232

Journal of Parapsychology © The Rhine Research Center

2019, Vol. 83, No. 2, 232-247 http://doi.org/10.30891/jopar.2019.02.08

Effects of Mood and Emotion on a Real-World Working Computer System

and Network Environment1

John G. Kruth
Rhine Research Center

Abstract. This study used a custom computer system designed to induce anxiety in participants and
determine if people who are anxious produce more errors in an independent working computer
network. Participants (N = 130) were asked to complete sixteen tasks on a computer in twenty
minutes to receive a reward. Each participant self-rated their anxiety levels during the tasks. In
addition, 130 sessions were run without a computer operator. The network ran independent of
the tasks, and operated continuously during the sessions. The first hypothesis predicted sessions
without operators would produce fewer network errors than sessions with operators, but it was not
supported (p = 0.35). The second hypothesis predicted that anxious operators would produce more
errors on the independent network than those less anxious. Initial analysis indicated an unsupport-
ed hypothesis, but the initial design did not properly identify anxious users. A post-hoc revised
grouping based on actual reported anxiety resulted in this hypothesis being supported (p = 0.04, d
= 0.45) indicating that anxious computer operators may affect network communication. There may
be other electronic effects as a result of human emotions. Additional research is necessary to con-
firm these results and explore whether the intensity of emotions affects electronics.
Keywords: electronics; emotion, network; signal fault; mind-matter interaction; PK

Since 1970, a number of studies have been published indicating that, through focused intention,
people are able to create an electronic disruption or have an influence on electronic systems or quantum
processes (e.g., Jahn, et. al. 1997; Morris, 1986; Radin, 1990; Schmidt, 1970). Unintentional effects on
larger objects have been observed in reports of apparent poltergeist activity, but instead of the activity
being attributed to a mischievous spirit or disruptive ghost, these formal investigations have focused on
unintentional effects produced by human agents who are regularly present when the activity is observed
(Pratt & Roll, 1957). Numerous investigations have reported unintentional effects on physical objects,
like photographs, blankets, trinkets, or bottles (e.g. Palmer, 1974; Roll & Storey, 2004), but other times
the effects are observed on electronic devices and phone systems (Kruth & Joines, 2015). Some of
these investigations have presented indications that anxiety and stress contribute to the demonstration
of these unintentional events (Kruth & Joines, 2015; Pratt & Roll, 1957; Roll, Burdick, & Joines, 1973,
1974). The previous studies lead to the proposition that unintentional activity may influence electronic
systems and that this activity may be exacerbated by stress and anxiety.

1 Send correspondence to: John G. Kruth, M. S., Rhine Research Center, 2741 Campus Walk Avenue, Building 500 , Durham, NC 27705,
USA, John.Kruth@rhine.org. This study was supported by a grant from the BIAL Foundation (489/14). It was preregistered with the Koestler
Parapsychology Unit’s Registry for Parapsychology Experiments.
 EFFECTS OF MOOD AND EMOTION 233

A study by Broughton and Perlstrom (1986) explored the performance of study participants play-
ing a Random Number Generator (RNG) based computer game that measured the effect of intention
and focus by the participants. The study included a measure of self-reported anxiety and questions
related to the practice of a mental discipline. It focused on performance in intentional tasks in which
participants were trying to win a computer game with focused intention. The only significant finding
was a negative correlation between higher anxiety measures and the intentional tasks. In other words,
when the participants were more anxious, they appeared to affect the RNG in the direction opposite to
their intention. Anxiety caused them to lose the game, which the authors interpreted to be the result of
an unconscious block or masked desire to avoid having the stated effect.

A field investigation by Kruth and Joines (2015) reported consistent electronic disturbances and
signal interruptions on telephones, electronic locks, computers, computer networks, and printers. These
events only occurred when a specific 11-year-old boy was present. The study also indicated that the
disturbances were reduced and eventually stopped after the boy learned to reduce his anxiety using
simple stress reduction exercises.

Anecdotal reports by people who have had Near-Death Experiences indicate that a fairly com-
mon aftereffect of a Near-Death Experience is unusual activity by electronic equipment (Atwater, 2007;
Fracasso & Friedman, 2012; Knittweis, 1997; Nouri, 2008). Despite a significant number of anecdotal
reports, no known published laboratory reports have tested these effects. That is, there are no known
published reports on the effect of human emotion and moods on the operation of computer systems
and networks in a real-world working environment.

The goals of this study were two-fold. The first goal was to create a monitored computer system
that would detect unintentional effects on the system and network stability. The second was to examine
the effects of human emotions and moods on the operation of computer systems and networks in a
simulated working environment. If human moods and emotions have an effect on the system, it is an
indication that unintentional electronic interference may result from certain emotions, and they may be
causing some computer and network malfunctions in real-world environments.

A custom computer system and network was created and the system was monitored for errors
while a sample of computer operators performed timed computer tasks in a stress-induced environ-
ment. An experimental group was purposely frustrated in their tasks by inoperative software, while a
control group performed the task without interference. The system was also run without operators to
produce a final no-operator control condition for the system operations.

The two hypotheses for this study were exploratory in nature. The design was registered as fully
specified, including the analysis techniques, in order to provide the most thorough review of hypotheses
based on the study design:

H1: Computer operators can have unintentional effects on the electronics of computer systems
and network connections, and these effects will be reduced when the computers are running automat-
ically, without operators.
234 KRUTH

H2: Anxiety and stress evoke unintentional effects on electronics and networks, and computer
operators under greater stress will demonstrate more software and network errors than operators in a
less stressful environment or an environment that does not require a computer operator.

