See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/231121608

The Difficulties in Emotion Regulation Scale: Factor Structure and Association

with Nonsuicidal Self-Injury in Adolescent Inpatients

Article in Journal of Psychopathology and Behavioral Assessment · September 2012

DOI: 10.1007/s10862-012-9292-7

CITATIONS READS

58 2,048

4 authors, including:

Amanda Venta Carla Sharp

Sam Houston State University University of Houston
58 PUBLICATIONS 855 CITATIONS 252 PUBLICATIONS 4,299 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Mentalising and social cognition in personality pathology View project

Maternal Depression View project

All content following this page was uploaded by Carla Sharp on 28 October 2015.

The user has requested enhancement of the downloaded file.

J Psychopathol Behav Assess
DOI 10.1007/s10862-012-9292-7

The Difficulties in Emotion Regulation Scale: Factor

Structure and Association with Nonsuicidal Self-Injury
in Adolescent Inpatients
Jose Perez & Amanda Venta & Sarah Garnaat &
Carla Sharp

# Springer Science+Business Media, LLC 2012

Abstract Research suggests that difficulties in emotion detecting NSSI among inpatient adolescents. Results pro-
regulation are an important correlate of nonsuicidal self- vide further support for the relation between emotion regu-
injury (NSSI) in adults. Research examining this link in lation difficulties and NSSI. The DERS appears to be a
adolescents is limited by the lack of comprehensive instru- useful measure of detecting NSSI in clinical samples of
ments to assess difficulties in emotion regulation. Against adolescents.
this background, the aims of the current study were to (a)
confirm the six-factor structure of the Difficulties in Emo- Keywords Nonsuicidal self-injury . Emotion regulation .
tion Regulation Scale (DERS; Gratz & Roemer, Journal of Adolescence . Psychiatric sample
Psychopathology and Behavioral Assessment, 23(4), 253–
263, 2004) in a sample of adolescent inpatients (N0218);
(b) explore the relation between different aspects of emotion Nonsuicidal self-injury (NSSI) among adolescents is an
dysregulation and lifetime NSSI while controlling for psy- increasing health concern. NSSI is defined as “the deliberate
chopathology and sex; and (c) assess the clinical utility of destruction of body tissue without conscious suicidal intent
the DERS in detecting lifetime NSSI status. Fit indices but resulting in injury severe enough for tissue damage to
obtained through Confirmatory Factor Analysis indicated occur” (Gratz and Roemer 2008). An estimated 21 % of
that the six-factor structure of the DERS fit the data ade- community adolescents (Brausch and Gutierrez 2010) and
quately and that most items loaded strongly on their respec- an estimated 61–68 % of inpatient adolescents (DiClemente
tive latent factor. All six latent factors were significantly et al. 1991; Sim et al. 2009) engage in self-injuring behavior.
correlated with each other, with the exception of lack of While several theoretical approaches to NSSI have been
emotional awareness and difficulties engaging in goal- proposed (see Suyemoto 1998 for a review), a common
directed behavioral when distressed. Regression analyses theme among these theories is that NSSI assists in the
revealed that only the limited access to emotion regulation escape, management, or regulation of emotion (Chapman
strategies subscale accounted for a significant portion of the et al. 2006). Accordingly, Chapman et al. (2006) suggest
variance in NSSI when controlling for other aspects of that NSSI is an emotional avoidance strategy or behavior
emotion dysregulation, sex, and psychopathology. Receiver (along with other maladaptive strategies such as substance
Operating Characteristic analysis indicated that the DERS abuse or thought suppression). In this model, individual
limited access to emotion regulation strategies subscale traits like difficulty regulating emotions when aroused, ten-
score has moderate diagnostic accuracy in detecting the dency to avoid, emotion regulation skills deficits, poor
presence of NSSI. The optimal cut-off score was 21.5 when distress tolerance, and high emotional intensity combine
with state vulnerabilities such as a stimulus that elicits an
emotional response to produce NSSI as a strategy for tem-
J. Perez : A. Venta : S. Garnaat : C. Sharp (*) porary relief. The Chapman model places the individual’s
Department of Psychology, The University of Houston,
maladaptive management of unwanted emotions at the cen-
126 Heyne Building,
Houston, TX 77024, USA ter of NSSI, as do most theories of NSSI as reviewed by
e-mail: csharp2@uh.edu Suyemoto (1998).
 J Psychopathol Behav Assess

Although multiple and complex reasons for self-injuring important to note that emotion regulation, as defined by Gratz
co-occur and should be assessed (Brown et al. 2002), sig- and Roemer and others does not include emotional intensity
nificant empirical evidence exists for the link between emo- per se. For instance, Mennin and colleagues (Mennin et al.
tion dysregulation and NSSI. A recent analysis of 18 studies 2007) explicitly differentiate difficulties with emotion regula-
which investigated various reasons for self-injuring found tion from the experience of heightened emotion. Their model
strong support for an overall affect-regulation function of of emotion dysfunction includes four components: heightened
NSSI (Klonsky 2007). The author noted that although other intensity of emotions, poor understanding of emotions, nega-
functions (such as self-punishment, anti-dissociation, and tive reactivity to emotions, and maladaptive emotional man-
interpersonal-influence) were endorsed as reasons for self- agement responses. They suggest that while heightened
injuring too, using NSSI as a means to regulate affect and intensity of emotions (“frequently experiencing negative af-
emotion remained a ubiquitous finding across studies. Lab- fect strongly and having emotional reactions that occur in-
oratory findings confirm self-report studies. For instance, tensely, easily, and quickly” Mennin et al. 2007, p. 286) may
research has shown that those who self-injure have higher make it more difficult to successfully manage emotions, it is
physiological arousal states when asked to imagine the distinct from the three other components.
period leading up to an episode of self-injury, and conse- Extant measures assessing emotion dysregulation in
quently lower levels of arousal when they imagine both the youth mirror differences in the various models of ER and
self-injury and the period immediately after (Haines et al. often tend to assess only one aspect of ER. For instance, the
1995; Brain et al. 1998). Further, several laboratory studies Children’s Sadness Management Scale (CSMS; Zeman et al.
have now used pain as a proxy for NSSI in self-injuring 2001) includes factors that mostly tap into impulse control
samples and concluded that pain, and NSSI, often serve and ER strategies (i.e. distracting oneself with other activi-
emotion regulation functions. For instance, Bresin and ties, calming oneself). The Emotion Regulation Index for
Gordon (2011) found that painful temperature stimulation Children and Adolescents (ERICA; MacDermott et al.
reduced negative affect in individuals with a history of 2010) emphasizes impulsivity. Other measures, such as the
NSSI. Similarly, Franklin and colleagues (Franklin et al. Regulation of Emotions Questionnaire (REQ; Phillips and
2010) showed that pain regulates cognitive processing and Power 2007) focus more on the internal (i.e. awareness and
affective valence in people with a history of NSSI. Taken reflection of emotions) and external (i.e. expression of emo-
together, these studies support the conclusion that, for many, tions) aspects of ER. As a final example, the Emotion
NSSI serves emotion regulation and negative affect reduc- Expression Scale for Children (EESC; Penza-Clyve and
tion purposes. Zeman 2002) focuses mainly on reluctance of expressing
The link between emotion regulation (ER) and NSSI negative emotions and difficulties with emotional clarity. In
described in the aforementioned adult literature has also addition to highlighting the variety of existing conceptual
been demonstrated for adolescents. For instance, affect reg- models and assessments of ER, the differences between
ulation has been cited as the reason for self-injury above aforementioned measures make comparing findings on ER
other functions in adolescents (Nock et al. 2010) and in the challenging.
majority of self-injuring adolescents (Nock and Prinstein In the present study, Gratz and Roemer’s (2004) model of
2004; Kumar et al. 2004; Nixon et al. 2002; Penn et al. ER was used as a theoretical basis because of its previous
2003). Moreover, just as in the adult literature, adolescents use in NSSI research. This model defines ER as “involving
report experiencing negative emotional states (such as an- the (a) awareness and understanding of emotions, (b) accep-
ger, depression, and loneliness and frustration) prior to self- tance of emotions, (c) ability to control impulsive behaviors
injuring, with a marked decrease in those emotions both and behave in accordance with desired goals when experi-
during and after the self-injuring episode, lending further encing negative emotions, and (d) ability to use situationally
support to an affect regulation and management model of appropriate emotion regulation strategies flexibly to modu-
NSSI (Laye-Gindhu and Schonert-Reichl 2005). late emotional responses as desired in order to meet individ-
While the above research clearly demonstrates the link ual goals and situational demands” (p. 42). Gratz and
between ER and NSSI in both adults and adolescents, it is Roemer’s (2004) broad definition of ER comes from cited
tempered by the reality that emotion regulation is a multi- evidence suggesting that ER includes both the capacity to
dimensional construct (Gratz and Roemer 2004; Chapman regulate emotional responses and the ability to experience
et al. 2006; Linehan 1993) and there is currently no consen- and distinguish a broad spectrum of emotions and accept
sus in the conceptualization of ER (Gratz and Roemer those emotions rather than trying to repress them. Therefore,
2004). Defined generally, emotion dysregulation refers to in the Gratz and Roemer model, adaptive ER includes
“maladaptive emotional responsiveness reflected in dys- having both a repertoire of ER strategies and sufficient
functional understanding, reactivity, and management” flexibility to use them. These emotion regulation strategies
(Mennin et al. 2007, pg. 295–296). In defining ER, it is allow the individual to regulate the intensity and duration of
J Psychopathol Behav Assess

