it said that men are more difficult to test because of their greater arm “strength” or

their unwillingness to allow their arms to go weak or relinquish “control” of their arm.

Thirdly, this is especially interesting because of the finding that kMMT was equally

accurate whether a TP’s dominant arm or non-dominant was used. All told, this leads me

to speculate that arm “strength” has little to do with kMMT and how accurate it can be.

In summary, while many intriguing results emerged from the data collected, the most

noteworthy finding of this study was that kMMT was used to accurately distinguish lies

from truth. The second most intriguing result was the failure to detect any correlation

with any of the potential influencing factors listed above.

© drannejensen 2014
 97

2.5.2 Strengths and Limitations

Noting the compelling nature of this new evidence, one must keep in mind the limitations

of this study. However, many of its features can be viewed as both a limitation and a

strength. For example, while every effort was made to simulate real clinical settings,

there were occasions when this was impossible: Certain testing locations were loud and

disruptive – and therefore, offered less than ideal testing conditions. However, since the

scores did not differ significantly, either a distracting environment does not make a

difference to kMMT accuracy, or other testing sites were equally distracting.

In addition, the heterogeneity of the samples can be considered a limitation and a

strength. Heterogeneity of Practitioners is a strength in that any type of kMMT

practitioner was recruited, regardless of practitioner-type, professional standing, kMMT-

style, kMMT technique practiced, or length or extent of practice experience. In addition,

in order to be as clinically authentic as possible, Practitioners were allowed to test the

way they usually test, within the few confines of this research scenario. While this is a

strength, this methodology may be criticised for not being a realistic clinical setting. It

may be further criticised for not strictly controlling kMMT methods, for instance, by not

utilising force plates to monitor the amount of force Practitioners apply during the

kMMT, or by not using strictly standardised muscle testing procedures. These concepts

were considered and rejected for the following reasons. Firstly, this study attempted to

reproduce a real clinical setting, and force plates are not routinely used in clinical

practice. Supporting this decision, previous studies using force plates showed a distinct

difference between muscles labelled “strong” and “weak”,3, 25, 26 making their use in this

study redundant. Secondly, in clinical practice, kMMT styles vary widely from

practitioner to practitioner, so again to simulate a real clinical setting, Practitioners were

© drannejensen 2014
 98

allowed to choose their own kMMT style. Prior research has shown that the shape of

force curves did not differ between styles of muscle testing, only between “strong” and

“weak” muscles.11 Therefore, for these reasons force plates and strict testing protocols

were purposely omitted.

Also, while a heterogeneous sample of TPs was sought (i.e. people from a variety of

educational, socioeconomic and age range groups), the use of only kMMT-naïve TPs

introduced a degree of homogeneity into this sample, and as a result, nothing is known

about kMMT accuracy using patients already familiar with its methods, which naturally

would account for a large percentage of a kMMT practitioner’s clientele.

Further to this, in actual clinical settings, patients normally seek out muscle testing

practitioners for specific real concerns, perhaps even having been referred to a specific

practitioner explicitly for kinesiology. In this study, TPs (1) were healthy (i.e. were not

actively seeking treatment); (2) did not have a choice of practitioner – pairs were

assembled at random and by convenience; (3) were naïve to kMMT whereas patients

typically present to kMMT-practitioners of their own volition, and are usually somewhat

aware of what will occur during a session; and (4) again, were kMMT-naïve, whereas

real patients are only kMMT-naïve on one occasion, and then on subsequent visits, they

would not be naïve. By far the majority of patients of kMMT practitioners would fall into

the category of non-naïve, which was not the case in this study. Therefore, for these

reasons, replication of an ideal clinical setting was not achieved.

Clear strengths of this study include a high degree of blinding and the choices of

reference standard and target condition. It is commonly thought that practitioners can

introduce a great deal of bias during kMMT, and therefore, care was taken to control this

© drannejensen 2014
 99

bias. One way to limit practitioner bias was accomplished through random blinding of

Practitioners. Similarly, it is thought that patients can introduce bias in kMMT by letting

their arms go weak at will, an example of response bias39, 109 or social desirability bias;110

however, much effort was made to keep TPs blind as well. Furthermore, the choices of

reference standard and target condition were clear and well-defined. In studies of

diagnostic test accuracy, it is presumed that the target condition is either present or

absent,7 and ideally, the best available method for detecting the presence or absence of

the target condition is used as the reference standard.47 Ideally, studies of diagnostic test

accuracy employ a true “gold” standard, which demonstrates perfect accuracy in

distinguishing the presence or absence of the target condition. However, perfect gold

standards are rare in medical testing, so an imperfect reference standard is normally

used.117 In this study, the reference standard was the actual verity of the spoken

statements, which was definitively known to be either true or false – a perfect reference

standard – and therefore, it may be considered a true “gold” standard.

A weakness of this study is that kMMT was only compared to one other index test (i.e.

Intuition), rather than to other widely-used methods of lie detection, such as polygraph118

and other methods used in forensic science119 and detecting “tells” in poker.120 Another

limitation of this study is its lack of generalisability to other applications of kMMT.

While kMMT may be useful in distinguishing false statements from true statements,

kMMT may not be useful for any other application, such as detecting a food allergy90, 94

or the need for foot orthotics,121 a specific mattress,122 homeopathy123 or a nutritional

supplement.34 This point is important to emphasise due to the widespread and varied use

of kMMT.

© drannejensen 2014
 100

2.5.3 Comparisons to Other Studies

There has only been one other study published that attempted to estimate the accuracy of

kMMT to distinguish truth from lies. The study done by Monti et al.25 was similar to this

study in that kMMT is used to detect true statements; however, there are dissimilarities as

well. TPs (n=89) only spoke 4 statements, 2 true and 2 false, and it appears that only one

practitioner performed all the muscle testing, however this was not stated explicitly. This

latter point is not apparent because the authors make reference to other individuals

involved in this study as “examiners” and “testers” (plural intended). This ambiguity

lends itself to confusion. Another limitation of this study is that the degree of practitioner

blinding is not clear.25 For instance, one statement used in this study was: “My name is

_____.” Each TP spoke two “My name is _____” statements, one true and one false. For

the false statements, males inserted “Alice” and females, “Ralph.” While it is not

reported if the practitioner in this study was blind to the verity of these statements, it is

suspected that blinding is unlikely. The two other statements spoken by the TP were “I

am an American,” and “I am a Russian.” Since being an American citizen was one

inclusion criterion for participants, it is also likely that the practitioner also knew the

verity of these statements as well; however this also was not stated explicitly. Finally, the

degree of TP blinding was not apparent either. While it was clearly stated that TPs were

all kMMT-naïve, it was not specified if the TPs knew the paradigm being used (i.e. True

→ Strong, False → Weak). These are important considerations which may have

significantly limited this study’s usefulness.

© drannejensen 2014
 101

2.5.4 Possible Explanations of Results

The results indicate that there is a statistically significant difference between kMMT and

Chance and between kMMT and Intuition, for detecting lies; however, showing there is a

difference, does not explain why there is a difference. I will now explore possible

explanations as to first why there might be such a difference, and then, other factors that

may have led to this difference.

Firstly, since my findings show a significant difference between kMMT and Chance, this

rules out the possibility of Chance or “luck” causing the difference. This may be an

obvious argument, but one that I felt is important to highlight. During the data collection

phase of this research, it was my experience that there are staunch critics of kMMT who

vehemently make this assertion. Yet, this blind conviction in many ways resembles that

of those kMMT practitioners on the other side of the fence, who “believe” in the process

without scientific proof, and during data collection, I have met a good deal of such

practitioners. In fact, it is the opposite side of the same coin. Perhaps one day it would be

interesting to revisit the argument that belief or faith are contributing factors, given that a

valid measure of these items can be applied.

Second, it has been asserted that it is not kMMT that is used to detect lies, and it is

simply that muscle-testing practitioners are skilled at “reading” people. Alternatively,

practitioners may be picking up “something else” from patients, like intuition or a gut

feeling – or something equally ethereal or indescribable. However, since a significant

difference was found between kMMT accuracy and Intuition accuracy, it seems

improbable that kMMT practitioners are drawing on other such signs of deceit. On the

other hand, using this line of reasoning, it would make sense that kMMT may be just

© drannejensen 2014
 102

another somatic reaction to lying. While it is known that specific physiological changes

occur when lying,86, 124-129 it is possible that a muscle weakening after speaking a false

statement could be one such change. This remains a question for future research.

Another possible explanation of these results is that the Practitioner and /or the TP may

be acting nonconsciously to bias the muscle test. This nonconscious modulation of

muscular movement was labelled the Ideomotor Effect by psychologist/physiologist

William B. Carpenter in 1852.130 It is common to attribute the Ideomotor Effect to any

unproven, puzzling phenomena, such as dowsing, Ouija boards, automatic writing, the

movement of the table and other objects in séances, the motion of a pendulum, Facilitated

Communication and muscle testing.116, 131 However, since the Practitioners were blind to

the verity of the spoken statement, it is unlikely that Practitioners could be unwittingly

responsible for an ideomotor action. Also, since there was no significant difference

between the pairs whose TPs reported guessing the paradigm, and those who did not, it is

unlikely that TPs caused an ideomotor response either.

It may be argued that other factors, such as fatigue and learning, may also play roles in

the accuracy of kMMT. However, since the accuracy of kMMT in the first block was no

different from the last block of regular testing (p=0.35), it is unlikely that either had

much of an influence. In addition, it was suggested that Practitioners may have “cheated”

by seeing a reflection of the TP’s screen in his glasses; however, since there was no

significant difference in kMMT between pairs whose TPs wore glasses from those that

did not, this claim is also unlikely.

Finally, it is also possible that kMMT accuracy in actual clinical settings may differ from

the results obtained in this study, as a result of a Hawthorne Effect.132 Factors such as test

© drannejensen 2014
 103

anxiety or the intensity or duration of assessment may have caused Practitioners to miss

subtle changes in muscular strength.133 All these factors may have contributed to a

reduction in the kMMT accuracy reported. Another example of a possible Hawthorne

Effect influence is my own presence in the testing room. There is a chance that my

presence may have influenced the results in some way; therefore, future research may

isolate the participants and remove any observers to avoid this possibility. Furthermore,

while every effort was made to keep testing conditions similar for all pairs, the nature of

the recruitment and data collection made this impossible. In fact, one testing location was

particularly disruptive; however, while there was a slight decline in the kMMT

accuracies of the group, the difference did not reach significance. Future studies may

want to control their testing environments more closely, if possible. Further to the matter

of the testing environment, a number of Practitioners commented in their Post-Testing

Questionnaires that water should have been made available for TPs, as they believed that

dehydration may affect kMMT accuracy. While this claim remains unsubstantiated for

kMMT, dehydration does seem to inhibit cognition.134-136 In some data collection

venues, water was available, in others, it was not. An avenue of future research might be

to investigate how hydration levels may influence kMMT accuracy.

2.5.5 Implications for clinical practice

These results, showing that kMMT may be useful for distinguishing false from true

statements by practitioners trained in kMMT, may have direct and indirect implication

for clinicians in various facets of healthcare.

Secondly, with the onset of the Evidence-based Practice movement, there has been a

strong emphasis placed on establishing the validity of healthcare practices, yes, among

© drannejensen 2014
 104

clinicians, but especially by policymakers. Since kMMT has been estimated to be used

by over 1 million practitioners worldwide, this may be indication enough to warrant

consideration. Now with some compelling evidence that kMMT may be a valid

assessment method, this might encourage a closer look. The result could be more

practitioners using kMMT to guide treatment, and policymakers making it more widely

available within national and private healthcare plans.

2.5.6 Unanswered questions and future research

The completion of this initial study in this series of 5 studies, which generated

meaningful and applicable results, indicates, first and foremost, that the usefulness of

kMMT as a diagnostic tool may indeed be assessed, similar to other diagnostic tools.

Despite scepticism from both camps – from within conventional medicine and also from

within alternative medicine – a robust methodology has now been developed that may

serve as a benchmark for future research in the kMMT field.

As is the case with any research, this first study has opened a Pandora’s Box of additional

questions. For instance,

 What factors influence kMMT accuracy?

 Why is there such a range in kMMT accuracies (40-91%)?

 What characteristics or conditions are required to achieve a 90% kMMT

accuracy – compared to 40% accuracy?

 Is it possible to “learn” to perform more accurate kMMT?

 How much influence do the Practitioner and TP each have on kMMT accuracy?

 Is Practitioner accuracy stable (reproducible / repeatable)?

© drannejensen 2014
 105

Future research may also wish to address any differences in accuracies between two TP

cohorts: kMMT-naïve vs. non-kMMT-naïve. Also, since this study investigated only one

application of kMMT, future research is needed to rigorously assess other commonly-

used applications of the various forms of kMMT, such as to detect the need for

nutritional supplementation11, 34-36, 137 and chiropractic subluxations.39, 52, 138 It is

acknowledged that clinical investigations of complementary and alternative medicine

techniques are difficult because of factors such as the use of complex, individualised

treatments, lack of standardisations,91 differing treatment outcome philosophies, lack of

appropriate outcome measures, and the importance of the therapeutic relationship.139

Therefore a consensus on basic diagnostic parameters of such conditions must first be

reached before rigorous studies of diagnostic test accuracy can be undertaken. I suggest

that future research on analogous conditions should first focus on establishing concrete

operational definitions, and then on developing a reference standard measure for meeting

diagnostic criteria.

It also could be argued that dissecting the kMMT out of a technique system and

examining it separately would render meaningless results. Obviously, this is not the case,

as the results of this study can be meaningfully applied to any system that uses kMMT to

detect lies (or truths). Furthermore, despite the apparent differences in kMMT technique

systems, they use kMMT in fundamentally similar ways. Therefore, I maintain that as

long as the Practitioner understands the paradigm being used, and agrees on the

interpretation of the dichotomous result (i.e. what a “strong” response means as opposed

to a “weak” response for each kMMT posed), that the utility of specific applications of

kMMT can be satisfactorily investigated using scientific methods. It could be further

© drannejensen 2014
 106

argued that each kMMT technique system should be evaluated for its effectiveness as a

whole, using rigorous scientific methods such as randomised controlled trials.

2.6 Summary

In this study kMMT was used with significant accuracy to distinguish lies from truths,

compared to both Intuition and Chance. This demonstrates that it is possible to develop a

robust methodology for assessing the value of kMMT as a diagnostic tool. Nevertheless,

despite tracking on a variety of testing characteristics, no one factor was identified that

seemed to influence kMMT accuracy. Strengths of this study include a high degree of

blinding, the heterogeneity of the samples, the choice of a clear target condition, and the

choice of a “gold standard” as the reference standard. The main limitation of this study is

its lack of generalisability to other applications of kMMT.

© drannejensen 2014
 107

2.7 Chapter 2 – List of Tables and Figures

2.7.1 Tables

TABLE 2.1 – Summary of study aims.

TABLE 2.2 - Participant enrolment criteria.

TABLE 2.3 - Summary of general methods.

TABLE 2.4 - Results of the pilot study. Comparison of Practitioner characteristics,

trained and untrained in kMMT.

TABLE 2.5 – Summary of statistical terms of the error-based measures of accuracy as

defined in the context of this study.

TABLE 2.6 - Demographics of Practitioners.

TABLE 2.7 - Diagnostic Accuracy while Practitioner is Blind (Blocks 1-4): The means
and 95% CIs for Overall Fraction Correct, Sensitivity, Specificity, Positive Predictive
Value, and Negative Predictive Value. (A) For kMMT; (B) For Intuition.

TABLE 2.8 - The Impact of Misleading the Practitioner on kMMT & Intuition
Accuracies (for n=48 pairs). (A) kMMT vs Intuition Accuracies for the Misled
Condition; (B) Misled vs. Not Misled kMMT Accuracies for Blind and Not Blind
Conditions.

TABLE 2.9.A - The influence of various categorical factors of the Test Patient on kMMT
accuracy. (1) The Test Patient guessing the paradigm , and (2) The Test Patient wearing
glasses during testing.

TABLE 2.9.B - The influence on various categorical characteristics of Practitioner on

kMMT Accuracy. (1) Practitioner Profession, (2) Practitioner's Practising Status, and (3)
Practitioner's Self-Ranked kMMT Expertise.

TABLE 2.9.C - The influence on various mixed categorical variables on kMMT

Accuracy. (1) Practitioner's Gender; (2) Test Patient's Gender, and (3) Sameness of
Gender.

TABLE 2.9.D - The influence on various categorical variables on kMMT Accuracy.

Choice of Practitioner's and Test Patient's arms.

TABLE 2.10 - Correlations (r) with p-values among kMMT Accuracies by Block.
(n=48).

© drannejensen 2014
 108

TABLE 2.11 - Correlations (r) among kMMT Accuracy and Practitioner demographics.
p(2-tailed)<0.05.

TABLE 2.12 - kMMT Accuracy by kMMT Technique System. Neuro Emotional

Technique vs. Applied Kinesiology vs. All Others Combined.

TABLE 2.13 - Post Hoc Analysis of Testing Locations. kMMT Accuracy Location X
compared to all other locations combined.

2.7.2 Figures

FIGURE 2.1 – Examples of kMMT testing positions. (A) Both Practitioner and Test
Patient seated, facing each other; (B) The Test Patient seated and the Practitioner
standing to one side; (C) The Test Patient seated and Practitioner standing in front; (D)
Both Test Patient and Practitioner standing, with the Practitioner behind; and (E) & (F)
Both the Test Patient and the Practitioner standing, with the Practitioner in front to one
side. Note also the variations of shoulder position, that in (B) the thumb of the Test
Patient is down, and that in (C) the elbow of the Test Patient is bent.

FIGURE 2.2 – Testing scenario layout. (A) Blocks 1-4; (B) Blocks 5-6. The Practitioner
(blue) viewed a monitor (also blue) which the Test Patient could not see and entered his
results on a keyboard. The Test Patient (red) viewed a monitor (also red) which the
Practitioner could see, has an ear piece in his ear through which he receives instructions,
and used a mouse to advance his computer to the next picture/statement. Note that while
in Blocks 1-4, the Practitioner was presented with either the same picture as the Test
Patient or a blank, black screen, while in Blocks 5-6, a third possibility is introduced: a
picture which was different from the Test Patient’s picture.

FIGURE 2.3 – Participant Flow Diagram : Study 1.

FIGURE 2.4 - Flowchart for one kMMT / Intuition. Participant pairs performed 10
repetitions of this series per Block. kMMT Blocks were alternated with Intuition Blocks,
and each pair performed 6 Blocks of each (kMMT & Intuition).

FIGURE 2.5 – Examples of visual stimuli. (A), (B) and (C) are examples that could have
been presented to either the Test Patient or the Practitioner, while (D) – a blank, black
screen – could have been presented only to the Practitioner.

FIGURE 2.6 - Histogram showing normal distributions of accuracies - indicating use of

t-test statistic is appropriate (n=48). (A) kMMT (40 tests), while Practitioner was Blind
(Blocks 1-4 only), (B) Intuition (40 Intuits), while Practitioner was Blind (Blocks 1-4
only).

© drannejensen 2014
 109

FIGURE 2.7 – Scatterplot of kMMT Accuracy vs. Intuition Accuracy – when the
Practitioner is Blind (Block 1-4 only).

FIGURE 2.8 – ROC Spaces and ROC Curve for kMMT. (A) Shadow area represents
better performances; (B) Those above the red diagonal had a tendency toward finding
strong kMMT responses, while those below the red diagonal had a tendency toward weak
kMMT responses; (C) ROC Curve. [In Blocks 1-4 only, when the Practitioner was Blind
(n=48 pairs)].

FIGURE 2.9 – kMMT Accuracy by Block with 95% confidence intervals.

© drannejensen 2014
 110

CHAPTER 3
Study 2 – Replication of Study 1

“No amount of experimentation can ever prove me right;

a single experiment can prove me wrong.”

Albert Einstein

© drannejensen 2014
 111

CHAPTER 3 : STUDY 2 – REPLICATION OF STUDY 1

3.1 ABSTRACT

Research Objectives: To replicate Study 1 (Chapter 2, page 47) using a simplified

methodology, and to estimate the accuracy (overall fraction correct) of kinesiology-style

manual muscle testing (kMMT) used to distinguish lies from truth in spoken statements.

Methods: A prospective study of diagnostic test accuracy was carried out. Twenty

Practitioners who routinely practised kMMT were paired with Test Patients (TPs) who may

or may not have been kMMT-naïve. The Pairs performed 40 kMMTs as TPs spoke True and

False statements. Blocks of kMMT alternated with blocks of Intuition. The verity of the

spoken statements was randomly assigned, with the prevalence of Lies fixed at 0.50.

Results: kMMT accuracy was found to be 0.594 (95% CI 0.541 - 0.647), which was

significantly different from Intuition accuracy (0.514; 95% CI 0.483 - 0.544; p=0.01) and

Chance (0.500; p<0.01). These results fell within or close to the 95% Confidence Intervals of

Study 1. Also, similar to the previous study (see page 88), testing for various factors that may

have consistently influenced kMMT accuracy failed to detect any correlations.

Summary: This study successfully replicated Study 1 by again finding that kMMT can be

used with significant accuracy to distinguish lies from truths, compared to both Intuition and

Chance. Moreover, this study further supports the concept that a simple yet robust

methodology for assessing the value of kMMT as a diagnostic tool can be developed and

implemented effectively. Comparable to Study 1, no factors were identified that seemed to

© drannejensen 2014
 112

consistently influence kMMT accuracy. Also similar to Study 1, the main limitation of this

study is its lack of generalisability to other applications of kMMT.

Keywords: sensitivity; specificity; kinesiology; muscle weakness; lie detection; deception;

lying; intuition; arm; upper extremity

© drannejensen 2014
 113

3.2 Introduction

When the results of an experiment are other than what was expected, a natural inclination is

to want to repeat the study to see if its results could be replicated. This was the main reason

for doing Study 2: To repeat Study 1 to see if kMMT could again be used to accurately

distinguish lies from truth. So, using the knowledge gained from the first study, a subsequent

study was designed which featured the salient points of Study 1, with some modifications

aimed at simplifying the methods while improving its robustness.

Through the experiences of Study 1, much was learned about the recruitment of participants

and the data collection processes. Firstly, since the primary outcome was measured only

when the Practitioner was blind, the use of two computers in Study 1 added awkward

logistics and unnecessary complexity. Therefore, in Study 2, the Practitioner’s computer was

removed, leaving him blind to the verity of the TP’s statement for the entire study. Secondly,

another problematic part of Study I was removed: Blocks 5 & 6. In these blocks, I attempted

to deceive the Practitioner into biasing the muscle test. Since the kMMT accuracy in these

blocks was similar to the earlier blocks, it appears that this study design failed to encourage

bias. Since the potential for practitioner bias is a recurring criticism of kMMT, this issue is

important, but is presently left for future research. Therefore, methodologically, Study 2 is a

simplified version of Study 1.

In addition, recruitment of TPs was widened to include those with prior kMMT experience

and those who knew their Practitioner. This was modified after having considerable difficulty

recruiting TPs who met the strict enrolment criteria of Study 1 (see page 52), in the locality of

the already-enrolled Practitioner. So, in Study 2, the TP enrolment criteria were relaxed for

convenience, and also I thought it would be interesting to see what results a stratification of

TPs would produce.

© drannejensen 2014
 114

One other change in the protocol was that whilst in Study 1 I was present, in Study 1 I was

not present during the testing (i.e. I left the room). It has been suggested that I, as a clinician

who uses kMMT regularly in practice, may possess a certain bias toward the validity of this

tool; therefore, it was possible that the results of Study 1 may have been influenced by my

presence. For instance, I may have unintentionally displayed subtle signs when the TP was

lying which I did not when the TP was telling the truth, or vice versa. Irrespective of the

legitimacy or otherwise of these claims, once the pair was ready to begin testing, I left the

room until they were finished.

Because in Study1 I observed that Practitioners often seemed anxious prior to testing, I was

curious if an anxious state influenced performance.140 Therefore, a question about anxiety

was added to the Practitioner’s Pre-testing Questionnaire. In this question, Practitioners were

asked to rate their level of subjective state anxiety using a 10cm Visual Analog Scale (VAS).

A VAS was chosen over more commonly-used anxiety measures, like the State-Trait Anxiety

Inventory (STAI), because of self-report ratings, VASs have been found to have the greatest

sensitivity and the least susceptibility to bias,141 plus they are simple to use.

Finally, the prevalence of Lies in this study was fixed at 0.50. Since the prevalence of the

target condition naturally varies with clinical setting and patient-mix, I wished to examine the

effect a fixed prevalence had on kMMT accuracy. This is different to Study 1, where the

prevalence of Lies varied from Pair to Pair around a mean prevalence of 0.47.

Each of these changes was successfully implemented into the protocol of Study 2. Besides

these, all other aspects were identical: the primary and secondary index tests, the

reference/gold standard test, the target condition, the testing positions and layout, the stimuli,

the remainder of the questionnaires, etc. Finally, the study aim was identical: To estimate the

© drannejensen 2014
 115

accuracy of kMMT to distinguish lies from truth when the practitioner was blind. In the next

section, I will briefly outline the methods again.

3.3 Methods

This study is a prospective study of diagnostic test accuracy. No participant was assessed

prior to enrolment. This protocol received ethics committee approval by the Oxford Tropical

Research Ethics Committee (OxTREC; Approval #41-10) and the Parker University

Institutional Review Board for Human Subjects (Approval # R15_10). Also, this study

protocol was registered with two clinical trials registries: the Australian New Zealand

Clinical Trials Registry (ANZCTR; www.anzctr.org.au), and US-based ClinicalTrials.gov.

Written informed consent was obtained from all participants, and all other tenets of the

Declaration of Helsinki were upheld. This paper was written in accordance with the

Standards for the Reporting of Diagnostic Test Accuracy Studies (STARD) guidelines (see

Appendix D, page 372, for the STARD Checklist).47, 66, 99

Fundamentally, the methodology of this study followed the same basic structure as Study 1

(See page 51), with the following exceptions:

(1) Throughout this study, the Practitioners in this study were not intermittently blind, but

completely blind to the verity of the TP’s statement;

(2) The sample size was reduced to 20 pairs;

(3) TPs who were not kMMT-naïve were included;

(4) Some TPs were acquainted with their Practitioners;

(5) The number of kMMTs & Intuitions were reduced from 60 each to 40 each;

(6) The blocks where the Practitioners were misled were removed;

(7) I, the Principal Investigator, was not present in the room during testing;

© drannejensen 2014
 116

(8) Practitioners were asked to rate their subjective state anxiety prior to testing; and

(9) The prevalence of Lies was fixed at 0.50.

Other than these features, the methodology of this study remained consistent to Study 1 (see

page 51).

3.3.1 Participants and Setting

Two groups of participants were recruited: (A) Healthcare practitioners (n=20) who routinely

use kMMT in practice (“Practitioners”), and (B) Test Patients (n=20; “TPs”). Practitioners

and TPs were recruited in the same manner as in Study 1 (see page 52), in the American

states of Texas, New York, Arizona and California. The enrolment criteria for this study were

similar as well, except also included were those volunteers who were not naïve to kMMT,

and also those that knew the Practitioner. See Table 3.1 for a summary of enrolment criteria.

Another difference was that when participants were ready to begin the testing phase of this

study, I left the room.

TABLE 3.1 – Participant enrolment criteria – Study 2

Practitioner Test Patient
(n=20) (n=20)

 Aged 18-65 years  Aged 18-65 years

 Fully functioning & painfree arms  Fully functioning & painfree arms
 Fluent in English  Fluent in English
 Not visually-impaired, deaf or mute  Not visually-impaired, deaf or mute
 Any type of healthcare professional
 Uses kMMT regularly in practice
 Uses any kMMT technique(s)
 Any amount of kMMT experience
 Any amount of kMMT expertise
 Any number of years in practice

© drannejensen 2014
 117

3.3.2 Test Methods

The target condition, index test, reference standard, the secondary index test (i.e. Intuition),

study population, and participant recruitment and sampling were identical to Study 1 (see

page 51).

3.3.2.1 The Testing Scenario

The main difference in the testing scenario of this study compared to Study 1 was that only

the TP viewed a computer screen: The Practitioner did not. Still, the Pairs were positioned in

such a way so that the Practitioner could not see the TP’s screen. One of the reasons I made

this change was that during Study 1, I had to repeatedly remind Practitioners to look at their

computer screen prior to testing, which many seemed to dislike. See Figure 3.1 for a layout of

the testing scenario for this study.

FIGURE 3.1 – Testing scenario layout: The Test Patient (red) viewed a monitor
which the Practitioner could see, had an ear piece in his ear through which he
received instructions. After the muscle test, the Practitioner (blue) entered his results
on a keyboard.

© drannejensen 2014
 118

Each Practitioner performed 40 kMMTs on the TP, broken up into 4 blocks of 10 tests each

and recorded their results in the same way as in Study 1 (see page 66). Four (4) Intuition

Blocks alternated with 4 kMMT Blocks. See Figure 3.2 for the Participant Flow Diagram.

FIGURE 3.2 – Participant Flow Diagram : Study 2

kMMT, kinesiology-style Manual Muscle Testing; *Touching wrist & observing

The stimuli presented were randomly selected from the same database of 100 affect-neutral

pictures/statements used in Study I. DirectRT Research Software (Empirisoft Corporation,

New York, NY) was programmed to present a unique sequence of stimuli for each Pair, while

randomizing the verity of the statements (i.e. True or False). Another difference of this study

© drannejensen 2014
 119

was that the prevalence of False Statements was set constant at 0.50. Finally, participants

completed similar pre- and post-testing questionnaires as in Study 1 (see page 68; and

Appendix A, page 244); however, a question about the Practitioner’s subjective state anxiety

was added to the Pre-Testing Questionnaire. This question used a similar 10cm VAS,

comparable the questions about Confidence ratings, with the left end anchored with “None,”

and the right end, “Worst Ever” (see Appendix A, page 245).

3.3.3 Statistical Methods

Using the results from Study 1 (see page 75), a new sample size calculation was performed.

Using a 95% confidence interval, it was determined that a study with 20 Practitioner-TP pairs

would have good statistical power to demonstrate the kMMT could be used to distinguish

truth from lies. In addition, subgroup analyses will be shown for completeness, although the

study was not powered to make these kinds of comparisons.

Again, since I was mainly interested in estimating how well kMMT can be used to detect lies,

I report error-based measures of accuracy: overall fraction correct, sensitivity and

specificity73 – and their 95% confidence intervals (95% CI). Error-based measures will also

be reported for Intuition.

Statistical advice was sought, during the design phase and for data analysis for this study. All

data were analyzed using STATA 17.0, specifically the commands ttest and pwcorr, sig.

© drannejensen 2014
 120

3.4 Results

3.4.1 Participants

Twenty unique Practitioner-TP pairs were enrolled between July and November 2011. There

were 13 female and 7 male Practitioners, and 8 female and 12 male TPs. Of the 20

Practitioners enrolled, there were 14 chiropractors, 2 mental health professionals, 1

acupuncturist, and 3 other health professionals. Fourteen Practitioners were in full-time

practice, 4 were in part-time practice, and 2 were not currently practising. The Practitioners’

mean (SD) number of years in practice was 17.6 (9.3) years. The mean age for Practitioners

was 49.3 (12.0) years, and for TPs, 40.8 (12.1) years. For a summary of Practitioner

demographics, see Table 3.2.

© drannejensen 2014
 121

TABLE 3.2 - Demographics of Practitioners

© drannejensen 2014
 122

3.4.2 Test Results

Pairs took between 10 and 40 minutes to complete their participation. All Pairs completed all

testing in full. Aside from TP arm fatigue, there were no adverse events reported from any

testing. Also, I noted that the removal of the Practitioner’s computer not only simplified the

testing methods, it also appeared to be a more natural and authentic clinical scenario

compared to Study 1 (see page 51). All accuracies were normally distributed (see Appendix

Figure B.3.1), so parametric statistics were used, mainly the Student t-test and ANOVA.

3.4.2.1 kMMT and Intuition Accuracies

The mean (95% CI) accuracy (i.e. overall fraction correct) for kMMT was 0.594 (0.541 –

0.647), which was significantly different from both the mean (95% CI) Intuition accuracy,

0.514 (0.483 – 0.544; p=0.01), and Chancei (p<0.01). The 2x2 tables for each Pair can be

found in Appendix Table B.3.1. To calculate sensitivity, specificity, PPV and NPV, I

calculated these statistics for each pair and report their means (95% CIs): sensitivity, 0.503

(0.421 - 0.584); specificity, 0.685 (0.616 - 0.754); PPV, 0.613 (0.553 - 0.673); and NPV,

0.583 (0.534 - 0.631). The ROC Curve for kMMT accuracy (i.e. [sensitivity] vs. [1-

specificity]) can be found in Figure 3.3.

The same mean statistics are reported for the Intuition Condition in Table 3.3. I again ran

several checks. I compared the sensitivity and specificity of kMMT with the mean kMMT

accuracies for False Statement only and True Statements only, and found them to be

identical, as they should be. See Appendix Table B.3.2. In addition, comparing Table 3.3 to

Table 2.6 of Study 1 (see page 74), these mean accuracy statistics were close to or within the

95% CIs of Study 1 (see page 78), and vice versa. That is, the mean kMMT accuracy and

i
Chance here refers to the hypothetical situation where either outcome was equally likely: 50-50.

© drannejensen 2014
 123

NPV obtained in Study 1 fell just above the 95% CIs of the mean kMMT accuracy and NPV

of this study, and the mean sensitivity, specificity and PPV of Study 1 (see page 78) fell

within the respective 95% CIs of this study, while the mean kMMT accuracy of this study fell

just below the 95% CI of the mean kMMT accuracy of Study 1.

TABLE 3.3 – Diagnostic accuracy – as sensitivity, specificity, positive predictive

value, and negative predictive value (n=20 Pairs) (A) For kMMT; (B) For Intuition.
n Mean 95% CI
(A) kMMT
20 0.594 0.541 – 0.647
Overall Fraction Correct 20 0.503 0.421 – 0.584
Sensitivity 20 0.685 0.616 – 0.754
Specificity 20 0.613 0.553 – 0.673
Positive Predictive Value 20 0.583 0.534 – 0.631
Negative Predictive Value
kMMT, kinesiology-style Manual Muscle Testing; CI, Confidence Interval
(B) Intuition n Mean 95% CI
20 0.514 0.483 – 0.544
Overall Fraction Correct
20 0.425 0.356 – 0.494
Sensitivity
20 0.603 0.555 – 0.650
Specificity
20 0.494 0.427 – 0.561
Positive Predictive Value
20 0.515 0.490 – 0.540
Negative Predictive Value
kMMT, kinesiology-style Manual Muscle Testing; CI, Confidence Interval

FIGURE 3.3 – ROC Curve for kMMT. Dots above the red line represent better performances.

© drannejensen 2014
 124

Looking at kMMT accuracy by Block, Figure 3.4 shows that there again appears to be no

distinct pattern, and in fact, no correlations reached significance (see Table 3.4)

FIGURE 3.4 – Mean kMMT accuracies by Block.

kMMT, kinesiology-style manual muscle testing; CI, Confidence Interval

TABLE 3.4 – Correlations (r ) with p-values among kMMT accuracies by Block

© drannejensen 2014
 TABLE 3.5 – The influence on various Practitioner categorical factors on kMMT accuracy. (1) Profession, (2) practising
status, and (3) self-ranked kMMT expertise.

© drannejensen 2014
125
 126

3.4.2.2 Potential Influencing Factors

Again, there may have been participant characteristics that influenced kMMT accuracy;

therefore I considered the same specific traits of Practitioners and TPs. First, I looked at

Practitioner Profession, Practitioner Practising Status and Practitioner Self-ranked kMMT

Expertise. See Table 3.5. Due to the smaller sample size of this study, I broke Practitioner

Profession down into only 2 categories: (2) Chiropractors (n=14) and (2) All Others (n=6).

Chiropractors had a mean (95% CI) kMMT accuracy of 0.607 (0.535 - 0.679), while All

Others achieved 0.563 (0.478 - 0.647), which were not statistically different (p= 0.36). The

breakdown of Practitioners’ Practising Status was: 14 Full Time, 4 Part Time, and 2 Not

Practising. Full Time Practitioners scored a mean (95% CI) kMMT accuracy of 0.561 (0.504

- 0.618), Part Time, 0.706 (0.508 - 0.905), and Not Practising, 0.600 (0.000 – 1.000), and the

difference between these groups did not reach significance (p=0.07). For Self-Ranked kMMT

Expertise, 7 Practitioners rated themselves as a “4” (i.e. “Expert”), 10 as a “3” and 3 as a “2,”

while none rated themselves as a “1” or “0.” The mean (95% CI) kMMT accuracy for “4”-

ranked Practitioners was 0.611 (0.470 – 0.751), for “3”-ranked Practitioners, 0.590 (0.518 -

0.662), and for “2”-ranked, 0.567 (0.387 - 0.746), and their differences by ANOVA did not

reach significance (p=0.86).

Stratifying by technique system practiced, I again grouped the Practitioners into 3 groups: (1)

NET Practitioners (n=14), (2) AK Practitioners (n=11), and (3) All Others (n=3). See Table

3.6. However, since it is common that kMMT practitioners have been trained in more than

one kMMT technique system, it was possible for Practitioners to be included in more than

one group. For instance, if Practitioner A practiced NET and TBM, he was put into class (1)

and (3), and if Practitioner B practiced NET, AK and CK, he was put into all 3 classes. The

NET Practitioners scored a mean (95% CI) kMMT accuracy of 0.579 (0.509 - 0.649), the AK

© drannejensen 2014
 127

Practitioners, 0.580 (0.510 - 0.650), and All Others, 0.605 (0.535 - 0.675), and their

differences were not significant (p=0.80). Once again, I have included this comparison for

completeness; however, due to this cross-pollination and the small sample sizes, this

comparison may not be meaningful.

TABLE 3.6 – The influence on kMMT technique system on kMMT accuracy.

Also, I once again compared Genders: (1) Practitioner Gender, (2) TP Gender, and (3)

Sameness of Gender (see Table 3.7). There were 7 male Practitioners who achieved a mean

(95% CI) accuracy of 0.593 (0.498 - 0.688), and 13 female Practitioners who achieved a

mean (95% CI) accuracy of 0.594 (0.520 - 0.668), and the difference between Practitioners

did not reach significance (p=0.98). Likewise, there were 12 male TPs and 8 female TPs, and

when their mean (95% CI) accuracies [0.575 (0.507 - 0.643) and 0.622 (0.519 - 0.725)

respectively], were compared, unlike these results of Study 1 (see page 87), no significant

difference was found (p=0.40). Finally, the sameness of genders of the Pairs was compared.