Method

Sample
The sample consisted of 130 participants, 65 in each group, including 89 who self-identified as
females, 39 males, and two who identified as other genders. The age of the participants ranged from
18-75 years with a mean age of 47.6 years. The sample participants were obtained through electronic
and paper advertisements, mostly distributed by hand in the neighborhood and town around the re-
search lab or at a nearby university. Groups were determined using an automated randomization pro-
cess. Neither the experimenters nor the participants were aware of the group assignments. Masking was
maintained throughout the analysis process to avoid any unintentional influence on the data analysis.
Participants were informed that they were taking part in a parapsychology study exploring the interac-
tion between people and computer systems. No additional details were provided to participants to de-
scribe exactly what was being measured or how the data were being collected until after the study was
completed. None of the participants were queried about their beliefs or expectations related to the
study. The Institutional Review Board of the Rhine Research Center gave Ethics approval to this study.

Procedure
Two standard commercial desktop computers (Lenovo Model K450E) were networked in a peer-
to-peer system utilizing custom software performing standard operations and communicating using
standard networking protocols. Each session consisted of four activities.

1. Computer operators interacted with custom software on computer #1. This user interactive soft-
ware provided instructions to the participants and asked them to perform a series of tasks on the
computer in a limited amount of time (20 minutes or less).
2. Computer #1 continuously sent data to Computer #2 throughout the session. These data were
not related to the activity being completed by the participants, and was not affected by the soft-
ware used by them. (See section below on Isolation of Software Processes.) The participants had
no knowledge of this network communication and did not have any direct interaction or effect
on the data being transferred between the computers.
3. Testing and error checking software was run on both Computer #1 and #2 to monitor the network
communication for errors and log all activity and errors detected in the network communication.
4. Participants self-assessed relaxation and anxiety before and after the interactive sessions to pro-
vide an assessment of the change in anxiety or stress experienced during the session.

Participants were provided with an identification code and log in information before they began
the study, and were told that if they completed the tasks associated with the study in less than 20 min-
utes, they would receive an award of a gourmet chocolate bar and an entry in a raffle for a $200 gift card.
The rewards were selected to motivate the participants to complete the tasks.

Each participant in the study was assigned to Group 1 or Group 2 by a random selection process
 EFFECTS OF MOOD AND EMOTION 235

performed by the computer system at the moment that the participant logged in to begin the session.
Group 1 was presented with a series of timed tasks to perform on the computer, and the user interactive
software presented to this group operated in a normal manner, enabling the tasks to be completed with
a series of simple operations like button pushes, matching photos, moving items with the mouse, locat-
ing hidden images, and typing in text fields. See Appendix for details of the tasks that were completed
by each participant.

Group 2 received the same instructions and tasks as those assigned to Group 1, and the user in-
teractive software looked exactly the same as the interface used for Group 1. Group 2 was instructed to
perform the same timed tasks, but the software used by this group purposefully introduced malfunctions
and errors into the process. For Group 2, button presses sometimes malfunctioned on purpose, text fields
did not immediately accept input from the operator, items were hidden longer, or additional delays were
introduced by popup dialog boxes or software pauses. The purpose of the malfunctions was to induce a
sense of urgency and to increase the stress and anxiety experienced by the participants while completing
the computer tasks. Group 1 was the control group and Group 2 was the experimental group.

While the software tasks were being performed on the client computer by the participants, a
constant stream of data was sent from computer #1 to computer #2 via an isolated, hard-wired peer-
to-peer network. Because many of the modern networking protocols (like TCP/IP) include a significant
amount of data checking and automatic correction, a specific networking protocol was used that does
not perform error checking. This protocol is called UDP and is commonly used internally in telecommu-
nication software processes. It is a very fast protocol that depends on the network software to perform
all error checking and to handle all issues that arise during network communication.

During network communication, a predefined set of data was encoded, packaged, and sent from com-
puter #1 to computer #2. On computer #2, the data were unpacked, reassembled, decoded, and written to
the hard drive. Data integrity and reliability testing were integrated into the networking software. The results
of the tests were logged and the number and location of the errors were recorded for each session.

At the beginning and end of each session, the participant was asked to assess their relaxation and
anxiety levels on a scale of 1 to 10 - 1 being relaxed and calm and 10 being least relaxed; 1 being very
low anxiety and 10 being highest anxiety.

Networking Software
The network communication software was run continuously throughout the study. Errors were
logged for the communication process 24 hours a day whether sessions were being run or not. The
network was designed to create an extremely sensitive system that would produce transient faults that
could be detected regularly while the network was operating. In order to create a thin threshold for er-
rors to occur in the system, the network communication process was tuned to produce transient faults
or errors regularly by adjusting the number of messages sent between the computers. The number of
messages sent between the systems was adjusted to produce a minimal number of synchronization
faults, where the messages collided due to the speed of the network, while the network continued to
produce errors due to other faults that were the result of other external influences on the network. (See
Common Errors on Computer Networks) The optimal level to produce this effect was determined to in-
236 KRUTH

clude over 767 lines being sent every second between the two computers, producing errors nearly every
day while the system was running. This resulted in an average of 339 errors per day, ranging from 0 to
3,799 in a single day.

