Content uploaded by Jayne Gackenbach
Author content
All content in this area was uploaded by Jayne Gackenbach on Sep 30, 2014
Content may be subject to copyright.
Video Game Play As Nightmare Protection:
A Preliminary Inquiry With Military Gamers
Jayne Gackenbach
Grant MacEwan University
Evelyn Ellerman and Christie Hall
Athabasca University
Soldiers who play video games to varying degrees were solicited to fill out a
survey on dreams and gaming. A prescreening filtered out those who were not
soldiers, who did not game, and who were suffering from various psycho-
logical problems in the last six months. The remaining soldiers filled out these
inventories: general and military demographics, history of video game play,
Emotional Reactivity and Numbing Scale (ERNS), and a Trauma Inventory.
They were then asked to provide two dreams, one recent and one that was
impactful from their military service. Following the military dream they filled
out Impactful Dreams Questionnaire (IDQ) about that dream only. Dream
content analysis was conducted using threat simulation, war content, and
lucid/control/gaming content. High- and low-end frequency gamer groups
were identified and compared on these dream content scales. Because the
nightmare literature shows that affect load and distress are predictors of
nightmare suffering, ERNS and Trauma history were covariates in the
ANCOVA’s on gamer group ⫻dream type. It was found that the high-end
gaming group exhibited less threat and war content in their military dreams
than the low-end group.
Keywords: nightmares, videogame, threat simulation, military
Gackenbach and colleagues have been investigating how video game play
affects dreams (summarized in Gackenbach, Kuruvilla, Dopko, & Le, 2010; Gack-
enbach, in press). One type of dream that they have studied is the nightmare
(Gackenbach & Kuruvilla, 2008a; Gackenbach & Kuruvilla, 2008b; Gackenbach et
al., 2009; Le & Gackenbach, 2009). In early inquiries they found through content
This article was published Online First August 22, 2011.
Jayne Gackenbach, Department of Psychology, Grant MacEwan University, Alberta, Canada;
Evelyn Ellerman, Associate Professor Communication Studies, Athabasca University, Athabasca, Al-
berta, Canada; and Christie Hall, Communication Studies, Athabasca University, Athabasca, Alberta,
Canada.
We’d like to extend our appreciation to Kris La Marca, John Bown, Katherine Wisniewski, and
Mary-Lynn Ferguson for their help with this project.
Correspondence concerning this article should be addressed to Jayne Gackenbach, PhD., Depart-
ment of Psychology, Grant MacEwan University, 6-323H, 10700 - 104 Street, Edmonton, Alberta T5J
4S2 Canada. E-mail: gackenbachj@macewan.ca
221
Dreaming © 2011 American Psychological Association
2011, Vol. 21, No. 4, 221–245 1053-0797/11/$12.00 DOI: 10.1037/a0024972
analysis that, while gamers did behave aggressively in dreams, this behavior oc-
curred far less often than with those who rarely gamed. On the other hand, when
dream aggression happened for gamers, it was more intense. Gamers were also less
likely to experience misfortune in dreams. This combination of selective but strong
aggression and the lack of misfortune suggested that gamers did not see threat in a
dream as frightening, but rather as empowering. This was supported in a follow-up
study looking specifically at the threat elements in the dreams of gamers versus
low-end players. Gackenbach and Kuruvilla (2008a) found that not only were
gamers’ dreams not associated with threat motifs, gamers did not consider these
dreams as nightmares, nor did they see them as scary. In their study, threatening
dream content was associated with watching violent TV or movies the night before
the dream.
In a second follow-up, the Gackenbach group has looked at nightmares versus
bad dreams, a distinction important in the dream literature (Zadra, Pilon, &
Donderi, 2006). Zadra (2010) pointed out that nightmares are more likely to be
characterized by physical aggression, whereas bad dreams are more often associ-
ated with interpersonal conflict. Thus this distinction seemed appropriate for
further examination of gamers’ nightmares. Le and Gackenbach (2009) reported
different emotional responses to these two types of dream as a function of gaming.
Specifically, gamers did not feel anxiety in their nightmares but they did in their bad
dreams. The rarely gaming group reported no difference in anxiety as a function of
negative dream type. Thus, as had been found before by this research group, the
unique response to nightmares evidenced by gamers seems not to be the same as
their response to bad dreams, but is specific to the type of dream which most closely
resembles their game life, the nightmare.