Those Pairs of the same gender (i.e. Male-Male or Female-Female; n=5) achieved a mean

(95% CI) kMMT accuracy of 0.625 (0.490 - 0.760), and those Pairs of different genders (i.e.

© drannejensen 2014
 128

Male-Female or Female-Male; n=15) achieved a mean (95% CI) kMMT accuracy of 0.583

(0.519 - 0.648), the difference of which was found to be not significant (p=0.49).

Again for completeness, I report the influence of arm choice on kMMT accuracy: (1)

Practitioners reported choice of preferred arm (of their own) for performing kMMT, and (2)

The use of TP’s dominant vs. non-dominant arm during testing. Sixteen Practitioners reported

preferring to use their own right arm for kMMT, and this group scored a mean (95% CI)

accuracy of 0.598 (0.543 - 0.654), while 4 Practitioners reported preferring to use their own

left arm for kMMT, scoring a mean (95% CI) accuracy of 0.575 (0.321 - 0.829), and there

was no significant difference in their accuracies (p=0.80). Similarly, during testing, 5 Pairs

used the TP’s dominant arm and 15 used the non-dominant arm, and achieved kMMT (95%

CI) means of 0.620 (0.408 - 0.832) and 0.585 (0.533 - 0.637), respectively, the difference of

which was insignificant (p=0.68). See Table 3.8.

TABLE 3.7 – The influence on fixed properties of Participants on kMMT accuracy. (1)
Practitioner’s gender; (2) Test Patient’s gender, and (3) sameness of gender.

© drannejensen 2014
 129

TABLE 3.8 – The influence on various Pair categorical variables on kMMT

accuracy. Choice of Practitioner’s and Test Patient’s arms.

Like Study 1, I compared those TPs who reported guessing the paradigm to those who did

not, and those TPs who wore glasses to those who did not. Those TPs who reported guessing

the paradigm (n=6), scored a mean (95% CI) kMMT accuracy of 0.621 (0.507 - 0.735), and

those who did not (n=14), 0.582 (0.515 - 0.650), and while they were both statistically

different from Chancexii (former, p=0.04, latter, p=0.02), there was no significant difference

between them (p=0.49). Likewise, those Pairs whose TPs wore glasses during testing (n=5)

achieved a mean (95% CI) kMMT of 0.560 (0.386 - 0.734), which was indistinguishable

from Chance (p=0.39), yet likely underpowered. Also, those Pairs whose TPs did not wear

glasses during testing (n=15) achieved a mean (95% CI) kMMT of 0.605 (0.546 - 0.664),

which was statistically different from Chance (p<0.01). Also, the difference between these 2

groups not reach significance (p=0.54). See Table 3.9.

Unlike Study 1, Study 2 also included some TPs who: (1) were non-kMMT-naïve, and (2)

knew their paired Practitioner (see Table 3.9). Comparing the kMMT accuracy of those Pairs

whose TP was not naïve to kMMT (n=9) to those Pairs whose TPs were kMMT-naïve

(n=11), while their difference approached significance, none was found (p=0.07). The mean

(95% CI) kMMT accuracy for those Pairs whose TPs were kMMT- naïve was 0.634 (0.555 -

© drannejensen 2014
 TABLE 3.9 – The influence of various categorical TP characteristics on kMMT accuracy. (1) The Test Patient guessing the paradigm, (2) The Test Patient
wearing glasses during testing, (3) TP experience with kMMT, and (4) TP knew practitioner.

© drannejensen 2014
130
 131

0.713), which was significantly different from Chance (p<0.01), and for those Pairs whose

TP was not naïve to kMMT, 0.544 (0.475 - 0.614), which was no different than Chance

(p=0.018). Correspondingly, those Pairs who knew each other (n=3) achieved a mean (95%

CI) kMMT accuracy of 0.658 (0.340 - 0.977), which was possibly underpowered, was found

to be no different from Chance (p=0.17). On the other hand, those Pairs who did not know

each other (n=17) achieved a mean (95% CI) kMMT accuracy of 0.582 (0.525 – 0.639),

which was statistically different from Chance (p<0.01). Also there was no significant

difference in the mean kMMT accuracies of these 2 groups (p=0.42).

3.4.2.3 Correlation Testing

Correlations were run among kMMT accuracy and various independent variables. Firstly, I

compared kMMT accuracy to the Practitioner demographics of age, number of years in

practice, number of years practising kMMT, and usual number of hours per day using

kMMT. No significant correlations were detected between kMMT accuracy and any of these

variables (see Table 3.10).

TABLE 3.10 – Correlations (r ) among kMMT accuracy and Practitioner demographics.

p(2-tailed)<0.05

© drannejensen 2014
 132

Then, I compared kMMT accuracy to Confidence ratings. Participants were asked to rate

their levels of confidence in a number of items. Practitioners rated:

(1) their Confidence in their own kMMT ability, and

(2) their Confidence in kMMT in general;

while TPs rated:

(1) their Confidence in kMMT in general,

(2) their Confidence in their paired Practitioner, and

(3) their Confidence in their paired Practitioner’s kMMT ability.

Conceiving that it may be the change in these ratings that may have some relationship to

kMMT accuracy, I calculated the differences between the post- and pre-testing ratings, and

compared them to kMMT accuracy. See Table 3.11. kMMT accuracy seemed to be modestly,

but significantly, negatively correlated to the change in Practitioner Confidence in kMMT in

General (Post‒ minus Pre‒testing; r = -0.6049; p<0.01), but to none of the other changes in

Confidence ratings. See Figure 3.5. Furthermore, there was no other significant relationship

detected between kMMT accuracy and any other Confidence score (see Appendix Table

B.3.3).

Finally, looking to see how Practitioner Subjective State Anxiety influenced kMMT

accuracy, I created a scatterplot, which showed no obvious relationship. Further analysis

revealed a correlation coefficient which was insignificant (r = 0.0737; p= 0.76). See Figure

3.6.

© drannejensen 2014
 133

FIGURE 3.5 – Scatterplot for correlation.

FIGURE 3.6 – Correlation among kMMT accuracy and Practitioner’s

subjective anxiety rating.

© drannejensen 2014
 TABLE 3.11 – The change in confidence ratings and correlations (r; with p-values) among kMMT accuracy and participant confidence
scores. p(2-tailed)<0.05

© drannejensen 2014
134
 135

3.5 Discussion

3.5.1 Statement of Principal Findings

This study successfully replicated the results of Study 1, finding that kMMT can be used to

distinguish lies from truth with an accuracy above that of Intuition or Chance. Also the mean

accuracies achieved in this study fell within or very close to the 95% CIs of Study I, and the

mean accuracies of Study 1 (see page 78) fell within or very close to the CIs of this study,

indicating that the results of these two studies were remarkably similar. Furthermore, there

again seems to be no correlations between kMMT accuracy and Intuition accuracy, or

between kMMT accuracy for True vs. False Statements. Furthermore, kMMT accuracy was

not affected by Block, which may mean that fatigue and/or learning appear not to contribute.

A sensitivity of 0.503 (95% CI 0.421 - 0.584) indicated that only half the Lies were detected,

while a specificity of 0.685 (95% CI 0.616 - 0.754) suggests that 69% of the Truths were

detected. These statistics were quite different to those achieved in Study 1 (see page 78).

However, interestingly, in both studies, a higher proportion of Truths were detected

compared to Lies, which may suggest that, when using kMMT, Truths are easier to detect

than Lies.

In this study, prevalence of Lies was fixed at 0.50, compared to Study I, where the prevalence

of Lies varied. From the mean PPV, this study found that if the kMMT result went “weak,”

there was a 61% chance that the statement was actually a Lie, and from the NPV, if the

kMMT result was “strong,” there’s a 58% chance that the statements was True. These

findings were distinctly different from those Pairs tweezed out of Study 1 whose prevalence

of Lies was exactly 0.50 (n=9; see Appendix Table B.2.8). Once again, since the prevalence

© drannejensen 2014
 136

of Lies is not actually known in a true clinical setting, the predictive values reported here

should not be applied universally.106

Because this study was powered to evaluate kMMT overall fraction correct (n=20), I am

cautious about assessing the credibility of any subgroup analysis, especially with the small

sample sizes of some of the subgroups.142 However, for completeness, I discuss them below.

The present study explored three additional factors that Study 1 did not consider: (1) The

effect of Practitioner’s Subjective Anxiety on kMMT accuracy, (2) the effect of using non-

kMMT-naïve TPs, and (3) the effect of using participants that knew each other. None of these

factors made a significant difference to kMMT accuracy.

Overall, analogous to Study 1, this study failed to find any consistently significant influencers

of kMMT, even the influencer that Study 1 foundii proved to lack significance in this study.

That is, the following characteristics failed to be related to kMMT accuracy:

 Practitioner profession

 Practitioner’s number of years in practice

 Practitioner’s number of years practising kMMT

 Practitioner’s usual number of hours/day using kMMT

 Practitioner’s kMMT technique(s) used

 Practitioner’s current practising status

 Practitioner’s self-ranked kMMT-expertise

 Practitioner age

 TP age

ii
Study 1 found that Pairs with male TPs achieved a higher mean accuracy than those with female TPs.

© drannejensen 2014
 137

 Practitioner’s gender

 TP’s gender

 Pair’s sameness of gender

 TP’s handedness

 Which arm the Practitioner used during testing

 If the TP reporting guessing the paradigm

 If the TP wore glasses during testing

 If the TP was kMMT-naïve

 If the TP knew their paired practitioner

 Any TP confidence rating

 All but one (1) Practitioner confidence rating

 Practitioner’s subjective anxiety, or

 Block of testing (Late vs. Middle vs. Early in the testing)

This failure to detect any factor that consistently impacts kMMT accuracy is curious indeed. I

can only speculate that no relationships between these variables actually exist, or that the

measures I used to detect them lack sufficient sensitivity or the study lack sufficient power.

3.5.2 Strengths and Limitations

Many of the strengths and limitations of Study 1 also apply to this study (see page 97).

However, some of the factors that were a source of limitation in Study 1 were modified to

improve this study. For example, the sample of TPs was more heterogeneous in that also

included were those that were not naïve to kMMT, as were those that knew their paired

Practitioner. These changes did not seem to affect the study outcome, and yet broadened the

generalisability.

© drannejensen 2014
 138

Other limitations of Study 1 could not be improved. For instance, while every effort was

made to keep the testing environments realistic, the nature of the research setting still could

not exactly replicate a true clinical setting. The other primary limitation common to both

studies is that their results cannot be generalised to other applications of kMMT. While

kMMT may be useful in distinguishing false statements from true statements, kMMT may

not be useful for any other application, such as detecting food allergy, the need for nutritional

supplementation, or foetal gender, despite being used to detect such targets in clinical

practice.34, 90, 94, 143 Another limitation of this study is the smaller sample size may have

underpowered subgroup analyses, since no significant relationships were detected once again.

One of the clear strengths of this study is that a simpler methodology achieved similar results.

Other strengths shared by other studies are the choice of reference standard (i.e. a “gold”

standard) and a high degree of blinding. In this study, the Practitioners were blind throughout,

which may more closely emulate a true clinical setting. Also, while some TPs reported

guessing the paradigm and therefore were (or during the course of testing, became) possibly

not blind to expectations, TP’s blindness did not seem to influence kMMT accuracy. In

addition, in this study TPs were again not blind to the verity of the statements they spoke,

which also does not necessarily match a true clinical situation.

3.5.3 Possible Explanations of Results

Since the results obtained in this study were similar to that of Study 1 (see page 75), a

number of explanations are in order. Firstly, despite Study 1 being replicated without

randomly blinding the Practitioner and without solely naïve TPs, the similarity in results

suggests that blinding in kMMT is less important than was previously supposed.

© drannejensen 2014
 139

The one significant correlation detected in this study was between kMMT accuracy and the

change in Practitioner’s Confidence in kMMT in General. This correlation was moderate and

negative, which oddly translates to this: Practitioners being more confident after testing than

before was related to a lower kMMT accuracy. Since in practice, this makes little sense, a

likely explanation of this significant correlation is chance, and at minimum, further

examination is warranted before any genuine relationship can be considered.144

Also, it was suspected that the level of anxiety that a Practitioner experienced prior to testing

might have a deleterious impact on being able to perform kMMT accurately.145, 146 However,

this was not the case. Reasons for this could be that anxiety does not actually affect kMMT

testing, or that the 10cm VAS used was not an adequate measure of subjective state anxiety.

Another reason may be that it is not anxiety per se that might be an influencer but rather

stress or a stress response.

3.5.4 Implications for clinical practice

The implications of these results for clinicians are the same as those described in Study 1

(Chapter 2, page 103). Also, the successful replication of Study 1 weakens the argument of

those detractors who suggest that kMMT is nonsense and lump it in together with fortune-

telling, divination, dowsing and other unsubstantiated phenomenon.

3.5.5 Unanswered questions and future research

The results of this study, naturally lead to further queries. Firstly, since blinding the

Practitioners did not seem to influence kMMT accuracy, I wonder if the same outcomes can

be achieved if the TPs are blind to the verity of statements they were speaking. This question

is important to address in future research since in practice, during the course of a kMMT

© drannejensen 2014
 140

session, TPs are often kept in the dark about what target condition(s) the kMMT Practitioner

is assessing. Furthermore, in essence, many times it is nonconscious beliefs that are

purportedly being uncovered with kMMT.24 The concept of testing for nonconscious beliefs

at this point in time leads to many significant methodological conundrums, not to mention the

ethical considerations. However, researchers may wish to explore this avenue in the future.

Since these two studies failed to detect any factors that consistently affect kMMT

performance, future research may also want to delve into other possible influencers. It may be

necessary to employ non-subjective (i.e. not self-report, and more objective) measures of

such qualities as confidence, trust, rapport or other characteristics of the doctor-patient dyad.

Following on from this, I noticed that the ranges of kMMT accuracy scores over both studies

were relatively wide. In Study 1, kMMT accuracies ranged from 0.400 to 0.917, a width of

over 50 points, and similarly, the range in Study 2 was 0.425 to 0.825, a 40-point difference.

Prospective researchers may want to study further those practitioners that achieved high

scores (e.g. >0.80) and compare them to those that scored low (e.g. <0.50). In addition, if

commonalities are noticed in high-scoring practitioners, it would be interesting to attempt to

teach these to low-scoring practitioners to see if their scores could improve.

Another curious outcome of this research is that Practitioner anxiety did not affect kMMT

accuracy. Future studies may want to use a more standard measure of anxiety, like the State-

Trait Anxiety Inventory (STAI), with both the Practitioners and the TPs. I suggest also to

measure TPs’ anxiety because a common, yet unsubstantiated, explanation of kMMT is that

lies (a form of stress147) cause the body to react by weakening muscles, such as the indicator

muscle. Therefore, it would be interesting to assess other psychophysiological measures of

© drannejensen 2014
 141

stress or anxietyiii of TPs during the testing (e.g. heart rate, heart rate variability, skin

conductance, even EEG or fMRI), and correlate them with the speaking of lies during

kMMT. Moreover, if a stress response caused by lying results in diminished muscular

strength, then the same result may be obtained if other tests of muscular strength are

employed. Of special interest may be those tests of muscular strength that are objective, and

free of an assessor’s subjectivity, such as grip-strength testing.

Finally, it also would be interesting if future research would vary other factors which I kept

constant in these 2 studies. For instance, perhaps if the valence of stimuli were modified to be

emotionally arousing (negatively or positively) this might evoke a stronger stress response,

and if stress is indeed a governing factor, kMMT accuracy might improve (as might Intuition

accuracy as well) if more stress was induced.

3.6 Summary

This study successfully replicated Study 1 by again finding that kMMT can be used with

significant accuracy to distinguish lies from truths, compared to both Intuition and Chance.

Moreover, this study further supports the concept that a simple yet robust methodology for

assessing the value of kMMT as a diagnostic tool can be developed and implemented

effectively. Comparable to Study 1, no factors were identified that seemed to consistently

influence kMMT accuracy. The strengths of this study were also parallel to Study 1 : (1) a

high degree of blinding, (2) the heterogeneity of the samples, (3) the choice of a clear target

condition, and (4) the choice of a “gold standard” as the reference standard. Finally, also

iii
While stress and anxiety are distinct conditions in psychological terms, they may be related. For instance,
experientially it is difficult to feel anxiety without some element of psychological stress. [Cox RH. Sport
psychology: Concepts and applications. 6th ed. New York: McGraw-Hill; 2006.]

© drannejensen 2014
 142

similar to Study 1, the main limitation of this study is its lack of generalisability to other

applications of kMMT.

© drannejensen 2014
 143

3.7 Chapter 3 – List of Tables and Figures

3.7.1 Tables

TABLE 3.1 – Participant enrolment criteria – Study 2.

TABLE 3.2 – Demographics of Practitioners.

TABLE 3.3 – Diagnostic accuracy – as sensitivity, specificity, positive predictive value, and
negative predictive value (n=20 Pairs) (A) For kMMT; (B) For Intuition

TABLE 3.4 – Correlations (r) with p-values among kMMT accuracies by Block.

TABLE 3.5 – The influence on various Practitioner categorical factors on kMMT accuracy.
(1) Profession, (2) practising status, and (3) self-ranked kMMT expertise.

TABLE 3.6 – The influence on kMMT technique system on kMMT accuracy.

TABLE 3.7 – The influence on fixed properties of participants on kMMT accuracy. (1)
Practitioner’s gender, (2) Test Patient’s gender, and (3) sameness of gender.

TABLE 3.8 – The influence on various Pair categorical variables on kMMT accuracy. Choice
of Practitioner’s and Test Patient’s arms.

TABLE 3.9 – The influence of various categorical TP characteristics on kMMT accuracy. (1)
The Test Patient guessing the paradigm, (2) The Test Patient wearing glasses during testing,
(3) TP experience with kMMT, and (4) TP knew Practitioner.

TABLE 3.10 – Correlations (r) among kMMT accuracy and Practitioner demographics. p(2-
tailed)<0.05

TABLE 3.11 – The change in confidence ratings and correlations (r; with p-values) among
kMMT accuracy and participant confidence scores. p(2-tailed)<0.05

3.7.2 Figures

FIGURE 3.1 – Testing scenario layout: The Test Patient (red) viewed a monitor which the
Practitioner could see, had an ear piece in his ear through which he received instructions.
After the muscle test, the Practitioner (blue) entered his results on a keyboard.

FIGURE 3.2 – Participant Flow Diagram : Study 2

FIGURE 3.3 – ROC Curve for kMMT. Those dots above the red line represents better
performances

© drannejensen 2014
 144

FIGURE 3.4 – Mean kMMT accuracies by Block.

FIGURE 3.5 – Scatterplots for correlation.

FIGURE 3.6 – Correlation among kMMT accuracy and Practitioner’s subjective anxiety
rating

© drannejensen 2014
 145

CHAPTER 4
Study 3 – Grip Strength Dynamometry for Lie Detection

“The most accurate diagnosing tool you can have is in your office—YOUR PATIENT, with

Innate intelligence the body language combined with muscle testing.”

George J. Goodheart, Jr. (Founder, Applied Kinesiology)

© drannejensen 2014
 146

CHAPTER 4 : STUDY 3 – GRIP STRENGTH DYNAMOMETRY FOR

LIE DETECTION

4.1 ABSTRACT

Research Objectives: To investigate if dynamometric muscle testing (DMT) could be used

to distinguish Lies from Truth.

Methods: A prospective study of diagnostic test accuracy was carried out. Twenty Test

Patients (TPs), aged 18-65 years, with fully functioning and painfree hands, were recruited.

After viewing a picture on a computer screen, TPs were instructed to speak a specific

statement about the picture and then squeeze a dynamometer for 5 seconds, giving a

maximum effort each time. The examiner recorded the grip strength (to the nearest 1 kg)

directly into the computer, which advanced the screen to the next picture/statement. Testing

proceeded in this manner until 20 DMTs were performed, 10 by each hand.

Results: The mean grip strength after True statements was found to be 24.9 kg (95% CI 20.3

to 29.6), and after False statements, 24.8 (95% CI 20.2 to 29.5), which were not statistically

different (p=0.61). No significant correlations were detected between difference in grip

strength (False – True) and age, gender, confidence in MMT (pre-testing or post-testing), or

change in confidence scores. Also compared were mean grip strengths by block and were

found to be stable throughout testing

Summary: DMT via hand-held grip strength dynamometry failed to distinguish Lies from

Truth. These results seem to suggest that strength, as measured by DMT, is not impacted by

deceit. However, some other yet undetermined quality may allow kMMT to accurately make

© drannejensen 2014
 147

this distinction unlike DMT. A limitation of this study is it is not generalisable to other

applications of muscle testing or other target conditions.

Keywords: sensitivity; specificity; kinesiology; muscle weakness; muscle contraction; lie

detection; deception; lying; grip strength; dynamometry.

© drannejensen 2014
 148

4.2 Introduction

The previous two studies in this series have found that kinesiology-style manual muscle

testing (kMMT) is better than Chance or Intuition at distinguishing Lies from Truth. One

possible explanation of these results is that lying, a known stress, causes specific

physiological changes,148, 149 150 and muscle weakening may be one such change. If this

hypothesis is true, then other types of muscle strength testing might also be useful at

distinguishing lies from truth.

It is widely thought that kMMT lacks sufficient subjectivity to be a valid test.151 If practising

clinicians are to have confidence in kMMT, it must be compared to and shown to agree with

a more objective measure of muscle strength. A quantitative, instrumented alternative to

manual muscle testing (MMT), hand-held grip strength dynamometry (see Figure 4.1) has

been shown to be an accurate, reliable and objective test of muscle strength, which correlates

well with other forms of MMT.56, 152-154 Dynamometric muscle testing (DMT) quantifies

MMT by recording the peak force generated by a muscle or a group of muscles when loaded

in tension or compression.56

Previous studies have attempted to use grip strength dynamometry to detect conditions other

than musculoskeletal. Radin successfully used grip strength to distinguish refined sugar

(purportedly, a toxic substance) from sand (hypothetically inert or nontoxic).155 However,

three replication studies that followed failed to support his findings.156-158

Notwithstanding these discouraging results, it is speculated that DMT might be a useful

alternative to kMMT for detecting deceit. With the intent of enhancing objectivity, the aim of

this study was to investigate if DMT could be used to distinguish Lies from Truth.

© drannejensen 2014
 149

4.3 Methods

Reported here is a prospective study of diagnostic test accuracy. No participant was assessed

prior to enrolment. This protocol received ethics committee approval by the Oxford Tropical

Research Ethics Committee (OxTREC; Approval #41-10) and the Parker University

Institutional Review Board for Human Subjects (Approval # R19_10). Also, this study

protocol was registered with two clinical trials registries: the Australian New Zealand

Clinical Trials Registry (ANZCTR; www.anzctr.org.au), and US-based ClinicalTrials.gov.

Written informed consent was obtained from all participants, and all other tenets of the
FIGURE 4.1 – Grip strength dynamometer. (A) Example of a grip
strength
Declaration of Helsinki dynamometer;
were (B)paper
upheld. This Face of
wasa grip strength
written dynamometer.
in accordance with the

Standards for the Reporting of Diagnostic Test Accuracy Studies (STARD) guidelines (see

Appendix D, page 374, for the STARD Checklist).47, 66, 99

© drannejensen 2014
 150

4.3.1 Participants and Setting

Unlike the other studies in this series, this study recruited only one group of participants: Test

Patients (TPs; n=20), who were aged 18-65 years, had fully functioning and painfree hands,

and were fluent in English. Volunteers were excluded if they had visual, auditory or speech

impairment. In addition, kMMT-naïve as well as non-kMMT-naïve TPs were enrolled.

Recruitment was by direct contact, social media and word of mouth. All recruitment took

place in America, in the states of Texas and New York.

Each participant was given a Participant Information Sheet (PIS) and gave written informed

consent. They also completed the same pre- and post-testing questionnaires used in Studies 1

and 2, omitting questions referring to a Practitioner (See Appendix A, page 245). In the post-

testing questionnaire, participants were asked if they noticed anything different in their tests

following True statements compared to False statements. This question was included to

ascertain if they guessed the aim of the study, which was to investigate if grip strength can be

used to distinguish Lies from Truth. In this study, non-kMMT-naïve TPs were enrolled in

addition to kMMT-naïve TPs. And it was likely that those with prior kMMT experience were

aware of the paradigm that kMMT following False statements resulted in a “weak” outcome,

and kMMT following a True statement resulted in a “strong” outcome. Therefore, it was

necessary to track on both the kMMT-naivety of the TPs and if they noticed a difference or

guessed the paradigm.

4.3.2 Test Methods

The target condition (i.e. deceit) and reference standard (i.e. the verity of the spoken

statement) remained consistent with Studies 1 and 2. However, the index test used to detect

deceit was hand-held grip strength DMT. Each participant performed 20 DMTs after

© drannejensen 2014
 151

speaking an instructed statement out loud, 10 with their dominant hand and 10 with their non-

dominant hand, broken up into blocks of five: 5 dominant, 5 non-dominant, 5 dominant, 5

non-dominant.

The stimuli presented were selected from the same database of 100 affect-neutral

pictures/statements used in Studies 1 and 2. DirectRT Research Software (Empirisoft

Corporation, New York, NY) was programmed to randomly present a unique sequence of

stimuli for each participant, while randomizing the verity of the statements (i.e. True or

False), and keeping the prevalence of False statements constant at 0.50.

4.3.2.1 Grip Strength Dynamometry

All DMT was performed using the same factory calibrated hydraulic JAMAR (Model

J00105, Lafayette, Indiana, USA) analogue hand-grip dynamometer, found to be an

accurate159 and reliable160 measure of grip strength. In addition, DMT correlates well with

other forms of muscle testing,56 and its intra-subject test–retest variability has been found to

be small.161 TPs were instructed to squeeze the dynamometer for 5 seconds, giving a

maximum effort each time. They could rest as needed. The examiner read the kilograms scale

on the dial face, and after recording the result, reset the peak-hold needle to zero, ready for

the next effort. Grip strength was measured to the nearest 1kg.

4.3.2.2 Procedures

The TP was seated comfortably in front of a computer and held the dynamometer vertically

in his hand, elbow at his side and bent to 90 degrees, and forearm and wrist in neutral (i.e.

palm facing medially; see Figure 4.2). The investigator (AJ) was seated in front and to the

side of the TP, positioned to that she could read the dial of the dynamometer (see Figure

4.1.B), which was facing away from the TP, and also not see the computer screen. For the

© drannejensen 2014
 152

testing scenario layout, see Figure 4.3. One repetition of DMT was similar to one repetition

of kMMT in previous studies: (1) TP viewed a picture, (2) TP was instructed (via an

earpiece) what to say in relation to the picture, (3) TP took the DMT position, (4) while

viewing the picture, TP spoke the instructed statement, (5) TP immediately performed the

DMT, and (6) the examiner recorded the grip strength result directly into the computer, which

advanced the screen to the next picture/statement. Testing proceeded in this manner until 2

blocks of 5 DMTs were performed by each hand.

FIGURE 4.2 - DMT testing position example. (A) Elbow flexed to approximately 90o,
(B) Elbow at side, gauge facing away from TP, toward assessor.

© drannejensen 2014
 153

FIGURE 4.3 – Testing scenario layout.

4.3.3 Statistical Methods

Mean grip strengths following False statements and True statements were calculated for each

TP and are reported with their 95% confidence intervals. Then these means were compared

using a paired t-test. The same analyses were made stratified by gender, stratified by

dominant and non-dominant hand, stratified by kMMT-naivety and stratified by reporting

having guessed the paradigm. In addition, the mean difference in grip strengths (MeanTrue –

MeanFalse) was calculated and compared to 0.0 (i.e. no difference) using a paired t-test. Then,

a comparison of group means was conducted with linear regression models using as

covariates the participant characteristics of age and self-reported confidence scores. Finally,

© drannejensen 2014
 154

correlation analyses were made between mean Grip Strengths and other participant

characteristics.

Statistical advice was sought during the design phase, after piloting, and before data analysis.

All data were analyzed using Stata/IC 12.1 (StataCorp LP, College Station, Texas),

specifically the commands “ttest” and “pwcorr.”

4.4 Results

4.4.1 Participants

Twenty TPs were enrolled between June and August 2011: 11 males and 9 females. The

mean (SD) age was 48.4 (12.1) years. Seventeen reported being right-hand dominant and 3

left-hand dominant, and 14 reported being kMMT-naïve and 6 reported having had some

prior experience with kMMT. For a summary of participant demographics, see Table 4.1.

Also, see Figure 4.4 for the Participant Flow Diagram.

TABLE 4.1 – Demographics of participants.

© drannejensen 2014
 155

FIGURE 4.4 - Participant Flow Diagram : Study 3 (Grip Strength).

4.4.2 Test Results

Participants took between 5 and 15 minutes to complete their participation, all completed all

DMT in full and there were no adverse events reported from any testing. Histograms of grip

strength scores showed that the data are normally distributed (see Appendix Figure B.4.1), so

parametric statistics have been applied.

© drannejensen 2014
 156

The mean grip strength after True statements was found to be 24.9 kg (95% CI 20.3 to 29.6),

and after False statements, 24.8 (95% CI 20.2 to 29.5), which was not statistically different

(p=0.61). Also calculated for each participant was the difference in mean grip strengths

between True and False statements (Difference = TrueMean Grip Strength – FalseMean Grip Strength),

and the mean difference was found to be 0.1 (95% CI -0.4 to 0.6), which was no different

from 0.0 (p=0.61). See Table 4.2.A and Appendix Table B.4.1.A.

I also looked at the dominant and the non-dominant hands independently. Using the data for

the dominant hand only, the mean grip strength after True statements was found to be 23.9 kg

(95% CI 19.3 to 28.5), and after False statements, 23.5 (95% CI 18.9 to 28.2), which was not

statistically different (p=0.21). For the non-dominant hand only, the mean grip strength after

True statements was found to be 26.0 kg (95% CI 21.3 to 30.7), and after False statements,

26.1 (95% CI 21.2 to 31.0), which was not statistically different (p=0.81). See Table 4.2.B

and Appendix Table B.4.1.B.

Next I looked at different groups separately: Males vs. Females, kMMT-naïve vs. non-

kMMT-naïve, and those TPs who reported guessing the paradigm vs. those who did not. For

males (n=11), the mean grip strengths after True and False statements were identical: both

31.3 kg (95% CI for False statements, 26.3 to 36.3, and for True statements, 26.2 to 36.4),

and therefore, not statistically different (p=0.98). Whereas for females (n=9), the mean grip

strength after False statements was 17.2 kg (95% CI 21.3 to 30.7), and for True statements

16.9 kg (95% CI 12.4 to 21.4), which were not statistically different (p=0.43). For the

kMMT-naïve subgroup (n=14), the

© drannejensen 2014
 157

TABLE 4.2 – Comparison of mean grip strengths (kg) for False vs. True
statements. (A) Combined data for both hands, and the mean difference, (B)
Dominant and non-dominant hands separately.

mean grip strength after False statements was 25.9 kg (95% CI 19.6 to 32.2), and for True

statements 25.8 kg (95% CI 19.4 to 32.2), while for the non-kMMT-naïve subgroup (n=6),

the mean grip strength after False statements was 22.3 kg (95% CI 14.7 to 29.9), and for True

statements 22.9 kg (95% CI 16.1 to 29.8), which were both not statistically different (p=0.74

and p=0.11, respectively). For those who did not report guessing the paradigm (n=12), the

mean grip strength after False statements was 26.6 kg (95% CI 20.1 to 33.0) and for True

statements 26.1 kg (95% CI 19.6 to 32.6), which were not statistically different (p=0.09).

However, for those who reported guessing the paradigm (n=8), the mean grip strength

difference reached significance (p=0.02): 22.2 kg (95% CI 14.1 to 30.3) after False

statements and 23.2 kg (95% CI 15.1 to 31.3) after True statements. See Table 4.3.

Also compared were mean grip strengths by block which were found to be consistent

throughout testing (see Figure 4.5).Finally, no significant correlations were detected between

© drannejensen 2014
 158

difference in grip strength (False – True) and age, gender, confidence in kMMT (pre-testing

or post-testing), or change in confidence scores (see Table 4.4).

© drannejensen 2014
 TABLE 4.3 – Comparison of mean grip strengths (kg) for False vs. True statements. By gender; (B) By naivety to kMMT; and
(C) By reported guessing the paradigm.

© drannejensen 2014
159
 160

TABLE 4.4 – Correlations among grip strengths and other participant characteristics.

FIGURE 4.5 – Mean grip strengths by Block. (A) Dominant hand, (B) Non-dominant hand.

© drannejensen 2014
 161

FIGURE 4.5 (con’t.)

4.5 Discussion

4.5.1 Statement of Principal Findings

Unlike previous studies in this series, the current study failed to demonstrate that deceit can

be detected by muscle testing, in this case DMT using grip strength dynamometry. This

seems to imply that the kMMT practitioner is an integral part of the practitioner-patient

complex, and cannot be removed. Another interesting finding is that the mean grip strengths

found in this study were lower that accepted normative values.162, 163

Comparing these DMT results to those of kMMT may be inappropriate for a number of

reasons. First of all, the DMT in this study tested the participants’ grip strength, whereas the

© drannejensen 2014
 162

prior studies in this series used kMMT on deltoid muscles, and it has not been established

that the results of the kMMT studies can be generalised to testing muscles other than the

deltoid. Secondly, while both types of muscle testing (DMT and kMMT) use isometrici

muscle contractions (at least initially), DMT uses a patient’s maximum effort, whereas

kMMT uses submaximal force.164

4.5.2 Possible Explanations of Results

One might be tempted to attribute the lack of efficacy found in this study to a lack of

sensitivity of the chosen testing instrument. However, this explanation is unlikely because

previous research found that DMT was actually more discriminating than MMT in

identifying small differences in muscle strength.56, 161, 165 Another plausible explanation of

these results may be that while DMT is measuring strength, deceit does not cause changes in

strength, per se, but in some other quality (or qualities) perceptible by kMMT but not by

DMT. More specifically, the strength of a muscle is the degree of force it can exert, and is

dependent upon the size of the muscle and its nerve supply, but not dependent upon time or

displacement.166 However, time and displacement (or velocity) seem to be important factors

in kMMT. 26, 27, 167 Therefore, strength (as assessed through DMT) may have been an

incongruous index test to compare to kMMT, and perhaps another measure, such as power,

might be more appropriate. A third explanation of these results might be that the practitioner

is an integral component of the muscle testing dynamic, and therefore, if s/he is removed, the

accuracy of the muscle test used to detect deceit becomes insignificant.

i
isometric contraction: muscular contraction not accompanied by movement of the joint. (Mosby's Medical
Dictionary, 8th edition. Oxford, UK: Elsevier, 2009.)

© drannejensen 2014
 163

Numerous internal and external factors exist that may impact a person’s ability to contract

maximally, such as fatigue, pain, volition, motivation and even time of day.168 It is unclear

why the participants in this study scored generally lower than the standard reference values;

however, it may be because of their high mean age (48.4 years, SD 12.1 years), or that all

data collection occurred in America, or that this sample was particularly unhealthy. Because

no information was collected on health status, fatigue, time of day or motivation level, it is

difficult to speculate as to the cause of this marked difference.

Furthermore, since in this study grip strengths were found to be block-wise stable throughout

testing, it is unlikely that learning, fatigue or other internal/external factors played a

significant role. Lastly, the DMT in this study was patient-initiated while kMMT is usually

tester-initiated. This may have had an influence as it seems that there are fundamental

differences between the two.169, 170

4.5.3 Strengths and Limitations

A strength of this study is its rigorous design which was kept consistent with other studies in

this series. In addition, since one examiner (AJ) performed all assessment, adherence was

high. A more explicit strength was the duration of participation was suitable: given that

maximum effort tests like DMT are limited by patient fatigue, and since the results show that

fatigue was not influencing factor. In addition, the inclusion of participants with and without

prior kMMT experience was a strength. While there was a significant difference between

those participants who reported guessing the paradigm and those who did not, there was no

significant difference found between kMMT-naïve and non-kMMT-naïve TPs. One limitation

of this study is that its results are not generalisable to other applications of DMT, or to any

MMT, or to other target conditions.

© drannejensen 2014
 164

4.5.4 Implications for Clinical Practice

Because these results failed to confirm our null hypothesis that DMT may be useful for

distinguishing Lies from Truth, the clinical implications of these results are limited: DMT is

not useful for detecting deceit.

4.5.5 Unanswered Questions and Future Research

Future research may want to use other quantifiable measures to compare to kMMT, such as

muscular power. Furthermore, the use of a digital / computerised dynamometer might give

additional information about force, time and displacement, which our analog dynamometer

could not.

4.6 Summary

DMT via hand-held grip strength dynamometry failed to distinguish Lies from Truth. One

explanation of this might be that strength, as measured by DMT, is not impacted by deceit.

For instance, perhaps it is not strength, but some other yet undetermined quality, that allows

kMMT to accurately make this distinction but not DMT. A limitation of this study is it is not

generalisable to other applications of muscle testing or other target conditions.

© drannejensen 2014
 165

4.7 Chapter 4 – List of Tables and Figures

4.7.1 Tables

TABLE 4.1 – Demographics of participants.

TABLE 4.2 – Comparison of mean grip strengths (kg) for False vs. True statements. (A)
Combined data for both hands, and the mean difference, (B) Dominant and non-dominant
hands separately.

TABLE 4.3 – Comparison of mean grip strengths (kg) for False vs. True statements. (A) By
gender; (B) By naivety to kMMT; and (C) By reported guessing the paradigm.

TABLE 4.4 – Correlations among grip strengths and other participant characteristics.

4.7.2 Figures

FIGURE 4.1 – Grip strength dynamometer. (A) Example of a grip strength dynamometer; (B)
Face of a grip strength dynamometer.

FIGURE 4.2 – DMT testing position example. (A) Elbow flexed to approximately 90o, (B)
Elbow at side, gauge facing away from TP, toward assessor.

FIGURE 4.3 – Testing scenario layout.

FIGURE 4.4 – Participant Flow Diagram : Study 3 (Grip Strength).

FIGURE 4.5 – Mean grip strengths by Block : (A) Dominant hand, (B) Non-dominant hand

© drannejensen 2014
 166

CHAPTER 5
Study 4 – Exploring the Variation in kMMT Accuracy
through Repeatability and Reproducibility

“The only relevant test of the validity of a hypothesis

is comparison of prediction with experience.”