Errors were recorded during each of the study sessions, and an equal number of sessions was
randomly selected when there were no computer operators present. These no-operator sessions were
used to determine the reliability and integrity of the network communication system without any hu-
man interaction and to test the first hypothesis (H1).

Common Errors on Computer Networks

Networking errors occur as a result of faults in the network system, and faults are typically classified
as permanent, intermittent, or transient. Permanent faults usually are not resolved until a repair action
is taken, for example a hardware failure or a disconnected network cable. Intermittent faults occur on
a periodic basis and reduce the reliability of a network. Transient faults are temporary, and are usually
corrected by error recovery software (Steinder & Sethi, 2004).

In a simple, peer-to-peer, hard-wired network, like the one used in this study, the most common
faults would be transient unless there are specific hardware problems. Intermittent failures are more
common in larger networks with multiple nodes or routers. Transient faults are most often the result of
external interference, like electromagnetic interference (Cha et. al., 1996), high-energy particles (Nor-
mand, 1996), or attacks, or internal dysfunctions like design flaws or software bugs (Huang et. al., 2014).
In addition, synchronization issues can produce transient faults until the temporal consistency is re-es-
tablished (Steinder & Sethi, 2004).

The networking software in this study integrated fault detection and logging directly into the net-
working system. Since the data sent over the network was fixed and consistent, the fault detection soft-
ware could be specifically designed and tested to eliminate internal dysfunctions and automatically log
any other errors that occurred in the data transmission process. Because the network was isolated from
any other computers there was no possibility of attacks, but faults could occur as a result of electro-
magnetic interference or high-energy particles. These intermittent data transmission errors were logged
without regard to the specific cause of the interference.

Isolation of Software Processes

When running multiple software processes on a single computer system, questions arise about the
isolation or interaction of these programs. In this system, the network software and the user interactive
software were both running on a single computer. Could these processes have affected each other?

To isolate computer processing and avoid interaction between programs, the programs must first
be run in different threads or different processes (Vokorokos, Balaz, & Mados, 2015). In addition, to
avoid competition for OS resources, access to the resources must be controlled or the programs must
not use the same files, network resources, locks, processes, CPU, among other resources (Liang, Venka-
takrishnan & Sekar, 2003; Vokorokos, Balaz, & Mados, 2015).

The two programs used in this study were carefully designed so that they were not using the same
files or processes, and they were run on separate OS threads. Only the networking software used the
 EFFECTS OF MOOD AND EMOTION 237

network resources; the user interactive software did not access the network. It could be argued that
since the software programs were being run on the same computer that they were sharing CPU resourc-
es and one process was affecting the operation of the other process through competition for CPU time.
In this case, neither program performed process intensive tasks or extensive mathematical calculations.
The networking software specifically was designed to perform extremely simple tasks that only included
accessing the file system, minor data processing, and network communication.

Even with these careful design considerations, it could be argued that having both processes run
on a single computer could affect the results of H1 where the network integrity was being compared in
conditions where the user interactive software was being run and other times when it was not. Given
this minor possibility of interaction between the programs, the results of the analysis for H1 should be
carefully considered before coming to conclusions.

Hardware Setup
The hardware consisted of two standard, commercially available PCs (Lenovo K450E desktop sys-
tems) running the operating system Windows 7. These computers were connected via a standard, hard-
wired networking cable, and neither computer was connected to any external network or the internet.
Wireless capabilities were turned off on both computers.

Software Design
All of the software used for this study was custom designed and built. There were seven major
software components: the controller software that ran each session, the user interactive software, the
network communication software, the software that sent data across the network, the software that
received data from the network, the test software used to determine networking reliability and data
integrity, and the data logging and recording software.

1. The controller software: This software is the containing structure for the entire session. This soft-
ware allows the participants to log in, randomly assigns them to an experimental group, starts the
interactive software, and logs all data collected throughout the session.
2. User Interactive software (See Appendix): This software provides instructions to the participants
and enables them to complete the timed tasks. There were two versions of this software, one for
each group. The first version used by Group #1 included a series of tasks familiar to most computer
users including typing text in text boxes, using the mouse to move items around, matching images,
and finding hidden objects. This version worked as is normally expected of error free software. The
second version, used by Group #2, looked identical to the software used by Group #1. The same
instructions and tasks were provided in this version, but the user interface purposely contained
mistakes and malfunctioning components. For example, buttons did not consistently respond to
mouse clicks, text fields put the wrong text into the fields, intentional obstructions were included
like built in delays or unnecessary popup dialog boxes. This second version of the interactive soft-
ware was designed to induce anxiety and frustration in the participant using the software.
3. Network Communication Software: This component implemented the UDP network communi-
cation protocol for the network. It provided the mechanisms to read data, pack it for transfer,
send data, retrieve data, unpack data, and write data to disk. This is standard networking soft-
ware that is familiar to most network programmers.
238 KRUTH

4. Network Sending Software: This component read a predefined set of data from the hard drive
and utilized the Network Communication Software to send the data across the network.
5. Network Receiving Software: This component accepted data from the network connection and
utilized the Network Communication Software to unpack and read the data sent. It also wrote
the data received to the hard drive.
6. Network Testing Software: This component was integrated into the network communication, send-
ing, and receiving software. It checked each step in the data transfer and receiving process to confirm
data integrity at each step. This component recorded all of the errors and logged the location and
nature of each error. It also maintained a count of the errors at each checkpoint for each session.
7. Data Logging and Recording Software: This component maintained the data security and integ-
rity for each session and for the study. The data logging and recording software was used by all
components that wrote to the disk for each session.