such emotions, such that the urgency associated with its (Gratz and Roemer 2008) and community-dwelling adoles-
expression, as well as the individual’s behavior and response cents (Neumann et al. 2010; Weinberg and Klonsky 2009),
to an emotionally charged situation, are controlled. our aim was to confirm the factor structure rather than build
Based on this model, Gratz and Roemer (2004) devel- a new theoretical model of a competing factor structure. It is
oped the Difficulties in Emotion Regulation Scale (DERS) well known that factor structures established in adults or
to map onto and assess their model of ER. Accordingly, the community samples of adolescents are not by default direct-
DERS encompasses various aspects of ER, including non- ly transferred to clinical adolescent samples (Sharp et al.
acceptance of emotional responses (nonacceptance), diffi- 2006). Demonstrating a similar factor structure for the
culties engaging in goal directed behavior (goals), impulse DERS in a clinical sample that has been previously shown
control difficulties (impulse), lack of emotional awareness for adult and adolescent community samples would have
(awareness), limited access to emotion regulation strategies direct relevance for the use of the DERS in clinical settings.
(strategies), and lack of emotional clarity (clarity). The Establishing the factor structure in the clinical sample of
DERS stands to improve existing research on ER because, the present study also enables us to further explore addi-
as a comprehensive measure with multiple subscales, it tional questions that speak to the relevance of the DERS in
explores several dimensions of ER at once. It is important to clinical settings. Therefore, our second aim was to ex-
note that using the Gratz and Roemer (2004) model of ER and amine which aspect(s) of difficulties in ER is/are most
the DERS represents only one approach to conceptualizing relevant to NSSI in adolescents because few studies
ER which, while valuable in the present study due to previous have explored all aspects of ER as defined by Gratz and
links to NSSI, does not reflect the existing variety of models Roemer (2004). Gratz and Roemer (2008) used the DERS in a
and assessments of ER. Moreover, Chapman et al. (2006), self-harming female college sample, and found that limited
who provide a model specifically addressing the role of emo- access to emotion regulation strategies and lack of emotional
tion in NSSI, highlight emotion regulation skills deficits, clarity accounted for greater variance in NSSI above and
avoidance, and difficulty regulating emotions when aroused beyond other aspects of ER measured by the DERS. Indeed,
as separate entities while the Gratz and Roemer (2004) model limited access to emotion regulation strategies was suggested
combines them under the heading of emotion regulation. by Chapman et al. (2006) as a potentially important ER factor
The downward extension of the DERS to adolescents that may lead to NSSI. Specifically, Chapman et al. (2006)
who display NSSI has not yet occurred. However, its poten- suggest that individuals with limited access to emotion regu-
tial for use in adolescent samples was demonstrated by two lation strategies that they perceive as effective, when faced
other studies. Weinberg and Klonsky (2009), for instance, with intense emotions, may be more likely to engage in
used exploratory factor analysis to confirm the six-factor maladaptive coping, such as avoidance which, in turn, may
structure of the DERS in a community-dwelling adolescent lead to NSSI (see model pg. 373). Importantly, the DERS
sample. The authors noted good to excellent internal con- strategies subscale contains items reflecting an individual’s
sistency in the 6 DERS subscales (although they noted only belief that emotion regulation strategies will be ineffective and
modest internal consistency in the “awareness” subscale), that negative emotions will continue regardless of what they
with overall good reliability and validity of the DERS scale. do. In other words, items do not probe whether a respondent
Similarly, the DERS’ factor structure has also been confirmed can use strategies such as breathing, distraction, etc. but instead
in a large sample of Dutch community dwelling adolescents probe whether the individual believes that he or she is able to
(Neumann et al. 2010). use strategies effectively to lessen negative emotions and avoid
Taken together, the DERS shows promise for investigat- being overwhelmed. We predicted that limited access to emo-
ing ER in adolescents, and, due to the comprehensive nature tion strategies would be most associated with NSSI, even
of the DERS, holds potential for examining the relation when controlling for possible confounds such as sex, age,
between various aspects of difficulties in ER and NSSI in and psychopathology, which are often neglected in stud-
a wide variety of populations. There is a great need to ies of ER and NSSI. Controlling for these variables is
examine difficulties in ER as it relates to NSSI in adoles- important, given findings indicating that adolescent girls
cents because of the high rates of NSSI amongst adolescents are more likely to self-injure than boys (Hawton and
(DiClemente et al. 1991; Sim et al. 2009) in addition to the Harriss 2008) and that the risk for NSSI increases with
fact that the age of onset for NSSI is typically during age within adolescence (Sourander et al. 2006) and with
adolescence (Kumar et al. 2004). a diagnosis of depression, anxiety (Andover et al. 2005), and
Against this background, the first aim of the present externalizing disorder (Nock et al. 2006).
study was to confirm the six-factor structure of the DERS Our third aim was to establish the clinical utility value of
(nonacceptance, goals, impulse, awareness, strategies, and the DERS for the presence of lifetime NSSI. This is accom-
clarity) in a sample of adolescent inpatients. Given the now plished by determining sensitivity, specificity, and the clin-
well-established six-factor structure of the DERS in adults ical cut-off score of the most predictive subscale of the
 J Psychopathol Behav Assess