Milton Friedman

© drannejensen 2014
 167

CHAPTER 5 : STUDY 4 – EXPLORING THE VARIATION IN KMMT

ACCURACY THROUGH REPEATABILITY AND REPRODUCIBILITY

5.1 ABSTRACT

Research Objectives: To explore the variation in mean kMMT accuracy and whether this

variation can be attributable to participant characteristics.

Methods: A prospective study of diagnostic test accuracy was carried out in a round-robin

fashion, similar in methodology to Study 2 (see page 111). Sixteen Practitioners tested each

of 7 Test Patients using 20 kMMTs broken into 2 blocks of 10 which alternated with 2 blocks

of 10 Intuitions. Mean kMMT accuracies (as overall percent correct) were calculated for each

unique pair. Reproducibility and repeatability was assessed using analyses of variance

(ANOVA) and scatter and Bland-Altman plots.

Results: The mean kMMT accuracy (95% CI) was 0.616 (0.578 - 0.654), which was

significantly different from both the mean Intuition accuracy, 0.507 (95% CI 0.484 - 0.530;

p<0.01) and Chance (p<0.01). Visual inspection of scatterplots of mean kMMT accuracies by

Practitioner and by TP suggest large variances among both subsets, and regression analysis

revealed that kMMT accuracy could not be predicted by TP (r= ‒0.14; p=0.19), nor by

Practitioner (r=0.01; p=0.90). A significant effect imposed by both Practitioners and TPs

individually and together was found at the p<0.05 level; however, together they account for

only 57.0% of the variance, with 43.0% of the variance unexplained by this model. From a a

statistical perspective, Bland-Altman Plots of mean kMMT accuracy by Practitioner do show

adequate repeatability since all scores fell within 2 SDs of the mean; however, the wide range

of scores also suggests insufficient repeatability from a clinical perspective. Finally,ANOVA

© drannejensen 2014
 168

demonstrated that an insignificant amount of variance could be explained by Block [F(1,21) =

0.02, p = 0.90].

Summary: The variation in the mean kMMT accuracy can only be explained 57% by

participant characteristics; therefore, there are other factors at play that could not be

explained by the model used. Additional research is needed to explain this variance.

Keywords: variability; stability; precision; reproducibility; repeatability; reliability; validity;

intra-examiner; inter-examiner; kinesiology; muscle weakness; lie detection; deception;

lying.

© drannejensen 2014
 169

5.2 Introduction

According to Bossuyt, the first question to ask in the evaluation a new diagnostic test is

“Does it measure what it is supposed to measure?” – otherwise known as its analytic

validity.73 Moreover, the first step in assessing a test’s analytic validity is to estimate its

accuracy, which can be measured as sensitivity and specificity, overall fraction correct,

positive predictive value and negative predictive value and others. The previous studies in

this series were aimed at exactly that: the estimation of the accuracy of kMMT to distinguish

Lies from Truth. However, a diagnostic test is only considered valid if it is

both accurate and precise; therefore, due to the wide range of kMMT accuracies found in

previous studies, assessing the precision of kMMT is an important next step.

Precision can be defined as “the degree to which repeated measurements under unchanged

conditions show the same results.”171 Just as there are numerous ways to quantify the

accuracy of a diagnostic test,73 there are several terms currently used to describe its precision,

such as reproducibility, repeatability, reliability (inter-tester and intra-tester), and stability. In

addition, some medical statisticians look at a test’s confidence intervals when determining its

precision.78 However, it is most common to gauge the precision of a test in terms of its

reproducibility and repeatability.48 Unfortunately, these two terms are also frequently

confused. For clarity, they are defined as:

Reproducibility : the variability of the average values obtained by several observers

while measuring the same item (interobserver variability).48

Repeatability : the variability of the measurements obtained by one person while

measuring the same item repeatedly (intraobserver variability).48

© drannejensen 2014
 170

Applying these terms to the context of kMMT, reproducibility, then, may be described as the

degree of variability in kMMT accuracy between different Practitioners testing the same TPs,

and repeatability may be described as the degree of variability in kMMT accuracy when a

Practitioner tests the same TP at different times.

Hence, the aim of this study was to assess the precision of kMMT used for distinguishing

Lies from Truth, so that that both reproducibility and repeatability could be evaluated. More

specifically, my research questions for this study became: (1) Is the kMMT accuracy that a

Practitioner’s achieves with one TP consistent over many TPs, or is it TP- or pair-specific?;

and (2) Is the kMMT accuracy obtained with one TP consistent over many Practitioners, or is

it Practitioner- or pair-specific?

5.3 Methods

This study is a prospective study of diagnostic test accuracy in a round-robin format. No

participant was assessed prior to enrolment. This protocol received ethics committee approval

by the Oxford Tropical Research Ethics Committee (OxTREC; Approval #41-10) and the

Parker University Institutional Review Board for Human Subjects (Approval # R16_10).

Also, this study protocol was registered with two clinical trials registries: the Australian New

Zealand Clinical Trials Registry (ANZCTR; www.anzctr.org.au), and US-based

ClinicalTrials.gov. Written informed consent was obtained from all participants, and all other

tenets of the Declaration of Helsinki were upheld. This paper was written in accordance with

the Standards for the Reporting of Diagnostic Test Accuracy Studies (STARD) guidelines

(see Appendix D, page 376, for the STARD Checklist).47, 66, 99

© drannejensen 2014
 171

5.3.1 Participants and Setting

Volunteer Practitioners were solicited from a group of muscle testing practitioners attending a

seminar in Dallas, Texas, in June 2012. Regarding the Practitioners, all recruitment for

participation, all enrolment and all data collection were done during the course of one

afternoon and evening. In addition, seven (7) TPs were recruited and enrolled in the few days

leading up to the event from a convenience sample of a mixture of kMMT-naïve and non-

kMMT-naïve individuals. Similar to previous studies in this series (see Chapter 2, page 52;

Chapter 3, page 116; and Chapter 4, page 150), participants were sought who were aged 18-

65 years, had fully functioning and painfree upper extremities, and were fluent in English.

Volunteers were excluded if they were markedly hearing-, sight- or speech-impaired.

Recruitment was by direct contact and word of mouth.

Once enrolled, Practitioners waited in a “holding room” where they were given a Participant

Information Sheet (PIS) and completed a written informed consent form. They also

completed similar Pre- and Post-Testing Questionnaires as were used in Studies 1 and 2 (see

Appendix A, page 245). Once enrolled, the TPs waited in the testing room (i.e. a different

room to the Practitioners), where they, too, were given a Participant Information Sheet (PIS)

and completed a written informed consent form and Pre-Testing Questionnaires similar to

what were used in Studies 1 and 2 (see Appendix A, page 245).

Aside from demographic information (e.g. gender, age, etc), the Pre-Testing Questionnaires

asked participants (both Practitioners and TPs) to rate various characteristics on a 10cm

Visual Analog Scale. For example, both Practitioners and TPs were asked to mark on the

VAS their level of confidence in kMMT in general, with the left of the VAS anchored with

“No Confidence Whatsoever (0%)” and on the right with “Complete Confidence (100%).”

© drannejensen 2014
 172

They were also asked to rate their current level of test anxiety on a similar scale, with the left

marked as “No Anxiety Whatsoever” and on the right with “Worst Anxiety Ever.” In

addition, before and after each round, participants were asked to rate other factors (e.g. how

well do they know this person, how much do they like this person, how much connection do

they feel with this person, etc.) on an ordinal scale of 0 to 10 (with 0 the lowest and 10 the

highest). To see the actual instruments employed, see Appendix A.5.

In the testing room, 7 complete testing stations were set up on 7 individual tables, evenly

spaced around the room. Each testing station had a computer loaded with the research

software, a keyboard, a mouse, an earpiece, and 2 chairs. The stations were configured in

such a way that the TP could only see his/her monitor and no other monitors, and so that the

Practitioner could not see the TP’s monitor. See Figure 5.1.

FIGURE 5.1 – Testing scenario layout: The Test Patient (red) viewed a
monitor which the Practitioner could see, had an ear piece in his ear through
which he received instructions. After the muscle test, the Practitioner (blue)
entered his results on a keyboard.

© drannejensen 2014
 173

5.3.2 Test Methods

Fundamentally, the methodology of this study followed the same basic structure as Study 2

(see page 115): The target condition was deceit, the reference standard was the actual verity

of the spoken statement, the primary index test used to detect deceit was kMMT, the

secondary index testing was Intuition (without using kMMT). Each Practitioner-TP pair

performed 20 kMMTs and 20 Intuits, broken up into 2 blocks of each: 10 kMMTs – 10

Intuits – 10 kMMTs – 10 Intuits. Each Practitioner tested all TPs and all TPs were tested by

all Practitioners. All Practitioners were blind to the verity of the TP statement.

The stimuli presented were selected from the same database of 100 affect-neutral

pictures/statements used in Studies 1 and 2. DirectRT Research Software (Empirisoft

Corporation, New York, NY) was programmed to randomly present a unique sequence of

visual and auditory stimuli for each TP, while randomizing the verity of the statements (i.e.

True or False), and keeping the prevalence of False statements constant at 0.50.

When testing began, the 7 TPs were seated at their respective tables (see Figure 5.2).

FIGURE 5.2 – Testing room flow.

© drannejensen 2014
 174

Practitioners entered the testing room and sat with a TP. Before starting, they each completed

a number of Pre-Test Ratings about each other, and the Practitioner could perform up to 5

practice kMMTs. Then the testing began. As soon as they finished each round, they

completed a number of Post-Test Ratings about that round of testing, and then the

Practitioner moved on to another testing station and another TP. Once a Practitioner

completed testing all 7 TPs, s/he completed a Post-Test Questionnaire (see Appendix A, page

245) and then was done with his/her participation. After each TP was tested by all 16

Practitioners, they completed a short Post-Test Questionnaire (again, see Appendix A, page

245) and were also done with their participation. See Figure 5.3 for a Participant Flow

Diagram.

© drannejensen 2014
 175

FIGURE 5.3 – Participant Flow Diagram : Study 4.

© drannejensen 2014
 176

5.3.3 Statistical Methods

5.3.3.1 Sample Size

Due to the nature of recruitment and the round-robin participation, a convenient sample of

volunteers from the seminar was enrolled as Practitioners. No additional sample size

calculation was performed. With 7 TPs confirmed as participating, it was hoped that at

minimum 7 Practitioners would also volunteer. With more than double that projected number,

I was confident that a meaningful analysis would result from the data collected.

5.3.3.2 Methods of Analysis

Because this is also a study of diagnostic test accuracy, for each Pair, I report error-based

measures of accuracy: overall fraction correct, sensitivity, specificity, PPV and NPV 73 – and

their 95% confidence intervals (95% CI). In addition, I report the mean measures grouped by

both the Practitioner and the TP. The same error-based measures will also be reported for

Intuition.

In regard to assessing reproducibility and repeatability, primarily graphical methods were

used for analysis, and since mean kMMT accuracies can be calculated grouping by either

Practitioners or TPs, scatterplots for each grouping are presented. Using scatterplots of

kMMT Accuracy vs. Practitioner and kMMT Accuracy vs. TP, reproducibility was assessed

visually by looking at the width of the range of mean kMMT accuracies, with smaller ranges

indicating better reproducibility. Repeatability was assessed visually by using scatterplots of

kMMT accuracies of Block 1 vs. kMMT accuracies of Block 2, and better repeatability was

suggested by plots close to a reference line with a slope of 1 (m=1). Repeatability was also

visually assessed using Bland-Altman plots of mean kMMT Accuracy vs. the Difference in

kMMT Accuracy (Block 2 – Block 1), the “bias,” and included in these plots were reference

© drannejensen 2014
 177

lines at the mean difference, and the mean difference  2SD (the limits of 95% agreement). It

is expected that 95% of differences between the two measurements will fall within 2SDs of

the mean.172 With Bland-Altman plots, good correlation (i.e. agreement, and therefore, good

repeatability) is a question of clinical judgment, not statistical. Nevertheless, the smaller the

range between these two limits (2SDs) the better the agreement, and the less the bias.

Further statistical analysis was carried out using an Analysis of Variance (ANOVA) to

determine how much of the variance in scores could be attributed to different models. One

model looked at the influence of Practitioners and TPs, and in another model, all participant

characteristics were included (i.e. age, gender, years of experience, confidence, willingness,

etc.).

Statistical advice was sought during the design phase and before data analysis. All data were

analyzed using Stata/IC 12.1 (StataCorp LP, College Station, Texas), specifically the

commands “twoway scatter” and “anova.”

The study of variance / agreement in clinical testing may seem simple in principle, but in

practice is extremely complex.173 In actuality, the data collected in this study is also

extremely complex – because it is not only nested (i.e. multiple tests by multiple Practitioners

testing multiple TPs), the primary outcome (i.e. mean kMMT accuracy) can be calculated and

averaged with respect to either the Practitioner or the TP. As a result, in-depth analyses of

this data would require complicated and elaborate statistical methods, the like of which would

be expected from a student writing a DPhil in statistics, but would be beyond the scope of a

student writing a DPhil in clinical research. Therefore, the statistical analyses presented in

this section reflect methods aligned with the level of knowledge expected by a clinical

researcher, not a statistician.

© drannejensen 2014
 178

5.4 Results

5.4.1 Participants

Sixteen Practitioners and 7 TPs were enrolled in late June 2012. Of the Practitioners, 8 were

male and 8 were female; 14 were chiropractors, 2 were acupuncturists; 9 were in full-time

practice, 6 were in part-time practice and 1 was not currently practising; their mean (SD) age

was 45.1 (12.4) years; and 15 reported being right-handed and 1, left-handed. Their mean

(SD) number of years in practice was 13.8 (10.0) years, their mean (SD) number of years of

using kMMT in practice was 12.5 (8.8) years, and their mean usual hours/day using kMMT

was 5.5 (3.4) hours. Seven Practitioners ranked their own kMMT expertise as “4,” seven

ranked their own kMMT expertise as “3,” and two ranked their own kMMT expertise as “1.”

No Practitioners ranked their own kMMT expertise as either “2” or “0.” The mean (SD) score

of self-ranked kMMT expertise was 3.2 (0.7) out of a possible 4 (0=”None” and 4=”Expert”).

As measured using the 10cm Visual Analog Scales (VAS), their mean (SD) degree of

confidence in own kMMT ability (pre-testing) was 8.4 (1.5), their mean (SD) degree of

confidence in kMMT in general (pre-testing) was 8.0 (1.6) and their median degree of test

anxiety (range) was 0.7 (0.0 to 5.0). For a summary of Practitioner demographics, see Table

5.1. Also, for the Participant Flow Diagram, see Figure 5.3.

One of the 7 TPs (originally called TP#5) failed to follow written and verbal instructions

which resulted in no data being collected by her computer; therefore, she was excluded from

all analyses. For convenience, the TP originally called TP#7 was renamed TP#5. Of the 6

remaining TPs, 3 were male and 3 were female; their mean (SD) age was 36.7 (15.0) years;

and all 6 reported being right-handed. As measured using 10cm VAS, their mean (SD; range)

degree of experience with kMMT ability was 5.3 (3.6; 0.2 to 9.9), their mean (SD) degree of

confidence in kMMT in general (pre-testing) was 8.0 (1.8) and their median degree of test

© drannejensen 2014
 179

anxiety (range) was 0.1 (0.0 to 3.5). Finally, 3 TPs reported guessing the paradigm, and 3 TPs

did not report guessing the paradigm.i

i
Paradigm: Lies resulted in a “weak” kMMT, Truth resulted in a “strong” kMMT.

© drannejensen 2014
 180

TABLE 5.1 – Demographics of Practitioners

© drannejensen 2014
 181

5.4.2 Test Results

Participants took between 10 and 20 minutes to complete one round of testing. The duration

of participation for Practitioners was 1 to 1 ¼ hours, and the duration of participation for the

TPs was approximately 3 hours, which including rounds of testing interspersed with short rest

periods of 5-10 minutes. There were no adverse events reported from any testing.

5.4.2.1 Accuracy

Accuracy scores were calculated for each Pair (n=96; see Appendix Table B.5.1), and mean

accuracy scores were also calculated for each Practitioner and each TP (n=16 and n=6

respectively; see Table 5.2). For kMMT, the mean overall fraction correct (95% CI) was

0.616 (0.578 - 0.654), the mean sensitivity (95% CI) was 0.595 (0.549 - 0.640), the mean

specificity (95% CI) was 0.638 (0.430 - 0.486), the mean Positive Predictive Value (PPV;

95% CI) was 0.632 (0.588 - 0.676), and the mean Negative Predictive Value (NPV; 95% CI)

was 0.609 (0.5673 - 0.652). For Intuition, the mean overall fraction correct (95% CI) was

0.507 (0.484 - 0.530), the mean sensitivity (95% CI) was 0.456 (0.424 - 0.487), the mean

specificity (95% CI) was 0.557 (0.527 - 0.588), the mean PPV (95% CI) was 0.502 (0.475 -

0.530), and the mean NPV (95% CI) was 0.514 (0.491 - 0.537). In all these 5 measures of

accuracy, kMMT accuracy was found to be significantly more than both Intuition accuracy

and Chance.ii See Table 5.3. Finally, there was no significant difference (p=0.91) in mean

kMMT accuracies between those Pairs containing a TP who reported guessing the paradigm

(mean 0.609, 95% CI 0.191 to 1.000) and those Pairs containing a TP who did not report

guessing the paradigm (mean 0.623, 95% CI 0.397 to 0.849).

ii
Chance here refers to the hypothetical situation where either outcome was equally likely: 50-50.

© drannejensen 2014
 182
TABLE 5.2- Mean Accuracy Data for each Practitioner and each TP individually: Acuracy (Overall Percent Correct), Sensitivity, Specificity, PPV and NPV; for kMMT and Intution.

Accuracy* Sensitivity Specificity PPV NPV

Mean 95% CI SD Range Mean 95% CI SD Mean 95% CI SD Mean 95% CI SD Mean 95% CI SD
kMMT
Practitioner 1 0.550 0.340 - 0.760 0.200 0.300 - 0.750 0.550 0.306 - 0.794 0.232 0.550 0.324 - 0.776 0.215 0.539 0.291 - 0.787 0.236 0.571 0.369 - 0.773 0.193
2 0.500 0.390 - 0.610 0.105 0.350 - 0.600 0.583 0.480 - 0.687 0.098 0.417 0.236 - 0.597 0.172 0.508 0.410 - 0.606 0.094 0.490 0.350 - 0.630 0.133
3 0.633 0.479 - 0.788 0.147 0.450 - 0.850 0.567 0.371 - 0.762 0.186 0.700 0.426 - 0.974 0.261 0.675 0.500 - 0.850 0.167 0.603 0.451 - 0.756 0.145
4 0.617 0.433 - 0.800 0.175 0.400 - 0.800 0.744 0.640 - 0.849 0.100 0.505 0.168 - 0.841 0.320 0.618 0.425 - 0.812 0.184 0.619 0.379 - 0.859 0.228
5 0.600 0.395 - 0.805 0.195 0.350 - 0.900 0.683 0.503 - 0.864 0.172 0.517 0.232 - 0.802 0.271 0.609 0.407 - 0.810 0.192 0.603 0.356 - 0.851 0.236
6 0.717 0.562 - 0.871 0.147 0.600 - 0.950 0.644 0.399 - 0.889 0.233 0.789 0.644 - 0.935 0.139 0.743 0.586 - 0.900 0.150 0.713 0.554 - 0.872 0.152
7 0.658 0.464 - 0.853 0.186 0.350 - 0.850 0.583 0.391 - 0.776 0.183 0.733 0.497 - 0.970 0.225 0.698 0.495 - 0.901 0.193 0.632 0.443 - 0.821 0.180
8 0.792 0.568 - 1.000 0.213 0.400 - 1.000 0.900 0.785 - 1.000 0.110 0.683 0.300 - 1.000 0.366 0.781 0.573 - 0.989 0.198 0.766 0.358 - 1.000 0.388
9 0.550 0.409 - 0.691 0.134 0.450 - 0.800 0.513 0.275 - 0.750 0.226 0.592 0.367 - 0.818 0.215 0.546 0.370 - 0.722 0.168 0.560 0.417 - 0.703 0.136
10 0.525 0.353 - 0.697 0.164 0.250 - 0.700 0.424 0.345 - 0.503 0.075 0.621 0.358 - 0.884 0.250 0.556 0.347 - 0.765 0.199 0.508 0.342 - 0.675 0.159
11 0.592 0.376 - 0.808 0.206 0.300 - 0.800 0.704 0.527 - 0.880 0.168 0.483 0.119 - 0.848 0.347 0.604 0.395 - 0.814 0.200 0.578 0.270 - 0.886 0.294
12 0.733 0.559 - 0.908 0.166 0.500 - 0.950 0.644 0.461 - 0.828 0.175 0.818 0.636 - 1.000 0.173 0.780 0.564 - 0.996 0.206 0.705 0.552 - 0.858 0.146
13 0.708 0.497 - 0.919 0.201 0.400 - 0.900 0.579 0.229 - 0.930 0.334 0.828 0.692 - 0.963 0.129 0.708 0.396 - 1.000 0.297 0.718 0.505 - 0.932 0.204
14 0.658 0.403 - 0.914 0.244 0.350 - 1.000 0.646 0.394 - 0.898 0.240 0.673 0.358 - 0.987 0.300 0.675 0.392 - 0.958 0.270 0.649 0.382 - 0.916 0.254
15 0.517 0.302 - 0.731 0.204 0.150 - 0.750 0.417 0.124 - 0.709 0.279 0.617 0.295 - 0.938 0.306 0.582 0.247 - 0.916 0.318 0.515 0.327 - 0.704 0.180
16 0.508 0.351 - 0.665 0.150 0.350 - 0.700 0.329 0.173 - 0.485 0.149 0.678 0.510 - 0.846 0.160 0.491 0.250 - 0.732 0.230 0.518 0.392 - 0.645 0.121
TP 1 0.728 0.613 - 0.843 0.216 0.350 - 1.000 0.681 0.539 - 0.823 0.266 0.775 0.642 - 0.908 0.249 0.772 0.653 - 0.892 0.224 0.718 0.594 - 0.842 0.233
2 0.572 0.480 - 0.664 0.173 0.250 - 0.900 0.648 0.539 - 0.758 0.205 0.498 0.371 - 0.626 0.239 0.555 0.466 - 0.643 0.166 0.596 0.482 - 0.709 0.213
3 0.672 0.595 - 0.749 0.145 0.350 - 0.850 0.527 0.421 - 0.633 0.199 0.814 0.736 - 0.892 0.146 0.731 0.629 - 0.833 0.191 0.645 0.579 - 0.712 0.125
4 0.569 0.501 - 0.637 0.128 0.350 - 0.800 0.494 0.383 - 0.605 0.208 0.644 0.576 - 0.711 0.126 0.560 0.471 - 0.650 0.168 0.571 0.510 - 0.632 0.114
5 0.419 0.355 - 0.482 0.120 0.150 - 0.600 0.475 0.373 - 0.577 0.191 0.359 0.229 - 0.490 0.245 0.415 0.349 - 0.481 0.124 0.376 0.284 - 0.467 0.172
6 0.738 0.672 - 0.803 0.123 0.550 - 0.950 0.742 0.660 - 0.824 0.154 0.736 0.626 - 0.847 0.208 0.760 0.674 - 0.845 0.161 0.751 0.676 - 0.826 0.141
Intution
Practitioner 1 0.542 0.411 - 0.672 0.124 0.400 - 0.750 0.600 0.412 - 0.788 0.179 0.483 0.344 - 0.623 0.133 0.531 0.405 - 0.656 0.120 0.553 0.410 - 0.696 0.136
2 0.500 0.395 - 0.605 0.100 0.350 - 0.600 0.420 0.258 - 0.582 0.154 0.573 0.354 - 0.791 0.208 0.501 0.342 - 0.660 0.151 0.498 0.408 - 0.589 0.086
3 0.392 0.261 - 0.522 0.124 0.250 - 0.600 0.283 0.204 - 0.362 0.075 0.500 0.312 - 0.688 0.179 0.379 0.211 - 0.548 0.160 0.403 0.293 - 0.514 0.105
4 0.542 0.458 - 0.626 0.080 0.400 - 0.600 0.500 0.434 - 0.566 0.063 0.583 0.461 - 0.706 0.117 0.551 0.459 - 0.643 0.088 0.535 0.456 - 0.615 0.075
5 0.550 0.462 - 0.638 0.084 0.450 - 0.650 0.459 0.324 - 0.595 0.129 0.641 0.544 - 0.738 0.093 0.549 0.428 - 0.671 0.116 0.554 0.478 - 0.629 0.072
6 0.525 0.433 - 0.617 0.088 0.400 - 0.650 0.574 0.459 - 0.689 0.110 0.477 0.332 - 0.622 0.138 0.519 0.413 - 0.625 0.101 0.534 0.453 - 0.616 0.078
7 0.533 0.470 - 0.597 0.061 0.450 - 0.600 0.450 0.362 - 0.538 0.084 0.617 0.462 - 0.771 0.147 0.554 0.456 - 0.652 0.093 0.525 0.475 - 0.576 0.048
8 0.542 0.435 - 0.649 0.102 0.400 - 0.700 0.456 0.289 - 0.622 0.159 0.623 0.544 - 0.702 0.075 0.529 0.404 - 0.654 0.119 0.549 0.449 - 0.649 0.095
9 0.517 0.352 - 0.681 0.157 0.250 - 0.700 0.483 0.332 - 0.635 0.144 0.547 0.300 - 0.794 0.236 0.522 0.324 - 0.721 0.189 0.519 0.345 - 0.693 0.166
10 0.475 0.420 - 0.530 0.052 0.400 - 0.550 0.424 0.381 - 0.467 0.041 0.527 0.429 - 0.626 0.094 0.468 0.399 - 0.537 0.066 0.483 0.442 - 0.524 0.039
11 0.508 0.396 - 0.620 0.107 0.400 - 0.700 0.437 0.225 - 0.649 0.202 0.574 0.530 - 0.619 0.042 0.480 0.345 - 0.615 0.129 0.529 0.410 - 0.648 0.114
12 0.492 0.370 - 0.613 0.116 0.350 - 0.650 0.450 0.305 - 0.595 0.138 0.533 0.362 - 0.705 0.163 0.491 0.345 - 0.638 0.140 0.487 0.376 - 0.599 0.106
13 0.533 0.350 - 0.717 0.175 0.300 - 0.800 0.424 0.220 - 0.628 0.194 0.642 0.449 - 0.836 0.185 0.529 0.267 - 0.790 0.249 0.536 0.394 - 0.678 0.136
14 0.500 0.293 - 0.707 0.197 0.150 - 0.650 0.574 0.309 - 0.840 0.253 0.426 0.222 - 0.630 0.194 0.486 0.297 - 0.676 0.181 0.534 0.277 - 0.792 0.246
15 0.442 0.311 - 0.572 0.124 0.250 - 0.600 0.383 0.229 - 0.538 0.147 0.500 0.385 - 0.615 0.110 0.429 0.285 - 0.573 0.137 0.451 0.331 - 0.571 0.114
16 0.525 0.470 - 0.580 0.052 0.450 - 0.600 0.374 0.243 - 0.505 0.125 0.673 0.592 - 0.753 0.077 0.518 0.449 - 0.588 0.066 0.529 0.477 - 0.581 0.050
TP 1 0.503 0.451 - 0.555 0.097 0.350 - 0.700 0.453 0.392 - 0.514 0.115 0.553 0.480 - 0.627 0.138 0.506 0.438 - 0.574 0.128 0.504 0.459 - 0.550 0.085
2 0.469 0.394 - 0.543 0.140 0.150 - 0.700 0.432 0.336 - 0.528 0.180 0.501 0.407 - 0.595 0.177 0.453 0.377 - 0.530 0.144 0.478 0.400 - 0.556 0.146
3 0.494 0.435 - 0.552 0.109 0.300 - 0.650 0.428 0.349 - 0.507 0.148 0.560 0.472 - 0.648 0.165 0.495 0.424 - 0.566 0.133 0.497 0.442 - 0.552 0.103
4 0.547 0.488 - 0.606 0.110 0.400 - 0.800 0.492 0.416 - 0.567 0.141 0.600 0.522 - 0.678 0.147 0.547 0.470 - 0.624 0.145 0.547 0.497 - 0.596 0.093
5 0.475 0.417 - 0.533 0.110 0.250 - 0.600 0.419 0.333 - 0.504 0.160 0.531 0.454 - 0.608 0.145 0.468 0.393 - 0.544 0.142 0.479 0.424 - 0.534 0.103
6 0.556 0.502 - 0.610 0.101 0.400 - 0.750 0.512 0.414 - 0.610 0.184 0.599 0.531 - 0.666 0.127 0.545 0.484 - 0.606 0.114 0.578 0.515 - 0.641 0.118
kMMT, kinesiology-style manual muscle testing; *Accuracy as Overall Fraction Correct; PPV, Positive Predictive Value; NPV, Negative Predictive Value; SD, Standard Deviation, SE, Standard Error, CI, Confidence Interval; TP, Test Patient.

© drannejensen 2013
 TABLE 5.3 – Diagnostic accuracy of kMMT vs. Intuition: Means, 95% and significance. Accuracy (as overall fraction correct),
sensitivity, specificity, positive predictive value, and negative predictive value.

© drannejensen 2014
183
 184

5.4.2.2 Reproducibility

In the context of this study, reproducibility was considered to be the amount of variance of

the mean kMMT accuracy, and can be considered in terms of the Practitioner or in terms of

the TP (see Figure 5.4).

FIGURE 5.4 – Scatterplot of kMMT accuracy by Pair.

Visual inspection of scatterplots of mean kMMT accuracies by Practitioner and by TP

(Figure 5.5.A and 5.5.B, respectively) suggest large variances among both subsets. In

addition, the scatterplot by Practitioner (Figure 5.5.A) suggests that when the Pair contained

TP#1 or TP#6, kMMT accuracies were regularly among the highest of the sample, and with

TP#5, among the lowest. I realise that there may be difficulties in interpreting the correlation

coefficients between repeated measures;48, 174 nonetheless, regression analysis revealed that

kMMT accuracy could not be predicted by TP (r= ‒0.14; p=0.19), nor by Practitioner

(r=0.01; p=0.90).

© drannejensen 2014
 185

FIGURE 5.5.A – Reproducibility of kMMT accuracy by Practitioner: in order of mean accuracy.

Highest

Lowest
FIGURE 5.5.B – Reproducibility of kMMT accuracy by TP.

© drannejensen 2014
 186

In addition, analyses of variance (ANOVA) were performed. There was a significant effect

imposed by Practitioners and TPs individually and together on kMMT accuracy at the p<0.05

level (see Table 5.4 for details). However, together they account for only 57.0% of the

variance (Practitioners, 21.6%; TPs, 35.4%), with 43.0% of the variance unexplained by this

model. Another ANOVA was run to assess if Block had an influence on kMMT accuracy,

and no significant influence was detected [F(1,1) = 0.02, p = 0.90] (see Appendix Table

B.5.2).

TABLE 5.4 – ANOVA results : Practitioner and TP.

Source Partial SS df F Prob>F %
Model 1.8963 20 4.97 <0.01 57%

Practitioner 0.7196 15 2.51 <0.01 21.6%

TP 1.1767 5 12.33 <0.01 35.4%

Residual 1.4312 75 43.0%

Total 3.3275 95 100.0%

ANOVA was also used to attempt to explain the residual factors. Using all variables

collected, a significant effect for the model was demonstrated at the p<0.05 level [F(1,88) =

4.15, p = 0.03] (see Table 5.5.A). In this analysis, the model could account for 98.1% of the

variance, leaving less than 2% unexplained. However, after obtaining statistical advice about

analysing these ANOVAs, it was determined that there were too many variables in the model

for the 98.1% result to be meaningful; therefore I include Table 5.5.A for completeness, but

caution is advised about over-interpreting these results. Following on from these discussions,

further ANOVAs were run using a univariate general linear model of the same variables, to

investigate which participant characteristics may have had an influence on kMMT accuracy

(see Table 5.5.B). No Practitioner characteristics’ influence reached significance, unlike 3

TPs characteristics: (1) TP age, (2) TP anxiety, and (3) Confidence in Practitioner's kMMT

Ability, with only TP age having a significant correlation, although negative (r = ‒0.0047;

© drannejensen 2014
 187

p<0.01). Pursuant to these results, correlations were run to detect if any other participant

factor could be associated to kMMT accuracy. Two Practitioner characteristics were found to

significantly negatively correlate to kMMT accuracy: (1) Age (r = ‒0.2986; p<0.01), and (2)

Year in practice (r = ‒0.2461; p=0.02). Also, 3 TP characteristics were found to significantly

positively correlate to kMMT accuracy: (1) Age (r = 0.2429; p=0.02), and (2) Confidence in

Practitioner (r = 0.2347; p=0.02), and (3) Confidence in Practitioner’s kMMT ability (r =

0.2362; p=0.02). See Table 5.6. However, similar to the ANOVA results, caution is advised

when interpreting these results because due to the many variables being analysed, they might

not be meaningful.

TABLE 5.5.A – ANOVA for all variables.

Source Partial SS df F Prob>F %
Model 3.2648 88 4.15 0.03* 98.1%

Practitioner characteristics
Connection with TP 0.4050 7 2.38 0.05 12.2%
Age 0.3503 10 3.92 0.04* 10.5%
Gender 0.0031 1 0.35 0.58 8.9%
How well Practitioner knew TP 0.2245 4 2.31 0.08 6.7%
How well Practitioner liked TP 0.2211 5 1.82 0.14 6.6%
Confidence in own ability to test TP 0.1325 6 0.91 0.50 4.0%
TP’s “willingness” to be tested 0.0646 3 0.89 0.46 1.9%
Years of kMMT Experience 0.0489 1 5.46 0.05 1.5%
Self-rated Anxiety 0.0349 5 0.29 0.92 1.1%
Profession 0.0177 1 1.98 0.20 0.5%
Number of Years in Practice 0.0070 2 0.39 0.69 0.2%
Currently Practising 0.0009 1 0.10 0.76 0.0%

TP characteristics
How well TP liked Practitioner 0.2529 8 1.30 0.29 7.6%
Self-rated Anxiety 0.2401 7 1.41 0.24 7.2%
Confidence in Practitioner 0.2045 8 1.05 0.42 6.1%
Confidence in Practitioner’s kMMT 0.1852 7 1.09 0.40 5.6%
Ability 0.1728 7 1.02 0.44 5.2%
Age 0.1295 4 3.62 0.07 3.9%
How well TP knew Practitioner 0.0495 2 1.02 0.37 1.5%
Gender 0.0268 1 2.99 0.13 0.8%

Residual 0.0626 7 1.9%

Total 3.3275 95 100.0%

SS, Sum of squares; df, degrees of freedom; F, f-statistic; Prob, probability; TP, Test Patient; *Reached significance

© drannejensen 2014
 188

TABLE 5.5.B – Univariate general linear model using all variables individually.

TABLE 5.6 – Significant Results of Correlations Testing

Participant Characteristic r p-value
Practitioner characteristics
Age -0.2986 <0.01
Number of Years in Practice -0.2461 0.02
TP characteristics
Age 0.2429 0.02
Confidence in Practitioner 0.2347 0.02
Confidence in Practitioner’s kMMT Ability 0.2362 0.02
r, correlation coefficient; TP, Test Patient

5.4.2.3 Repeatability

Repeatability was considered to be the amount of variance in the kMMT accuracy between

Block 1 and Block 2 for each Pair. It can be reported by Practitioner or by TP. In tests with

good repeatability, scatterplots of Block 1 scores vs. Block 2 scores will hover along a

diagonal line (slope=1) indicating identical scores (see Figure 5.6).

© drannejensen 2014
 189

FIGURE 5.6 – Blank repeatability scatterplot. The diagonal line represents

identical scores between Block 1 and Block 2. The red lines represent the likelihood
of Chance (0.500). Therefore, the better scores are to the right and towards the top,
and pairs with good repeatability will hover around the diagonal line.

Figure 5.7 shows scatterplots of mean kMMT accuracies by Practitioner (A) and by TP (B).

Visual inspection of these scatterplots suggests a reasonable amount of agreement between

mean kMMT accuracies in Blocks 1 and 2, especially with respect to TPs (B), which suggests

adequate repeatability.

In contrast, visual inspection of similar individual scatterplots of kMMT scores for each

Practitioner (Appendix Figure B.5.1) suggest that some Practitioners showed good

repeatability (e.g. #12 and #15), while others demonstrated poor repeatability (e.g. #7 and

#9). Likewise visual inspection of similar scatterplots for each TP, showed noticeably

superior repeatability for some TPs (e.g. TP #1) than others (e.g. TP #4). See Appendix

Figure B.5.2.

© drannejensen 2014
 190

FIGURE 5.7 – Repeatability scatterplots : Mean kMMT accuracy –

Block 1 vs Block 2. (A) by TP, and (b) by Practitioner.

1
.9
Mean kMMT Accuracy Block 2 - by TP

.8
.7
.6
.5
.4
.3
.2
.1
0

0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1
Mean kMMT Accuracy Block 1 - by TP

(A)
1
.9
.8
.7
.6
.5
.4
.3
.2
.1
0

0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1
Mean kMMT Accuracy Block1 - by Practitioner

(B)

Bland-Altman Plots were created for each TP, which plotted their mean kMMT accuracies

with each Practitioner vs. the difference in their mean accuracies (Block 2 – Block 1). See

Figures 5.10. For 3 of the 6 TPs, all scores fell within 2 SDs of the mean, indicating sufficient

© drannejensen 2014
 191

repeatability;175 however, from a clinical perspective, the large SDs of the mean scores (see

large band widths in Figures 5.10) suggest poor agreement by Block, and thus poor

repeatability. In addition, ANOVA demonstrated that an insignificant amount of variance

could be explained by Block alone [F(1,21) = 0.02, p = 0.90].

FIGURE 5.10 – Bland-Altman Plots by TP : Difference against mean for kMMT accuracies.
.6
.4

Mean + 2SD
.2

Mean
0
-.2

Mean - 2SD
-.4
-.6

0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1
Average kMMT Accuracy - TP1
.6

Mean + 2SD
.4
.2

Mean
0
-.2

Mean - 2SD
-.4
-.6

0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1
Average kMMT Accuracy - TP2

© drannejensen 2014
 192

FIGURE 5.10 (con’t.)