Interaction with Participants

Only the author interacted with the participants before, during, and after the experimental sessions. He
had a moderate belief that the study could produce results supporting both H1 and H2, but due to complete
masking of the data collected and the grouping of participants, there was no apparent method that would
enable the researcher to influence the opinion of the participants or their behavior during the sessions. Each
participant was presented with a consent form discussing the nature of the sessions, but these forms did not
include any information about the data collection technique or exactly what was being measured during the
sessions. The participants were only instructed to complete the tasks within the specified time limit in order
to receive the reward. Each participant was formally briefed by the researcher using the same language.

After the session was completed, the researcher interacted comfortably with each participant and
provided debriefing information describing the nature of the software they were using and describing
that there was a network running in the room during the session. Neither the participants nor the re-
searcher knew the results of any individual sessions or how many errors were collected for any specific
participant. This information was masked throughout the analysis process and continues to be com-
pletely masked. This was accomplished through an automated analysis process designed to avoid any
knowledge of individual sessions.

Data Collection
The data evaluated in this study were the location and number of network errors, the group associ-
ated with each session, and the self-rating of relaxation and anxiety collected at the beginning and end
of each session. Data for this study were automatically collected by the software programs. The programs
automatically collected all of the network errors that occurred during each session. It recorded at which
point an error occurred and how many errors occurred in each session. The program also associated the
data with the group that was automatically, randomly assigned when the user logged into the session. Fi-
nally, the program collected and stored the response to the self-rating of relaxation and anxiety provided
by each participant at the beginning and end of each session. Any data related to the users’ performance
on the mundane, user interactive computer tasks was ignored and was not collected by the program.

The saved data was stored on the hard drives of computer #1 and computer #2 which was only
accessible to researchers associated with the study. Besides the data collected electronically, a separate
 EFFECTS OF MOOD AND EMOTION 239

record that associated each participant with a unique User ID was stored in a log book, separate from
the computer data.

Performance Markers/Indicators
Data integrity was assessed by the testing program at different points in the network communica-
tion process. Data integrity was evaluated when the data were read from the hard drive on computer
#1, when the data were prepared for transfer, when the data were about to be sent over the network,
when the data were received on computer #2, when the data were unpacked on computer #2, and
when the data were decoded and stored to the hard drive on computer #2. At each checkpoint, the
number of errors recorded was tallied.

The markers used for evaluation and analysis were the total number of errors in a session, the
number of errors at each checkpoint, and the location of each error. Due to the potential vulnerability
of the data transmission process from Computer 1 to Computer 2, it was considered the most likely
location to see variance in error data, so the count of errors found during the transmission of data was
pre-defined as the primary analysis variable. Secondary factors related to reading, writing, packing, and
unpacking the data were recorded but turned out to be insignificant to the process (see Analysis below).

Analysis

Data preparation
The log files for the sessions were combined with the log files from the network software that collected
the number of errors that occurred on the system. Using the date and time of each session as a key, the network
error log was queried to determine the number of errors that were recorded on the independently running
network during each session. This information was separated into groups to further assist with the evaluation
of the second hypothesis (H2) that participants who experienced greater stress (higher anxiety) would uncon-
sciously affect the networking system to produce a greater number of errors than the control group.

The error data included records of errors when data were transferred between the computers
over the networking system, and the packing and unpacking processes on the sending and receiving
computers. While pilot testing the process, there were no errors detected in the processes of packing
and unpacking the data on the computers. During the study, just as was the case during the pilot stud-
ies, errors were recorded in the data transfer process on the network, but no errors were found in the
secondary processes on the sending or receiving computers. Since the primary analysis was related to
the transfer of data (see Data Collection above), this was the only error data considered. The secondary
factors measuring errors in the reading of the data, packaging of data, unpacking of the data, and writ-
ing of the data did not occur, so they were not considered in the analysis.

All calculations related to significance were processed using SPSS statistical analysis software.

To determine if the means between the two groups were significantly different, Levene’s test was
used to determine if the variance of the two groups were similar enough to conduct independent
t-tests by taking into account the number of groups, and the total number of cases in all groups.
240 KRUTH

The effect size, d, was calculated using Hedges’ g since the newly established groups were of dif-
ferent sizes. This effect size was corrected to remove a small bias to create an unbiased score using the
calculation proposed by Hedges and Olkin (1985, p. 81).

Results
The original experimental and control sessions were defined based on the groups that were using
the different versions of the tasking software, and each group contained 65 participants. The experi-
mental group used the software that included obstructions to induce anxiety, and the control group
used the software that did not include obstructions. The experimental group was expected to produce
a group that experienced higher anxiety during the session.

When the experimental sessions and control sessions were compared, the anxiety levels for the
two groups were nearly identical. The experimental group had a mean anxiety difference of 2.35 while
the control group had a mean difference of 2.36. The groups showed no significant differences in the
number of errors detected (p = .96). Upon further investigation, it was determined that the experimen-
tal sessions, which were designed to invoke higher anxiety in the participants, did not produce a higher
self-reported anxiety than the control sessions. This provoked further investigation.

The difference between the self-reported anxiety before and after the session was evaluated for
each of the 130 sessions, regardless of whether it was a control or experimental session. Self-rated
anxiety increased on average across all of the sessions with a mean difference of +2.35 on a 10-point
scale. Only 24 of the 65 participants in the group originally designated as the experimental group had
reported more anxiety than the mean value, while 17 participants of the 65 in the original control group
reported more anxiety than the mean value.