DERS when used with inpatient adolescents. One barrier to comorbidity. In this sample, the average number of Axis I
effective treatment of NSSI in adolescents is the extremely diagnoses was 2.38 (SD02.44). Diagnostic and demographic
limited number of reliable and valid assessment measures information for this sample is presented in Table 1.
for identified risk factors, particularly emotional regulation
skills (Zeman et al. 2006). Reporting biases in NSSI in
Measures
adolescence have furthermore been suggested (Nock
2010), which necessitates proxy measures of NSSI in hos-
Nonsuicidal Self-Injury The Deliberate Self Harm Inventory
pital settings. By establishing clinical cut-offs for a relevant
(DSHI; Gratz 2001) is a 17-item self-report measure that
DERS subscale, this measure can be used in clinical settings
assesses the frequency, severity, duration, and type of self-
as a valuable tool to identify those most at risk for NSSI
harm (i.e. NSSI) behavior. This measure specifies self-harm
based on deficits in ER. For instance, it is possible that
behaviors as those that are “deliberate, direct destruction or
adolescents may be reluctant to endorse NSSI items with
alteration of body tissue without conscious suicidal attempt,
high face validity. Proxy measures for NSSI (like a relevant
but resulting in injury severe enough for tissue damage (e.g.,
DERS subscale) may be extremely useful in this regard.
scarring) to occur” (Gratz 2001, p. 255). The DSHI has
Taken together, this study stands to add important validation
demonstrated high internal consistency (α0.82–.83), test-
information to the use of the DERS and its subscales in
retest reliability (φ0.68, p<.001 for dichotomous use and
investigating and detecting NSSI in the context of inpatient
r0.92, p<.001 for continuous use), construct validity, and
adolescent psychiatric problems.
concurrent validity (Gratz 2001). Following the guidelines
set forth in Gratz and Tull (2010) when comparing individuals
with a history of NSSI versus those without, a dichotomous
Method
variable was created in which adolescents who answered yes
to any item were assigned to the “NSSI” group and those who
Participants
did not answer yes to any item were assigned to the “No
NSSI” group. 64.7 % of the sample had self-injured during
All consecutive admissions (N0275) from a 16-bed adolescent
their lifetime and was therefore included in the “NSSI” group.
inpatient unit which usually serves adolescents with severe
treatment-refractory behavior, psychiatric, and substance dis-
Difficulties in Emotion Regulation The Difficulties in Emo-
orders were invited to participate. The average length of stay in
tion Regulation Scale (DERS; Gratz and Roemer 2004) is a
this program is 5–7 weeks. 53.2 % of the sample had previ-
self-report questionnaire measure that assesses emotion
ously been hospitalized for psychiatric reasons, including drug
and alcohol detoxification and rehabilitation. The median and
modal number of previous hospitalizations were 1 and 0, Table 1 Diagnostic and demographic characteristics of the sample
respectively. The treatment follows a milieu-based integrative
approach that includes both psychodynamic (mentalization- Percentage of sample
based) and cognitive-behavioral approaches.
Female sex 58.7 %
Of those approached for consent, 23 declined participa-
Any internalizing disorder 59.2 %
tion in the study, 9 consented but did not complete the
Depressive disordera 38.5 %
battery, 2 consented and later revoked consent, 10 were
Anxiety disorderb 47.7 %
discharged prior to being assessed, and 11 were excluded
Any externalizing disorder 41.1 %
on the basis of (a) diagnosis of schizophrenia or any psy-
ADHD 18.8 %
chotic disorder, and/or (b) diagnosis of mental retardation.
Conduct disorder 20.2 %
Inclusion criteria were age between 12 and 17 and English
Oppositional defiant disorder 19.7 %
fluency. On the grounds of these exclusion criteria, 11 were
Other
excluded from the study, leaving 220 adolescents in the
Eating disorderc 6.4 %
sample. Participants missing one or more complete meas-
ures were excluded, amounting to a total of 2 participants. Mania 4.6 %
The final sample consisted of 218 adolescents with a mean Hypomania 3.2 %
age of 15.93 years (SD01.41). The sample contained 128 Schizophrenia 3.2 %
(58.7 %) females and 90 (41.3 %) males and had the following a
Includes major depressive disorder and dysthymia
ethnic breakdown: 90.8 % White, 2.8 % Hispanic, 1.8 % b
Includes social phobia, separation anxiety, specific phobia, panic disorder,
Asian, 0.9 % Black, 1.8 % Mixed, 0.5 % Native Hawaiian agoraphobia, generalized anxiety disorder, post-traumatic stress disorder,
or Pacific Islander, and 1.4 % Other or Unreported. It is to be and obsessive compulsive disorder
c
expected that a severe inpatient sample display high diagnostic Includes anorexia nervosa and bulimia nervosa
J Psychopathol Behav Assess

dysregulation. It consists of 36 items that are scored on a 5 adolescent was obtained in person. During the consenting
point Likert scale, ranging from 1 (‘almost never (0–10 %)’) procedure, the limits of confidentiality were discussed. Spe-
to 5 (‘almost always (91–100 %)’). A higher score indicates cifically, adolescents were told that all information provided
greater emotion dysregulation. The measure assesses six would remain confidential with two exceptions, namely if
separate scales including: nonacceptance, goals, impulse, they (a) disclosed information suggesting that they were
awareness, strategies, and clarity. In the measure’s initial currently at risk of harming themselves or others, or (b)
publication, the DERS displayed good internal consis- disclosed abuse or neglect of a minor, elderly individual,
tency (α 0.93), construct and predictive validity, and or adult with a disability. It is important to note that adoles-
test-retest reliability across 4–8 weeks (p<.01) (Gratz cents were informed that retrospective information regard-
and Roemer 2004). In the present study, each subscale had ing suicide-related behaviors would not result in a breach of
good internal consistency (α: nonacceptance0.90, goals0.86, confidentiality.
impulse0.92, awareness0.85, strategies0.92, clarity0.81). Adolescents were then consecutively assessed by doc-
toral level clinical psychology students, licensed clini-
Psychopathology The Computerized Diagnostic Interview cians, and/or trained clinical research assistants. Diagnostic
Schedule for Children(C-DISC; Shaffer et al. 2000) is a highly interviews were conducted independently and in private
structured clinical interview used to diagnose psychiatric dis- with the adolescents according to the standard proce-
orders in children and adolescents between the ages of 9 and dures of the C-DISC previously described. All adolescents
17. While it is designed to be administered by lay inter- were assessed within the first two weeks following
viewers, all adolescents in this study were interviewed by admission.
doctoral psychology students or clinical research assistants
who had completed training and several practice sessions
administering the interview under the supervision of the fourth Results
author (CS). The interview is administered following comput-
erized prompts that the interviewer reads out loud. The C- The most common form of NSSI in the present sample was
DISC has previously demonstrated adequate test-retest reli- cutting, with 51.8 % of the sample endorsing this behavior
ability after 1 year for most diagnoses (κ0.25–.92) (Shaffer et in their lifetime. The other methods assessed and the per-
al. 2000). Though no validity information was provided in the centage of adolescents endorsing the behavior is as follows:
initial C-DISC publication, later studies have called into ques- burning with cigarette 11.9 %, burning with lighter or match
tion the validity of the C-DISC when compared with clinician 24.3 %, carving words into skin 28.9 %, carving pictures
diagnoses (Lewczyk et al. 2003) though it remains common into skin 22.5 %, severely scratched self 37.6 %, bit self
practice to opt for the C-DISC in research due to the structured 20.6 %, rubbed with sandpaper 4.6 %, dripped acid onto
nature of the assessment. skin 0.9 %, used cleaner to scrub skin 1.8 %, stuck sharp
For the purposes of this study, only diagnoses that met objects into skin 30.3 %, rubbed glass into skin 11.9 %,
full DSM-IV criteria on the clinical report of the C-DISC broken own bones 0.9 %, banged head 18.8 %, punched self
were considered. All positive diagnoses of anxiety (includ- 20.6 %, prevented wounds from healing 28.9 %, and other
ing social phobia, separation anxiety, specific phobia, panic method 19.3 %. The average age of onset for cutting was
disorder, agoraphobia, generalized anxiety disorder, post- 13.55 years (SD02.043) and the lifetime frequency of cut-
traumatic stress disorder, and obsessive compulsive disor- ting ranged between 0 and 40 times with a mean of 5 times
der), dysthymia, and major depressive disorder were (SD08.206). For 8.3 % of the sample, cutting had resulted
grouped into the “any internalizing” category. Finally, in hospitalization or injury severe enough to require medical
ADHD, conduct disorder, and oppositional defiant disorder treatment at some point.
were grouped into the “any externalizing” category. The
collapsing of overarching groups for analyses was motivat-
ed by a desire to avoid clumsiness and multiple comparisons Confirmatory Factor Analysis of the DERS
in the data analyses.
To confirm the factor structure reported by Neumann et al.
Procedures (2010) in a clinical adolescent sample, we used confirmato-
ry factor analysis (CFA) for ordered-categorical variables
The study was approved by the appropriate institutional based on polychoric correlations. CFA models with categor-
review board. All adolescents admitted to an inpatient ical indicators generally require large sample sizes in order
psychiatric unit were approached on the day of admission to obtain accurate test statistics, parameter estimates, and
about participating in this study. Informed consent from the standard errors (Brown 2006). However, for simpler models
parents was collected first, and if granted, assent from the with a modest number of indicators (as in the present
 J Psychopathol Behav Assess