.6
.4

Mean + 2SD
.2

Mean
0
-.2

Mean - 2SD
-.4
-.6

0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1
Average kMMT Accuracy - TP3
.6

Mean + 2SD
.4
.2

Mean
0
-.2
-.4

Mean - 2SD
-.6

0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1
Average kMMT Accuracy - TP4

© drannejensen 2014
 193

FIGURE 5.10 (con’t.)

.6

Mean + 2SD
.4
.2

Mean
0
-.2
-.4

Mean - 2SD
-.6

0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1
Average kMMT Accuracy - TP5
.6

Mean + 2SD
.4
.2

Mean
0
-.2
-.4

Mean - 2SD
-.6

0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1
Average kMMT Accuracy - TP6

© drannejensen 2014
 194

5.5 Discussion

5.5.1 Statement of Principal Findings

Supporting results from earlier studies, kMMT once again was found to distinguish Lies from

Truth significantly more accurately than either Chance or Intuition. Also, the finding that the

mean specificity, 0.638 (95% CI 0.430 to 0.486), was more than the mean sensitivity, 0.595 (95%

CI 0.549 to 0.640), suggests that Truths were generally easier to detect than Lies; however, the

difference between these means did not reach significance (p=0.1759), so caution is advised when

interpreting these results.

Only 57% of the between-Pairs variance in mean kMMT accuracy could be explained by the

Pair itself, leaving 43% attributable to unknown factors. The within-Pair variance of mean

kMMT accuracy suggests adequate repeatability. In addition, visual inspection of

reproducibility and repeatability scatterplots and ANOVA results seems to suggest that

stability of kMMT accuracy may be more TP-specific than Practitioner-specific. The results

for each Practitioner, individually, suggest that both reproducibility and repeatability may be

sufficiently stable for some Practitioners, and insufficient for others. The results for each TP,

individually, suggest the same. Overall, analyses of both reproducibility and repeatability of

kMMT accuracy showed variances that could not be explained by any single factor.

5.5.2 Comparisons to Other Studies

There are a number of studies published in the muscle testing literature which have attempted

to quantify the reliability of muscle testing used in various capacities. Using AK-MMT,

Conable reports finding fair intra-tester agreement (κ=0.54),164 Perot et al. report finding

good reliability, and176 Pollard et al. report finding good agreement between a novice

practitioner and an experienced practitioner when using AK-MMT.55 In 2 studies also using

© drannejensen 2014
 195

AK-MMT mixed inter-examiner agreement was found.6, 177 On the other hand, systematic

reviews of AK-MMT found that its reliability could not be adequately determined,14, 49, 151

and in 2 studies using kMMT both intra-tester and inter-tester reliability were found to be

insufficient.178, 179 The mixed results of these reports, along with the findings of this present

study, speak to the difficulty of exploring the variability of MMT.

5.5.3 Strengths and Limitations

A strength of this study is that its results support the findings of earlier studies, therefore,

demonstrating that kMMT accuracy can be adequately estimated using rigorous scientific

methods and that these methods produce durable results. Another strength are the breadth of

the types of the kMMT-practitioners enrolled, and the degrees of experience of all

participants. In addition, similar to other studies in this series, testing during this study was as

true to clinical practice as possible in a research setting. A limitation of this study is, again, its

generalisability to other applications of kMMT, to kMMT using muscles other than the

deltoid, and to other forms of MMT. Also, a weakness in the study design may have been the

duration of participation of the TPs: 3+ hours in total. This amount of time may have

adversely impacted their performance or their compliance to strict procedures.

5.5.4 Implications for Clinical Practice

Naturally, a clinician would like to know how stable or reliable kMMT accuracy is. S/He

would be interested in knowing if it can be relied upon from one patient to the next

(reproducibility) and with the same patient from one visit to the next (repeatability).

Therefore, the primary concern is whether the largest variability is small enough to be

clinically meaningful,175 and the results of this study suggest that this must be taken on a pair-

© drannejensen 2014
 196

by-pair basis. On the other hand, how reliable kMMT accuracy must be in order for it to be

useful is a clinical decision, not a statistical one.180

5.5.5 Unanswered Questions and Future Research

The results of this study suggest that the accuracy of kMMT used in this way may be

sufficiently reproducible and repeatable. However, since kMMT is a test used by practitioners

to guide treatment usually within the context of a specific protocol or technique system, to

assess the true usefulness of kMMT, randomised, controlled trials must be carried out to

assess the effectiveness of the various technique systems that employ kMMT. In other words,

aside from kMMT being accurate or precise in and of itself, it must be ascertained if its use

leads to improved patient outcomes, such as a better quality of life. Future research must

focus on responding to this concern.

5.6 Summary

The accuracy of kMMT used for distinguishing lies from truth and the reproducibility and

repeatability of kMMT accuracy were assessed and found to be sufficient; however, there

factors at play that could not be explained by the model used. Additional research is needed

to explain the variance.

© drannejensen 2014
 197

5.7 Chapter 5 – List of Tables and Figures

5.7.1 Tables

TABLE 5.1 – Demographics of Practitioners

TABLE 5.2 – Mean Accuracy data for each TP individually: Accuracy (overall percent
correct), sensitivity, specificity, PPV and NPV; for kMMT and Intuition.

TABLE 5.3 – Diagnostic accuracy of kMMT vs. Intuition: Means, 95% and significance.
Accuracy (as overall fraction correct), sensitivity, specificity, positive predictive value, and
negative predictive value.

TABLE 5.4 – ANOVA results : Practitioner and TP.

TABLE 5.5.A – ANOVA for all variables.

TABLE 5.5.B – Univariate general linear model using all variables individually.

TABLE 5.6 – Significant results of correlations testing.

5.7.2 Figures

FIGURE 5.1 – Testing scenario layout: The Test Patient (red) viewed a monitor which the
Practitioner could see, had an ear piece in his ear through which he received instructions.
After the muscle test, the Practitioner (blue) entered his results on a keyboard.

FIGURE 5.2 – Testing room flow.

FIGURE 5.3 – Participant Flow Diagram : Study 4.

FIGURE 5.4 – Scatterplots of mean kMMT accuracy by Pair : (A) By Practitioner, and (B)
By TP

FIGURE 5.5.A – Reproducibility of kMMT accuracy by Practitioner: in order of mean

accuracy.

FIGURE 5.5.B – Reproducibility of kMMT accuracy by TP.

FIGURE 5.6 – Blank repeatability scatterplot. The diagonal line represents identical scores
between Block 1 and Block 2. The red lines represent the likelihood of Chance (0.500).
Therefore, the better scores are to the right and towards the top, and pairs with good
repeatability will hover around the diagonal line.

© drannejensen 2014
 198

FIGURE 5.7 – Repeatability scatterplots : Mean kMMT accuracy – Block 1 vs. Block 2. (A)
by Practitioner, (B) by TP

FIGURE 5.8 – Repeatability scatterplots : Block 1 vs Block 2 – by Practitioner (#1-16).

FIGURE 5.9 – Repeatability scatterplots : Block 1 vs Block 2 – by TP (#1-6)

FIGURE 5.10 – Bland-Altman Plots by TP : Difference against mean for kMMT accuracies.

© drannejensen 2014
 199

CHAPTER 6
Study 5 – Using Emotionally-Arousing Stimuli

“It shows the truth - that the real meaning of a word

is only as powerful or harmless as the emotion behind it.”

Sarah Silverman

© drannejensen 2014
 200

CHAPTER 6 : STUDY 5 – USING EMOTIONALLY-AROUSING STIMULI

6.1 ABSTRACT

Research Objectives: To determine if using emotionally-arousing stimuli influences kMMT

accuracy compared to affect-neutral stimuli.

Methods: A prospective study of diagnostic test accuracy was carried out. Twenty

Practitioners who routinely practised kMMT were paired with Test Patients (TPs) who may

or may not have been kMMT-naïve. The Pairs performed 40 kMMTs as TPs spoke True and

False statements about a mix of affect-neutral and emotionally-arousing pictures. Blocks of

kMMT alternated with blocks of Intuition. The verity of the spoken statements was randomly

assigned, with the prevalence of Lies fixed at 0.50.

Results: kMMT accuracy using emotionally-arousing stimuli was no better or worse than

when using affect-neutral stimuli (p=0.35). However, using all stimuli, kMMT accuracy

(0.648; 95% CI 0.558 - 0.737) was found to be significantly better than Intuition accuracy

(0.526; 95% CI 0.488 - 0.564; p=0.01) and Chance (0.500; p<0.01). In addition, similar to

previous studies in this series, this study also failed to detect any characteristic that

consistently influenced kMMT accuracy.

Summary: This study found that using emotionally-arousing stimuli was no different from

using affect-neutral stimuli. However, this study would have been strengthened by adding

personally-relevant, high-stakes lies instead of lies instead of emotionally-arousing

(impersonal) stimuli. The primary limitation of this study is its lack of generalisability to

other applications of kMMT. The main strengths of this study were its choice of a “gold

standard” as the reference standard and its high degree of blinding. Finally, this study is

© drannejensen 2014
 201

further evidence that a simple yet robust methodology for assessing the value of kMMT as a

diagnostic tool can be developed and implemented effectively.

Keywords: sensitivity; specificity; kinesiology; muscle weakness; lie detection; deception;

lying; arm; upper extremity; emotional stress.

© drannejensen 2014
 202

6.2 Introduction

Polygraphs operate on the idea that lying causes certain physiological changes, and the higher

the stakes of the lie, the more accurate the polygraph reading.181 A thread of a similar concept

runs through many kMMT technique systems: It is a common theory among some

practitioners that a stress in the body is related to a weakening of a muscle.11, 24, 182 Therefore,

if kMMT has been found to achieve a certain degree of accuracy at detecting Lies using

affect-neutral stimuli (see previous chapters), then I hypothesise that the use of emotionally-

arousing stimuli may serve to increase its accuracy.

Considering what factors may influence the accuracy of a diagnostic test is important for both

clinicians and researchers. A fundamental element of clinical research is the use of some

measure to detect if a change has occurred. Therefore, the results of a study hinge upon the

validity of the selected measure: Choose the wrong measure, or one that lacks sufficient

sensitivity (for the sample population), and the study will be both meaningless and an

irresponsible waste of time and other resources. Therefore, knowing how to produce the best

possible results in kMMT testing will be of value to kMMT researchers and kMMT

practitioners alike. Furthermore, since a diagnostic test is only valuable to clinicians if it leads

to a correct diagnosis and effective treatment, kMMT practitioners will benefit from knowing

more about what type of stimuli will elicit a truer kMMT response.

Similar to the objects they represent, pictures and words can cause physiological and

behavioural reactions.183-185 The International Affective Picture System (IAPS) is a database

of over 1000 pictures that were rated by a large group of people (men and women) for the

feelings they evoked.101 Likewise, the Affective Norms for English Words (ANEW) is a

database of English words with associated emotional ratings, which complements the

IAPS.104 These sets of pictures and words have been rated for pleasure, arousal, and

© drannejensen 2014
 203

dominance, and the average reported emotional impact are listed along with its standard

deviation. Together these databases provide an emotional stimulus standardisation necessary

for rigorous scientific analysis.

This study used pictures and words from these databases, with specific valencesi known to be

either neutral or emotionally arousing. The primary aim of this study was to determine if the

use of emotionally-arousing stimuli had an impact on kMMT accuracy, compared to affect-

neutral stimuli.

6.3 Methods

This study is a prospective study of diagnostic test accuracy. No participant was assessed

prior to enrolment. This protocol received ethics committee approval by the Oxford Tropical

Research Ethics Committee (OxTREC; Approval #41-10) and the Parker University

Institutional Review Board for Human Subjects (Approval # R17_10). Also, this study

protocol was registered with two clinical trials registries: the Australian New Zealand

Clinical Trials Registry (ANZCTR; www.anzctr.org.au), and US-based ClinicalTrials.gov.

Written informed consent was obtained from all participants, and all other tenets of the

Declaration of Helsinki were upheld. This paper was written in accordance with the

Standards for the Reporting of Diagnostic Test Accuracy Studies (STARD) guidelines (see

Appendix D, page 378, for the STARD Checklist).47, 66, 99

Essentially, the methodology of this study followed closely to that of Study 2 (see page 115),

with the exception that emotionally-arousing pictures were mixed with the affect-neutral

pictures from the database used in previous studies. Otherwise, these components remained

i
See glossary

© drannejensen 2014
 204

the same: the target condition (Lying), index test (kMMT), reference standard (the actual

verity of the spoken statement), the secondary index test (Intuition), study population,

participant recruitment and sampling, enrolment criteria, sample size (n=20 Pairs), testing

scenario, testing methods, participant flow and prevalence of Lies.

6.3.1 Participants and Setting

Two groups of participants were recruited: (A) Healthcare practitioners (n=20) who routinely

use kMMT in practice (“Practitioners”), and (B) Test Patients (n=20; “TPs”). Practitioners

and TPs were recruited in the same manner as in Study 2 (see page 116), in the American

state of California; however, a mixture of kMMT-naïve and non-naïve TPs were included, as

were a mix of TPs who knew their paired Practitioner and those who did not.

6.3.2 The Stimuli

The visual stimuli presented were a mix of affect-neutral pictures from the database used in

previous studies and pictures that were emotionally-arousing. The emotionally arousing

pictures were chosen for their similarity to those in the International Affective Picture System

(IAPS; National Institute of Mental Health Center for Emotion and Attention, University of

Florida, Gainesville, FL)100 which had mean arousal levels above 7.0,39 and supplemented

with additional similar pictures. Whenever possible, the pictures were paired with statements

containing words from the Affective Norms for English Words (ANEW; National Institute of

Mental Health Center for Emotion and Attention, University of Florida, Gainesville, FL),

which had a mean arousal valence above 4.8.104 For examples of emotionally-arousing

pictures, see Figure 6.1.

© drannejensen 2014
 205

FIGURE 6.1 – Examples of affective visual stimuli used in this study. (A), (B), (C) and
(D) are examples that could have been presented to a Test Patient during either the kMMT or
the Intuition Blocks.

For this study, 40 emotionally-arousing and 40 affect-neutral pictures were placed into a

database. Of the 80 (2 x 40) pictures, half of each group was allocated to the kMMT Blocks

and half to the Intuition Blocks. In addition, half of these subgroups were paired with True

statements and half with False statements, so that the prevalence of Lies was again fixed at

0.50. For clarity of this breakdown, refer to Figure 6.2, which describes how the emotional

valence, verity and Blocks were distributed. The order of stimuli was randomly chosen and

presented using DirectRT Research Software (Empirisoft Corporation, New York, NY), so

that each Pair was presented with a unique sequence of stimuli.

© drannejensen 2014
 206

FIGURE 6.2 – Distribution of emotional valence of pictures.

6.3.3 The Testing Scenario

The testing scenario was the same as in Study 2 (see page 117). Each Practitioner performed

40 kMMTs on the TP, broken up into 4 blocks of 10 tests each and recorded their results in

the same manner. Four (4) Intuition Blocks alternated with 4 kMMT Blocks. All Practitioners

were blind to the verity of the TP statement.See Figure 6.3 for the Participant Flow Diagram.

Also, participants completed the same pre- and post-testing questionnaires.186

© drannejensen 2014
 207

FIGURE 6. 3 – Participant Flow Diagram : Study 5 – Emotional Stimuli.

6.3.4 Statistical Methods

No new sample size calculation was performed for this study, but kept at 20 Practitioner-TP

pairs. As in Study 2 (see page 126), this study was not powered for subgroup analyses but

will be shown for completeness. Again, I report error-based measures of accuracy for both

kMMT and Intuition: overall fraction correct73, sensitivity, specificity, Positive Predictive

Value (PPV) and Negative Predictive Value (NPV) – and their 95% confidence intervals

(95% CI). Statistical advice was sought, during the design phase and for data analysis for this

study. All data were analyzed using STATA 17.0, specifically the commands ttest and

pwcorr, sig.

© drannejensen 2014
 208

6.4 Results

6.4.1 Participants

Twenty unique Practitioner-TP pairs were enrolled in October and November 2011. Included

were 13 female and 7 male Practitioners, and 12 female and 8 male TPs. Of the 20

Practitioners, there were 14 chiropractors, 1 mental health professional, 1 acupuncturist, 1

naturopath, 1 other health professional and 2 other professionals (non-health).

Twelve Practitioners were in full-time practice, 8 were in part-time practice, and none

enrolled were currently not practicing. The Practitioners’ mean (SD) number of years in

practice was 16.8 (12.7) years. The mean age for Practitioners was 48.5 (12.0) years, and for

TPs, 37.9 (12.4) years. For a summary of Practitioner demographics, see Table 6.1. In

addition, there were 13 TPs who were kMMT-naïve and 7 who had prior experience with

kMMT, and 8 TPs knew their paired Practitioner and 12 who did not.

6.4.2 The Stimuli

As outlined in the Methods section, for emotionally-arousing stimuli, chosen for inclusion in

the database of stimuli were pictures with Arousal Scores over 7.0 and words with arousal

scores of over 4.8. The overall mean (SD) picture arousal score was 4.8 (1.3), for

emotionally-arousing stimuli, the mean (SD) was 7.6 (0.5), and for affect-neutral stimuli, the

mean (SD) was 4.9 (0.6). The overall mean (SD) word arousal score was 5.0 (1.1), for

emotionally-arousing stimuli, the mean (SD) was 6.0 (0.7), and for affect-neutral stimuli, the

mean (SD) was 4.8 (1.2). For both pictures and words, their arousal scores were normally

distributed (see Appendix Figure B.6.2). For picture-word pairs if the sum of their arousal

scores were in excess of 9.8 it was considered emotionally-arousing, otherwise it was

considered affect-neutral.

© drannejensen 2014
 209

6.4.3 Test Results

Pairs took between 10 and 45 minutes to complete their participation. All Pairs completed all

testing in full. Aside from TP arm fatigue, there were no adverse events reported from any

testing. All accuracies were normally distributed (see Appendix Figure B.6.1), so parametric

statistics were used, mainly the Student t-test and ANOVA.

6.4.3.1 kMMT and Intuition Accuracies

I first will compare accuracies using emotionally-arousing and affect-neutral stimuli, and then

I will report accuracies using all stimuli. Using only emotionally-arousing stimuli, the mean

(95% CI) kMMT accuracy (i.e. overall fraction correct) for kMMT was 0.632 (0.544 - 0.720),

which was significantly different from Chanceii (0.500; p=0.01) but not from the mean (95%

CI) Intuition accuracy, 0.545 (0.491 - 0.599; p=0.09). To calculate sensitivity, specificity,

PPV and

ii
Chance here refers to the hypothetical situation where either outcome was equally likely: 50-50.

© drannejensen 2014
 210

TABLE 6.1 – Demographics of Practitioners.

© drannejensen 2014
 211

NPV, I calculated these same statistics for each pair and report their means (95% CIs):

sensitivity was 0.589 (0.453 - 0.725), specificity, 0.681 (0.575 - 0.788), PPV, 0.669 (0.577 -

0.760) and NPV, 0.641 (0.534 - 0.748). See Table 6.2. Finally, the mean kMMT accuracy

using only emotionally-arousing stimuli was not significantly different from kMMT accuracy

using only affect neutral stimuli (p=0.35). See Table 6.3. The same mean statistics are

reported for the Intuition condition in Table 6.2. The 2x2 tables for each Pair can be found in

Appendix Table B.6.2.

Using only affect-neutral stimuli, the mean (95% CI) kMMT accuracy (i.e. overall fraction

correct) for kMMT was 0.659 (0.560 - 0.757), which was significantly different from both

Chance (p=0.01), and the mean (95% CI) Intuition accuracy, 0.508 (0.459 - 0.556; p=0.01).

The mean (95% CI) kMMT sensitivity was 0.645 (0.521 - 0.770), specificity, 0.671 (0.545 -

0.796), PPV, 0.654 (0.538 - 0.770) and NPV, 0.675 (0.581 - 0.769). See Tables 6.2 and 6.3.

The same mean statistics are reported for the Intuition condition in Table 6.2.

© drannejensen 2014
 TABLE 6.2 – Diagnostic accuracy : Overall fraction correct, sensitivity, specificity, positive predictive value, and negative predictive value (n=20
Pairs) for kMMT & Intuition. All stimuli compared to emotionally-arousing stimuli and affect-neutral stimuli.

© drannejensen 2014
212
 213

TABLE 6.3 – The influence of stimuli valence on accuracy.

© drannejensen 2014
 214

Using all stimuli, the mean (95% CI) accuracy (i.e. overall fraction correct) for kMMT was

0.648 (0.558 - 0.737), which was significantly different from the mean (95% CI) Intuition

accuracy, 0.526 (0.488 - 0.564; p=0.01), and Chance (p<0.01).

The other accuracy statistics for kMMT and Intuition are reported in Table 6.2, and True vs.

False comparisons can be found in Appendix Table B.6.1. The plot of kMMT Sensitivity vs.

Specificity can be found at Figure 6.4. In addition, it is noticed that these statistics for Study 1

(see page 78) fell within this study’s 95% CI, implying that the results of the 2 studies were

similar. Correlations of kMMT by Block were also run for both Emotionally-arousing and

Affect Neutral stimuli. No correlations reached significance for the Emotionally-arousing

stimuli. While some Block’s accuracies were correlated using Affect-neutral stimuli, the

correlations did not display a consistent pattern (see Table 6.4).

Figure 6.4 – Scatterplot graph of sensitivity of kMMT vs. specificity of kMMT. Those that
scored above the green line did better than Chance.

© drannejensen 2014
 215

TABLE 6.4 - Correlations (r) with p-values among kMMT accuracies by Block. Using (A)
emotionally-arousing stimuli, and (B) affect-neutral stimuli.
(A) Block 1 Block 2 Block 3
Block 1 1.0000

Block 2 0.0201 1.0000

p-value 0.93

Block 3 0.1548 0.3077 1.0000

p-values 0.51 0.19

Block 4 0.2125 0.1605 -0.1189

p-values 0.38 0.51 0.63
kMMT, kinesiology-style Manual Muscle Testing.

(B) Block 1 Block 2 Block 3

Block 1 1.0000

Block 2 0.3776 1.0000

p-value 0.10

Block 3 0.6343 0.5588 1.0000

p-values <0.01* 0.01*

Block 4 0.8118 0.2591 0.5485

p-values <0.01* 0.27 0.01*
kMMT, kinesiology-style Manual Muscle Testing; * Significance Reached.

6.4.3.2 Potential Influencing Factors

As mentioned above, this study was not powered to do subgroup analyses, but for

completeness, I compared kMMT accuracies for the different arousal scores by various

participant characteristics, and there results are outlined in Tables 6.5 to 6.9.

First, looking specifically at Practitioner characteristics that might have had an influence, I

grouped the pairs by Practitioner profession, Practitioner practising status and Practitioner

self-ranked kMMT expertise. Table 6.5 shows that none of these subgroupings were found to

be significantly different from each other, except for self-ranked kMMT expertise using only

affect-neutral stimuli: Twelve Practitioners rated themselves as a “4” (i.e. “Expert”), and 8 as

a “3,” “2,” or “1.” None ranked themselves as a “0.” The mean (95% CI) kMMT accuracy for

© drannejensen 2014
 (p=0.04).
TAB TABLE 6.5 - The influence on various Practitioner categorical factors on kMMT accuracy: (1) profession, (2) practising status, and
(3) self-ranked kMMT expertise. Using (A) emotionally-arousing stimuli, and (B) affect-neutral stimuli.
other Practitioners, 0.549 (0.429 - 0.669), which was found to be significantly different
4-ranked Practitioners using affect-neutral stimuli was 0.732 (0.592 - 0.872), and for the 8

© drannejensen 2014
216
 217

TABLE 6.6 - The influence on kMMT technique system on kMMT accuracy. Using (A)
emotionally-arousing stimuli, and (B) affect-neutral stimuli.

TABLE 6.7 - The influence of gender on kMMT accuracy. (1) Practitioner gender, (2) Test Patient
gender, and (3) Practitioner-Test Patient sameness of gender. Using (A) emotionally-arousing stimuli,
and (B) affect-neutral stimuli.

TABLE 6.8 - The influence Test Patient arm used on kMMT accuracy. Using
(A) emotionally-arousing stimuli, and (B) affect-neutral stimuli.

© drannejensen 2014
 218

Stratifying by technique system practiced also did not find any significant differences

between subgroups, using either emotionally-arousing stimuli (p=0.98) or affect-neutral

stimuli (p=0.85). See Table 6.6. Also, there was no significant difference between any of the

three Gender subgroupings or the choice of TP arm subgrouping, using either emotionally-

arousing stimuli or affect-neutral stimuli (see Tables 6.7 and 6.8, respectively). Likewise, if

the TP wore glasses did not seem to influence accuracy, using either emotionally-arousing

stimuli or affect-neutral stimuli (see Table 6.9).

Finally, 9 TPs reported guessing the paradigm, whereas 11 TPs did not report guessing the

paradigm. For all stimuli, for just emotionally-arousing stimuli and for just affect-neutral

stimuli, these 2 groups did not differ significantly in kMMT accuracy (p=0.10, p=0.06, and

p=0.21 respectively). In addition, comparing the mean kMMT accuracies of those Pairs

whose TPs were kMMT-naïve (n=13) to those Pairs whose TPs were not kMMT-naïve (n=7)

for all stimuli, for just emotionally-arousing stimuli and for just affect-neutral stimuli, also

found no significant differences (p=0.26, p=0.50, and p=0.21). Lastly, comparing the mean

kMMT accuracies of those Pairs who knew each other (n=8) to those Pairs did not know each

other (n=12) for all stimuli, for just emotionally-arousing stimuli and for just affect-neutral

stimuli, also found no significant differences (p= 0.48, p=0.22, and p=0.70).

© drannejensen 2014
 TABLE 6.9 - The influence of various categorical TP characteristics on kMMT accuracy. (1) The Test Patient guessing the paradigm, (2) The
Test Patient wearing glasses during testing, (3) Test Patient experience with kMMT, and (4) If Test Patient knew Practitioner. Using (A) emotionally-
arousing stimuli, and (B) affect-neutral stimuli.

© drannejensen 2014
219
 220

6.4.3.3 Correlation Testing

Like previous studies, this study did not find any significant correlations between kMMT

accuracy and the Practitioner demographics of age, number of years in practice, number of

years practising kMMT, and usual number of hours per day using kMMT (see Table 6.10).

Also, there was no significant correlation between kMMT accuracy and TP age (r = 0.1102,

p=0.65).

TABLE 6.10 – Correlations (r) among kMMT accuracies and Practitioner

demographics. p(2-tailed)<0.05

As one may anticipate, there was a significant positive correlation between kMMT accuracies

using only emotionally-arousing stimuli and kMMT accuracy using only affect-neutral

stimuli (see Table 6.11 and Appendix Figure B.6.3) That is, if a Pair scored high with

emotionally-arousing stimuli, they seemed to score high with affect-neutral stimuli, and/or

vice versa. Also, this study found that kMMT accuracy using only affect-neutral stimuli was

significantly and positively correlated to a change in Practitioner’s Confidence in Own

kMMT Ability (Post‒ minus Pre‒testing; p=0.01). No other significant relationship was

detected between kMMT accuracy and any other Confidence score using either emotionally-

arousing stimuli or affect-neutral stimuli (see Table 6.11).

© drannejensen 2014
 221

Finally, in this study also, Practitioner Subjective State Anxiety showed no obvious

relationship to kMMT accuracy using either emotionally-arousing stimuli or affect-neutral

stimuli (see Appendix Figures B.6.3.T & U). Further analysis of this relationship revealed

correlation coefficients which were insignificant (emotionally-arousing stimuli: r = -0.1153,

p= 0.63; affective-neutral stimuli: r = -0.1543; p= 0.52).

TABLE 6.11 - The change in confidence ratings and correlations (r; with p-values) among
kMMT accuracies and Participant confidence scores. p(2-tailed)<0.05.

© drannejensen 2014
 222

6.5 Discussion

6.5.1 Statement of Principal Findings

There were a number of unique findings from this study, plus many of the previous findings

were supported by this study. The main finding was that the use of emotionally-arousing

stimuli did not improve mean kMMT accuracy scores. In fact, the mean kMMT accuracy

using only emotionally-arousing stimuli was less than the mean kMMT accuracy using only

affect-neutral stimuli, but their difference did not reach significance. Furthermore, using only

emotionally-arousing stimuli, the mean kMMT accuracy was not significantly different from

Intuition either. One interesting result was that regardless of valence of stimuli (i.e. affect-

neutral, emotionally-arousing, or both combined), kMMT accuracy was consistently found to

be significantly better than Chance.

In addition, findings of this study also replicated those of previous studies, in that (when

using the data from all stimuli) kMMT accuracy was significantly better than Intuition and

Chance at distinguishing Lies from Truth. Plus, since the mean kMMT accuracy of this study

fell within the 95% CI of Study 1 (see page 75) and the mean kMMT of Study 1 fell within

this study’s 95% CI, this suggests that the results were indeed similar.

Using all stimuli, a mean (95% CI) sensitivity of 0.620 (0.500 - 0.740) indicated that 62% the

Lies were detected, while a mean (95% CI) specificity of 0.675 (0.570 - 0.780) suggests that

68% of the Truths were detected, once again suggesting Truths were easier to detect.

Separating stimuli into emotionally-arousing and affect-neutral also found mean specificities

were higher than mean sensitivities, also implying Truths were easier to detect, regardless of

stimuli valence.

© drannejensen 2014
 223

Since the prevalence of Lies is unknown in real clinical settings, it was fixed at 0.50 overall,

0.53 for emotionally-arousing stimuli and 0.48 for affect-neutral stimuli (see 2x2 Tables in

Appendix Table B.6.2). The kMMT PPV and NPV for all stimuli, emotionally-arousing

stimuli and affect-neutral stimuli were all analogous, ranging only from 0.641 to 0.675,

indicating that roughly 2/3 of the time, the result of a kMMT could be depended upon in this

study. Although these results are similar to the findings of previous studies in this series,

since PPV and NPV vary with prevalence, the predictive values reported here should not be

applied universally.106

Once again, since these methods were powered for a sample size of n=20, I am cautious

about assessing the credibility of any subgroup analysis.142 However, the following

participant characteristics were found to have no influence on kMMT accuracy:

 Emotional arousal score of stimuli

 Practitioner profession

 Practitioner’s number of years in practice

 Practitioner’s number of years practising kMMT

 Practitioner’s usual number of hours/day using kMMT

 Practitioner’s kMMT technique(s) used

 Practitioner’s current practising status

 Practitioner age

 TP age

 Practitioner’s gender

 TP’s gender

 Pair’s sameness of gender

© drannejensen 2014
 224

 TP’s handedness

 If the TP reporting guessing the paradigm

 If the TP wore glasses during testing

 If the TP was kMMT-naïve

 If the TP knew their paired practitioner

 Any TP confidence rating

 All but one (1) Practitioner confidence rating

 Practitioner’s subjective anxiety, or

 Block of testing (Late vs. Middle vs. Early in the testing)

This failure to detect any factor that consistently impacts kMMT accuracy is consistent with

the findings of previous studies.

6.5.2 Strengths and Limitations

Similar to previous studies, the main strengths include the heterogeneity of sample, the use of

a true gold standard as a reference standard, and its simple but robust methodology, while

limitations include its lack of similarity to a true clinical setting and generalisability to other

applications of kMMT. Another limitation of this study was the sample size may not have

allowed for meaningful subgroups analyses.

One aspect of this study that could have been improved is the type of lie that TPs were asked

to tell. In previous studies, TPs were asked to tell lies about affect-neutral stimuli (i.e.

pictures and words), so in this study, some of the stimuli were exchanged for emotionally-

arousing ones, with the expectation that the stress response would be enhanced, making it

easier for kMMT to distinguish Lies from Truth. Nevertheless, because kMMT accuracy was

© drannejensen 2014
 225

not influenced by valence of stimuli, this one change may not have been sufficient to invoke

a change in kMMT accuracy. O’Sullivan et al. found that police professionals were more

successful at detecting high-stakes lies than low-stakes lies.181 She defines a high-stakes lie as

a lie that has personal relevance or that is important to the liar, such as a strongly held

opinion or about a highly stressful personal event or one which had a significant consequence

(positive or negative). On the other hand, she defined a low-stakes lie as one that is relatively

trivial, such as a “white lie,” or one where the reward or punishment is immaterial. Therefore,

I believe this current study would have been strengthened if the lies that the TPs were asked

to tell were personally relevant to the TP.

6.5.3 Possible Explanations of Results

Using emotionally-arousing stimuli did not achieve kMMT accuracies that were any better or

worse than using affect-neutral stimuli. There are 2 possible general explanations for this

result: (1) that the arousal valence of stimuli actually does not influence kMMT accuracy, or

(2) that the arousal valence of stimuli actually does influence kMMT accuracy, and this study

did not detect the difference. Moreover, there may be a number of valid explanations for each

of these 2 possibilities.

First, regarding the former, it might be that, contrary to my study hypothesis, the degree of

emotional arousal of the stimuli is not a factor that affects kMMT accuracy. The TPs spoke

Lies and Truths about pictures they viewed on a computer screen. It may be that it is not the

valence of the picture being presented that is important, but how personally relevant the Lie

is to the TP that may be a mediator.181 Alternatively, it may be some other characteristic of

the Lie that has a more intense impact on kMMT accuracy than personal relevance. Either of

© drannejensen 2014
 226

these, I believe, is a more plausible justification of the results of this study than the latter

reason.

Regarding the latter reason, in my view, in the unlikely event that it is correct, there are again

2 possible scenarios: Arousal valence of the stimuli does influence kMMT accuracy, either

(1) positively or (2) negatively – and these study methods did not detect it. By positively, I

mean using emotionally-arousing stimuli improves kMMT accuracy, and by negatively, I

mean using emotionally-arousing stimuli worsens or reduces kMMT accuracy.

If using emotionally-arousing stimuli does have an impact on kMMT accuracy, positively or

negatively, then perhaps this study may have had methodological problems which led to this

failure to detect. For instance, this study may have been underpowered to detect the

difference, or the prevalence of Lies chosen for this study was suboptimal, hindering

detection of any difference. It is interesting to notice that kMMT accuracy was consistently 3-

6% lower when using emotionally-arousing stimuli than when using affect-neutral stimuli

(however, these differences never reached significance; see Table 6.3). On the other hand, if

using emotionally-arousing stimuli does negatively influence kMMT accuracy, this might

represent the opposite of my original hypothesis. In this instance, it could be that

emotionally-arousing stimuli did cause stress in the TP, which made it more difficult, not

easier, for the Practitioner to perform accurate kMMT. The reasons for this, then, might be

either that this stress caused a neuromuscular aberration which resulted in inconsistent

kMMT outcomes, or that the Practitioner sensed this extra stress in the TP which resulted in

him feeling extra stress himself, which negatively affected his kMMT accuracy. Regardless, I

am unconvinced that using emotionally-arousing stimuli impacts kMMT accuracy negatively,

or at all.

© drannejensen 2014
 227

6.5.4 Implications for clinical practice

From the results of this study, two recommendations for clinical practice surfaced. Firstly, I

suggest that kMMT practitioners have their own kMMT accuracies assessed, because

knowing their sensitivities and specificities would inform them if they are more accurate for

truths or for lies. Each can then use this information to adapt his kMMT sessions accordingly.

That is, if a practitioner knows he is more accurate with lies, he can introduce proportionately

more lies during a session. Second is something that just occurred to me, regardless of it

being a recurrent result. It is commonly thought that learning kMMT is a clinical skill, and as

such is learned just like any other clinical skill (e.g. taking blood pressure readings, taking a

spinal radiograph, or performing an otoscopic exam). However, the results of this study, just

as in previous studies of this series, suggest that length or extent of clinical experience with

kMMT did not correlate with kMMT accuracy. So, new practitioners and potential

practitioners should be encouraged by this, and not disheartened if it seems like they are

performing it inadequately – because it probably is not so.

In addition, this study supported the findings of previous studies in this series, in that once

again it demonstrated that rigorous methodology does exist that can consistently estimate the

diagnostic accuracy of kMMT for distinguishing lies from truth.

6.5.5 Unanswered questions and future research

Having just suggested to new practitioners that practice experience is not an influencer of

kMMT accuracy, I feel that I must say that one of the most important targets of future kMMT

research must still be to ascertain what factors promote more accurate kMMT. It still

intrigues me that kMMT accuracy scores can range from 20% correct to 100% correct within

a small sample of 20 Pairs. I cannot help but wonder what features does the 100% accurate

© drannejensen 2014
 228

Pair possess that the 20% accurate Pair does not – and vice versa. Also, it would be

interesting to know if these characteristics are innate or are acquired (i.e. learned).

Instead of focusing on the valence of the stimuli, future studies may want to compare kMMT

accuracy using personally-relevant and high-stakes lies to kMMT accuracy using low-stakes,

irrelevant lies, such as those used in previous studies. Researchers may want to use a multi-

prong approach toward devising high-stake lies: (1) by using highly stressful, personal

events, and/or (2) by amplifying the value of the reward and/or the severity of the

punishment, and/or (3) by a combination of both.181 If using high-stakes lies does turn out to

improve the accuracy of kMMT, then if a practitioner tailors the statements he asked a patient

to speak so that they are personally-relevant, I speculate that his accuracy score may improve.

6.6 Summary

This study repeated Study 2 using a mixture of emotionally-arousing stimuli and affect-

neutral stimuli, and the differences in accuracies were compared. The results of this study

suggest that kMMT using emotionally-arousing stimuli is no more or less accurate than

kMMT using affect-neutral stimuli, contradicting my initial hypothesis. However, it was

found that (using all stimuli combined) kMMT can be used with significant accuracy to

distinguish lies from truths, compared to both Intuition and Chance. Furthermore, this study

also failed to identify any factors that consistently influenced kMMT accuracy. The study

would have been strengthened by adding personally-relevant, high-stakes lies instead of lies

about emotionally-arousing stimuli. The primary limitation of this study is its lack of

generalisability to other applications of kMMT. The main strengths of this study were its

choice of a “gold standard” reference standard and its high degree of blinding. Finally, this

© drannejensen 2014
 229

study is further evidence that a simple yet robust methodology for assessing the value of

kMMT as a diagnostic tool can be developed and implemented effectively.

© drannejensen 2014
 230

6.7 Chapter 6 – Study 5 – List of Tables and Figures

6.7.1 Tables

TABLE 6.1 - Demographics of Practitioners.

TABLE 6.2 – Diagnostic accuracy : Overall fraction correct, sensitivity, specificity, positive
predictive value, and negative predictive value (n=20 Pairs) for kMMT & Intuition. All
stimuli compared to emotionally-arousing stimuli and affect-neutral stimuli.