A subgroup of all sessions was selected where the self-reported anxiety difference was greater
than the mean difference for all the sessions (i.e. anxiety > +2.35). This group included 41 participants of
the total 130 participants in the study (24 from the original experimental group and 17 from the origi-
nal control group). This group was considered the group that experienced the highest levels of anxiety
change during the session, and this group was redefined as the experimental group since it met the
criteria for evaluating H2 (greater stress conditions).

The number of errors in the high anxiety sessions was compared with the sessions in which par-
ticipants reported a smaller change in anxiety during the session. This included 89 sessions that were
considered the control group where participants self-reported a change in anxiety at or below the mean
change for all participants (i.e. anxiety <= +2.35).

The mean number of errors in the high-anxiety group (HI) = 12.20 errors per session. The mean
number of errors in the control group © was 3.57. Levene’s test for equal variances was used to compare
the two groups and the variances were different, F (1, 128) = 12.39, p = .001. A comparison of means
 EFFECTS OF MOOD AND EMOTION 241

was evaluated with an independent sample t-test which indicated a difference between the groups, p
= .038. Despite the significant difference in means, the effect size was small (d = 0.45), but the power of
the results was moderately high (0.61) (see Table 1 for descriptive statistics).

Table 1

Anxiety Group versus Non-Anxiety Group: Descriptive Statistics

Group N Mean Std. Dev. Std. Err. Mean

Non-Anxiety 89 3.57 16.673 1.767

Anxiety 41 12.20 23.479 3.667

The first hypothesis (H1) predicted that there would be more errors detected in the network sys-
tem when a participant was present than when there was no participant using the computer. To eval-
uate H1, all of the sessions for all participants (130) were compared with an equal number of sessions
that were randomly selected as a “no-operator group” when the networking system was running un-
monitored and no people were present in the room with the computers. The no-operator sessions were
selected to compare directly with the experimental sessions. The no-operation sessions had the same
duration as the experimental sessions and were selected at comparable times during the day. Some
no-operator sessions were on the same day but at a different time than the experimental sessions, and
some were at the exact same time but on days directly before or after the experimental session. The
random selection was achieved with a query to the random number generator at www.random.org that
produces a true random value. The random numbers indicated the hour for the no-operator session,
and the minute within the hour and the length of the sessions exactly corresponded with each experi-
mental session that included a participant.

The sessions that included a participant (SP) consisted of 130 sessions producing a mean of 6.29
(SD = 19.41) errors per session with a range of 0-126 errors per session and a standard deviation of SD
= 19.410. The randomly selected no-operator sessions (SR) consisted of 130 sessions with a mean of =
4.17 (SD = 17.30) errors, with a range of 0-137. There was no difference between the groups, F(1, 258) =
3.52, p = .35. Besides self-rating anxiety, the groups also rated their level of relaxation change during the
sessions. The relation between the level of relaxation and the number of errors recorded in the sessions
was not significant, F(1, 128) = 3.65, p = .32.

Discussion
The study hypotheses predicted that human interaction with machines would have an effect on
the number of errors detected in the operation of the computer processing and network communica-
tions. In addition, they predicted that participants who experienced greater stress (were more anxious)
would produce more errors than those less anxious. It was predicted that the group that did not in-
volve any computer operators would produce the fewest errors. Hypothesis 1 predicted that sessions
involving participants would produce more errors in the independently operating network than random
sessions where no participant was present. This hypothesis was not supported with equal groups that
242 KRUTH

included sessions with participants using the computer and sessions where no one was in the room with
the computers.

Hypothesis 2 predicted that participants who experienced greater stress (higher anxiety) would
produce more errors in the computer network providing evidence that the mood of a computer user
could have an effect on the operation of a computer network. The testing of this hypothesis was initially
planned for testing with two equal groups of participants who were randomly assigned to use software
specially designed to evoke anxiety in one of the groups. Participants who were in both the experimen-
tal and control groups reported anxiety at equal rates. Because of this, the grouping was modified to
include participants who reported the highest change in anxiety during the sessions as the experimen-
tal group, and the remaining participants as the control group. This change in grouping resulted in 41
participants in the high-anxiety group and 89 participants in the control group, and the difference was
significant. An analysis of the revised groups found a difference in the number of network errors detect-
ed in sessions completed by participants who experienced anxiety using a computer system and the
sessions of those who did not. These results were supported by a small effect size and a moderate power
for the study, and they met the criteria initially proposed as critical values for this study.

Recommendations
The experimental software was designed to induce higher anxiety in the participants, but par-
ticipants self-reported higher anxiety in both versions of the software. It is obvious that the software
differences in the user interactive software did not achieve the desired effect, but there may be another
factor. Self-reporting can be unreliable, and it is possible that some participants might have under-re-
ported their anxiety level or over-reported due to a lack of awareness of their emotions, a desire to
suppress their own anxieties, or any number of other reasons. A better measure of anxiety could be
achieved with physiological monitoring of the participants during the sessions to determine if there are
any correlations between factors that indicate a change in mood (blood pressure, heart rate, skin con-
ductance, etc.) and the number of errors detected in the network system. Other mood states could also
be examined to determine their effects on the system.