investigation), sample sizes of 150–200 have been found to found to differ significantly by sex, internalizing diagnosis,
be sufficient (Brown 2006; Flora and Curran 2004). and externalizing diagnosis, all three variables were con-
This confirmatory model contained six latent factors: trolled for in subsequent analyses. In order to determine the
nonacceptance, goals, impulse, awareness, strategies, and relation between NSSI, age, and DERS subscale scores,
clarity. All latent factors were allowed to correlate freely. independent samples t-tests comparing those who did and
The number of individual items loading on each factor did not endorse NSSI were conducted. These results show
varied between five and eight items (for the specific items that age did not differ significantly by NSSI group. All of
loading on each factor, see Table 2). One item (33) was these results are presented in Table 4.
allowed to cross-load on two distinct factors, as reported by
Neumann et al. (2010), and no item-residuals were allowed The Relation Between DERS and NSSI Controlling for Sex,
to correlate. This factor analysis was conducted with the Age, and Psychopathology
Mplus 6.0 software program (Muthén and Muthén 1998–
2010) using the weighted-least squares multivariate estima- To explore which aspects of emotion dysregulation were
tor. Model fit was evaluated using established recommen- most relevant to NSSI at the cross-sectional level, while
dations for two well-known fit indices, Comparative Fit controlling for internalizing, externalizing, and sex, it was
Index (CFI), Root Mean Square Error of Approximation necessary to first establish whether multicollinearity existed
(RMSEA) and Weighted Root Mean Square Residual for the all predictor variables given the high correlations
(WRMR). Good model fit is indicated by a CFI greater than between DERS subscale scores (see Table 3) and the previ-
or equal to .95, RMSEA values close to .06, and WRMR<.90 ously established relation between psychopathology and
(Hu and Bentler 1999). Additional recommendations advise ER. Multicollinearity was assessed by calculating the formal
that RMSEA values of .08 or less indicate acceptable fit detection-tolerance and the variance inflation factor (VIF).
and values greater than .08 indicate poor model fit (Brown Because multicollinearity was not a problem, with tolerance
and Cudeck 1993). Inspection of most fit indices indicated greater than 0.1 (Menard 1995) and a VIF less than 10
that this model fit the data acceptably well (CFI0.95; (Myers 1990) for every predictor, centering the predictor
RMSEA0.08), although WRMR01.37. The vast majority variables was not necessary (Aiken and West 1991; Holmbeck
of items loaded strongly on their respective latent factor 2002).
(specific item loadings are available in Table 2). Additionally, Next, sex, internalizing disorder, externalizing disorder,
all six latent factors (corresponding to the six subscales of the nonacceptance, goals, impulse, awareness, strategies, and
DERS) were significantly correlated with each other, with the clarity were explored as predictor variables in binary logistic
exception of the correlation between awareness and goals regression with the DSHI dichotomous variable as the out-
which was nonsignificant; these correlations are presented in come variable. Sex, internalizing, and externalizing were
Table 3. entered in the first step and the aforementioned DERS sub-
scales were entered in the second step. These results are
Preliminary Analyses of Bivariate Relations Between Key presented in Table 5. Only sex, internalizing disorder, and
Study Variables the DERS subscale assessing limited access to emotion
regulation strategies retained significance in the second
The second aim of the current study was to investigate the step. Addition of this subscale to the variables already
relations between aspects of difficulties in ER and NSSI. included in the first step incrementally improved the model
Because several other variables apart from difficulties in ER (Δ-2 Log Likelihood012.322, p<.001).
may account for this relation, we first explored potential
confounding variables, including sex, age, and diagnoses of The Clinical Utility Value of the DERS in Detecting
internalizing or externalizing disorders. The purpose of Lifetime Presence of NSSI
these analyses was to identify confounds that should be
included in analyses related to the second aim. Chi-square To establish the clinical utility value of the DERS in detect-
analyses by sex revealed that females were significantly ing the presence of NSSI, we used Receiver Operating
more likely to self-injure than males. Chi-square analyses Characteristics (ROC) analysis. Specifically, we sought to
by internalizing diagnosis were also conducted and showed assess the performance of the DERS limited access to emo-
that individuals with an internalizing diagnosis were signif- tion regulation subscale score in detecting lifetime NSSI
icantly more likely to self-injure than those without an status, thereby establishing a clinical cut-off score. A ROC
internalizing diagnoses. Finally, chi-square analyses by di- curve is created when the true positive rate (sensitivity) is
agnosis of externalizing disorder revealed that individuals plotted against the false positive (1—specificity) rate. The
diagnosed with an externalizing disorder were more likely to area under the curve (AUC) can then be calculated using the
self-injure than those without. Because NSSI status was non-parametric trapezoid method (Hanley and McNeil
J Psychopathol Behav Assess

Table 2 CFA item factor

loadings Factor Item Loading (SE)