TABLE 6.3 - The influence of stimuli valence on accuracy.

TABLE 6.4 - Correlations (r) with p-values among kMMT accuracies by Block. Using (A)
emotionally-arousing stimuli, and (B) affect-neutral stimuli.

TABLE 6.5 - The influence on various Practitioner categorical factors on kMMT accuracy:
(1) profession, (2) practising status, and (3) self-ranked kMMT expertise. Using (A)
emotionally-arousing stimuli, and (B) affect-neutral stimuli.

TABLE 6.6 - The influence on kMMT technique system on kMMT accuracy. Using (A)
emotionally-arousing stimuli, and (B) affect-neutral stimuli.

TABLE 6.7 - The influence of gender on kMMT accuracy. (1) Practitioner gender, (2) Test
Patient gender, and (3) Practitioner-Test Patient sameness of gender. Using (A) emotionally-
arousing stimuli, and (B) affect-neutral stimuli.

TABLE 6.8 - The influence Test Patient arm used on kMMT accuracy. Using (A)
emotionally-arousing stimuli, and (B) affect-neutral stimuli.

TABLE 6.9 - The influence of various categorical TP characteristics on kMMT accuracy. (1)
The Test Patient guessing the paradigm , (2) The Test Patient wearing glasses during testing,
(3) Test Patient experience with kMMT, and (4) If Test Patient knew Practitioner. Using (A)
emotionally-arousing stimuli, and (B) affect-neutral stimuli.

TABLE 6.10 - Correlations (r) among kMMT accuracies and Practitioner demographics.

TABLE 6.11 - The change in confidence ratings and correlations (r; with p-values) among
kMMT accuracies and participant confidence scores. p(2-tailed)<0.05

6.7.2 Figures

FIGURE 6.1 – Examples of affective visual stimuli. (A), (B), (C) and (D) are examples that
could have been presented to a Test Patient during either the kMMT or the Intuition Blocks.

© drannejensen 2014
 231

FIGURE 6.2 – Distribution of emotional valence of pictures.

FIGURE 6.3 – Participant Flow Diagram : Study 5 – Emotional Stimuli.

Figure 6.4 – Scatterplot graph of sensitivity of kMMT vs. specificity of kMMT. Those that
scored above the green line did better than Chance.

FIGURE 6.5 - Mean kMMT & Intuition accuracies by Block. (A) Emotionally-arousing
stimuli only, and (B) affect-neutral stimuli only.

© drannejensen 2014
 232

CHAPTER 7
Discussion

“All truth passes through three stages. First, it is ridiculed.

Second, it is violently opposed. Third, it is accepted as being self-evident.”

Arthur Schopenhauer

© drannejensen 2014
 233

CHAPTER 7 : DISCUSSION

7.1 Statement of principal findings

A primary finding of this series of studies is that it is indeed possible to estimate the accuracy

of kMMT using rigorous evidence-based scientific methodology. This has been demonstrated

repeatedly by generating meaningful and useful results in which the accuracy (as overall

fraction correct), sensitivity, specificity, PPV and NPV of kMMT have been calculated and

compared to controlled conditions. Plus, I have used the widely-accepted standards of the

STARD Statement to design, implement and analyse the results of these studies. Also

following evidence-based health care standards, these studies have been reviewed and

approved by ethics boards and have been registered with clinical trials registries.

Another main finding is that kMMT accuracy seems to be repetitively better than Intuition or

Chance at distinguishing Lies from Truth. A summary of these results can be found in Figure

7.1, which compares the accuracy of kMMT compared to Intuition alone across Studies 1, 2,

4 and 5. [Note that Study 3 (Chapter 4) is not included because its primary outcome was not

comparable to the other studies.] For similar comparisons, see also Appendix Figure B.7.1. It

seems that kMMT accuracy is greater than 10% better than Intuition and can be

conservatively estimated to be at least 60%. Furthermore, both the Practitioner and the TP are

integral parts of the kMMT-complex, each responsible for a significant amount of the

variation in accuracy, with the TP being moderately more influential.

FIGURE 7.1 – Forest Plot : Difference of the mean accuracies (kMMT – Intuition)

© drannejensen 2014
 234

There were also some interesting findings regarding other measures of accuracy. For each

trial the specificity was higher than the sensitivity, meaning that the proportion of Truths

detected was consistently better than the proportion of Lies. This suggests that, using kMMT,

truths may be easier to detect than deceit. Similarly, since the PPVs and the NPVs were

consistently found in the 60%-range, it seems that (at this prevalence of Liesi) there is

approximately a 60% chance that a weak kMMT result predicts a Lie and strong kMMT

result predicts a Truth. These results parallel the aforementioned findings for the kMMT

accuracy estimation. Furthermore, despite the consistently wide variation, the accuracy of

kMMT seems to be sufficiently stable within and between groups.

Other notable findings were that no participant characteristic was identified which impacted

accuracy consistently, and that no correlations were found that dependably predicted

accuracy. In fact, even intuitiveness did not seem to correlate, nor did the blindness of the

practitionerii. Therefore, what influences kMMT accuracy remains wholly unknown.

However, the outcome of a kMMT appears not to be the result of an ideomotor effect, as

some propose.116, 131 Furthermore, since the blindness of the practitionerii did not impact

kMMT accuracy, this seems to suggest practitioner were unable to be persuaded to bias the

muscle test.

7.2 Strengths and limitations

A main strength of this series of studies is the rigorousness of the methods, such as high

degrees of blinding, high adherence to procedures, and increased rigour as series progressed.

Examples of the increased rigour include the degrees of blinding improved and my potential

i
In all studies, the prevalence of Lies was held relatively constant around 0.50.
ii
“... blindness of the practitioner...” – whether the Practitioner was blind to the verity of the TP’s spoken
statement, or not, did not seem to influence kMMT accuracy (see Chapter 2, page 80).

© drannejensen 2014
 235

influence was removed. Other strengths of this series include the clearly defined target

condition,187 which enabled the use of an objective reference standard (i.e. an error-free

reference standard, a true gold standard).188 Another strength was the inclusion of a

secondary index test (i.e. Intuition) as a comparator. Other noteworthy strengths were the

heterogeneity of the samples (both the Practitioners and TPs), including those with varied

kMMT experience and backgrounds, and a high degree of blinding.

A limitation of these studies is the lack of generalisability to other applications of kMMT.

Tempting as it is to speculate, firm caution is advised against over-generalising. Another

limitation is that kMMT was only compared to Intuition and Chance, and not to other tests

used to distinguish Lies from Truth, such as polygraph. In addition, a limitation of this series

is that these studies may have been underpowered to identify factors or characteristics that

may influence or predict kMMT accuracy. A final limitation is that this series of studies did

not ascertain if kMMT is clinically useful, or if a 60% accuracy is “good enough” in a clinical

context.

7.3 Comparison of results to other studies

Prior to completion of this series of studies, evidence in support of kMMT and AK-MMT as

valid diagnostic tests was weak and confusing, despite various attempts at their validation

dating back to the 1970s.6, 14, 25, 30-39, 49, 55, 58, 60, 90, 94, 143, 151, 164, 176-179, 189-201 The confusion

about the usefulness of these tests seems to stem from the uncertainty about what methods of

evaluation to employ and how best to minimise bias.73, 202 Since the evaluation of diagnostic

tests appreciably lags behind that of treatments, the quality of the reporting of diagnostic

techniques has yet to catch up.46, 66, 71, 73 With the development of the STARD Checklist, the

© drannejensen 2014
 236

methods for designing and implementing evaluations of diagnostic procedures, including

those of chiropractic and other CAMiii interventions, will undoubtedly mature as well.66

While detecting deceit may be of interest to some CAM practitioners, it is more widely

thought of to be a topic of concern in the field of forensic science. Typically, polygraph

equipment is used in criminal investigations and in employment and security screenings, by

monitoring a subject’s physiological changes, such as heart rate, respiratory rate, palmar

sweating, brain activity, vocal changes, facial expressions, and body language.203 In a

statement by the National Research Council (2003), the polygraph is reported to produce

accuracies (as overall fractions correct) ranging from 69% to 82%iv, a range comparable to

that of kMMT established in this series of studies.203, 204 Yet, also similar to kMMT, “... (the

polygraph) continues to be the subject of a great deal of scientific and public

controversy...”204

It also might be useful to compare the accuracy results obtained in this series to estimated

accuracies of other commonly-used dichotomous diagnostic tests. To do this, I have chosen a

number of comparator studies with similar sample sizes and populations.111, 202 For instance,

with a sample of 72 participants, Sutlive et al.205 evaluated the predictive validity of

FABER’s Test, used by chiropractors, physiotherapists and orthopaedists for diagnosing hip

osteoarthritis (OA). They found this test had a sensitivity (95% CI) of 57% (34-77%) and a

specificity (95% CI) of 71% (56-82%). Similarly, using a sample size of 40, Youdas et al.206

found that another dichotomous hip OA test, Trendelenburg’s Signv, has a sensitivity of 55%

iii
CAM, Complementary and Alternative Medicine
iv
No error statistics were provided in this report, which causes one to question its rigour. Also, these studies
generally employed high-stakes lies, which are thought to produce higher lie detection accuracy than low-stakes
lies [O'Sullivan M, Frank MG, Hurley CM, Tiwana J. Police lie detection accuracy: The effect of lie scenario.
Law and Human Behavior 2009; 33(6):530-38.]
v
Trendelenburg’s Sign & Kinetic Finger Wiggle : also taught to and commonly-used by chiropractors.

© drannejensen 2014
 237

and a specificity of 70% (no CIs reportediv). Likewise, Kerr et al.207 investigated the

diagnostic accuracy of visual field testing on a 172 patients (332 eyes), and found that a test

called Kinetic Finger Wigglev had an overall sensitivity (95% CI) of 39.0% (28.4-50.4%) and

an overall specificity (95% CI) of 97.2 (90.2-99.7%) in detecting visual field deficits. In light

of this evidence, it is helpful to know that the estimation of kMMT diagnostic accuracy found

in this series of studies is aligned with accuracy estimations of other diagnostic tests used by

similar professionals.

7.4 Implications for clinical practice

The direct implications that these results have on clinical practice are limited. However, one

implication is that they may serve to heighten the confidence that clinicians have in kMMTvi.

Another implication for clinical practice is that since specificity was consistently better than

sensitivity, for improved accuracy clinicians may wish to use a majority of true statements,

rather than false.

The accuracy of kMMT used in the context of these studies was found to be 60 to 70%.

Practically speaking, in regard to diagnostic tests, one is inclined to ask the question: How

good is good enough? Conventionally, this will depend upon a number of factors, such as the

prevalence of the target condition, the gravity of the target condition, whether the test is being

used to rule a condition in or out, and the importance of a false positive compared to a false

negative.106 The truth of the matter is there is no one accuracy statistic that fits all situations,

nor is there no one hard-and-fast rule for interpreting these statistics.73, 111

vi
As they did for me: My confidence in my own kMMT ability in clinical practice has grown as a result of
undertaking this line of research.

© drannejensen 2014
 238

So is 60% correct “good enough” for kMMT? This still remains to be seen, as the true value

of a diagnostic test ultimately lies in how it impacts patient outcomes.76

7.5 Unanswered questions and future research

This report describes some first steps taken to estimate kMMT accuracy; however there are

clearly still large gaps in the literature, ranging from specifically about the use of kMMT in

the same context employed in this series of studies, to broader questions about kMMT’s

clinical utility.

Specifically, I found myself asking these questions: What factors (if any) influence kMMT

accuracy? Is it possible for accuracy to improve? Is it necessary for accuracy to improve?

What are potential sources of variability? Would kMMT accuracy change if the paradigm

were reversedvii? What if different target conditions and/or population/s were chosen? How

does kMMT accuracy using other muscles compare? What would happen if the TPs were

completely blind as well? And is this even possible??

With knowledge about factors that can influence accuracy, it may be possible to maximise

and stabilise kMMT accuracy to a point where it may be more universally clinically

dependable. On the other hand, it is unclear if improved accuracy is a prerequisite for

improving patient health.76

Additionally, research is needed to assess the usefulness of kMMT for detecting other

commonly-used target conditions, such as the need for nutritional supplementation190 or in

the identification of an allergy or hypersensitivity or toxicity.32, 90, 191, 208 In the past, muscle

testing failed to successfully identify these conditions;34-36, 90, 94, 95, 137, 190, 191, 209-212 however,

vii
Reversed paradigm: Truth → weak; Lie → strong.

© drannejensen 2014
 239

in many of these studies strict evidence-based health care protocols may not have been

rigorously applied, and methodologies in general were vague.

Future research in the diagnostic usefulness of kMMT should employ rigorous methods,

including: (1) a clear and specific research objective, (2) a well-defined target condition, (3)

explicit outcomes that are easy to interpret, (4) an appropriate sample of the target population

(who were objectively selected), (5) an objective reference standard, (6) an adequate sample

size, and (7) appropriate blinding.202

On the other hand, it might not be as simplistic as this. It could be that CAM researchers are

trying to evaluate muscle testing using an allopathic yardstick, which may not be

appropriate.213 An example of this may be with allergies. The medical approach of the

diagnosis of an allergy is commonly through a skin prick test or a blood test of the IgE

antibody.214 However, it is questionable if these are appropriate measures for a CAM

perspective. In any case, the design of future diagnostic accuracy studies using kMMT for

other target conditions must be given in-depth and broad-minded consideration.

One might be tempted to suggest to future researchers to explore the mechanism/s of action

of kMMT, that is, to explore how kMMT worksviii. Making such a suggestion is beyond the

scope of an evidence-based health care approach. This question falls more the realm of

researchers in the basic sciences.

Finally, and most importantly, kMMT’s true clinical value must be explored.73, 76, 84, 117, 215, 216

Toward this end, the efficacy of kMMT technique systems must be investigated via

viii
“How does muscle testing work?” is a question I get asked repeatedly. However, the ones asking this
question are usually chiropractors or other muscle testing practitioners, and not research scientists, such as my
colleagues in my department. If there is one thing I learned through this process, it is that clinical researchers are
mainly concerned IF a test or an intervention works, and not HOW it works: If it works, then does it really
matter how?

© drannejensen 2014
 240

rigorously-designed randomised, controlled trials (RCTs). For example, future researchers

could compare the effectiveness of a naturopathic approach for irritable bowel syndrome

using two groups of naturopaths: (1) those that use kMMT to guide treatment (e.g. to

prescribe supplementation and offer dietary advice), and (2) those who do not use kMMT.

Another question that future researchers may want to explore is how effective are alternative

emotional healing techniques, such as HeartSpeakix, for such conditions as depression, panic

attacks or obsessive compulsive disorder, compared to traditional psychological approaches,

such as cognitive behavioural therapy.

Finally, aside from health-related quality of life queries, it will also be important for future

researchers to consider another aspect of clinical utility: Does using kMMT lead to more

efficient use of health care resources?73

7.6 Summary

kMMT has repeatedly been found to be significantly more accurate than both Intuition and

Chance, for one application of this common assessment method: distinguishing lies from

truths. Practitioners appear to be an integral part of the kMMT dynamic yet factors that

contribute or detract from its effectiveness could not be identified. A limitation of this series

of studies is a lack of generalisability to other applications of kMMT. Another limitation is

that with these results it is not ascertained if 60% correct is “good enough” in a clinical

context. Additionally, future investigators may want to explore kMMT accuracy to detect

other target conditions and in different population/s, and what factors (if any), such as

Practitioner and TP characteristics, influence kMMT accuracy. However, improved accuracy

may not result in improved patient outcomes. Therefore, an important next step in

ix
www.HeartSpeak.me

© drannejensen 2014
 241

establishing the validity of kMMT is assessing its clinical utility within kMMT technique

systems as a whole through effectiveness trails.

No test is perfect: 100% accurate, easy to use, risk-free and low cost.202, 217 However, the

results of this series of studies are encouraging. It is hoped that this report will urge

practitioners, researchers and health policy makers to explore kMMT, as a potential method

of assessment and patient management.

© drannejensen 2014
 242

7.7 Chapter 7 – List of Tables and Figures

7.7.1 Figures

FIGURE 7.1 – Forest Plot : Difference of the mean accuracies (kMMT – Intuition)

© drannejensen 2014
 243

APPENDICES

“An important scientific innovation rarely makes its way by gradually winning over and

converting its opponents: What does happen is that the opponents gradually die out.”

Max Planck

© drannejensen 2014
 244

APPENDIX A
Participant Forms

“There's a world of difference between truth and facts. Facts can obscure the truth.”

Maya Angelou

© drannejensen 2014
 245

APPENDIX A : PARTICIPANT FORMS

Participant Information Sheet (PIS) / Informed Consent : Studies 1,2,4,5

page 1 of 2

© drannejensen 2014
 246

Participant Information Sheet (PIS) / Informed Consent : Studies 1,2,4,5

page 2 of 2

© drannejensen 2014
 247

Participant Information Sheet (PIS) / Informed Consent : Study 3

page 1 of 2

© drannejensen 2014
 248

Participant Information Sheet (PIS) / Informed Consent : Study 3

page 2 of 2

© drannejensen 2014
 249

Practitioner Instruction Sheet : Study 1

page 1 of 2

© drannejensen 2014
 250

Participant Instruction Sheet – Practitioner : Study 1

page 2 of 2

© drannejensen 2014
 251

Participant Instruction Sheet – Test Patient : Study 1

page 1 of 2

© drannejensen 2014
 252

Participant Instruction Sheet – Test Patient : Study 1

page 2 of 2

© drannejensen 2014
 253

Participant Instruction Sheet - Practitioner : Studies 2,5

page 1 of 2

© drannejensen 2014
 254

Participant Instruction Sheet – Practitioner : Studies 2, 5

page 2 of 2

© drannejensen 2014
 255

Participant Instruction Sheet – Test Patient : Studies 2,5

page 1 of 2

© drannejensen 2014
 256

Participant Instruction Sheet – Test Patient : Studies 2, 5

page 2 of 2

© drannejensen 2014
 257

Participant Instruction Sheet : Study 3 – Grip Strength

page 1 of 1

© drannejensen 2014
 258

Participant Instruction Sheet – Practitioner : Study 4

page 1 of 2

© drannejensen 2014
 259

Participant Instruction Sheet – Practitioner : Study 4

page 2 of 2

© drannejensen 2014
 260

Participant Instruction Sheet – Test Patient : Study 4

page 1 of 2

© drannejensen 2014
 261

Participant Instruction Sheet – Test Patient : Study 4

page 2 of 2

© drannejensen 2014
 262

Pre-Testing Questionnaire – Practitioner : Studies 1,2,5

page 1 of 1

© drannejensen 2014
 263

Pre-Testing Questionnaire – Test Patient : Studies 1,2,5

page 1 of 1

© drannejensen 2014
 264

Pre-Testing Questionnaire – Test Patient : Study 3

page 1 of 1

© drannejensen 2014
 265

Pre-Testing Questionnaire – Practitioner : Study 4

page 1 of 1

© drannejensen 2014
 266

Pre-Testing Questionnaire – Test Patient : Study 4

page 1 of 1

© drannejensen 2014
 267

Post-Testing Questionnaire – Practitioner : All Studies

page 1 of 1

© drannejensen 2014
 268

Post-Testing Questionnaire – Test Patient : All Studies

page 1 of 1

© drannejensen 2014
 269

Practitioner Score Sheets : Study 4 – Reproducibility & Repeatability

page 1 of 2

© drannejensen 2014
 270

Practitioner Score Sheets : Study 4 – Reproducibility & Repeatability

page 2 of 2

© drannejensen 2014
 271

Test Patient Score Sheets : Study 4 – Reproducibility & Repeatability

page 1 of 2

© drannejensen 2014
 272

Test Patient Score Sheets : Study 4 – Reproducibility & Repeatability

page 2 of 2

© drannejensen 2014
 273

APPENDIX B
Extra Figures & Tables

“Fast is fine, but accuracy is everything.”

Wyatt Earp

© drannejensen 2014
 274

APPENDIX B : EXTRA FIGURES & TABLES : CONTENTS

List of Extra Figures & Tables

Chapter 1 – Background

APPENDIX TABLE B.1.1 – Literature search strategy.

Chapter 2 – Study 1

APPENDIX TABLE B.2.1 - 2x2 Table for kinesiology-style Manual Muscle Testing
(kMMT) for each Pair (n=48). Blocks 1-4, Practitioner blind.

APPENDIX TABLE B.2.2 - kMMT & Intuition accuracies for all statements, True
statements and False statements (for n=48 Pairs).

APPENDIX TABLE B.2.3 - kMMT accuracy by profession, and kMMT accuracy

correlations among professions.

APPENDIX TABLE B.2.4 - Table of correlations among all continuous variables (r).

APPENDIX TABLE B.2.5 - kMMT accuracy by kMMT technique system - detailed.

APPENDIX TABLE B.2.6 - Table of correlations (and p-values) among kMMT accuracy and
the difference in Test Patient confidence ratings (r).

APPENDIX TABLE B.2.7 - Table of correlations (and p-values) among kMMT accuracy and
the difference in Practitioner confidence ratings (r).

APPENDIX TABLE B.2.8 - kMMT accuracy for those Pairs in Study 1 whose prevalence of
Lies was 0.50.

APPENDIX FIGURE B.2.1 – Histograms showing normal distributions.

APPENDIX FIGURE B.2.2 – Scatterplots showing correlations and the correlation

coefficients (r).

© drannejensen 2014
 275

Chapter 3 – Study 2 (Replication of Study 1)

APPENDIX TABLE B.3.1 - 2x2 Tables for kMMT for each Pair (n=20). Each Pair
performed 40 kMMTs.

APPENDIX TABLE B.3.2 - kMMT & Intuition accuracies for all statements, True
statements and False statements (for n=20 Pairs).

APPENDIX TABLE B.3.3 - Correlations among kMMT accuracies and confidence ratings.

APPENDIX FIGURE B.3.1 - Histograms showing normal distribution.

Chapter 4 – Grip Strength Study

APPENDIX TABLE B.4.1 - Mean grip strengths (SD) by participant. (A) False vs. True
statements, and (B) dominant hand vs. non-dominant hand.

APPENDIX FIGURE B.4.1 – Histograms.

Chapter 5 – Reproducibility Study

APPENDIX TABLE B.5.1- Accuracy data for each Pair: Accuracy (overall fraction correct),
sensitivity, specificity, PPV and NPV; for kMMT and Intuition.

APPENDIX TABLE B.5.2 - ANOVA results : Practitioner, TP & Block.

APPENDIX FIGURE B.5.1 - Repeatability scatterplots : Block 1 vs Block 2 – by Practitioner

(#1-16).

APPENDIX FIGURE B.5.2 - Repeatability scatterplots : Block 1 vs Block 2 – by TP (#1-6).

© drannejensen 2014
 276

Chapter 6 – Emotional Stimuli Study

APPENDIX TABLE B.6.1 - kMMT & Intuition accuracies for all statements compared to
True and False statements. Using all stimuli, for n=20 Pairs.

APPENDIX TABLE B.6.2 - 2x2 Tables for kMMT for each Pair (n=20). Each Pair
performed 40 kMMTs. Using (A) emotionally-arousing, and (B) affect-neutral stimuli.

APPENDIX TABLE B.6.3 - Correlations among accuracy scores. kMMT & Intuition,
emotionally-arousing & affect-neutral stimuli.

APPENDIX FIGURE B.6.1 – Histograms for overall kMMT & Intuition accuracies.

APPENDIX FIGURE B.6.2 – Histograms of the distribution of (A) picture and (B) word
arousal levels – showing normal distributions.

APPENDIX FIGURE B.6.3 – Scatterplots.

Chapter 7 – Discussion / Summary

APPENDIX FIGURE B.7.1 : Additional Forest Plots

© drannejensen 2014
 277

CHAPTER 1

APPENDIX TABLE B.1.1 – Literature search strategy. The following search strategy was applied
in [TITLE/ABSTRACT] :

1st: “kinesiology”

AND

(‘diagnos*’ OR ‘accuracy’ OR ‘sensitivity’ OR ‘specificity’ OR ‘precision’ OR

‘validity’

OR ‘reproducibility’ OR ‘repeatability’ OR ‘utility’ OR ‘inter*examiner’ OR

‘inter*rater OR ‘intra*examiner’ OR ‘intra*rater’ OR ‘predict*’)

2nd: “kinesiology” OR “applied kinesiology”

AND

(‘diagnos*’ OR ‘accuracy’ OR ‘sensitivity’ OR ‘specificity’ OR ‘precision’ OR

‘validity’ OR ‘reproducibility’ OR ‘repeatability’ OR ‘utility’ OR ‘inter*examiner’

OR ‘inter*rater OR ‘intra*examiner’ OR ‘intra*rater’ OR ‘predict*’)

3rd: “kinesiology” OR “applied kinesiology” OR “manual muscle testing”

AND

(‘diagnos*’ OR ‘accuracy’ OR ‘sensitivity’ OR ‘specificity’ OR ‘precision’ OR

‘validity’ OR ‘reproducibility’ OR ‘repeatability’ OR ‘utility’ OR ‘inter*examiner’

OR ‘inter*rater OR ‘intra*examiner’ OR ‘intra*rater’ OR ‘predict*’)

© drannejensen 2014
 278

CHAPTER 2

APPENDIX TABLE B.2.1 - 2x2 Table for kinesiology-style Manual Muscle Testing (kMMT);
for each Pair (n=48). Blocks 1-4, Practitioner blind.

© drannejensen 2014
 279

APPENDIX TABLE B.2.1 (con’t.)

© drannejensen 2014
 280

APPENDIX TABLE B.2.1 (con’t.)

© drannejensen 2014
 281

APPENDIX TABLE B.2.1 (con’t.)

© drannejensen 2014
 282

APPENDIX TABLE B.2.1 (con’t.)

© drannejensen 2014
 APPENDIX TABLE B.2.2 - kMMT & Intuition accuracies for all statements, True statements and False statements (for n=48 Pairs).

© drannejensen 2014
283
 APPENDIX TABLE B.2.3 - kMMT accuracy by profession, and kMMT accuracy correlations among professions.

© drannejensen 2014
284
 285

TABLE B.2.4 - Table of correlations among all continuous variables (r ). (p <0.05)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
1. Practitioner's Age (years) 1.0000
2. Years in Practice 0.6030 1.0000
3. Years of kMMT Experience 0.5157 0.6908 1.0000
4. Average Hours/day using kMMT 0.1420 0.4680 0.2748 1.0000
5. Practitioner's Confidence in Own kMMT Ability - Pre-testing 0.3149 0.2657 0.5348 0.2837 1.0000
6. Practitioner's Confidence in MMT in General - Pre-testing 0.0204 0.0152 0.0493 0.2143 0.3540 1.0000
7. Practitioner's Confidence in Own kMMT Ability - Post-testing -0.0309 -0.1885 0.1812 0.1023 0.5311 0.3453 1.0000
8. Practitioner's Confidence in MMT in General - Post-testing -0.2016 -0.3984 -0.2122 0.1440 0.1783 0.6572 0.5564 1.0000
9. Practitioner's Confidence in kMMT in General - Post-testing -0.1054 -0.1206 0.0168 0.1165 0.0560 0.1994 0.4585 0.2659 1.0000
10. Practitioner's Confidence in Testing this TP -0.3177 -0.3042 -0.0714 0.1708 0.1035 0.1044 0.4897 0.4271 0.5013 1.0000
11. Test Patient Age (years) 0.0506 0.0632 0.0284 0.0090 -0.0181 0.1726 -0.0005 0.2185 -0.1318 -0.1248 1.0000
12. TP's Confidence in kMMT in General - Pre-testing -0.1409 -0.0131 -0.0374 -0.0674 0.0802 -0.1468 0.0458 -0.1373 0.0718 -0.1618 -0.0889 1.0000
13. TP's Confidence in Paired Practitioner - Pre-testing -0.2304 -0.1486 -0.2064 0.0598 0.0621 0.0262 0.0895 0.1170 0.1526 -0.0077 -0.2378 0.5143 1.0000
14. TP's Confidence in Practitioner's kMMT - Pre-testing -0.2145 -0.0921 -0.1188 0.0483 -0.0111 -0.0053 0.0344 0.0675 0.1460 -0.0100 -0.1751 0.5404 0.9018 1.0000
15. TP's Confidence in kMMT in General - Post-testing 0.0204 0.1189 0.0990 -0.0655 0.1194 -0.1376 -0.1210 -0.2193 0.0373 -0.1423 -0.0739 0.7545 0.4288 0.4713 1.0000
16. TP's Confidence in Paired Practitioner - Post-testing -0.1776 0.0038 0.0485 0.2175 0.0219 -0.0890 0.0738 0.1081 0.3011 0.2498 -0.2187 0.3906 0.6652 0.6642 0.4971 1.0000
17. TP's Confidence in Practitioner's kMMT - Post-testing -0.1300 0.0506 0.1174 0.2033 -0.0016 -0.1567 0.0703 0.0132 0.3003 0.2860 -0.2360 0.3842 0.5785 0.6100 0.4986 0.9691 1.0000
18. MMT Accuracy (Overall Fraction Correct) -0.0175 -0.0080 -0.0175 -0.0133 -0.0043 -0.0811 0.0546 0.0407 -0.0788 0.0477 0.0602 0.0300 0.1290 0.0863 0.1189 0.1851 0.1639 1.0000
19. The Sum of Practitioner's Confidence in Own kMMT Ability: Pre- + Post-testing 0.1477 0.0251 0.3940 0.2193 0.8546 0.3989 0.8939 0.4353 0.3106 0.3549 -0.0099 0.0705 0.0877 0.0152 -0.0110 0.0568 0.0422 0.0312 1.0000
20. The Difference of Practitioner's Confidence in Own kMMT Ability: Pre- - Post-testing 0.3327 0.4592 0.3082 0.1986 0.3650 -0.0436 -0.5950 -0.4423 -0.4507 -0.4399 -0.0167 0.0257 -0.0395 -0.0483 0.2462 -0.0603 -0.0787 -0.0641 -0.1715 1.0000
21. The Sum of Practitioner's Confidence in MMT in General: Pre- + Post-testing -0.1046 -0.2200 -0.0955 0.1950 0.2883 0.9022 0.5001 0.9180 0.2571 0.2993 0.2158 -0.1558 0.0807 0.0358 -0.1978 0.0150 -0.0749 -0.0194 0.4590 -0.2760 1.0000
22. The Difference of Practitioner's Confidence in MMT in General: Pre- - Post-testing 0.2761 0.5167 0.3227 0.1131 0.1845 0.3287 -0.2994 -0.4958 -0.1035 -0.4148 -0.0749 0.0029 -0.1164 -0.0907 0.1162 -0.2379 -0.1971 -0.1444 -0.0859 0.5040 -0.1108 1.0000
23. The Sum of ALL Practitioner's Ranked Confidences -0.1208 -0.2081 0.0847 0.2728 0.4534 0.5815 0.8092 0.7192 0.7216 0.7137 -0.0076 -0.0552 0.1149 0.0700 -0.1062 0.2101 0.1815 -0.0119 0.7358 -0.4591 0.7175 -0.2312 1.0000
24. The Sum of TP's Confidence in kMMT in General: Pre- + Post-testing -0.0651 0.0558 0.0322 -0.0713 0.1063 -0.1518 -0.0393 -0.1899 0.0584 -0.1624 -0.0870 0.9378 0.5039 0.5404 0.9354 0.4734 0.4707 0.0791 0.0322 0.1441 -0.1886 0.0630 -0.0859 1.0000
25. The Difference of TP's Confidence in kMMT in General: Pre- - Post-testing -0.2317 -0.1870 -0.1939 -0.0054 -0.0534 -0.0168 0.2370 0.1124 0.0507 -0.0318 -0.0236 0.3730 0.1341 0.1116 -0.3276 -0.1404 -0.1517 -0.1248 0.1170 -0.3110 0.0555 -0.1602 0.0706 0.0276 1.0000
26. The Sum of TP's Confidence in Paired Practitioner: Pre- + Post-testing -0.2304 -0.1486 -0.2064 0.0598 0.0621 0.0262 0.0895 0.1170 0.1526 -0.0077 -0.2378 0.5143 1.0000 0.9018 0.4288 0.6652 0.5785 0.1290 0.0877 -0.0395 0.0807 -0.1164 0.1149 0.5039 0.1341 1.0000
27. The Difference of TP's Confidence in Paired Practitioner: Pre- - Post-testing -0.0798 -0.1913 -0.3168 -0.1674 0.0521 0.1380 0.0254 0.0194 -0.1639 -0.3047 -0.0406 0.1851 0.4713 0.3490 -0.0481 -0.3450 -0.4175 -0.0564 0.0432 0.0216 0.0837 0.1347 -0.1037 0.0742 0.3344 0.4713 1.0000
28. The Sum of TP's Confidence in Practitioner's kMMT: Pre- + Post-testing -0.0944 -0.1615 -0.2674 -0.1645 -0.0107 0.1720 -0.0410 0.0612 -0.1763 -0.3361 0.0704 0.1735 0.3608 0.4359 -0.0343 -0.3510 -0.4473 -0.0887 -0.0307 0.0349 0.1255 0.1215 -0.1272 0.0753 0.2982 0.3608 0.8681 1.0000
29. The Difference of TP's Confidence in Practitioner's kMMT: Pre- - Post-testing -0.2145 -0.0921 -0.1188 0.0483 -0.0111 -0.0053 0.0344 0.0675 0.1460 -0.0100 -0.1751 0.5404 0.9018 1.0000 0.4713 0.6642 0.6100 0.0863 0.0152 -0.0483 0.0358 -0.0907 0.0700 0.5404 0.1116 0.9018 0.3490 0.4359 1.0000
30. The Sum of ALL TP's Ranked Confidences -0.1783 -0.0634 -0.1120 0.0003 0.0948 -0.0637 0.0352 -0.0270 0.1261 -0.0904 -0.1944 0.8156 0.8904 0.8482 0.7610 0.6651 0.6095 0.1223 0.0718 0.0513 -0.0490 -0.0395 0.0265 0.8419 0.0983 0.8904 0.3336 0.2651 0.8482 1.0000
31. The Sum of ALL Ranked Confidences (Practitioner & TP) -0.2070 -0.1935 -0.0134 0.1818 0.3925 0.3794 0.6108 0.5039 0.6080 0.4575 -0.1355 0.5056 0.6791 0.6183 0.4320 0.5975 0.5395 0.0732 0.5820 -0.2988 0.4880 -0.1942 0.7435 0.5009 0.1170 0.6791 0.1479 0.0851 0.6183 0.6883
kMMT, kinesiology-style manual muscle testing; TP, Test Patient; = Correlation (r ) reached significance (p <0.05); = kMMT Accuracy - no correlations (r) reached significance.

© drannejensen 2013
 APPENDIX TABLE B.2.5 - kMMT accuracy by kMMT technique system - detailed. NOTE: Those p-values below the line may not be meaningful.

© drannejensen 2014
286
 APPENDIX TABLE B.2.6 - Table of correlations (and p-values) among kMMT accuracy and the difference in Test Patient confidence ratings (r).

© drannejensen 2014
287
 APPENDIX TABLE B.2.7 - Table of correlations (and p-values) among kMMT accuracy and the difference in Practitioner
confidence ratings (r).

© drannejensen 2014
288
 289

APPENDIX TABLE B.2.8 - kMMT accuracy for those Pairs in

Study 1 whose prevalence of Lies was 0.50.
95% CI
n Mean Upper Lower
kMMT Accuracy 9 0.654 0.567 0.741
sensitivity 9 0.649 0.560 0.737
specificity 9 0.693 0.577 0.808
PPV 9 0.754 0.616 0.893
NPV 9 0.554 0.361 0.746
kMMT, kinesiology-style Manual Muscle Testing; CI, Confidence Interval.

© drannejensen 2014
 290

APPENDIX FIGURE B.2.1 – Histograms showing normal distributions.

8
6
Density

4
2
0

0.20 0.30 0.40 0.50 0.60 0.70

Prevalence of False Statements (Blocks 1-4, n=48)

(A)
4
3
Density

2
1
0

.4 .5 .6 .7 .8 .9
kMMT Accuracy (Practitioner Blind, Blocks 1-4)

(B)

© drannejensen 2014
 291

APPENDIX FIGURE B.2.1 (con’t.)

Intuition Accuracy (Practitioner Blind; Blocks 1-4)

(C)
.05
.04
.03
Density

.02
.01
0

20 30 40 50 60 70
Practitioner's Age (years)

(D)

© drannejensen 2014
 292

APPENDIX FIGURE B.2.1 (con’t.)

.06
.04
Density

.02
0

0 10 20 30 40 50
Practitioner's Number of Years in Practice

(E)
.06
.04
Density

.02
0

0 10 20 30 40
Practitioner's Number of Years of kMMT Experience

(F)

© drannejensen 2014
 293

APPENDIX FIGURE B.2.1 (con’t.)

.2
.15
Density

.1
.05
0

0 5 10 15 20
Practitioner's usual number of hours per day performing kMMT

(G)
.4
.3
Density

.2
.1
0

4 6 8 10
Practitioner's Confidence in Own kMMT Ability - Pre-testing

(H)

© drannejensen 2014
 294

APPENDIX FIGURE B.2.1 (con’t.)

.4
.3
Density

.2
.1
0

2 4 6 8 10
Practitioner's Confidence in Own kMMT Ability - Post-testing

(I)
.2
.15
Density

.1
.05
0

10 12 14 16 18 20
The Sum of the Practitioner's Confidence in Own kMMT Ability: Pre + Post-testing

(J)

© drannejensen 2014
 295

APPENDIX FIGURE B.2.1 (con’t.)

.5
.4
.3
Density

.2
.1
0

-2 0 2 4 6
The Difference in Practitioner's Confidence in Own kMMT Ability: Pre - Post-testing

(K)
.8
.6
Density

.4
.2
0

4 6 8 10
Practitioner's Confidence in kMMT in General - Pre-testing

(L)

© drannejensen 2014
 296

APPENDIX FIGURE B.2.1 (con’t.)

.3
.2
Density

.1
0

10 12 14 16 18 20
The Sum of Practitioner's Confidence in kMMT in General: Pre + Post-testing

(M)
.6
.4
Density

.2
0

-4 -2 0 2 4 6
The Difference in Practitioner's Confidence in kMMT in General: Pre - Post-testing

(N)

© drannejensen 2014
 297

APPENDIX FIGURE B.2.1 (con’t.)