Participants were motivated in this study with rewards that were designed to encourage them
to complete the study within the 20-minute time limit, including a gourmet chocolate bar and a raffle
for a gift card. Some participants stated that they were very excited to win the rewards, especially the
chocolate, and wanted to complete the tasks on time, but others actually stated that they did not care
about the rewards at all and were not in any hurry to finish. This variation in motivation seemed to have
an effect on the anxiety levels experienced by the participants, and more significant rewards might have
made the original experimental and control groups more likely to be influenced by the user interactive
software.

In addition, selecting participants with similar computer experience or expertise would provide
a more evenly distributed sample. In this study, some participants completed the sessions in 10 min-
utes while others took over 90 minutes to complete the same tasks. Despite this difference in time, the
participants’ self-reported anxiety difference did not correlate with the amount of time it took to com-
 EFFECTS OF MOOD AND EMOTION 243

plete the tasks. This suggests that some participants have little or no interest in completing the tasks
to receive the reward while others are highly motivated. Ideally, this study would have included highly
motivated participants to encourage the completion of the tasks in the specified time while evoking an
emotional response when their progress was obstructed.

Finally, the predefined critical values and success factors for this study may have been overly op-
timistic in requiring an effect size > .2 and a power > .5 with a significance level p < .05. It is clear that a
study that explores unconscious effects on a working computer network may produce a very small effect
size. Also, unless there is a very large sample planned, the power is likely to drop below 0.5. Future stud-
ies of this type should consider reducing the predicted effect size to d > 0.1 and the sample should be
increased significantly to provide a higher powered study.

Conclusions
This exploratory study provides preliminary evidence that the mood of the participants can pro-
duce an unconscious effect that will result in more errors in a computer system. With a more specific
method to rate the mood of the participants, more pre-qualification of participants, and a larger sample
size, future studies could shed additional light on these hypotheses and determine if this effect is strong
enough to merit additional applications.

For example, as these findings suggest that computer operators may affect network communi-
cations when they are experiencing higher anxiety or frustrations, businesses that employ computer
operators or provide technical support to computer users may be prudent to consider the comfort and
mood of computer operators while they work. Though this study only indicated that network errors
were increased by the computer operators with higher anxiety, there might be other effects on electron-
ics that increase as anxiety rises.

Many people have experienced computer errors when they are anxious or frustrated, and they
often attribute these errors to a lack of attention or stupid mistakes. Though many errors may be the
result of simple mistakes, additional errors may be the result of an unconscious effect that appears or
is amplified when computer users are anxious. Regardless of the cause, the same series of actions may
resolve the situation. Walk away from the computer and go down the hall to get a drink. Take a break,
calm your mind, and take a deep breath before going back to your computer. The problems just might
go away when the anxiety is reduced.

References

Atwater, P.M.H. (2007). The big book of near death experience. Newburyport, MA: Hampton Roads Pub-
lishing.
Broughton, R.S. & Perlstrom, J.R. (1986). PK experiments with a competitive computer game. Journal of
Parapsychology, 50, 193-211.
Cha, H., Rudnick, E. M., Patel, J. H., Iyer, R. K., Choi, G. S. (1996). A gate-level simulation environment for
alpha-particle-induced transient faults. IEEE Transactions on Computers, 45, 1248-1256. https://doi.
org/10.1109/12.544481
244 KRUTH

Fracasso, C. & Friedman, H. (2012). Electromagnetic aftereffects of near-death experiences: A preliminary

report on a series of studies currently under way. Journal of Transpersonal Research, 4(2), 34-55.
Hedges, L.V. (1981), Distribution theory for Glass’s estimator of effect size and related estimators, Journal
of Educational Statistics, 6(2), 106-128. https://doi.org/10.3102/10769986006002107
Hedges, L.V., & Olkin, I. (1985), Statistical methods for meta-analysis. London, UK: Academic Press.
Jahn, R., Dunne, B., Nelson, R., Dobyns, Y., & Bradish, G. (1997). Correlations of random binary sequences
with pre-stated operator intention: A review of a 12-year program. Journal of Scientific Exploration,
11, 345-367.
Knittweis, J. (1997). Electrical sensitivity of near-death experiencers. Journal of Near-Death Studies, 15,
223-225. https://doi.org/10.17514/JNDS-1997-15-3-p223-225.
Kruth, J.G. & Joines, W.T. (2015). Taming the ghost within: An approach toward addressing apparent elec-
tronic poltergeist activity. Journal of Parapsychology, 80, 70-86.
Liang, Z., Venkatakrishnan, V.N., & Sekar, R. (2003). Isolated program execution: An application transparent
approach for executing untrusted programs. Proceedings of ACSAC ‘03. Washington, DC: IEEE Comput-
er Society.
Morris, R. L. (1986). Psi and human factors: The role of psi in human-equipment interactions. In B. Shap-
in & L. Coly (Eds.), Current trends in psi research, 1–26. New York: Parapsychology Foundation.
Normand, E. (1996). Single-event effects in avionics. IEEE Transactions on Nuclear Science, 43, 461-474.
https://doi.org/10.1109/23.490893
Nouri, F. M. (2008). Electromagnetic after-effects of near-death experiences. Dissertation International Ab-
stracts, UMI No. 3352121. Retrieved from Proquest Dissertations and Thesis Database.
Palmer, J. (1974). A case of RSPK involving a ten-year-old boy: The Powhatan poltergeist. The Journal of
the American Society for Psychical Research, 68, 1-33.
Pratt, J.G. & Roll, W.G. (1957). The Seaford disturbances. Journal of Parapsychology, 22, 79-124.
Radin, D. (1990). Testing the plausibility of psi-mediated computer systems failures. Journal of Parapsy-
chology, 54, 2-19.
Roll, W. G., Burdick, D., & Joines, W. T. (1973). Radial and tangential forces in the Miami poltergeist. Journal
of the American Society for Psychical Research, 67, 267–281.
Roll, W. G., Burdick, D., & Joines, W. T. (1974). The rotating beam theory and the Olive Hill poltergeist. In Roll,
W. G., Morris, R. L., & Morris J. (Eds.), Research in Parapsychology, 1973, 64–67. Metuchen, NJ: Scarecrow.
Roll, W.G. & Storey, V. (2004). Unleashed: Of poltergeists and murder: The curious story of Tina Resch. New
York: Paraview Pocket Books.
Schmidt, H. (1970). A PK test with electronic equipment. Journal of Parapsychology, 34, 175-181.
Steinder, M.I., & Sethi, A. S. (2004). A survey of fault localization techniques in computer networks. Science
of Computer Programming, 53, 165-194. https://doi.org/10.1016/j.scico.2004.01.010
Vokorokos, L., Valaz, A., & Mados, B. (2015). Application security through sandbox virtualization. Acta Poly-
technica Hungarica, 12, 83-101. https://doi.org/10.12700/APH.12.1.2015.1.6
Huang, X.-F.., Zhou, C.-J., Huang, S., Huang, K.-X., & Li, X. (2014). Transient fault detection in networked con-
trol systems. International Journal of Distributed Sensor Networks. https://doi.org/10.1155/2014/346269