A. Awareness
2 I pay attention to how I feel 0.85 (0.025)
6 I am attentive to my feelings 0.93 (0.022)
8 I care about what I am feeling 0.76 (0.035)
10 When I’m upset, I acknowledge my emotions 0.63 (0.049)
17 When I’m upset, I believe my emotions are valid and important 0.67 (0.045)
34 I take time to figure out what I am really feeling 0.59 (0.053)
B. Clarity
1 I am clear about my feelings 0.71 (0.041)
4 I have no idea how I am feeling 0.74 (0.037)
5 I have difficulty making sense out of my feelings 0.84 (0.031)
7 I know exactly how I am feeling 0.76 (0.037)
9 I am confused about how I am feeling 0.71 (0.039)
C. Impulse
3 I experience my emotions as overwhelming and out of control 0.82 (0.036)
14 When I’m upset, I become out of control 0.93 (0.011)
19 When I’m upset, I feel out of control 0.94 (0.011)
24 When I’m upset, I feel I can remain in control over my behavior 0.62 (0.045)
27 When I’m upset, I have difficulty controlling my behavior 0.95 (0.010)
32 When I’m upset, I lose control over my behavior 0.97 (0.008)
D. Goals
13 When I’m upset, I have difficulty getting work done 0.85 (0.027)
18 When I’m upset, I have difficulty focusing on other things 0.92 (0.019)
20 When I’m upset, I can still get things done 0.54 (0.052)
26 When I’m upset, I have difficulty concentrating 0.93 (0.020)
33a When I’m upset, I have difficulty thinking about anything else 0.56 (0.051)
E. Nonacceptance
11 When I’m upset, I become angry at myself for feeling that way 0.85 (0.028)
12 When I’m upset, I become embarrassed 0.76 (0.036)
21 When I’m upset, I feel ashamed with myself 0.90 (0.020)
23 When I’m upset, I feel like I am weak 0.79 (0.041)
25 When I’m upset, I feel guilty 0.80 (0.032)
29 When I’m upset, I become irritated with myself 0.93 (0.018)
F. Strategies
15 When I’m upset, I believe I’ll remain that way for a long time 0.85 (0.022)
16 When I’m upset, I believe that I’ll end up very depressed 0.82 (0.024)
22 When I’m upset, I know that I can find a way to feel better 0.63 (0.043)
28 When I’m upset, I believe there is nothing I can do to feel better 0.82 (0.026)
30 When I’m upset, I start to feeling very bad about myself 0.90 (0.020)
31 When I’m upset, I believe that wallowing in it is all I can do 0.80 (0.029)
35 When I’m upset, it takes me a long time to feel better 0.84 (0.023)
a
Item loads on two distinct fac- 36 When I’m upset, my emotions feel overwhelming 0.89 (0.022)
tors. All factor loadings are fully 33a When I’m upset, I have difficulty thinking about anything else 0.34 (0.058)
standardized

1982) that yields an index of accuracy which has been and high accuracy when greater than .9 (Swets and
used in several other studies to establish criterion valid- Pickett 1982). The measure’s cut-off score can be estab-
ity (Thapar and McGuffin 1998; Fombonne 1991). A lished by finding the intersection of the measure’s sen-
measure is thought to have low diagnostic accuracy if sitivity and specificity curves. These analyses were
its AUC is below .7, moderate accuracy from .7 to .9, completed using SPSS (2009), Release 17.0.2.
 J Psychopathol Behav Assess

Table 3 CFA factor covariances and correlations

Factor Emotional awareness Emotional clarity Impulsivity Goal-directed behavior Nonacceptance ER strategies

Awareness 0.71 0.74 0.22 0.10* 0.31 0.21

Clarity 0.44 0.50 0.44 0.40 0.46 0.58
Impulse 0.15 0.25 0.68 0.64 0.51 0.73
Goals 0.07 0.24 0.45 0.72 0.52 0.75
Nonacceptance 0.22 0.28 0.37 0.38 0.73 0.73
Strategies 0.15 0.32 0.51 0.54 0.52 0.72

Factor correlations are presented above the diagonal in bold, with factor covariances below the diagonal and factor variances listed on the diagonal.
All correlations are significant (p<.01) unless otherwise noted. *Correlation not significant at .05

The ROC curve with DERS limited access to emotion the cut-off had been hospitalized for NSSI before; χ2 09.08,
regulation strategies subscale and NSSI status is shown in p0.003, df01, φ0.204, p0.003), though the two groups did
Fig. 1. Both the AUC and standard error were significant not differ with regard to age of onset of self-injury (Mabove 0
(p<.001), with an AUC of 0.728, indicating moderate diag- 13.46, Mbelow 013.52, t0.171, p0.865).
nostic accuracy. Additionally, plotting sensitivity and spec-
ificity (Fig. 2) at different cut-off scores on the DERS
subscale score indicated that the optimal cut-point, the in- Discussion
tersection of sensitivity and specificity, is 21.5 (Se0.69,
Sp0.70) when assessing NSSI. Independent samples t-test The current study had three aims. Since the underlying
and Chi-square analyses were used to explore differences in factor structure of the DERS had not yet been established
NSSI behavior using this cut-off score and revealed that in a clinical adolescent sample, our first aim was to confirm
adolescents above the cut-off used a greater number of the six-factor structure previously reported for adults (Gratz
methods to self-injure (M above 04.82, M below 01.57, and Roemer 2008) and for community-dwelling adolescents
t 0−8.04, p<.001, df0199.22, d0−1.07) and were more (Neumann et al. 2010). Given the adequate model fit indi-
likely to have been hospitalized as a result (13.30 % of cated by two of the three fit indices and the previous work
those above the cut-off compared with 2.0 % of those below supporting this model, we interpret this model to show

Table 4 Comparison of NSSI and no NSSI groups (t-tests of Chi-square analyses) with regard to demographics, psychopathology, and emotion regulation

Chi-Square analyses No NSSI NSSI χ2 (df) φ

% (n) endorsed % (n) endorsed
Female sex 25.00 (32) 75.00 (96) 14.46*** (1) −0.258***
Male sex 50.00 (45) 50.00 (45)
With internalizing disorder 22.90 (30) 77.10 (101) 24.72*** (1) 0.346***
Without internalizing disorder 57.33 (43) 42.67 (32)
With externalizing disorder 24.10 (20) 75.90 (63) 7.55** (1) 0.193**
Without externalizing disorder 42.86 (51) 57.14 (68)

Independent samples t-tests No NSSI NSSI t d df rpb

M (SD) M (SD)
Age 15.95 (1.46) 15.92 (1.39) 0.113 0.02 216 −0.0008
DERS
A. Nonacceptance 12.00 (5.96) 15.36 (6.81) −3.78*** −0.53 174.67a 0.240***
B. Goals 16.69 (5.68) 19.15 (4.89) −3.35** −0.46 216 0.222**
C. Impulse 13.30 (5.76) 17.38 (6.94) −4.40*** −0.64 216 0.286***
D. Awareness 16.12 (6.01) 17.97 (5.59) −2.28* −0.32 216 0.153*
E. Strategies 18.68 (8.34) 25.65 (8.29) −5.93*** −0.84 216 0.374***
F. Clarity 12.36 (5.23) 14.55 (4.74) −3.14** −0.44 216 0.209**

C-DISC data was not available for the full sample

a
Degrees of freedom are decreased because equal variances were not assumed due to a significant Levene Test for Equality of Variances. rpb 0 Point
Biserial correlation. *p<.05 **p<.01 ***p<.001
J Psychopathol Behav Assess