.4
.3
Density

.2
.1
0

0 2 4 6 8 10
Practitioner's Confidence in Own kMMT Ability with Paired TP - Post-testing

(O)
.05
.04
.03
Density

.02
.01
0

20 30 40 50 60
Test Patient's Age (years)

(P)

© drannejensen 2014
 298

APPENDIX FIGURE B.2.1 (con’t.)

.4
.3
Density

.2
.1
0

4 6 8 10
Test Patient's Confidence in kMMT in General - Pre-testing

(Q)
.25
.2
.15
Density

.1
.05
0

4 6 8 10
TP's Confidnce in kMMT in General - Post-testing

(R)

© drannejensen 2014
 299

APPENDIX FIGURE B.2.1 (con’t.)

.15
.1
Density

.05
0

5 10 15 20
The Sum of Test Patient's Confidence in kMMT in General: Pre + Post-testing

(S)
.6
.4
Density

.2
0

-4 -2 0 2 4
Difference in Test Patient's Confidence in kMMT in General: Pre - Post-testing

(T)

© drannejensen 2014
 300

APPENDIX FIGURE B.2.1 (con’t.)

.4
.3
Density

.2
.1
0

4 6 8 10
Test Patient's Confidence in Paired Practitioner - Pre-testing

(U)
.4
.3
Density

.2
.1
0

2 4 6 8 10
Test Patient's Confidence in Paired Practioner - Post-testing

(V)

© drannejensen 2014
 301

APPENDIX FIGURE B.2.1 (con’t.)

.25
.2
.15
Density

.1
.05
0

5 10 15 20
The Sum of the Test Patient's Confidence in Paired Practitioner: Pre + Post-testing

(W)
.25
.2
.15
Density

.1
.05
0

-4 -2 0 2 4 6
The Difference in the Test Patient's Confidence in Paired Practitioner: Pre - Post-testing

(X)

© drannejensen 2014
 302

APPENDIX FIGURE B.2.1 (con’t.)

.4
.3
Density

.2
.1
0

4 6 8 10
Test Patient's Confidence in Practitioner's kMMT Ability - Pre-testing

(Y)
.5
.4
.3
Density

.2
.1
0

0 2 4 6 8 10
Test Patient's Confidence in Practitioner's kMMT Ability - Post-testing

(Z)

© drannejensen 2014
 303

APPENDIX FIGURE B.2.1 (con’t.)

.2
.15
Density

.1
.05
0

5 10 15 20
The Sum of the Test Patient's Confidence in Practitioner's kMMT Ability: Pre + Post-testing

(AA)
.3
.2
Density

.1
0

-5 0 5 10
The Difference in Test Patient's Confidence in Practitioner's kMMT Ability: Pre - Post-testing

(AB)

© drannejensen 2014
 304

APPENDIX FIGURE B.2.1 (con’t.)

.06
.04
Density

.02
0

20 30 40 50 60
The Sum of all Practitioner-ranked Confidences

(AC)
.08
.06
Density

.04
.02
0

15 20 25 30 35 40
The Sum of All Test-Patient-ranked Confidences

(AD)

© drannejensen 2014
 305

APPENDIX FIGURE B.2.1 (con’t.)

.05
.04
.03
Density

.02
.01
0

50 60 70 80 90 100
The Sum of all Confidence Scores - both Practitioner-ranked & TP-ranked

(AE)

© drannejensen 2014
 306

APPENDIX FIGURE B.2.2 – Scatterplots showing correlations and the

correlation coefficients (r).
.9
.8
MMTAccuracy

.7
.6
k
.5

r = -0.0175
.4

20 30 40 50 60 70
Practitioner's Age (years)

(A)
.9

r = -0.0175
.8
MMTAccuracy

.7
.6
k
.5
.4

0 10 20 30 40
Years of kMMTExperience

(B)

© drannejensen 2014
 307

APPENDIX FIGURE B.2.2 (con’t.)

r = -0.0080

(C)
.9
.8

r = -0.0133
MMTAccuracy

.7
.6
k
.5
.4

0 5 10 15 20
Average hrs/day using kMMT

(D)

© drannejensen 2014
 308

APPENDIX FIGURE B.2.2 (con’t.)

.9
.8

r = -0.0043
MMTAccuracy

.7
.6
k
.5
.4

4 6 8 10
Practitioner's Confidence in Own kMMT Ability (Pre-testing)

(E)
.9
.8

r = -0.0811
MMTAccuracy

.7
.6
k
.5
.4

4 6 8 10
Practitioner's Confidence in kMMT in General (Pre-testing)

(F)

© drannejensen 2014
 309

APPENDIX FIGURE B.2.2 (con’t.)

.9

r = 0.0407
.8
MMTAccuracy

.7
.6
k
.5
.4

2 4 6 8 10
Practitioners Confidence in Own kMMT Ability (Post-testing)

(G)
.9

r = 0.0546
.8
MMTAccuracy

.7
.6
k
.5
.4

2 4 6 8 10
Practitioner's Confidence in kMMT in General (Post-testing)

(H)

© drannejensen 2014
 310

APPENDIX FIGURE B.2.2 (con’t.)

.9
.8
MMTAccuracy

.7
.6
k
.5

r = 0.0602
.4

20 30 40 50 60
TP Age (years)

(I)
.9

r = 0.0300
.8
MMTAccuracy

.7
.6
k
.5
.4

4 6 8 10
TP Confidence in kMMT in General (Pre-testing)

(J)

© drannejensen 2014
 311

APPENDIX FIGURE B.2.2 (con’t.)

.9

r = 0.1189
.8
MMTAccuracy

.7
.6
k
.5
.4

4 6 8 10
TP Confidence in kMMT in General (Post-testing)

(K)
.9

r = 0.1290
.8
MMTAccuracy

.7
.6
k
.5
.4

4 6 8 10
TP Confidence in Practitioner (Pre-testing)

(L)

© drannejensen 2014
 312

APPENDIX FIGURE B.2.2 (con’t.)

.9

r = 0.1851
.8
MMTAccuracy

.7
.6
k
.5
.4

2 4 6 8 10
TP Confidence in Practitioner (Post-testing)

(M)
.9

r = 0.0573
.8
MMTAccuracy

.7
.6
k
.5
.4

4 6 8 10
TP's Confidence in Practitioner's kMMT Ability - Pre-Testing

(N)

© drannejensen 2014
 313

APPENDIX FIGURE B.2.2 (con’t.)

.9

r = 0.1639
.8
MMTAccuracy

.7
.6
k
.5
.4

0 2 4 6 8 10
TP Confidence in Practitioner's kMMT (Post-testing)

(O)

(P) Correlation of kMMT accuracy and Intuition accuracy,

with 95% confidence intervals and fitted values

© drannejensen 2014
 314

APPENDIX FIGURE B.2.2 (con’t.)

1
.8
.6

r = -0.1761
.4

0 .2 .4 .6 .8 1
kMMT Accuracy for False Statements

(Q)

© drannejensen 2014
 315

CHAPTER 3
APPENDIX TABLE B.3.1 - 2x2 Tables for kMMT for each Pair (n=20). Each Pair
performed 40 kMMTs.

© drannejensen 2014
 316

APPENDIX TABLE B.3.1 (con’t.)

© drannejensen 2014
 APPENDIX TABLE B.3.2 - kMMT & Intuition accuracies for all statements, True statements and False statements (for n=20 Pairs).

© drannejensen 2014
317
 318

APPENDIX TABLE B.3.3 - Correlations among kMMT and Confidence Ratings.

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16.

1. Practitioner Confidence in OWN kMMT Ability: Pre-testing 1.0000

2. Practitioner Confidence in OWN kMMT Ability: Post-testing 0.5674 1.0000

0.01
3. Practitioner Confidence in kMMT in General: Pre-testing 0.2221 0.2341 1.0000
0.35 0.32
4. Practitioner Confidence in kMMT in General: Post-testing 0.0942 0.3186 0.8279 1.0000
0.69 0.17 0.00
5. TP Confidence in kMMT in General: Pre-testing -0.1099 -0.1505 -0.4230 -0.2232 1.0000
0.64 0.53 0.06 0.34
6. TP Confidence in Practitioner: Pre-testing 0.2042 0.1625 -0.0989 -0.0069 0.5278 1.0000
0.3878 0.4935 0.6782 0.9768 0.0168
7. TP Confidence in Practitioner's kMMT Ability: Pre-testing 0.1331 0.0571 -0.2041 -0.1513 -0.0309 -0.0418 1.0000
0.58 0.81 0.39 0.52 0.90 0.86
8. TP Confidence in kMMT in General: Post-testing 0.1815 0.0699 -0.2721 0.0718 0.3435 0.1094 -0.2126 1.0000
0.44 0.77 0.25 0.76 0.14 0.65 0.37
9. TP Confidence in Practitioner: Post-testing -0.0268 0.0619 0.0207 0.2981 0.3493 0.3288 -0.1304 0.6457 1.0000
0.91 0.80 0.93 0.20 0.13 0.16 0.58 <0.01
10. TP Confidence in Practitioner's kMMT Ability: Post-testing 0.1945 0.2371 0.0934 0.2836 0.2971 0.5095 -0.1297 0.5842 0.7810 1.0000
0.41 0.31 0.70 0.23 0.20 0.02 0.59 0.01 <0.01
11. Difference in Practitioner Confidence in OWN kMMT Ability: Post- – Pre-testing -0.2919 0.6220 0.0608 0.2806 -0.0703 -0.0054 -0.0602 -0.0914 0.0973 0.0905 1.0000
0.21 <0.01 0.80 0.23 0.77 0.98 0.80 0.70 0.68 0.70
12. Difference in Practitioner Confidence in kMMT in General: Post- – Pre-testing -0.2560 0.0563 -0.5520 0.0106 0.4224 0.1661 0.1389 0.5919 0.4063 0.2551 0.3088 1.0000
0.28 0.81 0.01 0.96 0.06 0.48 0.56 0.01 0.08 0.28 0.19
13. Difference in TP Confidence in kMMT in General: Post- – Pre-testing 0.2523 0.1943 0.1497 0.2611 -0.6074 -0.3815 -0.1521 0.5374 0.2325 0.2274 -0.0142 0.1213 1.0000
0.28 0.41 0.53 0.27 <0.01 0.10 0.52 0.01 0.32 0.34 0.95 0.61
14. Difference in TP Confidence in Practitioner: Post- – Pre-testing -0.1819 -0.0665 0.0955 0.2860 -0.0764 -0.4581 -0.0907 0.5238 0.6889 0.3440 0.0958 0.2549 0.5117 1.0000
0.44 0.78 0.69 0.22 0.75 0.04 0.70 0.02 <0.01 0.14 0.69 0.28 0.02
15. Difference in TP Confidence in Practitoner's kMMT Ability: Post- – Pre-testing -0.0738 0.0081 0.2146 0.2145 0.1057 0.1709 -0.9665 0.3496 0.3238 0.3799 0.0796 -0.0636 0.2007 0.1737 1.0000
0.76 0.97 0.36 0.36 0.66 0.47 <0.01 0.13 0.16 0.10 0.74 0.79 0.40 0.46
16. kMMT Accuracy 0.0885 -0.1439 0.2757 -0.0763 -0.2971 0.0151 -0.3695 -0.1451 -0.2646 0.0379 -0.2513 -0.6049 0.1441 -0.2606 0.3546 1.0000
0.71 0.55 0.24 0.75 0.20 0.95 0.11 0.54 0.26 0.87 0.29 <0.01 0.54 0.27 0.13
kMMT, kinesiology-style manual muscle testing; TP, Test Patient; = Correlation (r ) reached significance (p <0.05); = kMMT Accuracy - no correlations (r) reached significance.

© drannejensen 2013
 319

APPENDIX FIGURE B.3.1 - Histograms showing normal distribution.

4
3
Density

2
1
0

.4 .5 .6 .7 .8
kMMT Accuracy

(A)

(B)

© drannejensen 2014
 320

APPENDIX FIGURE B.3.1 (con’t.)

4
3
Density

2
1
0

.2 .4 .6 .8 1
kMMT Accuracy for TRUE Statements

(C)
2.5
2
1.5
Density

1
.5
0

.2 .4 .6 .8
kMMT Accuracy for FALSE Statements

(D)

© drannejensen 2014
 321

APPENDIX FIGURE B.3.1 (con’t.)

(E)

(F)

© drannejensen 2014
 322

APPENDIX FIGURE B.3.1 (con’t.)

4
3
Density

2
1
0

.2 .4 .6 .8 1
kMMT Accuracy during Block 1

(G)
3
2
Density

1
0

.3 .4 .5 .6 .7 .8
kMMT Accuracy during Block 2

(H)

© drannejensen 2014
 323

APPENDIX FIGURE B.3.1 (con’t.)

2
1.5
Density

1
.5
0

.2 .4 .6 .8 1
kMMT Accuracy during Block 4
(I)
2.5
2
1.5
Density

1
.5
0

.2 .4 .6 .8 1
kMMT Accuracy during Block 3

(J)

© drannejensen 2014
 CHAPTER 4

APPENDIX TABLE B.4.1 - Mean grip strengths (SD) by participant. (A) False vs. True statements, and (B) dominant hand vs. non-dominant hand.

© drannejensen 2014
324
 325

APPENDIX FIGURE B.4.1 – Histograms.

5
4
Frequency

3
2
1
0

10 20 30 40
Mean Grip Strength (kg) after False Statements
4
3
Frequency

2
1
0

10 20 30 40
Mean Grip Strength (kg) after True Statements

© drannejensen 2014
 326

APPENDIX TABLE B.5.1- Accuracy Data for each Pair: Accuracy (Overall Percent Correct), Sensitivity, Specificity, PPV and NPV; for kMMT and Intuition.

kMMT Intution
Pair TP Practitioner Accuracy* Sensitivity Specificity PPV NPV Accuracy* Sensitivity Specificity PPV NPV
1 1 1 0.750 0.600 0.900 0.857 0.692 0.400 0.300 0.500 0.375 0.417
2 1 2 0.350 0.500 0.200 0.385 0.286 0.450 0.600 0.300 0.462 0.429
3 1 3 0.750 0.600 0.900 0.857 0.692 0.350 0.300 0.400 0.333 0.364
4 1 4 0.550 0.600 0.500 0.545 0.556 0.550 0.500 0.600 0.556 0.545
5 1 5 0.900 1.000 0.800 0.833 1.000 0.600 0.500 0.700 0.625 0.583
6 1 6 0.950 0.900 1.000 1.000 0.909 0.600 0.700 0.500 0.583 0.625
7 1 7 0.850 0.800 0.900 0.889 0.818 0.450 0.400 0.500 0.444 0.455
8 1 8 1.000 1.000 1.000 1.000 1.000 0.500 0.400 0.600 0.500 0.500
9 1 9 0.500 0.300 0.700 0.500 0.500 0.700 0.500 0.900 0.833 0.643
10 1 10 0.450 0.400 0.500 0.444 0.455 0.450 0.400 0.500 0.444 0.455
11 1 11 0.500 0.600 0.400 0.500 0.500 0.450 0.400 0.500 0.444 0.455
12 1 12 0.950 0.900 1.000 1.000 0.909 0.500 0.400 0.600 0.500 0.500
13 1 13 0.900 1.000 0.800 0.833 1.000 0.450 0.444 0.455 0.400 0.500
14 1 14 1.000 1.000 1.000 1.000 1.000 0.650 0.600 0.700 0.667 0.636
15 1 15 0.600 0.200 1.000 1.000 0.556 0.400 0.300 0.500 0.375 0.417
16 1 16 0.650 0.500 0.800 0.714 0.615 0.550 0.500 0.600 0.556 0.545
17 2 1 0.300 0.300 0.300 0.300 0.300 0.600 0.700 0.500 0.583 0.625
18 2 2 0.400 0.500 0.300 0.417 0.375 0.450 0.222 0.636 0.333 0.500
19 2 3 0.650 0.600 0.700 0.667 0.636 0.250 0.200 0.300 0.222 0.273
20 2 4 0.450 0.778 0.182 0.438 0.500 0.500 0.500 0.500 0.500 0.500
21 2 5 0.550 0.700 0.400 0.538 0.571 0.550 0.556 0.545 0.500 0.600
22 2 6 0.650 0.667 0.636 0.600 0.700 0.500 0.700 0.300 0.500 0.500
23 2 7 0.650 0.500 0.800 0.714 0.615 0.500 0.500 0.500 0.500 0.500
24 2 8 0.800 1.000 0.600 0.714 1.000 0.700 0.700 0.700 0.700 0.700
25 2 9 0.600 0.778 0.455 0.538 0.714 0.500 0.333 0.636 0.429 0.538
26 2 10 0.250 0.300 0.200 0.273 0.222 0.500 0.400 0.600 0.500 0.500
27 2 11 0.450 0.800 0.100 0.471 0.333 0.550 0.500 0.600 0.556 0.545
28 2 12 0.700 0.700 0.700 0.700 0.700 0.450 0.500 0.400 0.455 0.444
29 2 13 0.900 0.900 0.900 0.900 0.900 0.550 0.400 0.700 0.571 0.538
30 2 14 0.700 0.778 0.636 0.636 0.778 0.150 0.200 0.100 0.182 0.111
31 2 15 0.550 0.700 0.400 0.538 0.571 0.250 0.200 0.300 0.222 0.273
32 2 16 0.550 0.375 0.667 0.429 0.615 0.500 0.300 0.700 0.500 0.500
33 3 1 0.600 0.700 0.500 0.583 0.625 0.500 0.500 0.500 0.500 0.500
34 3 2 0.600 0.500 0.700 0.625 0.583 0.350 0.300 0.400 0.333 0.364
35 3 3 0.550 0.300 0.800 0.600 0.533 0.300 0.200 0.400 0.250 0.333
36 3 4 0.800 0.800 0.800 0.800 0.800 0.600 0.500 0.700 0.625 0.583
37 3 5 0.750 0.600 0.900 0.857 0.692 0.650 0.500 0.800 0.714 0.615
38 3 6 0.600 0.300 0.900 0.750 0.563 0.400 0.444 0.364 0.364 0.444
39 3 7 0.600 0.400 0.800 0.667 0.571 0.500 0.400 0.600 0.500 0.500
40 3 8 0.850 0.800 0.900 0.889 0.818 0.550 0.400 0.700 0.571 0.538
41 3 9 0.450 0.200 0.700 0.400 0.467 0.450 0.500 0.400 0.455 0.444
42 3 10 0.700 0.500 0.900 0.833 0.643 0.550 0.500 0.600 0.556 0.545
43 3 11 0.800 0.667 0.909 0.857 0.769 0.450 0.300 0.600 0.429 0.462
44 3 12 0.800 0.667 0.909 0.857 0.769 0.400 0.500 0.300 0.417 0.375
45 3 13 0.700 0.500 0.900 0.833 0.643 0.650 0.500 0.800 0.714 0.615
46 3 14 0.850 0.700 1.000 1.000 0.769 0.650 0.800 0.500 0.615 0.714
47 3 15 0.750 0.600 0.900 0.857 0.692 0.400 0.300 0.500 0.375 0.417
48 3 16 0.350 0.200 0.500 0.286 0.385 0.500 0.200 0.800 0.500 0.500
49 4 1 0.650 0.700 0.600 0.636 0.667 0.450 0.600 0.300 0.462 0.429
50 4 2 0.600 0.700 0.500 0.583 0.625 0.600 0.400 0.800 0.667 0.571
51 4 3 0.550 0.400 0.700 0.571 0.538 0.450 0.300 0.600 0.429 0.462
52 4 4 0.800 0.700 0.900 0.875 0.750 0.400 0.400 0.400 0.400 0.400
53 4 5 0.500 0.600 0.400 0.500 0.500 0.450 0.300 0.600 0.429 0.462
54 4 6 0.650 0.600 0.700 0.667 0.636 0.650 0.600 0.700 0.667 0.636
55 4 7 0.650 0.600 0.700 0.667 0.636 0.550 0.600 0.500 0.545 0.556
56 4 8 0.750 0.800 0.700 0.727 0.778 0.600 0.600 0.600 0.600 0.600
57 4 9 0.500 0.500 0.500 0.500 0.500 0.600 0.600 0.600 0.600 0.600
58 4 10 0.500 0.400 0.600 0.500 0.500 0.500 0.400 0.600 0.500 0.500
59 4 11 0.700 0.600 0.800 0.750 0.667 0.400 0.222 0.545 0.286 0.462
60 4 12 0.600 0.500 0.700 0.625 0.583 0.650 0.600 0.700 0.667 0.636
61 4 13 0.400 0.100 0.700 0.250 0.438 0.800 0.700 0.900 0.875 0.750
62 4 14 0.450 0.300 0.600 0.429 0.462 0.450 0.444 0.455 0.400 0.500
63 4 15 0.450 0.200 0.700 0.400 0.467 0.600 0.600 0.600 0.600 0.600
64 4 16 0.350 0.200 0.500 0.286 0.385 0.600 0.500 0.700 0.625 0.583
65 5 1 0.300 0.222 0.364 0.222 0.364 0.550 0.700 0.400 0.538 0.571
66 5 2 0.500 0.600 0.400 0.500 0.500 0.550 0.600 0.500 0.545 0.556
67 5 3 0.450 0.700 0.200 0.467 0.400 0.600 0.400 0.800 0.667 0.571
68 5 4 0.400 0.700 0.100 0.438 0.250 0.600 0.600 0.600 0.600 0.600
69 5 5 0.350 0.500 0.200 0.385 0.286 0.600 0.600 0.600 0.600 0.600
70 5 6 0.600 0.500 0.700 0.625 0.583 0.500 0.500 0.500 0.500 0.500
71 5 7 0.350 0.400 0.300 0.364 0.333 0.600 0.400 0.800 0.667 0.571
72 5 8 0.400 0.800 0.000 0.444 0.000 0.400 0.300 0.500 0.375 0.417
73 5 9 0.450 0.600 0.300 0.462 0.429 0.250 0.300 0.200 0.273 0.222
74 5 10 0.600 0.444 0.727 0.571 0.615 0.450 0.400 0.500 0.444 0.455
75 5 11 0.300 0.556 0.091 0.333 0.200 0.500 0.400 0.600 0.500 0.500
76 5 12 0.500 0.400 0.600 0.500 0.500 0.350 0.200 0.500 0.286 0.385
77 5 13 0.550 0.375 0.667 0.429 0.615 0.300 0.100 0.500 0.167 0.357
78 5 14 0.350 0.500 0.200 0.385 0.286 0.450 0.500 0.400 0.455 0.444
79 5 15 0.150 0.100 0.200 0.111 0.182 0.450 0.400 0.500 0.444 0.455
80 5 16 0.450 0.200 0.700 0.400 0.467 0.450 0.300 0.600 0.429 0.462
81 6 1 0.700 0.778 0.636 0.636 0.778 0.750 0.800 0.700 0.727 0.778
82 6 2 0.550 0.700 0.400 0.538 0.571 0.600 0.400 0.800 0.667 0.571
83 6 3 0.850 0.800 0.900 0.889 0.818 0.400 0.300 0.500 0.375 0.417
84 6 4 0.700 0.889 0.545 0.615 0.857 0.600 0.500 0.700 0.625 0.583
85 6 5 0.550 0.700 0.400 0.538 0.571 0.450 0.300 0.600 0.429 0.462
86 6 6 0.850 0.900 0.800 0.818 0.889 0.500 0.500 0.500 0.500 0.500
87 6 7 0.850 0.800 0.900 0.889 0.818 0.600 0.400 0.800 0.667 0.571
88 6 8 0.950 1.000 0.900 0.909 1.000 0.500 0.333 0.636 0.429 0.538
89 6 9 0.800 0.700 0.900 0.875 0.750 0.600 0.667 0.545 0.545 0.667
90 6 10 0.650 0.500 0.800 0.714 0.615 0.400 0.444 0.364 0.364 0.444
91 6 11 0.800 1.000 0.600 0.714 1.000 0.700 0.800 0.600 0.667 0.750
92 6 12 0.850 0.700 1.000 1.000 0.769 0.600 0.500 0.700 0.625 0.583
93 6 13 0.800 0.600 1.000 1.000 0.714 0.450 0.400 0.500 0.444 0.455
94 6 14 0.600 0.600 0.600 0.600 0.600 0.650 0.900 0.400 0.600 0.800
95 6 15 0.600 0.700 0.500 0.583 0.625 0.550 0.500 0.600 0.556 0.545
96 6 16 0.700 0.500 0.900
0.900 0.833 0.643 0.550 0.444 0.636 0.500 0.583
Mean 0.616 0.595 0.638 0.632 0.609 0.507 0.456 0.557 0.502 0.514
SD 0.186 0.224 0.258 0.215 0.207 0.114 0.155 0.150 0.135 0.112
95% CI 0.578 - 0.654 0.549 - 0.640 0.585 - 0.690 0.588 - 0.676 0.567 - 0.652 0.484 - 0.530 0.424 - 0.487 0.527 - 0.588 0.475 - 0.530 0.491 - 0.537
* Accuracy as Overall Percent Correct; kMMT, kinesiology-style manual muscle testing; TP, Test Patient; PPV, Positive Predictive Value; NPV, Negative Predictive Value; SD, Standard Deviation; CI, Confidence Interval.

© drannejensen 2013
 327

APPENDIX TABLE B.5.2 - ANOVA results : Practitioner, TP & Block.

Source Partial SS df F Prob>F %

Model 3.7930 21 6.31 <0.01 43.8%

Practitioner 1.4391 15 3.35 <0.01 16.6%

TP 2.3534 5 16.44 <0.01 27.2%
Block 0.0005 1 0.02 0.90 0.0%

Residual 4.8670 170 56.2%

Total 8.6599 191 100.0%

© drannejensen 2014
 328

APPENDIX FIGURE B.5.1 - Repeatability Scatterplots : Block 1 vs Block 2 – by

Practitioner (#1-16).
1
.9
.8
.7
.6
.5
.4
.3
.2
.1
0

0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1
Mean kMMT Accuracy Block1 - Practitioner1
1
.9
.8
.7
.6
.5
.4
.3
.2
.1
0

0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1
Mean kMMT Accuracy Block1 - Practitioner2

© drannejensen 2014
 329

APPENDIX FIGURE B.5.1 (con’t.)

1
.9
.8
.7
.6
.5
.4
.3
.2
.1
0

0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1
Mean kMMT Accuracy Block1 - Practitioner3
1
.9
.8
.7
.6
.5
.4
.3
.2
.1
0

0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1
Mean kMMT Accuracy Block1 - Practitioner4

© drannejensen 2014
 330

APPENDIX FIGURE B.5.1 (con’t.)

1
.9
.8
.7
.6
.5
.4
.3
.2
.1
0

0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1
Mean kMMT Accuracy Block1 - Practitioner5
1
.9
.8
.7
.6
.5
.4
.3
.2
.1
0

0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1
Mean kMMT Accuracy Block1 - Practitioner6

© drannejensen 2014
 331

APPENDIX FIGURE B.5.1 (con’t.)

1
.9
.8
.7
.6
.5
.4
.3
.2
.1
0

0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1
Mean kMMT Accuracy Block1 - Practitioner7
1
.9
.8
.7
.6
.5
.4
.3
.2
.1
0

0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1
Mean kMMT Accuracy Block1 - Practitioner8

© drannejensen 2014
 332

APPENDIX FIGURE B.5.1 (con’t.)

1
.9
.8
.7
.6
.5
.4
.3
.2
.1
0

0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1
Mean kMMT Accuracy Block1 - Practitioner9
1
.9
.8
.7
.6
.5
.4
.3
.2
.1
0

0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1
Mean kMMT Accuracy Block1 - Practitioner10

© drannejensen 2014
 333

APPENDIX FIGURE B.5.1 (con’t.)

1
.9
.8
.7
.6
.5
.4
.3
.2
.1
0

0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1
Mean kMMT Accuracy Block1 - Practitioner11
1
.9
.8
.7
.6
.5
.4
.3
.2
.1
0

0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1
Mean kMMT Accuracy Block1 - Practitioner12

© drannejensen 2014
 334

APPENDIX FIGURE B.5.1 (con’t.)

1
.9
.8
.7
.6
.5
.4
.3
.2
.1
0

0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1
Mean kMMT Accuracy Block1 - Practitioner13
1
.9
.8
.7
.6
.5
.4
.3
.2
.1
0

0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1
Mean kMMT Accuracy Block1 - Practitioner14

© drannejensen 2014
 335

APPENDIX FIGURE B.5.1 (con’t.)

1
.9
.8
.7
.6
.5
.4
.3
.2
.1
0

0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1
Mean kMMT Accuracy Block1 - Practitioner15
1
.9
.8
.7
.6
.5
.4
.3
.2
.1
0

0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1
Mean kMMT Accuracy Block1 - Practitioner16

© drannejensen 2014
 336

APPENDIX FIGURE B.5.2 - Repeatability scatterplots : Block 1 vs Block 2 – by TP (#1-6).

1
.9
.8
.7
.6
.5
.4
.3
.2
.1
0

0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1
Mean kMMT Accuracy Block1 – TP1
1
.9
.8
.7
.6
.5
.4
.3
.2
.1
0

0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1
Mean kMMT Accuracy Block1 – TP2

© drannejensen 2014
 337

APPENDIX FIGURE B.5.2 (con’t.)

1
.9
.8
.7
.6
.5
.4
.3
.2
.1
0

0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1
Mean kMMT Accuracy Block1 – TP3
1
.9
.8
.7
.6
.5
.4
.3
.2
.1
0

0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1
Mean kMMT Accuracy Block1 – TP4

© drannejensen 2014
 338

APPENDIX FIGURE B.5.2 (con’t.)

1
.9
.8
.7
.6
.5
.4
.3
.2
.1
0

0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1
Mean kMMT Accuracy Block1 – TP5
1
.9
.8
.7
.6
.5
.4
.3
.2
.1
0

0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1
Mean kMMT Accuracy Block1 – TP6

© drannejensen 2014
 CHAPTER 6

APPENDIX TABLE B.6.1 - kMMT & Intuition accuracies for all statements compared to True and False statements. Using all stimuli, for n=20 Pairs.

© drannejensen 2014
339
 340

APPENDIX TABLE B.6.2 - 2x2 Tables for kMMT for each Pair (n=20). Each Pair performed
40 kMMTs. Using (A) emotionally-arousing, and (B) affect-neutral stimuli.

© drannejensen 2014
 341

APPENDIX TABLE B.6.2 (cont’d.)

© drannejensen 2014
 342

APPENDIX TABLE B.6.2 (cont’d.)

© drannejensen 2014
 343

APPENDIX TABLE B.6.2 (cont’d.)

© drannejensen 2014
APPENDIX TABLE B.6.3 - Correlations amoung accuracy scores. kMMT & Guessing, Emotionally-arousing & Affect-Neutral Stimuli.

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11.
1. kMMT Accuracy using Emotionally-Arousing Pictures 1.0000

2. kMMT Accuracy using Affect-Neutral Pictures 0.7599 1.0000

<0.01
3. Guessing Accuracy using Emotionally-Arousing Pictures -0.2222 0.0124 1.0000
0.35 0.96
4. Guessing Accuracy using Affect-Neutral Pictures 0.3030 0.2399 0.1294 1.0000
0.19 0.31 0.59
5. kMMT Accuracy using Emotionally-Arousing Spoken Words 0.9565 0.8385 -0.2030 0.2305 1.0000
<0.01 <0.01 0.39 0.33
6. kMMT Accuracy using Affect-Neutral Spoken Words 0.7997 0.9746 0.0130 0.3044 0.8135 1.0000
<0.01 <0.01 0.96 0.19 <0.01
7. Guessing Accuracy using Emotionally-Arousing Spoken Words 0.0500 0.0679 0.5960 0.5766 0.0255 0.0922 1.0000
0.83 0.78 0.01 0.01 0.92 0.70
8. Guessing Accuracy using Affect-Neutral Spoken Words 0.0969 0.2161 0.3938 0.6435 0.0569 0.2649 0.0991 1.0000
0.68 0.36 0.09 <0.01 0.81 0.26 0.68
9. kMMT Accuracy using Combined Emotionally-Arousing Simuli 0.9668 0.8061 -0.1876 0.1811 0.9713 0.8110 -0.0089 0.0515 1.0000
<0.01 <0.01 0.43 0.44 <0.01 <0.01 0.97 0.83
10. kMMT Accuracy using Combined Affect-Neutral Stimuli 0.8124 0.9832 -0.0214 0.3296 0.8553 0.9832 0.1077 0.2459 0.8119 1.0000
<0.01 <0.01 0.93 0.16 <0.01 <0.01 0.65 0.30 <0.01
11. Guessing Accuracy using Combined Emotionally-Arousing Stimuli -0.0419 0.2366 0.8763 0.3672 0.0049 0.2222 0.7703 0.3651 -0.0107 0.2087 1.0000
0.86 0.32 <0.01 0.11 0.98 0.35 <0.01 0.11 0.96 0.38
12. Guessing Accuracy using Combined Affect-Neutral Stimuli 0.2013 0.0290 0.0852 0.8880 0.0808 0.1219 0.3625 0.7116 0.0540 0.1353 0.1137
0.39 0.90 0.72 <0.01 0.73 0.61 0.12 <0.01 0.82 0.57 0.63
kMMT, kinesiology-style Manual Muscle Testing; = Significance reached.

© drannejensen 2013
 345

APPENDIX FIGURE B.6.1 – Histograms for overall kMMT & Intuition accuracies.

8
6
Frequency

4
2
0

.4 .6 .8 1
kMMT Accuracy - ALL stimuli
6
4
Frequency

2
0

.4 .45 .5 .55 .6 .65

Guessing Accuracy - ALL stimuli

© drannejensen 2014
 346

APPENDIX FIGURE B.6.2 – Histograms of the distribution of (A) picture and (B) word arousal
levels – showing normal distributions.

(A) (B)

© drannejensen 2014
 347

APPENDIX FIGURE B.6.3 – Scatterplots. * Reached significance.

r = 0.1581
p = 0.51
Intuition Accuracy – All stimuli

Intuition Accuracy

(A)
Intuition Accuracy using Affect-Neutral Pictures

r = -0.2222
p = 0.35

Intuition

(B)

© drannejensen 2014
 348

APPENDIX FIGURE B.6.3 (cont’d.)

Intuition Accuracy using Emotionally-Arousing Pictures

r = 0.2399
p = 0.31

Intuition

(C)
1

r = 0.7599
p < 0.01*
.8
.6
.4
.2

.2 .4 .6 .8 1
kMMT Accuracy using Emotionally-Arousing Pictures

Fitted values kMMT Accuracy using Affect-Neutral Pictures

(D)

© drannejensen 2014
 349

APPENDIX FIGURE B.6.3 (cont’d.)

r = 0.1294
p = 0.59
Intuition Accuracy using Affect-Neutral Pictures

Intuition Accuracy using Emotionally-Arousing Pictures

Intuition

(E)

r = 0.2649
p = 0.26
Intuition

Intuition

(F)

© drannejensen 2014
 350

APPENDIX FIGURE B.6.3 (cont’d.)

r = 0.0255
p = 0.92
Intuition

Intuition

(G)
1

r = 0.8135
p < 0.01*
.8
.6
.4
.2

.4 .6 .8 1
kMMT Accuracy using Emotionally-Arousing Spoken Words

Fitted values kMMT Accuracy using Affect-Neutral Spoken Words

(H)

© drannejensen 2014
 351

APPENDIX FIGURE B.6.3 (cont’d.)

r = 0.0991
p = 0.68
Intuition

Intuition
Intuition

(I)

r = 0.1353
p = 0.57
Intuition

Intuition

(J)

© drannejensen 2014
 352

APPENDIX FIGURE B.6.3 (cont’d.)

r = -0.0107
p = 0.96
Intuition

Intuition

(K)
1
.8
.6
.4

r = 0.8119
p < 0.01*
.2

.4 .6 .8 1
kMMT Accuracy using Combined Emotionally-Arousing Stimuli

Fitted values kMMT Accuracy using Combined Affect-Neutral Stimuli

(L)

© drannejensen 2014
 353

APPENDIX FIGURE B.6.3 (cont’d.)

r = 0.1137
Intuition

p = 0.63

Intuition

Intuition

(M)
50

r = -0.0772
p = 0.75
40
30
20
10
0

.4 .6 .8 1
kMMT Accuracy using Combined Emotionally-Arousing Stimuli

Fitted values Years in Profession

(N)

© drannejensen 2014
 354

APPENDIX FIGURE B.6.3 (cont’d.)

50

r = -0.0519
p = 0.83
40
30
20
10
0

.2 .4 .6 .8 1
kMMT Accuracy using Combined Affect-Neutral Stimuli

Fitted values Years in Profession

(O)
40

r = -0.0663
p = 0.78
30
20
10
0

.4 .6 .8 1
kMMT Accuracy using Combined Emotionally-Arousing Stimuli

Fitted values Years of kMMT Experience

(P)

© drannejensen 2014
 355

APPENDIX FIGURE B.6.3 (cont’d.)

40

r = 0.0261
p = 0.91
30
20
10
0

.2 .4 .6 .8 1
kMMT Accuracy using Combined Affect-Neutral Stimuli

Fitted values Years of kMMT Experience

(Q)
40

r = 0.1238
p = 0.60
30
20
10
0

.4 .6 .8 1
kMMT Accuracy using Combined Emotionally-Arousing Stimuli

Fitted values Usual Hours/Day Using kMMT

(R)

© drannejensen 2014
 356

APPENDIX FIGURE B.6.3 (cont’d.)

40
30

r = 0.0527
p = 0.83
20
10
0

.2 .4 .6 .8 1
kMMT Accuracy using Combined Affect-Neutral Stimuli

Fitted values Usual Hours/Day using kMMT

(S)
10

r = -0.1153
p = 0.63
8
6
4
2
0

.4 .6 .8 1
kMMT Accuracy using Combined Emotionally-Arousing Stimuli

Fitted values Practitioner Anxiety

(T)

© drannejensen 2014
 357

APPENDIX FIGURE B.6.3 (cont’d.)

10

r = -0.1543
p = 0.52
8
6
4
2
0

.2 .4 .6 .8 1
kMMT Accuracy using Combined Affect-Neutral

Fitted values Practitioner Anxiety

(U)

© drannejensen 2014
 358

CHAPTER 7
APPENDIX FIGURE B.7.1 : Additional Forest Plots.

(A) Forest Plot for kMMT Accuracies.

(B) Forest Plot for Intuition Accuracies.

© drannejensen 2014
 359

APPENDIX C
The Prevalence of Use of kMMT

“There may be fairies at the bottom of the garden. There is no evidence for it, but you can't

prove that there aren't any, so shouldn't we be agnostic with respect to fairies?”