Appendix

User Interactive Software Tasks and Obstructions

Sixteen (16) tasks were presented to every participant in this study. The tasks were common tasks
that are familiar to most computer users including simple games, addition, and typing in text boxes. The
participants were instructed to complete all of the tasks in 20 minutes to receive a reward. There were
 EFFECTS OF MOOD AND EMOTION 245

obstructions for some of the participants to induce a sense of urgency and possibly anxiety. The tasks
and obstructions are described below.

General Obstruction:
When participants are using the “obstructed” version of the software, between each task screen,
a dialog box appears indicating how much time has gone by and how much time is left. The user must
click away the dialog box before they can go onto the next screen.

Task 1. Move a ball from one side of the screen to the other using the mouse. Click on the ball and drag
it to the hole.
Obstruction: none
Task 2. Type a single word that is on the left side of the screen in a text box on the right side of the
screen.
Obstruction: The text disappears when the user moves their mouse over the “Continue” button. It must
be retyped. To complete the task.
Task 3. Choose the photos that contain water.
Obstruction: One of the photos that contain water must be clicked three times before the click registers.
Task 4. Add two six-digit numbers.
Obstruction: When typing the result in a text box, the 7 character is initially entered as the 6 character.
It must be corrected.
Task 5. Choose the box that contains the darkest shade of grey.
Obstruction: Clicking on the correct box selects another box on the screen. This occurs two times before
the correct box will register the click.
Task 6. Type a sentence from the left side of the screen in a text box on the right side.
Obstruction: Some characters are changed during typing and the sentence must be retyped until it is
correct.
Task 7. Click the boxes to reveal a face. Find two matching faces to clear the boxes from the screen.
When all the boxes are gone, move on to the next page.
Obstruction: Boxes reveal the face for less than one second. When matches are found, the boxes stay
on the screen and make the user wait for 3.5 seconds before allowing them to click to find another pair.
Task 8. Twenty-four cards on displayed face down. Find the Jack of Hearts by click to reveal each card.
Obstruction: The user must click every card before the Jack of Hearts is displayed. It’s always the last card.
Task 9. Type a paragraph in a box at the top of the screen into a text box at the bottom. This task was
difficult, so a button labeled “Give Up” was added to be displayed after 5 minutes had passed.
Obstruction: Some letters are purposefully mistypes and must be corrected before the “Continue” but-
ton can be pressed.
Task 10. Select the image of a key to open an image of a door. Five keys are presented and three doors
are presented. The user had to find the correct keys for the doors.
Obstruction: none.
Task 11. An image of a face is flashed on the screen for less than one second. Select the hair color, eye
color, and gender of the person in the photo. A button is provided to show the photo again. The same
photo is shown each time.
246 KRUTH

Obstruction: The photo changes each time the button is pressed until five different photos have been
shown. Also, the photos appear on the screen for a very short period of time, and can barely be seen.
Task 12. There are twenty-four cards shown on the screen, but only the back of the card is displayed.
Find the five cards that have “wavy lines” on them (like the ESP test cards). Clicking a wrong card adds
three seconds to your time.
Obstruction: The final card with wavy lines is always the last card that is turned over.
Task 13. Create change to make 92 cents using coins. A set number of each coin is available to make
this total.
Obstruction: When the first try is completed, it is always wrong, and the amount changes from 92 cents
to 91 cents without any obvious indication of the change.
Task 14. Click on a running digital clock to stop it at exactly 10.0 seconds. You can reset the clock if you
get it wrong. After five tries, it automatically says “Close enough” and lets the user continue.
Obstruction: The first five tries, it always misses by at least 0.1 seconds, and the user must try again. After
fifteen tries it automatically allows the user to continue.
Task 15. Find a hidden object behind trees and tents in the forest. Click on an image to reveal what’s
behind. Clicking the wrong object adds five seconds to the time.
Obstruction: When you click the wrong item, it pauses the screen for three seconds before another item
can be selected.
Task 16. Same as Task 1. Move the ball to the hole.
Obstruction: none