This lack of emotional responsivity to the nightmare may be due to the
numbing-to-violence which has been found in the gaming and violence literature
(Bushman & Anderson, 2009). But gamers are also active in these dreams. Thus a
possible chain of events is that heavy gamers who play first person shooter or
action/adventure type genres, practice reacting quickly and violently to threat in a
game. When they experience threat in a dream, rather than being intimidated by
the threat, they are empowered and fight back. Here is a dream showing this
scenario from Gackenbach et al. (2009):
So I went outside with my cat and shot these criminals that were trying to eat my dad and
they were on top of my dad trying to eat his arms and he was fighting them off, and they were
trying to hold him down and bite his shoulders and there was blood and stuff. And it was a
very graphic shootout for a dream; it was very blood and guts ya know? And when I ran out
of ammunition there was like pistol whipping and stuff going on. (Subject 002, dream 6)
In part this response in dreams may be due to gamers’ reports of more dream
control and lucidity, knowing they are dreaming while they are dreaming, than has
been found in those who rarely game (Gackenbach, 2006, 2009). In any case this
unique gamer response to the classic nightmare scenario—a sudden and unex-
pected threat appears in a dream; the dreamer feels helpless in the face of it and
awakens due to the fear it induces—offers potential nightmare inoculation to those
who may face such threats in the real world. The segments of society who regularly
face such threats are first responders, such as police and fire fighters, and those in
the military.
222 Gackenbach, Ellerman, and Hall
Video game play may offer not only a type of training for learning to shoot
targets or attend to peripheral cues in a dangerous environment, but also some
inoculation to threats in nightmares. These dreams are the most common element
of posttraumatic stress due to trauma. In a review and theoretical conceptualization
of the nightmare literature, Levin and Nielsen (2007, 2009) pointed out that
nightmares occur not only due to a daily stressor but also to affect distress
predispositions. In their model they consider affect load, or situational events like
interpersonal conflict and trauma, and affect distress, or dispositional traits which
may be genetic or due to life history like attachment issues or unresolved trauma,
as interacting to result in the experience of a nightmare. At its worst this process
can become pathological by disturbing sleep in order to avoid the nightmare or in
response to the nightmare. In addition, the psychological distress resulting from the
nightmare can retraumatize (Barrett, 2001).
In a more clinically oriented review of the posttraumatic nightmare literature,
Phelps, Forbes, and Creamer (2007) point out that memory for traumatic events is
stored in a different way than for every day events. These trauma memories are
unprocessed. Thus the hallmark of the Post Traumatic Stress Disorder (PTSD)
nightmare is “a repetitive, replay of the traumatic event, complete with accompa-
nying cognitive, affective, physiological, and behavioral response” (p. 342) and can
occur for decades. Therapeutically, Phelps et al. point out that “various studies
emphasize the importance of ‘facing and conquering’ the feared nightmare, in order
to eliminate it” (p. 352). This echoes what the Gackenbach group has found to be
well learned and thus spontaneous for gamers.
The idea that virtual worlds might act as therapy and/or protection against such
trauma affects has been addressed as well. Bushman and Anderson (2009) talk
about how the numbing effects of violent media, including gaming and movies,
inhibit helping. And on the flip side Anderson and Dill (2000) concluded that there
was an increase in aggressive thoughts, feelings, and behavior associated with
violent video game play. While not without criticisms (Ferguson, 2009), this work
certainly supports the finding of some bleed of aggressive behavior from gaming
into dreams found by the Gackenbach group. Relatedly, virtual worlds have also
been used in stress inoculation training for PTSD by Wiederhold and Wiederhold
(2008) and thus is consistent with the current suggested inquiry.
In the present study of military gamers, it is hypothesized that those who game
the most will experience less threat in dreams that occurred while in the military
when affect distress and load are controlled. Additionally, more dream control is
expected among high-end military gamers.
METHOD
Participants
Individuals who are currently serving in the military or have served in the
military constituted the research participants (N⫽377). They were contacted
through announcements posted through social organizations or online websites; or
they wrote the first author in response to media articles about her research
program. Upon contact they were asked to participate in a research study exam-
Video Games as Nightmare Protection 223
ining the relationship between playing video games and being in the military. They
were referred to the data collection website where they were prescreened. Those
who fit the prescreening criteria moved on to the research website. All individuals
who served or are currently serving in the military were eligible. Neither credit, nor
payment was offered for participation.