Richard Dawkins

© drannejensen 2014
 360

APPENDIX C : THE PREVALENCE OF USE OF KMMT

ABSTRACT

Research Objectives: To investigate the prevalence of use of kinesiology-style manual

muscle testing (kMMT).

Methods: First, a search of internet databases, textbooks and expert opinion were used to

compile a list of known kMMT technique systems. Then, direct contact was attempted via

email and telephone to representatives of each individual kMMT technique system. Once

contacted, the representative was asked to provide a conservative estimate of the number of

people trained in their form of kMMT. For those organisations unable to provide an estimate,

additional expert opinion was sought to approximate the numbers trained.

Results: Seventy-nine kMMT technique systems were identified, 46 of which provided the

requested estimate and 33 did not (for various reasons). From the information collected,

kMMT was estimated to be used by over 1 million people worldwide.

Summary: With over 1 million people trained worldwide, the widespread use of kMMT

merits further consideration, and proper exploration of its usefulness in clinical settings. This

estimation might be amplified due to the possibility of redundancies or attrition. Likewise, it

might be low due to misclassification or too narrow search methods.

Keywords: prevalence; education; kinesiology; muscle weakness; muscle contraction.

© drannejensen 2014
 361

Introduction

Kinesiology-style manual muscle testing (kMMT) is used by a variety of healthcare

professionals to gain more information about patients. Different from both manual muscle

testing (MMT) and the MMT used in Applied Kinesiology (AK-MMT) since kMMT usually

only uses one “indicator muscle” to detect a specified target condition. These target

conditions can range from physiological dysfunction (e.g. organ or system dysfunction) to

meridian imbalance to a patient’s level of stress, as well as to the integrity of the

neuromuscular system.

It has been reported that AK-MMT is used by approximately 40% of American

chiropractors.12, 13, 218 However it is uncertain how many people use kMMT. Therefore, the

initial purpose of this study was to estimate the number of people that use kMMT.

This aim seemed straight forward at first; however, when fully explored, it became quite

complex. First of all, more than just chiropractors use kMMT in practice. For example, health

care practitioners such as some psychologists, acupuncturists, naturopaths and massage

therapists use kMMT, but not all of these types of practitioners do. Also, more than just

health care professionals use kMMT, for example, some educators, coaches and parents. But

not all of these do either. Furthermore, it is widely known that there are possibly hundreds of

different kMMT technique systems and various professional kMMT organisations, the

memberships of which may overlap. Finally, since kMMT is not widely accepted and perhaps

even thought of as quackery, it is also possible that people that use kMMT do not want to be

known to be using it (i.e. “closet” practitioners), and so may not appear on any formal

registry.

© drannejensen 2014
 362

Therefore, it quickly became clear that the prevalence of use of kMMT would be challenging

to determine exactly. Consequently, the aim of the study then changed to estimating the

number of people trained to use kMMT. From this information, one can then make an

informed inference as to the prevalence of use of kMMT.

Methods

The first step taken to estimate the number of people trained to use kMMT was to create a list

of all organisations that offer or have offered training in kMMT or a system that uses kMMT.

Electronic searches were conducted using Google and MEDLINE (September 2009). No time

or language restrictions were used. Search terms were [“muscle test*” OR “kinesiology”]. In

addition, books on chiropractic techniques were consulted,219, 220 and experts in the field were

contacted via telephone, email and social media.

After the list of kMMT techniques / educators was compiled, contact was attempted by both

email and telephone, and two specific questions were asked: (1) “Do you use kinesiology or

muscle testing in (their technique)?” and if yes, then: (2) “In your best conservative estimate,

how many people have been trained in (their technique)?” The technique system was

included if they responded “Yes” to the first question. Data was collected between May 2008

and November 2009.

For completeness, a list of kMMT professional associations was also compiled, but no

membership information (e.g. size) was sought.

Results

Seventy-nine technique systems were identified to use some form of kMMT. Despite

attempts being made to contact all organisations, only 46 provided estimations of the number

© drannejensen 2014
 363

of people trained in their technique. See Appendix Table C.1. Of the 33 organisations that did

not provide estimates, some were not contactable, some did not respond to contact, some

stated that they could not provide an estimate, and some refused to answer. Instead,

experienced practitioners in these techniques were consulted to provide best guess estimates

of the number of people trained in their respective technique systems, and these numbers are

also listed in Appendix Table C.1.

From the information provided by the 46 contactable technique systems, it can be inferred

that over 900,000 people were trained to use kMMT. In addition, it was estimated that

another 110,000 people were trained in the use of kMMT in the 33 technique systems which

did not provide information. Therefore, it was estimated that over 1 million people were

trained to use kMMT. In addition, 65 professional associations or schools of kMMT were

identified globally. For completeness, a list can be found in Appendix Table C.2.

© drannejensen 2014
 364

FIGURE C.1 - Estimated number of practitioners trained in various Muscle Testing Techniques. The estimates
provided can be considered conservative estimates, meaning at least thess numbers of people trained. The information in the top
part of the table was provided by respective organisations via personal email, telephone or from the respective website. In the
bottom part of the table, no information was provided by the respective organisation, but was estimated by experts in these fields.

Estimated
# Trained Technique System Name Anacronym

1 150,000 Touch for Health TFH

2 110,000 Applied Kinesiology AK
3 100,000 Sacro-Occipital Technique SOT
4 80,000 Thought Field Therapy TFT
5 65,000 Contact Reflex Analysis CRA
6 65,000 PSYCH-K
7 60,000 Total Body Modification TBM
8 45,000 Yuen Method
9 35,000 Educational Kinesiology / Brain Gym Edu-K
10 30,000 Bio-Energetic Synchronization Technique BEST
11 30,000 Nambudripad's Allergy Elimination Technique NAET
12 25,000 RESET TMJ
13 20,000 Health Kinesiology HK
14 15,000 Advanced Energy Psychology™ ARP
15 12,000 Applied Physiology AP
16 12,000 Neuroenergetic Psychology NEP
17 10,000 Neural Organization Technique NOT
Technique representative provided information

18 6,000 Neuro Emotional Technique NET

19 3,500 Intuitive Kinesiology
20 3,000 Kinergetics
21 3,000 Metabolics - Functional Biochemistry
22 2,800 Chirodontics
23 2,500 Human Ecology Balancing Science HEBS
24 2,500 Manual Kinesiology MAK
25 2,500 Neuro Impulse Protocol NIP
26 2,200 Psychosomatic Energetics PSE
27 2,000 Chiro Plus Kinesiology CPK
28 2,000 Cranial Release Technique CRT
29 2,000 Neuro Energetic Kinesiology NEK
30 1,500 Dobson Muscle Testing Technique DMT
31 1,500 Matrix Response Testing MRT
32 1,500 One Brain (aka 3-in-1 Concepts)
33 1,200 Integrative Kinesiology IK
34 1,000 Foundation Clinical Kinesiology
35 1,000 Neuro-Modulation Technique NMT
36 1,000 Zahnärztliche PhysioEnergetik" (Dental Physioenergetics) ZÄPE
37 700 Aromatic Kinesiology
38 500 (The) Vickery Method TVM
39 500 Integrated Biodynamics IBD
40 500 Systematic Kinesiology
41 250 Synergistic Kinesiology
42 150 Allergy Pathway
43 120 Extreme Kinesiology XK
44 120 HoloDynamic Kinesiology HDK
45 60 Kinesiologie nach Gauer
46 50 Chirokinetic Therapy CKT
909,650 Subtotal

47 20,000 Emotional Code

48 15,000 Clinical Kinesiology CK
49 10,000 BodyTalk
50 10,000 NeuroLink
51 5,000 Be Set Free Fast BSFF
52 5,000 Learning Enhancement Advanced Program LEAP
53 5,000 Nutritional Response Testing NRT
54 5,000 Wholistic Kinesiology
55 3,000 Power vs. Force system
Technique representative did not provide information*

56 3,000 Professional Kinesiology Practice PKP

57 3,000 Wellness Kinesiology
58 2,000 Biokinesiology BK
59 2,000 Integrative Manual Therapy IMT
60 2,000 NeuroLinguistic Kinesiology NLK
61 2,000 Progressive Kinesiolgy
62 1,000 Advanced Allergy Therapeutics
63 1,000 Applied Psychoneurobiology APN
64 1,000 Autonomic Response Testing ART
65 1,000 Balance Kinesiology
66 1,000 Brain Integration Technique BIT
67 1,000 Cyberkinetics - Cybernetic Kinesiology
68 1,000 Energetic Kinesiology
69 1,000 Energy Consciousness Therapy ECM
70 1,000 Energy Diagnostic & Treatment Methods / Advanced Energy Psychology EDxTM
71 1,000 EnergyField Kinesiology
72 1,000 Negative Affect Erasing Method NAEM
73 1,000 Neural Systems Kinesiology
74 1,000 Neuro Organization Work NOW
75 1,000 Neurobiology / Neural Therapy / Psycho-Kinesiology
76 1,000 Physioenergetik
77 1,000 Riddler Reflex Technique
78 1,000 Stress Indicator Point System
79 1,000 Transformational Kinesiology TK
110,000 Subtotal
1,019,650 TOTAL
*Reasons for not providing information include: (1) Not be contactable, (2) Not responding to contact, (3) Not being able to provide
estimate, and (4) refusing to provide estimate.
NOTE: To provide updated information or make corrections, please email: dranne@drannejensen.com.

© drannejensen 2013
 365

APPENDIX TABLE C.2 - Kinesiology Organisations and Schools

Professional Association or School Country Website
1. Association of Specialised Kinesiologists - KwaZulu-Natal South Africa www.kinesiology.co.za
2. Association of Specialised Kinesiologists South Africa South Africa www.kinesiologysa.co.za
3. Australasian College of Kinsesiology Mastery Australia www.kinesiologymastery.com
4. Australian Kinesiology Association Australia http://www.kinesiology.org.au/
5. Berner Institut für Kinesiologie / Institut Bernois de Kinésiologie Switzerland www.bik.ch
6. Biokinesiolog Skolen Denmark www.kbhkinesiologiskole.dk
7. College of Complementary Medicine - Australia Australia www.complementary.com.au
8. Dansk Pædagogisk Kinesiologiskole Denmark www.kinesiologi-uddannelse.dk
9. Danske Kinesiologer Denmark www.kinesiologi.dk/
10. Den Norske Kinesiologi Forening Norway www.dnkf.org
11. Den Norske Kinesiologi Skolen Norway
12. Deutsche Gesellschaft für Angewandte Kinesiologie Germany www.dgak.de
13. Deutschen Ärztegesellschaft für Applied Kinesiology Germany www.daegak.de
14. Energy Kinesiology Association USA USA www.ask-us.org
15. Fédération Belge de Kinésiologie Belgium www.kinesiologybelgium.org
16. Health Umbrella Kinesiology Practitioners UK www.healthumbrella.co.uk
17. I.K.S.E.N. Italy www.iksen.it
18. Institut Belge de Kinesiologie Belgium www.ibk.be
19. Institut für Angewandte Kinesiologie Germany www.iak-freiburg.de
20. Institut für Kinesiologie Zürich Switzerland www.kinesiologie.edu
21. Integrated Practitioner Training UK www.integrated-kinesiology.co.uk
22. International Association of Specialized Kinesiology Worldwide www.iask.org
23. International College of Applied Kinesiology Worldwide www.icak.com
24. International College of Applied Kinesiology - Australasia Australia www.icak-australasia.com
25. International College of Applied Kinesiology - Austria Austria www.icak-d.de
26. International College of Applied Kinesiology - Benelux Belgium, Netherlands, Luxembourg www.icakbenelux.com
27. International College of Applied Kinesiology - Brasil Brazil www.icak.com.br
28. International College of Applied Kinesiology - Canada Canada www.icakcanada.com
29. International College of Applied Kinesiology - Germany Germany www.icak-d.de
30. International College of Applied Kinesiology - Korea Korea www.ak.or.kr
31. International College of Applied Kinesiology - UK UK www.icak.co.uk
32. International College of Applied Kinesiology - USA USA www.icakusa.com
33. International Institute of Kinesiology Australia www.iikinesiology.com
34. International Kinesiology College Australia / Worldwide www.ikc-info.org; www.tfhka.org
35. International Medical Society for Applied Kinesiology Austria www.imak.co.at
36. International NeuroKinesiology Institute Poland
37. Internationale Kinesiologie Akademie Germany
38. Japan Touch for Health Association Japan www.touch4health.ne.jp
39. KinAP Switzerland www.kinap-verband.ch
40. Kinesiologiforeningen Denmark www.kinesiologiforeningen.dk
41. Kinesiology College of Canada Canada www.kinesiologycollegeofcanada.com
42. Kinesiology College of Ireland Ireland http://www.kinesiologycollege.com/
43. Kinesiology College of Ireland Ireland http://www.kinesiologyireland.com/
44. Kinesiology Federation of UK UK www.kinesiologyfederation.org
45. Kinesiology Institute USA www.kinesiologyinstitute.com
46. KineSuisse Switzerland www.kinesuisse.ch
47. Klinghardt Academy - Germany Germany http://www.ink.ag/
48. Klinghardt Academy - UK UK http://www.klinghardtacademy.com/
49. Klinghardt Academy - USA USA
50. Nordiska Praktorskolan Sweden www.praktor.com
51. Österreichischen Berufsverband der Kinesiologen Austria www.kinesiologie-oebk.at
52. Praxis Integrative Achberg Germany www.integrative.de
53. Sammenslutningen af Alternative Behandlere Denmark www.alternativ-behandling.dk
54. Schweizerischen Berufsverbandes der Kinesiologinnen und Kinesiologen Switzerland www.kinesiologie-ch.ch
55. Schweizerischer Berufsverband für Kinesiologie Switzerland www.iask.ch
56. Schweizerischer Verband Nicht-Medizinische Kinesiologie Switzerland www.svnmk.ch
57. Svenska Kinesiologiskolan - Swedish School of Manual Kinesiology Sweden www.kinesiologi.se
58. Sveriges Yrkesutbildade Kinesiologer Sweden www.kinesiolog.se
59. The Academy of Systematic Kinesiology UK www.kinesiology.co.uk
© drannejensen 2013

60. The Association of Systematic Kinesiology, ASK UK www.systematic-kinesiology.co.uk

61. The British Kinesiology Centre UK www.britishkinesiology.co.uk
62. Topping International Institute Inc USA www.wellnesskinesiology.com
63. Touch For Health Instructors Association - Australia Australia www.touch4health.org.au
64. Touch For Health Kinesiology Association - USA USA www.tfhka.org
65. Vida Kinesiología Spain www.vidakine.org/
 366

Discussion

It is conservatively estimated that over 1 million people worldwide were trained in some form

of kMMT technique system. However, there are several limitations in this study that may be

sources of either overestimation or underestimation in the actual figure. Firstly, there are a

number of potential sources of overestimation that must be noted. For example, there are

likely redundancies in this calculation, since many kMMT practitioners undertake training in

more than one kMMT technique system. Therefore, it is likely that a kMMT trainee has been

counted repeatedly. Consequently, this may have inflated the estimation. In addition, it is also

likely that not all those trained actually practice or routinely use the kMMT technique system

they were trained in, which may also be a source of overestimation of the prevalence of use.

Similarly there are a various potential sources of underestimation. For example, if an

organisation did not have a presence on the World Wide Web, then it is likely that it was not

included in the list (see Appendix Table C.1), and therefore, not contacted. Such would be the

case with small or local kMMT educators, not part of a larger organisation. Also not included

were organisations that do not use kMMT as part of the formal training, but whose

practitioners routinely use kMMT within the technique system in practice. One example of

such an organisation is BodyTalk. With over 100,000 people trained in BodyTalk to date, it is

a noteworthy omission. However, BodyTalk does not officially teach kMMT, but uses

another similar dichotomous test to navigate through a session (John Veltheim, personal

communication, 2010). Nevertheless, kMMT is used routinely by BodyTalk practitioners, as

can be evidenced by a simple search for “BodyTalk” on the website www.YouTube.com.

Another example of this is with Emotional Freedom Technique (EFT), which is practiced

widely around the world and is rapidly growing in popularity. Like BodyTalk, EFT

purportedly does not teach seminar attendees to use kMMT, but EFT practitioners have been

© drannejensen 2014
 367

known to use kMMT in practice with EFT protocols. Therefore, these may be other sources

of gross underestimation of the prevalence of use of kMMT.

It might be noted by some critics that there are some technique systems included in the list

that some practitioners would argue do not use kMMT. One example would be Applied

Kinesiology (AK), which mainly uses MMT in the way Kendall and Kendall describe.1, 10, 11

However, many AK practitioners use one indicator muscle for therapy localisation, which can

be considered a form of kMMT; and therefore, AK and AK practitioners were included in

this survey. Likewise, Sacro Occipital Technique (SOT), a commonly-used chiropractic

techniquei, is not considered a kMMT-technique per se. Nevertheless, SOT practitioners use

the “arm fossa test” and another test called “body language” during assessment of a patient,

which can also be considered forms of kMMT.58, 179, 201 Therefore, SOT and SOT

practitioners were also included in this report.

Taking into account the results of this survey and these potential sources of over- and

underestimation, the prevalence of use of kMMT can be conservatively inferred to be over 1

million practitioners worldwide.

The implications of these results are significant. The prevalence of use is extensive, and yet

kMMT is not accepted as a valid assessment tool and even considered by some to be

charlatanism.21, 87-91, 93-95, 221 This suggests an urgency to undertake rigorous research to

explore the true usefulness of kMMT in clinical settings. The first step in this process should

be to determine its clinical validity by undertaking diagnostic test accuracy studies.73 A

second step would be to determine its precision (i.e. reproducibility and repeatability);73 that

i
SOT has been found to be used by approximately 40% of American chiropractors [Christensen, M. G.,
Kerkhoff, D., Kollasch, M. W., & L., C. ( 2000). Job analysis of chiropractic, 2000: A project report, survey
analysis and summary of the practice of chiropractic within the United States. Greeley, CO: National Board of
Chiropractic Examiners.]

© drannejensen 2014
 368

is, whether it can be relied upon in different clinical settings, with the same and different

patients, and over various timeframes. Finally, its clinical utility must be assessed, which

means answering the question: Does incorporating kMMT in patient management improve

patient outcomes or overall quality of life?73 This last step entails assessing the effectiveness

of the various technique systems (see Appendix Table C.1) using randomised controlled

trials.

The process of validating kMMT is in its early stages. However, the results of this study

indicate that the prevalence of use of kMMT is widespread enough to warrant further

investigation.

Summary

Through internet searches, personal communication and expert opinion, kMMT has been

estimated to be used by over 1 million people worldwide. This estimation might be amplified

due to the possibility of redundancies or attrition. Likewise, it might be low due to

misclassification or too narrow search methods. Regardless, the widespread use of kMMT

merits further consideration, and proper exploration of its usefulness in clinical settings.

© drannejensen 2014
 369

APPENDIX D
STARD Checklists

“It is discouraging how many people are shocked by honesty

and how few by deceit.”

Noel Coward

© drannejensen 2014
 370

APPENDIX D – STARD CHECKLISTS

STARD Checklist – Study 1 – Estimating the Accuracy of kMMT (Chapter 2)

Section and Topic Item On

# page #
TITLE/ABSTRACT/ 1 Identify the article as a study of diagnostic accuracy (recommend
47
KEYWORDS MeSH heading 'sensitivity and specificity').
INTRODUCTION 2 State the research questions or study aims, such as estimating
diagnostic accuracy or comparing accuracy between tests or across 49
participant groups.
METHODS 51
Participants 3 The study population: The inclusion and exclusion criteria, setting
52
and locations where data were collected.
4 Participant recruitment: Was recruitment based on presenting
symptoms, results from previous tests, or the fact that the 52
participants had received the index tests or the reference standard?
5 Participant sampling: Was the study population a consecutive
series of participants defined by the selection criteria in item 3 and 52
4? If not, specify how participants were further selected.
6 Data collection: Was data collection planned before the index test
and reference standard were performed (prospective study) or after 51
(retrospective study)?
Test methods 7 The reference standard and its rationale. 57
8 Technical specifications of material and methods involved
including how and when measurements were taken, and/or cite 55
references for index tests and reference standard.
9 Definition of and rationale for the units, cut-offs and/or categories
55
of the results of the index tests and the reference standard.
10 The number, training and expertise of the persons executing and
52
reading the index tests and the reference standard.
11 Whether or not the readers of the index tests and reference
standard were blind (masked) to the results of the other test and 67
describe any other clinical information available to the readers.
Statistical methods 12 Methods for calculating or comparing measures of diagnostic
accuracy, and the statistical methods used to quantify uncertainty 70
(e.g. 95% confidence intervals).
13 Methods for calculating test reproducibility, if done. N/A
RESULTS 72
Participants 14 When study was performed, including beginning and end dates of
72
recruitment.
15 Clinical and demographic characteristics of the study population
(at least information on age, gender, spectrum of presenting 72
symptoms).

© drannejensen 2014
 371

16 The number of participants satisfying the criteria for inclusion who

did or did not undergo the index tests and/or the reference
N/A
standard; describe why participants failed to undergo either test (a
flow diagram is strongly recommended).
Test results 17 Time-interval between the index tests and the reference standard,
75
and any treatment administered in between.
18 Distribution of severity of the target condition. 75
19 A cross tabulation of the results of the index tests (including
indeterminate and missing results) by the results of the reference
88
standard; for continuous results, the distribution of the test results
by the results of the reference standard.
20 Any adverse events from performing the index tests or the
92
reference standard.
Estimates 21 Estimates of diagnostic accuracy and measures of statistical
76
uncertainty (e.g. 95% confidence intervals).
22 How indeterminate results, missing data and outliers of the index
72
tests were handled.
23 Estimates of variability of diagnostic accuracy between subgroups
83
of participants, readers or centers, if done.
24 Estimates of test reproducibility, if done. N/A
DISCUSSION 25 Discuss the clinical applicability of the study findings. 92,103

© drannejensen 2014
 372

STARD Checklist – Study 2 – Replication of Study 1 (Chapter 3)

Section and Topic Item On

# page #
TITLE/ABSTRACT/ 1 Identify the article as a study of diagnostic accuracy (recommend
111
KEYWORDS MeSH heading 'sensitivity and specificity').
INTRODUCTION 2 State the research questions or study aims, such as estimating
diagnostic accuracy or comparing accuracy between tests or across 113
participant groups.
METHODS 115
Participants 3 The study population: The inclusion and exclusion criteria, setting
116
and locations where data were collected.
4 Participant recruitment: Was recruitment based on presenting
symptoms, results from previous tests, or the fact that the 116
participants had received the index tests or the reference standard?
5 Participant sampling: Was the study population a consecutive
series of participants defined by the selection criteria in item 3 and 116
4? If not, specify how participants were further selected.
6 Data collection: Was data collection planned before the index test
and reference standard were performed (prospective study) or after 115
(retrospective study)?
Test methods 7 The reference standard and its rationale. 117
8 Technical specifications of material and methods involved
including how and when measurements were taken, and/or cite 117
references for index tests and reference standard.
9 Definition of and rationale for the units, cut-offs and/or categories
117
of the results of the index tests and the reference standard.
10 The number, training and expertise of the persons executing and
116
reading the index tests and the reference standard.
11 Whether or not the readers of the index tests and reference
standard were blind (masked) to the results of the other test and 116
describe any other clinical information available to the readers.
Statistical methods 12 Methods for calculating or comparing measures of diagnostic
accuracy, and the statistical methods used to quantify uncertainty 119
(e.g. 95% confidence intervals).
13 Methods for calculating test reproducibility, if done. N/A
RESULTS 120
Participants 14 When study was performed, including beginning and end dates of
120
recruitment.
15 Clinical and demographic characteristics of the study population
(at least information on age, gender, spectrum of presenting 120
symptoms).
16 The number of participants satisfying the criteria for inclusion who
did or did not undergo the index tests and/or the reference
N/A
standard; describe why participants failed to undergo either test (a
flow diagram is strongly recommended).

© drannejensen 2014
 373

Test results 17 Time-interval between the index tests and the reference standard,
122
and any treatment administered in between.
18 Distribution of severity of the target condition. 122
19 A cross tabulation of the results of the index tests (including
indeterminate and missing results) by the results of the reference
131
standard; for continuous results, the distribution of the test results
by the results of the reference standard.
20 Any adverse events from performing the index tests or the
155
reference standard.
Estimates 21 Estimates of diagnostic accuracy and measures of statistical
155
uncertainty (e.g. 95% confidence intervals).
22 How indeterminate results, missing data and outliers of the index
154
tests were handled.
23 Estimates of variability of diagnostic accuracy between subgroups
155
of participants, readers or centers, if done.
24 Estimates of test reproducibility, if done. N/A
DISCUSSION 25 Discuss the clinical applicability of the study findings. 161,164

© drannejensen 2014
 374

STARD Checklist – Study 3 – Grip Strength Study (Chapter 4)

Section and Topic Item On

# page #
TITLE/ABSTRACT/ 1 Identify the article as a study of diagnostic accuracy (recommend
146
KEYWORDS MeSH heading 'sensitivity and specificity').
INTRODUCTION 2 State the research questions or study aims, such as estimating
diagnostic accuracy or comparing accuracy between tests or across 148
participant groups.
METHODS 149
Participants 3 The study population: The inclusion and exclusion criteria, setting
150
and locations where data were collected.
4 Participant recruitment: Was recruitment based on presenting
symptoms, results from previous tests, or the fact that the 150
participants had received the index tests or the reference standard?
5 Participant sampling: Was the study population a consecutive
series of participants defined by the selection criteria in item 3 and 150
4? If not, specify how participants were further selected.
6 Data collection: Was data collection planned before the index test
and reference standard were performed (prospective study) or after 149
(retrospective study)?
Test methods 7 The reference standard and its rationale. 150
8 Technical specifications of material and methods involved
including how and when measurements were taken, and/or cite 150
references for index tests and reference standard.
9 Definition of and rationale for the units, cut-offs and/or categories
151
of the results of the index tests and the reference standard.
10 The number, training and expertise of the persons executing and
150
reading the index tests and the reference standard.
11 Whether or not the readers of the index tests and reference
standard were blind (masked) to the results of the other test and 151
describe any other clinical information available to the readers.
Statistical methods 12 Methods for calculating or comparing measures of diagnostic
accuracy, and the statistical methods used to quantify uncertainty 153
(e.g. 95% confidence intervals).
13 Methods for calculating test reproducibility, if done. N/A
RESULTS 154
Participants 14 When study was performed, including beginning and end dates of
154
recruitment.
15 Clinical and demographic characteristics of the study population
(at least information on age, gender, spectrum of presenting 154
symptoms).
16 The number of participants satisfying the criteria for inclusion who
did or did not undergo the index tests and/or the reference
N/A
standard; describe why participants failed to undergo either test (a
flow diagram is strongly recommended).

© drannejensen 2014
 375

Test results 17 Time-interval between the index tests and the reference standard,
151
and any treatment administered in between.
18 Distribution of severity of the target condition. 155
19 A cross tabulation of the results of the index tests (including
indeterminate and missing results) by the results of the reference
155
standard; for continuous results, the distribution of the test results
by the results of the reference standard.
20 Any adverse events from performing the index tests or the
155
reference standard.
Estimates 21 Estimates of diagnostic accuracy and measures of statistical
155
uncertainty (e.g. 95% confidence intervals).
22 How indeterminate results, missing data and outliers of the index
154
tests were handled.
23 Estimates of variability of diagnostic accuracy between subgroups
155
of participants, readers or centers, if done.
24 Estimates of test reproducibility, if done. N/A
DISCUSSION 25 Discuss the clinical applicability of the study findings. 161,164

© drannejensen 2014
 376

STARD Checklist – Study 4 – Stability Study (Chapter 5)

Section and Topic Item On

# page #
TITLE/ABSTRACT/ 1 Identify the article as a study of diagnostic accuracy (recommend
167
KEYWORDS MeSH heading 'sensitivity and specificity').
INTRODUCTION 2 State the research questions or study aims, such as estimating
diagnostic accuracy or comparing accuracy between tests or across 169
participant groups.
METHODS 170
Participants 3 The study population: The inclusion and exclusion criteria, setting
171
and locations where data were collected.
4 Participant recruitment: Was recruitment based on presenting
symptoms, results from previous tests, or the fact that the 171
participants had received the index tests or the reference standard?
5 Participant sampling: Was the study population a consecutive
series of participants defined by the selection criteria in item 3 and 171
4? If not, specify how participants were further selected.
6 Data collection: Was data collection planned before the index test
and reference standard were performed (prospective study) or after 170
(retrospective study)?
Test methods 7 The reference standard and its rationale. 173
8 Technical specifications of material and methods involved
including how and when measurements were taken, and/or cite 173
references for index tests and reference standard.
9 Definition of and rationale for the units, cut-offs and/or categories
173
of the results of the index tests and the reference standard.
10 The number, training and expertise of the persons executing and
171
reading the index tests and the reference standard.
11 Whether or not the readers of the index tests and reference
standard were blind (masked) to the results of the other test and 173
describe any other clinical information available to the readers.
Statistical methods 12 Methods for calculating or comparing measures of diagnostic
accuracy, and the statistical methods used to quantify uncertainty 176
(e.g. 95% confidence intervals).
13 Methods for calculating test reproducibility, if done. 176
RESULTS 178
Participants 14 When study was performed, including beginning and end dates of
178
recruitment.
15 Clinical and demographic characteristics of the study population
(at least information on age, gender, spectrum of presenting 178
symptoms).
16 The number of participants satisfying the criteria for inclusion who
did or did not undergo the index tests and/or the reference
N/A
standard; describe why participants failed to undergo either test (a
flow diagram is strongly recommended).

© drannejensen 2014
 377

19 A cross tabulation of the results of the index tests (including

indeterminate and missing results) by the results of the reference
181
standard; for continuous results, the distribution of the test results
by the results of the reference standard.
20 Any adverse events from performing the index tests or the
181
reference standard.
Estimates 21 Estimates of diagnostic accuracy and measures of statistical
181
uncertainty (e.g. 95% confidence intervals).
22 How indeterminate results, missing data and outliers of the index
178
tests were handled.
23 Estimates of variability of diagnostic accuracy between subgroups
181
of participants, readers or centers, if done.
24 Estimates of test reproducibility, if done. 184
DISCUSSION 25 Discuss the clinical applicability of the study findings. 194,195

© drannejensen 2014
 378

STARD Checklist – Study 5 – Emotional Stim Study (Chapter 6)

Section and Topic Item On

# page #
TITLE/ABSTRACT/ 1 Identify the article as a study of diagnostic accuracy (recommend
200
KEYWORDS MeSH heading 'sensitivity and specificity').
INTRODUCTION 2 State the research questions or study aims, such as estimating
diagnostic accuracy or comparing accuracy between tests or across 202
participant groups.
METHODS 203
Participants 3 The study population: The inclusion and exclusion criteria, setting
204
and locations where data were collected.
4 Participant recruitment: Was recruitment based on presenting
symptoms, results from previous tests, or the fact that the 204
participants had received the index tests or the reference standard?
5 Participant sampling: Was the study population a consecutive
series of participants defined by the selection criteria in item 3 and 204
4? If not, specify how participants were further selected.
6 Data collection: Was data collection planned before the index test
and reference standard were performed (prospective study) or after 203
(retrospective study)?
Test methods 7 The reference standard and its rationale. 203
8 Technical specifications of material and methods involved
including how and when measurements were taken, and/or cite 204
references for index tests and reference standard.
9 Definition of and rationale for the units, cut-offs and/or categories
206
of the results of the index tests and the reference standard.
10 The number, training and expertise of the persons executing and
204
reading the index tests and the reference standard.
11 Whether or not the readers of the index tests and reference
standard were blind (masked) to the results of the other test and 206
describe any other clinical information available to the readers.
Statistical methods 12 Methods for calculating or comparing measures of diagnostic
accuracy, and the statistical methods used to quantify uncertainty 207
(e.g. 95% confidence intervals).
13 Methods for calculating test reproducibility, if done. N/A
RESULTS 208
Participants 14 When study was performed, including beginning and end dates of
208
recruitment.
15 Clinical and demographic characteristics of the study population
(at least information on age, gender, spectrum of presenting 208
symptoms).

© drannejensen 2014
 379

16 The number of participants satisfying the criteria for inclusion who

did or did not undergo the index tests and/or the reference
N/A
standard; describe why participants failed to undergo either test (a
flow diagram is strongly recommended).
Test results 17 Time-interval between the index tests and the reference standard,
209
and any treatment administered in between.
18 Distribution of severity of the target condition. 208
19 A cross tabulation of the results of the index tests (including
indeterminate and missing results) by the results of the reference
220
standard; for continuous results, the distribution of the test results
by the results of the reference standard.
20 Any adverse events from performing the index tests or the
209
reference standard.
Estimates 21 Estimates of diagnostic accuracy and measures of statistical
209
uncertainty (e.g. 95% confidence intervals).
22 How indeterminate results, missing data and outliers of the index
208
tests were handled.
23 Estimates of variability of diagnostic accuracy between subgroups
215
of participants, readers or centers, if done.
24 Estimates of test reproducibility, if done. N/A
DISCUSSION 25 Discuss the clinical applicability of the study findings. 222,227

© drannejensen 2014
 380

POSTFACE

“Imagination is more important than knowledge.”

Albert Einstein

© drannejensen 2014
 381

Final Thoughts

Throughout the implementation and analysis of these studies, I found myself pondering a

number of important questions about this research, such as: What is Truth? What is a Lie?

Are all lies conscious? Can one lie nonconsciously? Is truth absolute or relative? Universal or

personal? Answering these questions is beyond the scope of this dissertation; however, from

my experiences over the last few years, I will posit these speculations: Truth is personal.

Truth is dynamic and Truth is transient. Truth is conscious and Truth is nonconscious – and

these two truths may differ. True lying, paradoxically, requires intent to deceive. Which may

be nonconscious. All told, Truth is complex.

Lastly, before embarking on this DPhil programme, I was warned by a number of colleagues

that my “life would never be the same again.” Little did I know what exactly this meant. And

luckily I did not, because had I known, it is likely I never would have taken that first step.

Throughout these 7 years, as I progressed past each mile marker on this journey, things and

people and principles and ideas that I held dear began one by one to drop away, until now, at

the end, I feel as if I arrive naked, unencumbered, seeing things as they really are.

Getting to this point had its share of difficulties. At various times during this odyssey, I have

lost jobs, friends, mentors, and my memory. I have changed continents a half dozen times. I

have been heckled, harassed and hated, dismissed, disregarded and disparaged – by those,

surprisingly enough, from within my own faction (i.e. alternative medicine) and by those

from the “other side” (i.e. conventional medicine). I have been asked to answer to broad,

sweeping attacks on the chiropractic profession as a whole – by my colleagues in primary

© drannejensen 2014
 382

carei. I have also had to answer to blatant assaults on the usefulness of an evidence-based

practice approach and scientific methodology in general – by my esteemed colleagues in

chiropracticii. The Universe has truly tested my resolve.

On the other hand, I have been truly blessed as well. I have learned how to stand on my own

two feet. I learned I have a voice and how and when to use it. I have learned the value of true

friendship. I have learned to follow the light, to always follow the lightness. I have loved and

laughed more than I thought possible. I have expanded in ways previously inconceived. I

have grown up, grown into myself. And for this I am delighted – and infinitely grateful.

Yes, it has been an interesting and evolutionary personal journey: One I would not change for

the world.

i
For examples: “So, tell me, what evidence IS there for chiropractice anyway??” (Name withheld, personal
communication, 20 May 2009); And “Chiropractors are as useful as a boar with ...” (Name withheld, personal
communication, 2011).
ii
For example: "I'm not part of the Priestly Class of the Church of EBM. I'm sorry Anne, that ‘House of Cards’
does not apply to chiropractic ... your confidence that RCTs capture MUCH -- if ANY -- of the clinical realities
actually faced by real doctors and real patients may be misplaced...or a religious belief – ‘Scientism’ ... Still, AK
and MMT groups must play that game to get published in the Holy Roman Apostolic Catholic Church of today's
EBM Religion...even though I know its sheerest nonsense." (Name withheld, personal communication, 4 April
2014)

© drannejensen 2014
 383

REFERENCES

“Our duty is to believe that for which we have sufficient evidence,

and to suspend our judgment when we have not.”

John Lubbock

© drannejensen 2014
 384

REFERENCES

1. Kendall FK, McCreary EK. Muscles: Testing & Function. 4th ed. Baltimore, MD:

Williams & Wilkins; 1993.

2. Magee DJ, Sueki D. Orthopedic physical assessment atlas and video: Selected special

tests and movements. St. Louis, MO: Elsevier Saunders; 2011.

3. Conable K, Corneal J, Hambrick T, Marquina N, Zhang J. Electromyogram and force

patterns in variably timed manual muscle testing of the middle deltoid muscle. J

Manipulative Physiol Ther 2006; 29(4): 305-14.

4. Bohannon RW. Manual muscle test scores alone provide a limited indication of the

functional independence of home care patients. Int J Rehabil Res 1999; 22(2): 151-2.

5. Bohannon RW. Manual muscle testing: Does it meet the standards of an adequate

screening test? Clin Rehabil 2005; 19(6): 662-7.

6. Lawson A, Calderon L. Interexaminer agreement for Applied Kinesiology manual

muscle testing. Percept Mot Skills 1997; 84(2): 539-46.

7. Martin EG, Lovett RW. A method of testing muscular strength in infantile paralysis.

JAMA 1915; LXV(18): 1512-3.

8. Lovett RW, Martin EG. Certain aspects of infantile paralysis with a description of a

method of muscle testing. JAMA 1916; LXVI(10): 729-33.

9. Schmitt WH, Cuthbert SC. Common errors and clinical guidelines for manual muscle

testing: "The arm test" and other inaccurate procedures. Chiropr Osteopat 2008; 16.

© drannejensen 2014
 385

10. Kendall FP, Kendall HO. Muscles: Testing and function. Baltimore: Williams &

Wilkins; 1949.

11. Walther DS. Applied Kinesiology: Synopsis. 2nd ed. Pueblo, CO: Systems DC; 2000.

12. Christensen MG, Kerkhoff D, Kollasch MW, L. C. Job analysis of chiropractic, 2000:

A project report, survey analysis and summary of the practice of chiropractic within the

United States. Greeley, CO: National Board of Chiropractic Examiners, 2000.

13. Christensen MG, Kollasch MW. Job analysis of chiropractic, 2005: A project report,

survey analysis and summary of the practice of chiropractic within the United States.