Une Exploration des Effets de l’humeur et de l’émotion sur un Système Informatique

et un Environnement de Réseau Fonctionnant sur le Monde Réel

Résumé. Cette étude emploie un système et un réseau informatiques spécifiques conçus pour induire de
l’anxiété chez des participants et déterminer si les personnes anxieuses produisent plus d’erreurs dans
un réseau informatique fonctionnant indépendamment. Les participants (N = 130) ont complété seize
tâches sur un ordinateur en vingt minutes afin de recevoir une récompense. Chaque participant évaluait
ses niveaux d’anxiété durant les tâches. En parallèle, 130 sessions étaient lancées sans aucun opérateur
informatique. Le réseau fonctionnait indépendamment des tâches, et opérait continuellement durant
les sessions. La première hypothèse prédisait que les sessions sans opérateurs produiraient moins d’er-
reurs de réseau que les sessions avec opérateurs, mais cela ne fut pas vérifié (p = 0.35). La seconde
hypothèse prédisait que les opérateurs anxieux produiraient plus d’erreurs sur le réseau indépendant
que ceux avec moins d’anxiété. L’analyse initiale indiquait une hypothèse non-vérifiée, mais le design
initial n’identifiait pas correctement les utilisateurs anxieux. Un regroupement révisé post-hoc basé sur
la véritable anxiété rapportée a permis de vérifier cette hypothèse (p = 0.04, d = 0.45), indiquant que les
opérateurs informatiques anxieux pouvaient affecter le réseau de communication. Il y a pu avoir d’au-
tres effets électroniques produits par les émotions humaines. D’autres recherches sont nécessaires pour
confirmer ces résultats et explorer si l’intensité des émotions affecte l’électronique.
 EFFECTS OF MOOD AND EMOTION 247

Eine Untersuchung über die Auswirkungen von Stimmung und Emotion auf als
Reale-Welt funktionierendes Computersystem und Netzwerkumgebung

Zusammenfassung. Diese Studie verwendete ein kundenspezifisches Computersystem und Netzwerk,

das entwickelt wurde, um bei Teilnehmern Angst zu induzieren und um festzustellen, ob ängstliche Men-
schen mehr Fehler in einem unabhängigen Computernetzwerk produzieren. Die Teilnehmer (N = 130)
wurden gebeten, in zwanzig Minuten sechzehn Aufgaben an einem Computer zu erledigen, um eine
Belohnung zu erhalten. Jeder Teilnehmer bewertete selbst sein Angstniveau während der Aufgaben.
Darüber hinaus wurden 130 Sitzungen ohne einen Computeroperator durchgeführt. Das Netzwerk lief
unabhängig von den Aufgaben und war während der Sitzungen kontinuierlich in Betrieb. Die erste Hy-
pothese prognostizierte, dass Sitzungen ohne Operatoren weniger Netzwerkfehler produzieren würden
als Sitzungen mit Operatoren, was sich nicht bestätigte (p = 0,35). Die zweite Hypothese prognostizierte,
dass Ängstliche mehr Fehler im unabhängigen Netzwerk produzieren würden als weniger Ängstliche.
Eine erste Analyse bestätigte die Hypothese nicht, aber das anfängliche Design identifizierte ängstliche
Benutzer nicht richtig. Eine post-hoc vorgenommene Gruppeneinteilung, die auf der Grundlage der
tatsächlich berichteten Angst basierte, führte zu einer Bestätigung dieser Hypothese (p = 0,04, d = 0,45),
was darauf hindeutet, dass ängstliche Computeroperatoren die Netzwerkkommunikation beeinflussen
können. Es können andere elektronische Effekte als Folge menschlicher Emotionen auftreten. Weitere
Forschungen sind notwendig, um diese Ergebnisse zu bestätigen und zu untersuchen, ob die Intensität
von Emotionen elektronische Geräte beeinflussen kann.

Una Exploración de los Efectos del Estado de Ánimo y Emoción en un Entorno

Realists de Sistema y Red Informáticos

Resumen. Este estudio utilizó un sistema informático y una red diseñados para inducir ansiedad en los
participantes y determinar si las personas ansiosas producen más errores en una red informática que
funciona independientemente. Se pidió a los participantes (N = 130) que completaran 16 tareas en una
computadora en veinte minutos para recibir una recompensa. Cada participante calificó sus niveles de
ansiedad durante las tareas. Además se corrieron 130 sesiones sin un operador en la computadora. La
red funcionaba independientemente de las tareas y operaba continuamente durante las sesiones. La
primera hipótesis pronosticaba que las sesiones sin operadores producirían menos errores de red que
las sesiones con operadores, pero no fue así (p = 0.35). La segunda hipótesis predijo que los operadores
ansiosos producirían más errores en la red independiente que aquellos menos ansiosos. El análisis inicial
no respaldó la hipótesis, pero el diseño inicial no identificó adecuadamente a los usuarios ansiosos. Un
agrupamiento post hoc basado en la ansiedad real reportada apoyó esta hipótesis (p = 0.04, d = 0.45)
indicando que los operadores de computadora ansiosos pueden afectar al sistema de la red. Pueden
haber otros efectos electrónicos afectados por las emociones humanas. Se necesita más investigación
para confirmar estos resultados y explorar si la intensidad de las emociones afectan a la electrónica.
Reproduced with permission of copyright owner. Further reproduction
prohibited without permission.