Table 5 Regression weights when predicting NSSI status have parceled out important features of ER while ignoring
Predictor B SE Exp(B) a
95% CI b
p others when investigating NSSI. Sim et al. (2009), for
instance, limited their exploration of difficulties in ER
Step 1: Covariates to emotional awareness and expression, thereby exclud-
Sex −.99 0.33 0.37 0.20–.72 0.003** ing several components of Gratz and Roemer’s (2004)
Internalizing disorder 1.34 0.33 3.83 2.00–7.35 >.001*** definition of ER. Similarly, Mikolajczak et al. (2009)
Externalizing disorder 0.75 0.35 2.12 1.07–4.21 0.031* explored the relation between emotional coping and
Step 2: DERS subscales NSSI, but did not directly investigate the behavioral
Nonacceptance −0.03 0.04 0.97 0.90–1.04 0.454 component of ER. By using multiple dimensions of
Goals −0.08 0.05 0.92 0.84–1.01 0.081 ER, we demonstrated that only one subscale (limited
Impulse 0.01 0.04 1.01 0.94–1.09 0.795 access to emotion regulation strategies, defined as the
Awareness 0.05 0.04 1.05 0.97–1.13 0.208 “belief that there is little that can be done to regulate
Strategies 0.12 0.04 1.12 1.04–1.21 0.002** emotions effectively once an individual is upset”; Gratz
Clarity −0.02 0.05 0.99 0.90–1.09 0.733 and Roemer 2004; p. 47) remained significantly associ-
ated with NSSI status after controlling for other aspects
Sex and internalizing disorder retained significance in Step 2. Change of emotion dysregulation, sex and psychopathology.
in −2 Log Likelihood012.322, p<.001
This finding is consistent with previous research that
a
Exp(B) 0 Odds ratio for each predictor
b
found that this factor is an independent statistical predictor
Confidence interval is created around Exp(B), statistically significant of NSSI when assessing difficulties in ER (Gratz and
if 1 is not in the interval. *p<.05 **p<.01
Roemer 2004, 2008; Gratz and Tull 2010). Our findings
are also in line with research indicating that NSSI is often
used as a maladaptive alternative (in the absence of adaptive
adequate fit. However, this should be replicated in other strategies) to cope with unwanted feelings and emotions
clinical adolescent samples. (Lloyd-Richardson et al. 2007; Rodham et al. 2004), as well
Our second aim was to examine which of the components as several prominent theories which posit that one of the
of difficulties in ER was most strongly associated with NSSI functions of NSSI is as a means of emotion or affect regu-
status in a clinical sample of adolescents, while controlling lation, such as the Experiential Avoidance Model mentioned
for possible confounding variables. As such, this is the first previously (Chapman et al. 2006), Nock’s (2010) intraper-
study to consider multiple dimensions of ER in the same sonal model, and general affect-regulation models (Klonsky
sample of inpatient adolescents. Adrian et al. (2010) exam- 2007). Because the DERS strategies subscale targets beliefs
ined only overall ER difficulties, while other researchers that emotion regulation strategies will not be effective and that
negative emotions will continue and become overwhelming,
the aforementioned research may suggest that rather than not
having access to emotion regulation strategies, NSSI is associ-
ated with beliefs of inefficacy in emotion regulation. This
distinction is highly important clinically, because it points to
the possibility of restructuring cognitions surrounding emotion
regulation efficacy as an important aspect of NSSI treatment.
Lastly, the third aim of the study was to determine
the clinical utility value of the DERS in a clinical
sample. This was accomplished by determining the sen-
sitivity, specificity, and clinical cut-off score of the
measure when used with inpatient adolescents. The
DERS strategies subscale demonstrated adequate sensi-
tivity and specificity and served as a moderate predictor
in detecting the presence of NSSI. Additionally, plotting
sensitivity and specificity at different cut-off scores on
the DERS subscale score provided the optimal cut-point,
Fig. 1 ROC curve of DERS limited access to emotion regulation strat- the intersection of sensitivity and specificity, when
egies subscale in detecting presence of lifetime self-injurious behavior. assessing NSSI. When this cutoff was explored further,
Note. There were 141 cases positive for self-injury and 77 cases negative
for self-injury in this analysis. The AUC is 0.714 (SE0.037, p<.001),
it revealed group differences with regard to whether
indicating moderate accuracy in discriminating adolescents who engage NSSI had ever resulted in hospitalization and the num-
in self-injurious behaviors ber of methods of NSSI employed.
 J Psychopathol Behav Assess

Fig. 2 Sensitivity and 1.00

specificity plotted against
different cut-off scores on the 0.90
DERS limited access to emo- 0.80
tion regulation strategies sub-
scale in reference to self-injury. 0.70
Note. The optimal cut-off score
0.60
is determined by the intersec-
tion of the sensitivity and spec- 0.50 Sensitivity
ificity lines. In predicting self- Specificity
injury, the optimal cut-off score 0.40
for the DERS limited access to 0.30
emotion regulation strategies is
21.5 (Sensitivity0.69, 0.20
Specificity0.70)
0.10
0.00
7.00 10.00 13.00 16.00 19.00 22.00 25.00 28.00 31.00 34.00 37.00 40.00

Taken together, the findings reported here are signif- temporal and potentially causal relation between ER strate-
icant for several reasons. First, we confirm the factor gies and NSSI. Similarly, the study design focused on the
structure and clinical utility of the DERS for the con- utility of the DERS limited access to ER strategies subscale
current assessment of lifetime NSSI in clinical samples. in predicting lifetime NSSI, and therefore cannot speak to
Secondly, we provide further support for the relation the validity of the DERS for predicting current NSSI nor the
between difficulties in ER and NSSI in adolescents course of NSSI. Furthermore, the present sample was com-
and we provide information on the clinical utility of promised of adolescents in a psychiatric hospital, which
the limited access to emotion regulation strategies sub- represent only a small, albeit extreme, part of the overall
scale of the DERS. Methodologically, the study focused self-injuring population. Thus, our findings may not gener-
on a relatively large clinical sample and built upon alize to populations in which NSSI is presumably not as
existing research with community samples. This allows extreme. Additionally, the measure of NSSI employed in
us to place our results within the context of previous this study, while widely used and validated, does not collect
work with non-clinical adolescents with regard to meas- information about individuals’ motivations for self-injuring
ures of ER and NSSI. Furthermore, confirming the which is of great theoretical importance given findings that
factor structure of the DERS in a clinical sample serves adolescents self-injure for a variety of reasons (see Nock
to expand its utility for clinical research. Similarly, the 2010). Thus, we are unable to draw conclusions about
thorough investigation of many aspects of ER provided potential differences in the relation between limited ER
by the DERS allowed us to build upon existing research strategies and NSSI that may have emerged when motiva-
previously limited only to specific components of ER. tion for self-injuring is examined more closely. Also, al-
Also, determining psychopathology on the basis of a though we found that limited access to ER strategies was
structured clinical interview allowed us to confidently the only subscale of the DERS to independently predict
control for psychopathology in addition to demographic history of NSSI, the exact nature of the dependent variable
factors in our analyses. For these reasons, the present in this study needs to be considered. It is very possible that
study makes a valuable contribution to the existing other components of emotion dysregulation may relate dif-
research on NSSI in adolescents while pointing to the ferently to other NSSI outcomes, for instance, severity of
importance of subsequent research in the area of ER NSSI. Therefore, these findings cannot be used to recom-
strategies in order to understand its causal relation to mend that clinicians exclusively focus on improving access
NSSI and the extent to which it should be a target of to ER strategies at the cost of other components of ER. We
clinical treatment. merely suggest that limited access to ER strategies appears
However, the current study has several limitations of to be a very relevant facet of ER in determining the presence
note. Perhaps most importantly, the present study seeks to of NSSI. Relatedly, the present study made use of only
draw conclusions about the role of limited ER strategies in Gratz and Roemer’s (2004) model and assessment of ER
NSSI while relying upon a cross-sectional design and self- and did not compare this model to others proposed in the
report data. The fact that data is not collected at multiple existing literature. Therefore, these findings cannot speak to
time points prohibits conclusions about the causal impor- the validity of the Gratz and Roemer (2004) model of ER
tance of adequate ER strategies. Thus, the present study above another model of ER. Though this was not the aim of
highlights the importance of further research to explore the the present study it is a valuable area for future research.
J Psychopathol Behav Assess