Greeley, CO: National Board of Chiropractic Examiners, 2005.

14. Cuthbert SC, Goodheart Jr GJ. On the reliability and validity of manual muscle

testing: A literature review. Chiropr Osteopat 2007; 15.

15. Schmitt WH, Yanuck SF, Garten H. Applied Kinesiology as functional neurological

method of examination (Part II). Applied Kinesiology als funktionelle neurologische

untersuchung (Teil II) 1999; 40(11): 756-64.

16. Schmitt WH, Yanuck SF, Garten H. Applied Kinesiology as functional neurological

methods of examination. Part III. Applied Kinesiology als funktionelle neurologische

untersuchung (Teil III) 1999; 40(12): 850-9.

17. Schmitt WH, Yanuck SF, Garten H. Applied Kinesiology as functional neurological

examination (Part I). Applied Kinesiology als funktionelle neurologische untersuchung (Teil

I) 1999; 40(10): 702-13.

© drannejensen 2014
 386

18. Walther DS. Applied Kinesiology. Volume 1: Basic procedures and muscle testing.

2nd ed. Pueblo, CO: Systems DC; 1981.

19. Haas M, Cooperstein R, Peterson D. Disentangling manual muscle testing and

Applied Kinesiology: Critique and reinterpretation of a literature review. Chiropr Osteopat

2007; 15.

20. Diamond J. Applied Kinesiology: An opinion. Journal of the International Academy

of Preventive Medicine 1977; 4(1): 97-9.

21. Feinstein AR. Misguided efforts and future challenges for research on "diagnostic

tests". J Epidemiol Community Health 2002; 56(5): 330-2.

22. Hawkins DR. Power vs. force: The hidden determinants of human behavior. Carlsbad

(CA): Hay House; 2002.

23. Thie J, Thie M. Touch for health: A practical guide to natural health. Camarillo (CA):

DeVorss Publications; 2005.

24. Walker SW. Neuro Emotional Technique® Certification Manual. Encinitas (CA):

Neuro Emotional Technique, Inc.; 2004.

25. Monti DA, Sinnott J, Marchese M, Kunkel EJS, Greeson JM. Muscle test

comparisons of congruent and incongruent self-referential statements. Percept Mot Skills

1999; 88(3): 1019-28.

26. Caruso W, Leisman G. The clinical utility of force/displacement analysis of muscle

testing in Applied Kinesiology. Int J Neurosci 2001; 106(3-4): 147-57.

© drannejensen 2014
 387

27. Caruso W, Leisman G. A force/displacement analysis of muscle testing. Percept Mot

Skills 2000; 91(2): 683-92.

28. Leisman G, Shambaugh P, Ferentz AH. Somatosensory evoked potential changes

during muscle testing. Int J Neurosci 1989; 45(1-2): 143-51.

29. Leisman G, Zenhausern R, Ferentz A, Tefera T, Zemcov A. Electromyographic

effects of fatigue and task repetition on the validity of estimates of strong and weak muscles

in applied kinesiological muscle-testing procedures. Percept Mot Skills 1995; 80(3 Pt 1): 963.

30. Pollard H, Bablis P, Bonello R. Can the ileocecal valve point predict low back pain

using manual muscle testing? Chiropr J Austr 2006; 36: 58-62.

31. Peterson KB. A preliminary inquiry into manual muscle testing response in phobic

and control subjects exposed to threatening stimuli. J Manipulative Physiol Ther 1996; 19(5):

310-6.

32. Schmitt Jr WH, Leisman G. Correlation of Applied Kinesiology muscle testing

findings with serum immunologobulin levels for food allergies. Int J Neurosci 1998; 96(3-4):

237-44.

33. Friedman MH, Weisberg J. Applied Kinesiology: Double-blind pilot study. J Prosthet

Dent 1981; 45(3): 321-3.

34. Triano JJ. Muscle strength testing as a diagnostic screen for supplemental nutrition

therapy: A blind study. J Manipulative Physiol Ther 1982; 5(4): 179-82.

35. Kenney JJ, Clemens R, Forsythe KD. Applied Kinesiology unreliable for assessing

nutrient status. J Am Diet Assoc 1988; 88(6): 698-704.

© drannejensen 2014
 388

36. Pothmann R, Von Frankenberg S, Hoicke C, Weingarten H, Lüdtke R. Evaluation of

Applied Kinesiology in nutritional intolerance of childhood. [Evaluation der klinisch

Angewandten Kinesiologie bei nahrungsmittel-unverträglichkeiten im kindesalter.].

Forschende Komplementärmedizin und Klassische Naturheilkunde 2001; 8(6): 336-44.

37. Jacobs GE, Franks TL, Gilman PG. Diagnosis of thyroid dysfunction: Applied

Kinesiology compared to clinical observations and laboratory tests. J Manipulative Physiol

Ther 1984; 7(2): 99-104.

38. Jacobs GE. Applied Kinesiology: An experimental evaluation by double blind

methodology. J Manipulative Physiol Ther 1981; 4: 141-5.

39. Haas M, Peterson D, Hoyer D, Ross G. Muscle testing response to provocative

vertebral challenge and spinal manipulation: A randomized controlled trial of construct

validity. J Manipulative Physiol Ther 1994; 17(3): 141-8.

40. Walker SW. Neuro Emotional Technique®: Neuro Emotional Xtra Techniques

(NEXT). Encinitas ( CA): Neuro Emotional Technique, Inc.; 2004.

41. Millet K. Total Body Modification. 2013c. http://www.tbmseminars.com/ (accessed

11 June 2013.

42. Walker SW. Neuro Emotional Technique®: NET Basic Seminar Manual. Encinitas

(CA): Neuro Emotional Technique, Inc.; 2004.

43. Thie JF. Touch for Health Kinesiology Association. 2013c.

http://www.touch4health.org.au/ (accessed 11 June 2013).

© drannejensen 2014
 389

44. Touch for Health Instructors Association of Australia. Touch for Health Instructors

Association Australia. 2013c (accessed 11 Jue 2013).

45. Moyer K. Establish Communication in PSYCH-K. Enhanced Potenial LLC; 2013.

46. Knottnerus JA, Buntinx F. The evidence base of clinical diagnosis: Theory and

methods of diagnostic research. Oxford: Blackwell Publishing Ltd.; 2009.

47. Bossuyt PMM, Reitsma JB, Bruns DE, et al. Towards complete and accurate

reporting of studies of diagnostic accuracy: The STARD initiative. Br Med J 2003; 326: 41-4.

48. Slezák P, Waczulíková I. Reproducibility and repeatability. Physiol Res 2011; 60(1):

203-4.

49. Klinkoski B, Leboeuf C. A review of the research papers published by the

International College of Applied Kinesiology from 1981 to 1987. J Manipulative Physiol

Ther 1990; 13(4): 190-4.

50. Drouin JM, Valovich-McLeod TC, Shultz SJ, Gansneder BM, Perrin DH. Reliability

and validity of the Biodex System 3 Pro isokinetic dynamometer velocity, torque and position

measurements. Eur J Appl Physiol 2004; 91(1): 22.

51. Florence JM, Pandya S, King WM, et al. Intrarater reliability of manual muscle test

(Medical Research Council scale) grades in Duchenne's muscular dystrophy. Phys Ther 1992;

72(2): 115.

52. Haas M, Peterson D, Hoyer D, Ross G. The reliability of muscle testing response to a

provocative vertebral challenge. Journal of Chiropractic Technique 1993; 5(95-100).

© drannejensen 2014
 390

53. Jepsen JR, Laursen LH, Larsen AI, Hagert CG. Manual strength testing in 14 upper

limb muscles: A study of inter-rater reliability. Acta Orthop Scand 2004; 75(4): 442-8.

54. Perry J, Weiss WB, Burnfield JM, Gronley JK. The supine hip extensor manual

muscle test: A reliability and validity study. Arch Phys Med Rehabil 2004; 85(8): 1345-50.

55. Pollard H, Lakay B, Tucker F, Watson B, Bablis P. Interexaminer reliability of the

deltoid and psoas muscle test. J Manipulative Physiol Ther 2005; 28(1): 52.

56. Wadsworth CT, Krishnan R, Sear M. Intrarater reliability of manual muscle testing

and hand-held dynametric muscle testing. Phys Ther 1987; 67(9): 1342-7.

57. Ladeira CE, Hess LW, Galin BM, Fradera S, Harkness MA. Validation of an

abdominal muscle strength test with dynamometry. J Strength Cond Res 2005; 19(4): 925.

58. Leboeuf C, Jenkins DJ, Smyth RA. Sacro-Occipital Technique: The so-called arm-

fossa test. Intra-examiner agreement and post-treatment changes. . Journal of the Australian

Chiropractic Association 1988; 18: 67-8.

59. Perry J, Ireland ML, Gronley J, Hoffer M. Predictive value of manual muscle testing

and gait analysis in normal ankles by dynamic electromyography. Foot Ankle 1986; 6(5):

254-9.

60. Bohannon RW. Internal consistency of manual muscle testing scores. Percept Mot

Skills 1997; 85(3 PART I): 736-8.

61. Nahmani L, Serviere F, Dubois JM. Kinésiologie de l'articulation

temporomandibulaire: un nouveau test musculaire pour contrôler la normalité de l'occlusion.

© drannejensen 2014
 391

[Kinesiology of the temporomandibular joint: a new muscle test to evaluate the normality of

occlusion.]. Les Cahiers de prothese 1984; 12(48): 139-48.

62. Omura Y. New simple early diagnostic methods using Omura's 'Bi-digital O-ring

Dysfunction Localization Method' and acupuncture organ representation points, and their

applications to the 'drug & food compatibility test' for individual organs and to auricular

diagnosis of internal organs - Part I. Acupunct Electrother Res 1981; 6(4): 239-54.

63. Schmitt WH, Leisman G. Correlation of Applied Kinesiology muscle testing findings

with serum immunologobulin levels for food allergies. Int J Neurosci 1998; 96(3-4): 237.

64. Tiekert CG. Applied Kinesiology: Its use in veterinary diagnosis. Vet Med Small Anim

Clin 1981; 76(11): 1621-3.

65. Centre for Reviews and Dissemination. Systematic Reviews: CRD's guidance for

undertaking reviews in health care. York (UK): York Publishing Services; 2009.

66. Bossuyt PMM, Reitsma JB, Bruns DE, et al. The STARD statement for reporting

studies of diagnostic accuracy: Explanation and elaboration. Clin Chem 2003; 49(1): 7-18.

67. The Cochrane Collaboration. What is a Cochrane Diagnostic Test Accuracy

Review?2008.

http://www3.interscience.wiley.com/homepages/106568753/what_diagnostic.pdf (accessed.

68. Streiner DL, Norman GR. Health measurement scales: A practical guide to their

developement and use. 4th ed. Oxford (UK): Oxford University Press; 2008.

69. Hagino C. How to appraise research: A guide for chiropractic students and

practitioners. New York: Churchill Livingstone; 2003.

© drannejensen 2014
 392

70. Wright AJ. Conducting psychological assessment: A guide for practitioners.

Hoboken, NJ (USA): John Wiley & Sons, Inc.; 2011.

71. Sackett DL. A primer on the precision and accuracy of the clinical examination.

JAMA 1992; 267(19): 2638-44.

72. Straus SE, Richardson WS, Glasziou P, Haynes RB. Evidence-based medicine: How

to practice and teach EBM. 3rd ed. London: Elsevier Churchill Livingstone; 2005.

73. Bossuyt PMM. Defining biomarker performance and clinical validity. J Med Biochem

2011; 30(3): 193-200.

74. Albert PS. Estimating diagnostic accuracy of multiple binary tests with an imperfect

reference standard. Stat Med 2009; 28(5): 780-97.

75. Bossuyt PMM. Interpreting Diagnostic Test Accuracy Studies. Semin Hematol 2008;

45(3): 189-95.

76. Ferrante Di Ruffano L, Hyde CJ, McCaffery KJ, Bossuyt PMM, Deeks JJ. Assessing

the value of diagnostic tests: A framework for designing and evaluating trials. BMJ (Online)

2012; 344(7847).

77. Aalami OO, Fang TD, Song HM, Nacamuli RP. Physiological features of aging

persons. Arch Surg 2003; 138(10): 1068-76.

78. Sim J, Reid N. Statistical inference by confidence intervals: Issues of interpretation

and utilization. Phys Ther 1999; 79(2): 186-95.

© drannejensen 2014
 393

79. American Psychiatric Association. Diagnostic and Statistical Manual of Mental

Disorders - Text Revision (DSM-IV-TR). 4th ed. Washington, DC: American Psychiatric

Association; 2000.

80. Lord SJ, Staub LP, Bossuyt PMM, Irwig LM. Target practice: Choosing target

conditions for test accuracy studies that are relevant to clinical practice. BMJ (Online) 2011;

343(7822).

81. Truth. n.d. http://oxforddictionaries.com/definition/english/truth?q=truth (accessed 12

January 2013).

82. Lie. n.d. Dictionary.com website: http://dictionary.reference.com/browse/lie (accessed

6 August 2013).

83. Williams R. About PSYCH-K®. 2012. http://psych-k.com/ (accessed 6 August 2013).

84. Bossuyt PMM, Reitsma JB, Linnet K, Moons KGM. Beyond diagnostic accuracy:

The clinical utility of diagnostic tests. Clin Chem 2012; 58(12): 1636-43.

85. Sporer SL, Schwandt B. Paraverbal indicators of deception: A meta-analytic

synthesis. Appl Cogn Psychol 2006; 20(4): 421-46.

86. Yamaoka K. A psychophysiological study of determinants for detection of deception.

Bull Tokyo Med Dent Univ 1976; 23(1): 11-22.

87. Applied Kinesiology--what does the term mean? J Am Diet Assoc 1989; 89(4).

88. Eversaul GA. Dental Kinesiology. Basal Facts 1985; 7(2): 127-31.

© drannejensen 2014
 394

89. Farkas J, Garten H. Diagnosis and therapy of functional hypothyroidism by means of

Applied Kinesiology [Diagnostik und therapie funktioneller hypothyreose mit applied

kinesiology]. Arztezeitschrift fur Naturheilverfahren 1999; 40(1): 24.

90. Garrow JS. Kinesiology and food allergy. Br Med J 1988; 296(6636): 1573-4.

91. Nahin RL, Straus SE. Research into complementary and alternative medicine:

Problems and potential. Br Med J 2001; 322(7279): 161-4.

92. Omura Y. Acupuncture (with possible roles of serotonin & melatonin) and related

unorthodox methods of diagnosis and treatment: Non-invasive spheno-palatine ganglionic

block, abrasion of naso-pharyngeal mucosa, and Applied Kinesiology. Acupunct Electrother

Res 1979; 4(2): 69-89.

93. Schissel MJ, Dodes JE. Dentistry and alternative therapy. NY State Dent J 1997;

63(3): 32-7.

94. Teuber SS, Porch-Curren C. Unproved diagnostic and therapeutic approaches to food

allergy and intolerance. Curr Opin Allergy Clin Immunol 2003; 3(3): 217-21.

95. Wüthrich B. Unproven techniques in allergy diagnosis. J Investig Allergol Clin

Immunol 2005; 15(2): 86.

96. Ekman P. Telling lies: Clues to deceit in the marketplace, politics, and marriage. New

York: W.W. Norton & Co.; 2001.

97. Ekman P, Friesen WV. Detecting deception from the body or face. J Pers Soc Psychol

1974; 29(3): 288-98.

© drannejensen 2014
 395

98. Robinson ME, Dannecker EA. Critical issues in the use of muscle testing for the

determination of sincerity of effort. Clin J Pain 2004; 20(6): 392.

99. Bossuyt PMM, Leeflang MM. Chapter 6: developing criteria for including studies.

Cochrane Handbook of Systematic Reviews of Diagnostic Test Accuracy Version 40

[updated September 2008]: The Cochrane Collaboration; 2008.

100. Lang PJ, Bradley MM, Cuthbert BN. International Affective Picture System (IAPS):

Affective ratings of pictures and instruction manual. Technical Report A-8. Gainesville, FL:

University of Florida.; 2008.

101. Bradley MM, Lang PJ. The International Affective Picture System (IAPS) in the

study of emotion and attention. In: Coan JA, Allen JJB, eds. Handbook of Emotion

Elicitation and Assessment. Oxford: Oxford University Press; 2007: 29-46.

102. Davis WJ, Rahman MA, Smith LJ, et al. Properties of human affect induced by static

color slides (IAPS): Dimensional, categorical and electromyographic analysis. Biol Psychol

1995; 41(3): 229-53.

103. Lang PJ, Bradley MM, Cuthbert BN. International affective picture system (IAPS):

Technical manual and affective ratings. Gainesville: University of Florida, Center for

Research in Psychophysiology.; 1997.

104. Bradley MM, Lang PJ. Affective Norms for English Words (ANEW): Stimuli,

instruction manual and affective ratings. Technical report C-1. Gainesville, FL: The Center

for Research in Psychophysiology, University of Florida.; 1999.

105. Stevenson RA, Mikels JA, James TW. Characterization of the Affective Norms for

English Words by discrete emotional categories. Behav Res Methods 2007; 39(4): 1020-4.

© drannejensen 2014
 396

106. Altman DG. Practical statistics for medical research. London: Chapman & Hall/CRC;

1999.

107. Leeflang MMG, Bossuyt PMM, Irwig L. Diagnostic test accuracy may vary with

prevalence: Implications for evidence-based diagnosis. J Clin Epidemiol 2009; 62(1): 5-12.

108. Sackett DL. Bias in analytic research. J Chronic Dis 1979; 32(1-2): 51-68.

109. McGrath RE, Mitchell M, Kim BH, Hough L. Evidence for response bias as a source

of error variance in applied assessment. Psychol Bull 2010; 136(3): 450-70.

110. King MF, Bruner GC. Social desirability bias: A neglected aspect of validity testing.

Psychology and Marketing 2000; 17(2): 79-103.

111. Linnet K, Bossuyt PMM, Moons KGM, Reitsma JB. Quantifying the accuracy of a

diagnostic test or marker. Clin Chem 2012; 58(9): 1292-301.

112. Flach PA. The geometry of ROC space: Understanding machine learning metrics

through ROC isometrics. In: Fawcett T, Mishra N, editors.; 2003; Washington, DC; 2003. p.

194-201.

113. Hall J, Philip RCM, Marwick K, et al. Social cognition, the male brain and the autism

spectrum. PLoS ONE 2012; 7(12): e49033.

114. Hall JA. Gender effects in decoding nonverbal cues. Psychol Bull 1978; 85(4): 845-

57.

115. Dillon KM, Fenlason JE, Vogel DJ. Belief in and use of a questionable technique,

Facilitated Communication, for children with autism. Psychol Rep 1994; 75(1 Pt 2): 459-64.

© drannejensen 2014
 397

116. Hyman R. The mischief-making of ideomotor action: How people are fooled by

ideomotor action. Scientific Review of Alternative Medicine 1999; 3(2): 34-43.

117. Glasziou P, Irwig L, Deeks JJ. When should a new test become the current reference

standard? Ann Intern Med 2008; 149(11): 816-21.

118. Grubin D, Madsen L. Lie detection and the polygraph: A historical review. J Forens

Psychiatry Psychol 2005; 16(2): 357-69.

119. Porter S, Ten Brinke L. The truth about lies: What works in detecting high-stakes

deception? Legal and Criminological Psychology 2010; 15(1): 57-75.

120. Caro M. Caro's book of poker tells: The psychology and body language of poker. New

York: Cardoza Publishing; 2003.

121. Rosner AL, Conable KM, Edelmann T. Influence of foot orthotics upon duration of

effects of spinal manipulation in chronic back pain patients: A randomized clinical trial. J

Manipulative Physiol Ther 2014; 37(2): 124-40.

122. Kuo TBJ, Li JY, Lai CT, Huang YC, Hsu YC, Yang CCH. The effect of bedding

system selected by manual muscle testing on sleep-related cardiovascular functions. BioMed

Res Int 2013; 2013.

123. Moncayo R, Moncayo H, Ulmer H, Kainz H. New diagnostic and therapeutic

approach to thyroid-associated orbitopathy based on Applied Kinesiology and homeopathic

therapy. J Altern Complement Med 2004; 10(4): 643-50.

124. Verschuere B, Crombez G, Koster E. The International Affective Picture System: A

Flemish validation study. Psychol Belg 2001; 41(4): 205-17.

© drannejensen 2014
 398

125. McKenzie F, Scerbo M, Catanzaro J, Phillips M. Nonverbal indicators of malicious

intent: Affective components for interrogative virtual reality training. Int J Hum Comput Stud

2003; 59(1-2): 237-44.

126. Sporer SL, Schwandt B. Moderators of nonverbal indicators of deception: A meta-

analytic synthesis. Psychol Public Policy Law 2007; 13(1): 1-34.

127. Webb DE, Fagan J. The impact of dream interpretation using Psychological

Kinesiology on the frequency of recurring dreams. Psychother Psychosom 1993; 59(3-4):

203-8.

128. Ben-Shakhar G, Elaad E. The validity of psychophysiological detection of

information with the guilty knowledge test: A meta-analytic review. J Appl Psychol 2003;

88(1): 131-51.

129. Williams JA, Burns EL, Harmon EA. Insincere utterances and gaze: Eye contact

during sarcastic statements. Percept Mot Skills 2009; 108(2): 565-72.

130. Carpenter WB. On the influence of suggestion in modifying and directing muscular

movement, independently of volition. Proceedings of the Royal Institution of Great Britain;

1852; London; 1852. p. 147-53.

131. Carroll RT. The Skeptic's Dictionary: A collection of strange beliefs, amusing

deceptions, and dangerous delusions. Applied Kinesiology. 2012.

http://www.skepdic.com/akinesiology.html (accessed 8 August 2013).

132. Fernald DH, Coombs L, DeAlleaume L, West D, Parnes B. An assessment of the

Hawthorne Effect in practice-based research. J Am Board Fam Med 2012; 25(1): 83-6.

© drannejensen 2014
 399

133. Conable KM, Rosner AL. A narrative review of manual muscle testing and

implications for muscle testing research. J Chiropr Med 2011; 10(3): 157-65.

134. Tomporowski PD, Beasman K, Ganio MS, Cureton K. Effects of dehydration and

fluid ingestion on cognition. Int J Sports Med 2007; 28(10): 891-6.

135. D'Anci KE, Mahoney CR, Vibhakar A, Kanter JH, Taylor HA. Voluntary dehydration

and cognitive performance in trained college athletes. Percept Mot Skills 2009; 109(1): 251-

69.

136. Lieberman HR, Bathalon GP, Falco CM, Kramer FM, Morgan Iii CA, Niro P. Severe

decrements in cognition function and mood induced by sleep loss, heat, dehydration, and

undernutrition during simulated combat. Biol Psychiatry 2005; 57(4): 422-9.

137. Goodheart GJ. Muscle strength testing as a diagnostic screen for supplemental

nutrition therapy: A blind study [Letter]. J Manipulative Physiol Ther 1983; 6: 87.

138. Dobson GJ. Manual muscle testing combined with specific head positioning, and

other articular challenges, as an assessment of vertebral subluxation of the upper cervical

spine: a descriptive paper. J Vertebral Subluxation Res 1999; 3(2): 1-7.

139. Mason S, Tovey P, Long AF. Evaluating complementary medicine: Methodological

challenges of randomised controlled trials. Br Med J 2002; 325(7368): 832-4.

140. Cox RH. Sport Psychology: Concepts and Applications. 6th ed. New York: McGraw-

Hill; 2006.

141. Tang J, Gibson SJ. A psychophysical evaluation of the relationship between trait

anxiety, pain perception, and induced state anxiety. Journal of Pain 2005; 6(9): 612-9.

© drannejensen 2014
 400

142. Wang R, Lagakos SW, Ware JH, Hunter DJ, Drazen JM. Statistics in medicine -

Reporting of subgroup analyses in clinical trials. N Engl J Med 2007; 357(21): 2189-94+08.

143. Peterson KB, Peterson CD. A case series evaluating the accuracy of manual muscle

testing for predicting fetal sex. J Chiropr Med 2012; 11(1): 1-6.

144. Brookes ST, Whitely E, Egger M, Smith GD, Mulheran PA, Peters TJ. Subgroup

analyses in randomized trials: Risks of subgroup-specific analyses; power and sample size for

the interaction test. J Clin Epidemiol 2004; 57(3): 229-36.

145. Bonaccio S, Reeve CL. The nature and relative importance of students' perceptions of

the sources of test anxiety. Learn Individ Differ 2010; 20(6): 617-25.

146. Liebert RM, Morris LW. Cognitive and emotional components of test anxiety: A

distinction and some initial data. Psychol Rep 1967; 20(3): 975-8.

147. Pasricha T, Sallam H, Chen J, Pasricha PJ. The moment of truth: Stress, lying and the

GI tract. Expert Rev Gastroenterol Hepatol 2008; 2(3): 291-3.

148. Kugelmass S, Lieblich I. Effects of realistic stress and procedural interference in

experimental lie detection. J Appl Psychol 1966; 50(3): 211-6.

149. Podlesny JA, Raskin DC. Physiological measures and the detection of deception.

Psychol Bull 1977; 84(4): 782-99.

150. Ulatowska J. Indicators of deception in different lie scenarios. Roczniki

Psychologiczne / Annals of Psychology 2013; 16(1): 127-46.

151. Hall S, Lewith G, Brien S, Little P. A review of the literature in Applied and

Specialised Kinesiology. Forsch Komplementarmed 2008; 15(1): 40-6.

© drannejensen 2014
 401

152. Hayes KW, Falconer J. Reliability of hand-held dynamometry and its relationship

with manual muscle testing in patients with osteoarthritis in the knee. J Orthop Sports Phys

Ther 1992; 16(3): 145-9.

153. Aufsesser P, Horvat M, Croce R. A critical examination of selected hand-held

dynamometers to assess isometric muscle strength. Adapt Phys Activ Q 1996; 13(2): 153-65.

154. Aitkens S, Lord J, Bernauer E, Fowler Jr WM, Lieberman JS, Berck P. Relationship

of manual muscle testing to objective strength measurements. Muscle Nerve 1989; 12(3):

173-7.

155. Radin DI. A possible proximity effect on human grip strength. Percept Mot Skills

1984; 58(3): 887-8.

156. Braud WG. A possible proximity effect on human grip strength: an attempted

replication. Percept Mot Skills 1989; 68(1): 157-8.

157. Kendler BS, Keating TM. Re-evaluation of a possible proximity effect on muscular

strength. Percept Mot Skills 2003; 97(2): 371-4.

158. Quintanar AF, Hill TV. Sugar proximity and human grip strength. Percept Mot Skills

1988; 67(3): 853-4.

159. Mathiowetz V, Weber K, Volland G, Kashman N. Reliability and validity of grip and

pinch strength evaluations. J Hand Surg (USA) 1984; 9(2): 222-6.

160. Peolsson A, Hedlund R, Oberg B. Intra- and inter-tester reliability and reference

values for hand strength. J Rehabil Med 2001; 33(1): 36-41.

© drannejensen 2014
 402

161. Colombo R, Mazzini L, Mora G, et al. Measurement of isometric muscle strength: A

reproducibility study of maximal voluntary contraction in normal subjects and amyotrophic

lateral sclerosis patients. Med Eng Phys 2000; 22(3): 167-74.

162. Günther CM, Bürger A, Rickert M, Crispin A, Schulz CU. Grip Strength in Healthy

Caucasian Adults: Reference Values. J Hand Surg (USA) 2008; 33(4): 558-65.

163. Massy-Westropp NM, Gill TK, Taylor AW, Bohannon RW, Hill CL. Hand Grip

Strength: Age and gender stratified normative data in a population-based study. BMC Res

Notes 2011; 4.

164. Conable KM. Intraexaminer comparison of Applied Kinesiology manual muscle

testing of varying durations: a pilot study. J Chiropr Med 2010; 9(1): 3-10.

165. Schreuders TAR, Selles RW, Roebroeck ME, Stam HJ. Strength measurements of the

intrinsic hand muscles: A review of the development and evaluation of the Rotterdam

Intrinsic Hand Myometer. J Hand Ther 2006; 19(4): 393-402.

166. Guyton AC, Hall JE. Human physiology and mechanisms of disease, 10th edition.

10th ed. Philadelphia: W.B. Saunders Co.; 2000.

167. Sorenson EJ. A comparison of muscle strength testing techniques in amyotrophic

lateral sclerosis. Neurology 2003; 61(11): 1503-7.

168. McGarvey SR, Morrey BF, Askew LJ, An KN. Reliability of isometric strength

testing. Temporal factors and strength variation. Clin Orthop 1984; NO. 185: 301-5.

169. Hsieh CY, Phillips RB. Reliability of manual muscle testing with a computerized

dynamometer. J Manipulative Physiol Ther 1990; 13(2): 72-82.

© drannejensen 2014
 403

170. Conable K, Corneal J, Hambrick T, Marquina N, Zhang J. Investigation of methods

and styles of manual muscle testing by AK practitioners. J Chiropr Med 2005; 4(1): 1-10.

171. Taylor JR. An introduction to error analysis: The study of uncertainties in physical

measurements. Saucilito, CA: University Science Books; 1999.

172. Myles PS, Cui J. I. Using the Bland-Altman method to measure agreement with

repeated measures. Br J Anaesth 2007; 99(3): 309-11.

173. Bland JM. How do I analyse observer variation studies? 2004. http://www-

users.york.ac.uk/~mb55/meas/observer.pdf. (accessed 16 March 2014).

174. Bland JM, Altman DG. Measurement error and correlation coefficients. Br Med J

1996; 313(7048): 41-2.

175. Bland JM. Is there an acceptable value for the limits of agreement or the

repeatability? 2007. http://www-users.york.ac.uk/~mb55/meas/accept.htm. (accessed 16

March 2014).

176. Perot C, Meldener R, Goubel F. Objective measurement of proprioceptive technique

consequences on muscular maximal voluntary contraction during manual muscle testing.

Agressologie 1991; 32(10 Spec No): 471-4.

177. Pollard H, Calder D, Farrar L, Ford M, Melamet A, Cuthbert S. Interexaminer

reliability of manual muscle testing of lower limb muscles without the ideomotor effect.

Chiropr J Austr 2011; 41(1): 23-30.

178. Lüdtke R, Kunz B, Seeber N, Ring J. Test-retest-reliability and validity of the

kinesiology muscle test. Complement Thr Med 2001; 9(3): 141-5.

© drannejensen 2014
 404

179. Leboeuf C. The reliability of specific Sacro-Occipital Technique diagnostic tests. J

Manipulative Physiol Ther 1991; 14(9): 512-7.

180. Bland JM. Interpreting the limits of agreement: do I have good or bad agreement?

2009. http://www-users.york.ac.uk/~mb55/meas/interlim.htm. (accessed 16 March 2014).

181. O'Sullivan M, Frank MG, Hurley CM, Tiwana J. Police lie detection accuracy: The

effect of lie scenario. Law Hum Behav 2009; 33(6): 530-8.

182. Lundberg U, Kadefors R, Melin B, et al. Psychophysiological stress and EMG activity

of the trapezius muscle. Int J Behav Med 1994; 1(4): 354-70.

183. Brouwer AM, van Wouwe N, Mühl C, van Erp J, Toet A. Perceiving blocks of

emotional pictures and sounds: Effects on physiological variables. Front Hum Neurosci 2013;

(JUN).

184. Kuperman V, Estes Z, Brysbaert M, Warriner AB. Emotion and language: Valence

and arousal affect word recognition. J Exp Psychol Gen 2014.

185. Holmes EA, Mathews A, Mackintosh B, Dalgleish T. The causal effect of mental

imagery on emotion assessed using picture-word cues. Emotion 2008; 8(3): 395-409.

186. Absolute Threshold. In: Strickland B, ed. Gale Encyclopedia of Psychology. 2nd ed.

ed. London: Gale Group; 2001.

187. Knottnerus JA, Van Weel C, Muris JWM. Evidence base of clinical diagnosis:

Evaluation of diagnostic procedures. Br Med J 2002; 324(7335): 477-80.

188. De Groot JAH, Bossuyt PMM, Reitsma JB, et al. Verification problems in diagnostic

accuracy studies: Consequences and solutions. BMJ (Online) 2011; 343(7821).

© drannejensen 2014
 405

189. Schmitt Jr WH, Yanuck SF. Expanding the neurological examination using functional

neurologic assessment: Part II neurologic basis of Applied Kinesiology. Int J Neurosci 1999;

97(1-2): 77-108.

190. Buhler CF, Burgess PR, VanWagoner E. Changes in physical strength during

nutritional testing. Journal of Scientific Exploration 2008; 22(4): 495-515.

191. Schwartz SA, Utts J, Spottiswoode SJP, et al. A double-blind, randomized study to

assess the validity of Applied Kinesiology (AK) as a diagnostic tool and as a nonlocal

proximity effect. Explore (NY) 2014; 10(2): 99-108.

192. Lawson A, Calderon L. Interexaminer reliability of Applied Kinesiology manual

muscle testing. Percept Mot Skills 2004.

193. Farella M, Michelotti A, Pellegrino G, Giani U, Martina R. Interexaminer reliability

and validity for diagnosis of temporomandibular disorders of visual leg measurements used in

Dental Kinesiology. J Orofac Pain 2005; 19(4): 285-90.

194. Cuthbert SC, Rosner AL, McDowall D. Association of manual muscle tests and

mechanical neck pain: Results from a prospective pilot study. J Bodywork Movement Ther

2011; 15(2): 192-200.

195. Cuthbert S, Rosner AL. Menopause, chiropractic examination and treatment, and

salivary hormone levels: Correlations and outcomes. Chiropr J Austr 2013; 43(3): 107-20.

196. Carpenter SA, Hoffman J, Mendel R. An investigation into the effect of organ

irritation on muscle strength and spinal mobility. Bull Eur Chiropr Union 1977; 25(2): 42-60.

© drannejensen 2014
 406

197. Tsai LL, Liu HM. Effects of bedding systems selected by manual muscle testing on

sleep and sleep-related respiratory disturbances. Appl Ergon 2008; 39(2): 261-70.

198. Motyka TM, Yanuck SF. Expanding the neurological examination using functional

neurologic assessment - Part I: Methodological considerations. Int J Neurosci 1999; 97(1-2):

61-76.

199. Waxenegger I, Endler PC, Wulkersdorfer B, Spranger H. Individual prognosis

regarding effectiveness of a therapeutic intervention using pre-therapeutic "kinesiology

muscle test". The Scientific World Journal 2007; 7: 1703-7.

200. Tschernitschek H, Fink M. "Applied Kinesiology" in medicine and dentistry - A

critical review ["Angewandte Kinesiologie" in medizin und zahnmedizin - Eine kritische

übersicht]. Wien Med Wochenschr 2005; 155(3-4): 59-64.

201. Evans AC, Yelverton C. The validity and specificity of the arm fossa test. European

Chiropractors’ Union (ECU) 2014 Convention. Dublin, Ireland; 2014.

202. Peeling RW, Smith PG, Bossuyt PM. A guide for diagnostic evaluations. Nature

reviews Microbiology 2010; 8(12 Suppl): S2-6.

203. Bull R, Baron H, Gudjonsson G, Hampson S, Rippon G, Vrij A. A review of the

current scientific status and fields of application of polygraphic deception detection.

Leicester, UK: British Psychological Society, 2004.

204. National Research Council. The polygraph and lie detection. Committee to Review

the Scientific Evidence on the Polygraph. Washington, DC: The National Academic Press,

2003.

© drannejensen 2014
 407

205. Sutlive TG, Lopez HP, Schnitker DE, et al. Development of a clinical prediction rule

for diagnosing hip osteoarthritis in individuals with unilateral hip pain. J Orthop Sports Phys

Ther 2008; 38(9): 542-50.

206. Youdas JW, Madson TJ, Hollman JH. Usefulness of the Trendelenburg test for

identification of patients with hip joint osteoarthritis. Physiother Theory Pract 2010; 26(3):

184-94.

207. Kerr NM, Chew SSL, Eady EK, Gamble GD, Danesh-Meyer HV. Diagnostic

accuracy of confrontation visual field tests. Neurology 2010; 74(15): 1184-90.

208. Staehle HJ, Koch MJ, Pioch T. Double-blind study on materials testing with Applied

Kinesiology. J Dent Res 2005; 84(11): 1066-9.

209. Banis U. Diagnosis of allergy with kinesiology - A critical view. Allergiediagnostik

mit der kinesiologie - Eine kritische betrachtung 2001; 42(6): 414-7.

210. Gerez IFA, Shek LPC, Chng HH, Lee BW. Diagnostic tests for food allergy.

Singapore Med J 2010; 51(1): 4-9.

211. Kleine-Tebbe J, Herold DA. Inappropriate test methods in allergy. Ungeeignete

Testverfahren in der Allergologie 2010; 61(11): 961-6.

212. Scopp AL. An experimental evaluation of kinesiology in allergy and deficiency

disease diagnosis. Journal of Orthomolecular Psychiatry 1978; 7(2): 137-8.

213. Benor DJ. Applied kinesiology. Explor J Sci Heal 2014; 10(4): 217.

214. Holgate ST, Church MK, Broide DH, Martinez FD. Allergy. 4th ed. Oxford: Elsevier

Saunders; 2011.

© drannejensen 2014
 408

215. Fryback DG, Thornbury JR. The efficacy of diagnostic imaging. Med Decis Making

1991; 11(2): 88-94.

216. Schünemann HJ, Oxman AD, Brozek J, et al. GRADE: Grading quality of evidence

and strength of recommendations for diagnostic tests and strategies. BMJ 2008; 336(7653):

1106-10.

217. Riedel M. Diagnosing pulmonary embolism. Postgrad Med J 2004; 80(944): 309-19.

218. Christensen MG, Kollasch MW, Hyland JK. Practice analysis of chiropractic, 2010: A

project report, survey analysis, and summary of the practice of chiropractic within the United

States. Greeley, CO: National Board of Chiropractic Examiners, 2010.

219. Homola S. Chiropractic: History and overview of theories and methods. Clin Orthop

2006; (444): 236-42.

220. Cooperstein R, Gleberzon BJ. Technique Systems in Chiropractic. 1 ed. London:

Churchill Livingstone; 2004.

221. Wüthrich B, Bachert C. Electroacupuncture, Applied Kinesiology, bioresonance

therapy. [Elektroakupunktur, angewandte kinesiologie, bioresonanztherapie}. Allergologie

2006; 29(12): 483-4.

© drannejensen 2014