Finally, future work may benefit from testing the hypotheses and initial validation of the difficulties in emotion regulation scale.
Journal of Psychopathology and Behavioral Assessment, 26(1), 41–
in the current study with more sophisticated data analytic
54. doi:10.1023/b:joba.0000007455.08539.94.
approaches like Structural Equation Modeling. Gratz, K. L., & Roemer, L. (2008). The relationship between emotion
dysregulation and deliberate self-harm among female undergrad-
uate students at an urban commuter university. Cognitive Behav-
References iour Therapy, 37(1), 14–25. doi:10.1080/16506070701819524.
Gratz, K. L., & Tull, M. T. (2010). The relationship between emotion
dysregulation and deliberate self-harm among inpatients with
Adrian, M., Zeman, J., Erdley, C., Lisa, L., & Sim, L. (2010). Emotion substance use disorders. Cognitive Therapy and Research, 34
dysregulation and interpersonal difficulties as risk factors for (6), 544–553. doi:10.1007/s10608-009-9268-4.
nonsuicidal self-injury in adolescent girls. Journal of Abnormal Haines, J., Williams, C. L., Brain, K. L., & Wilson, G. V. (1995). The
Child Psychology. doi:10.1007/s10802-010-9465-3. psychophysiology of self-mutilation. Journal of Abnormal Psy-
Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and chology, 104, 471–489. doi:10.1037/0021-843X.104.3.471.
interpreting interactions. Newbury Park: Sage. Hanley, J. A., & McNeil, B. J. (1982). The meaning and use of the area
Andover, M. S., Pepper, C. M., Ryabchenko, K. A., Orrico, E. G., & Gibb, under a receiver operating characteristic (ROC) curve. Radiology,
B. E. (2005). Self-mutilation and symptoms of depression, anxiety, 143, 29–36. http://radiology.rsna.org/.
and borderline personality disorder. Suicide & Life-Threatening Be- Hawton, K., & Harriss, L. (2008). The changing gender ratio in
havior, 35(5), 581–591. doi:10.1521/suli.2005.35.5.581. occurrence of deliberate self-harm across the lifecycle. Crisis:
Brain, K. L., Haines, J., & Williams, C. L. (1998). The psycho- The Journal of Crisis Intervention and Suicide Prevention, 29
physiology of self-mutilation: evidence of tension reduction. (1), 4–10. doi:10.1027/0227-5910.29.1.4.
Archives of Suicide Research, 4(3), 227–242. doi:10.1023/ Holmbeck, G. N. (2002). Post-hoc probing of significant moderational
A:1009692507987. and mediational effects in studies of pediatric populations. Jour-
Brausch, A. M., & Gutierrez, P. M. (2010). Differences in non-suicidal nal of Pediatric Psychology, 27, 87–96. doi:10.1093/jpepsy/
self-injury and suicide attempts in adolescents. Journal of Youth and 27.1.87.
Adolescence, 39(3), 233–242. doi:10.1007/s10964-009-9482-0. Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in
Bresin, K., & Gordon, K. H. (2011). Changes in negative affect covariance structure analysis: conventional criteria versus new
following pain (vs. nonpainful) stimulation in individuals with alternatives. Structural Equation Modeling, 6, 1–55. doi:10.1080/
and without a history of nonsuicidal self-injury. Personality Dis- 10705519909540118.
orders: Theory, Research, and Treatment. doi:10.1037/a0025736. Klonsky, E. D. (2007). The functions of deliberate self-injury. A
Brown, T. A. (2006). Confirmatory factor analysis for applied re- review of the evidence. Clinical Psychology Review, 27, 226–
search. New York: Guilford Press. 239. doi:10.1016/j.cpr.2006.08.002.
Brown, M. W., & Cudeck, R. (1993). Alternative ways of assessing Kumar, G., Pepe, D., & Steer, R. A. (2004). Adolescent psychiatric
model fit. In K. A. Bollen & J. S. Long (Eds.), Testing structural inpatients’ self-reported reasons for cutting themselves. The Jour-
equation models (pp. 136–162). Beverly Hills: Sage. nal of Nervous and Mental Disease, 192(12), 830–836.
Brown, M. Z., Comtois, K. A., & Linehan, M. M. (2002). Reasons for doi:10.1097/01.nmd.0000146737.18053.d2.
suicide attempts and nonsuicidal self-injury in women with bor- Laye-Gindhu, A., & Schonert-Reichl, K. A. (2005). Nonsuicidal self-
derline personality disorder. Journal of Abnormal Psychology, harm among community adolescents: understanding the “whats”
111, 198–202. doi:10.1037/0021-843X.111.1.198. and “whys” of self-harm. Journal of Youth and Adolescence, 34
Chapman, A. L., Gratz, K. L., & Brown, M. Z. (2006). Solving the (5), 447–457. doi:10.1007/s10964-005-7262-z.
puzzle of deliberate self-harm: the experiential avoidance model. Lewczyk, C. M., Garland, A. F., Hurlburt, M. S., Gearity, J., & Hough,
Behaviour Research and Therapy, 44, 371–394. doi:10.1016/j.brat. R. L. (2003). Comparing DISC-IV and clinician diagnoses among
2005.03.005. youths receiving public mental health services. Journal of the
DiClemente, R. J., Ponton, L. E., & Hartley, D. (1991). Prevalence and American Academy of Child & Adolescent Psychiatry, 42(3),
correlates of cutting behavior: risk for HIV transmission. Journal 349–356. doi:10.1097/00004583-200303000-00016.
of the American Academy of Child and Adolescent Psychiatry, 30 Linehan, M. M. (1993). Cognitive-behavioral treatment for borderline
(5), 735–739. doi:10.1016/S0890-8567(10)80007-3. personality disorder. New York: Guildford Press.
Flora, D. B., & Curran, P. J. (2004). An empirical evaluation of Lloyd-Richardson, E. E., Perrine, N., Dierker, L., & Kelley, M. L.
alternative methods of estimation for confirmatory factor analysis (2007). Characteristic and functions on non-suicidal self-injury in
with ordinal data. Psychological Methods, 9, 466–491. a community sample of adolescents. Psychological Medicine: A
doi:10.1037/1082-989X.9.4.466. Journal of Research in Psychiatry and the Allied Sciences, 37(8),
Fombonne, E. (1991). The use of questionnaires in child psychiatry 1183–1192. doi:10.1017/s003329170700027x.
research: measuring their performance and choosing an optimal MacDermott, S. T., Gullone, E., Allen, J., King, N. J., & Tonge, B.
cut-off. Journal of Child Psychology and Psychiatry, 32(4), 677– (2010). The emotion regulation index for children and adolescents
693. doi:10.1111/j.1469-7610.1991.tb00343.x. (ERICA): a psychometric investigation. Journal of Psychopathol-
Franklin, J. C., Hessel, E. T., Aaron, R. V., Arthur, M. S., Heilbron, N., ogy and Behavioral Assessment, 32(3), 301–314. doi:10.1007/
& Prinstein, M. J. (2010). The functions of nonsuicidal self- s10862-009-9154-0.
injury: support for cognitive–affective regulation and opponent Menard, S. (1995). Applied logistic regression analysis. Sage univer-
processes from a novel psychophysiological paradigm. Journal of sity paper series on quantitative applications in the social scien-
Abnormal Psychology. doi:10.1037/a0020896. ces (pp. 7–106). Thousand Oaks: Sage.
Gratz, K. (2001). Measurement of deliberate self-harm: preliminary data Mennin, D. S., Holaway, R. M., Fresco, D. M., Moore, M. T., & Heimberg,
on the Deliberate Self-Harm Inventory. Journal of Psychopathology R. G. (2007). Delineating components of emotion and its dysregula-
and Behavioral Assessment, 23(4), 253–263. doi:10.1023/ tion in anxiety and mood psychopathology. Behavior Therapy, 38(3),
A:1012779403943. 284–302. doi:10.1016/j.beth.2006.09.001.
Gratz, K. L., & Roemer, L. (2004). Multidimensional assessment of Mikolajczak, M., Petrides, K. V., & Hurry, J. (2009). Adolescents
emotion regulation and dysregulation: development, factor structure, choosing self-harm as an emotion regulation strategy: the
 J Psychopathol Behav Assess

protective role of trait emotional intelligence. British Journal community sample of adolescents. Journal of the American Acad-
of Clinical Psychology, 48(2), 181–193. doi:10.1348/ emy of Child and Adolescent Psychiatry, 43(1), 80–87.
014466508x386027. doi:10.1097/00004583-200401000-00017.
Muthén, L. K., & Muthén, B. O. (1998–2010). Mplus users’ guide Shaffer, D., Fisher, P., Lucas, C. P., Dulcan, M. K., & Schwab-Stone,
(Sixth Edition). Los Angeles, CA: Muthén & Muthén. M. E. (2000). NIMH Diagnostic Interview Schedule for Children
Myers, R. (1990). Classical and modern regression with applications. Version IV (NIMH DISC-IV): description, differences from pre-
Pacific Grove: Duxbury Press. vious versions, and reliability of some common diagnoses. Jour-
Neumann, A., van Lier, P. A. C., Gratz, K. L., & Koot, H. M. (2010). nal of the American Academy of Child and Adolescent Psychiatry,
Multidimensional assessment of emotion regulation difficulties in 39(1), 28–38. doi:10.1097/00004583-200001000-00014.
adolescents using the Difficulties in Emotion Regulation Scale. Sharp, C., Goodyer, I. M., & Croudace, T. J. (2006). The Short Mood
Assessment, 17(1), 138–149. doi:10.1177/1073191109349579. and Feelings Questionnaire (SMFQ): a unidimensional item re-
Nixon, M. K., Cloutier, P. F., & Aggarwal, S. (2002). Affect regulation sponse theory and categorical data factor analysis of self-report
and addictive aspects of repetitive self-injury in hospitalized ado- ratings from a community sample of children aged 7 through 11.
lescents. Journal of the American Academy of Child and Adoles- Journal of Abnormal Child Psychology, 34(3), 365–377.
cent Psychiatry, 41(11), 1333–1341. doi:10.1097/00004583- doi:10.1007/s10802-006-9027-x.
200211000-00015. Sim, L., Adrian, M., Zeman, J., Cassano, M., & Friedrich, W. N. (2009).
Nock, M. K. (2010). Self-injury. Annual Review of Clinical Psychol- Adolescent deliberate self-harm: linkages to emotion regulation and
ogy, 6339–363. doi: 10.1146/annurev.clinpsy.121208.131258 family emotional climate. Journal of Research on Adolescence, 19
Nock, M. K., & Prinstein, M. J. (2004). A functional approach to the (1), 75–91. doi:10.1111/j.1532-7795.2009.00582.x.
assessment of self-mutilative behavior in adolescents. Journal of Sourander, A., Aromaa, M., Pihlakoski, L., Haavisto, A., Rautava, P.,
Consulting and Clinical Psychology, 72, 885–890. doi:10.1037/ Helenius, H., & Sillanpää, M. (2006). Early predictors of delib-
0022-006X.72.5.885. erate self-harm among adolescents. A prospective follow-up study
Nock, M. K., Joiner, T. E., Gordon, K. H., Lloyd-Richardson, E., & from age 3 to age 15. Journal of Affective Disorders, 93(1–3), 87–
Prinstein, M. J. (2006). Non-suicidal self-injury among adolescents: 96. doi:10.1016/j.jad.2006.02.015.
diagnostic correlates and relation to suicide attempts. Psychiatry SPSS for Windows, Rel. 17.0.3. (2009). Chicago: SPSS, Inc.
Research, 144(1), 65–72. doi:10.1016/j.psychres.2006.05.010. Suyemoto, K. L. (1998). The functions of self-mutilation. Clinical
Nock, M. K., Prinstein, M. J., & Sterba, S. K. (2010). Revealing the form Psychology Review, 18, 531–554. doi:10.1016/S0272-7358(97)
and function of self-injurious thoughts and behaviors: a real-time 00105-0.
ecological assessment study among adolescents and young adults. Swets, J. A., & Pickett, R. M. (1982). Evaluation of diagnostic sys-
Psychology of Violence, 1(S), 36–52. doi:10.1037/2152-0828.1.S.36. tems: Methods from signal detection theory. Orlando: Academic.
Penn, J. V., Esposito, C. L., Schaeffer, L. E., Fritz, G. K., & Thapar, A., & McGuffin, P. (1998). Validity of the shortened Mood and
Spirito, A. (2003). Suicide attempts and self-mutilative behavior in Feelings Questionnaire in a community sample of children and
a juvenile correctional facility. Journal of the American Academy of adolescents: a preliminary research note. Psychiatry Research, 81
Child and Adolescent Psychiatry, 42(7), 762–769. doi:10.1097/ (2), 259–268. doi:10.1016/s0165-1781(98)00073-0.
01.CHI.0000046869.56865.46. Weinberg, A., & Klonsky, E. D. (2009). Measurement of emotion
Penza-Clyve, S., & Zeman, J. (2002). Initial validation of the dysregulation in adolescents. Psychological Assessment, 21(4),
Emotion Expression Scale for Children (EESC). Journal of Clinical 616–621. doi:10.1037/a0016669.
Child and Adolescent Psychology, 31(4), 540–547. doi:10.1207/ Zeman, J., Shipman, K., & Penza-Clyve, S. (2001). Development and
153744202320802205. initial validation of the Children’s Sadness Management Scale.
Phillips, K. V., & Power, M. J. (2007). A new self-report measure of Journal of Nonverbal Behavior, 25(3), 187–205. doi:10.1023/
emotion regulation in adolescents: the regulation of emotions ques- A:1010623226626.
tionnaire. Clinical Psychology & Psychotherapy, 14(2), 145–156. Zeman, J., Cassano, M., Perry-Parrish, C., & Stegall, S. (2006). Emo-
doi:10.1002/cpp.523. tion regulation in children and adolescents. Journal of Develop-
Rodham, K., Hawton, K., & Evans, E. (2004). Reasons for deliberate mental and Behavioral Pediatrics, 27(2), 155–168. doi:10.1097/
self-harm: comparison of self-poisoners and self-cutters in a 00004703-200604000-00014.

View publication stats