The first level of prescreening inquired about whether the potential participant
had played video games. All but two responded affirmatively. They went on to the
second level of prescreening where several demographic questions were asked (e.g.,
age), as well as questions designed to determine whether the individual had
suffered stress-related symptoms in the last six months. If potential participants did
not meet the demographic requirements or indicated stress-related symptoms in the
previous six months, they were not moved on to the full questionnaire. Of the 377
who started the prescreening, 115 indicated that they had not been in the mili-
tary, 31 indicated that they were not yet 18 years of age, and 16 indicated that they
did not have a high school education. Responses that were affirmative to the
questions about stress in the previous six months indicated that 29 had been
diagnosed with a mental disorder; eight had considered suicide; 58 had engaged in
risky behaviors without concern for their mortality; and 20 had been addicted to
drugs or alcohol. Thus 279 potential research subjects were eliminated: two as
nongamers; 162 due to demographic requirements; and 115 because of stress-
related symptoms in the previous six months. The research questionnaire was
completed by the remaining 98 soldiers.
Measures
These are listed in the order they were administered online.
Prescreening Questionnaire. Eight questions constituted the two prescreen-
ing levels. The first level simply asked whether the potential participants had ever
played video games. A positive response allowed participants to proceed to the
second prescreening level where they were asked three demographic questions (i.e.,
military, age, education) and four mental health questions derived from Davis,
Byrd, Rhudy, and Wright (2007). Skip logic on these questions did not allow
moving on if a question was answered incorrectly. As noted, criteria for exclusion
were less than 18 years of age, not having a high school education and not being in
the military as well as indications of apparent diagnosed mental disorder, “active
suicidality or recent parasuicidal behaviors, or current alcohol or drug dependence”
(p. 190). This list is in line with a previous study on nightmares in trauma-exposed
individuals seeking treatment (Davis et al., 2007; p. 190).
Demographic Questionnaire. This was adapted from Smith et al., (2007) and
Hoge, Auchterlonie, and Milliken (2006). Both studies examined health behaviors
associated with military deployment. Thus these 11 items were asked for general
demographic information; gender, age, education, marital status, race or ethnicity,
and military specific information: country where served, military pay grade, service
type component, branch of the service, occupational service category, deployment
history, and deployment experience.
Video Game Play History Questionnaire. This 18-item questionnaire was
adapted from Gackenbach (2006), 2009; Gackenbach & Rosie, 2009; Gackenbach,
224 Gackenbach, Ellerman, and Hall
Rosie, Bown, & Sample, 2011). Questions inquired into the participants’ life history
of playing video games, including genres of preferred games. The questionnaire
also asked about games played prior to filling out the questionnaire in order to get
a recent sample of play behavior. Types of question included frequency of play,
length of play, number of games played, age begun play, age of peak play, and
genres preferred at various times in the life span. These questions were followed by
others inquiring into game(s) played immediately prior to filling out the question-
naires. Questions were also asked about physical game apparatus used as well as
social elements of play.
Gackenbach and Bown (2011a) report the following validity information for
this scale. In their study, the four game group-defining variables were frequency of
play, duration of typical play, number of games played in a lifetime, and age begun
playing with younger coded as a higher number.
Validity for these general history of game play questions was determined in terms of their
relationship to questions about their game play immediately prior to the research partici-
pation. The number of games subjects reported playing prior to participating in the study was
associated with typical gaming session duration r⫽.247, p⬍.009 and number of different
games played in lifetime r⫽.204, p⬍.032. In terms of these four group-defining history of
play item responses to the length of the prior to research participation gaming sessions (game
frequency r⫽.294, p⬍.0001; gaming session duration r⫽.496, p⬍.0001; number of
different games played in lifetime r⫽.325, p⬍.0001). Thus, history of gaming was related
in various ways to actual play behavior, in the 24 hours prior to filling out the research
inventories. (p. 3)
Emotional Reactivity and Numbing Scale (ERNS). This 62-item scale was
developed by Orsillo, Theodore-Oklota, Luterek, and Plumb (2007) because hy-
perarousability and numbing are known to be part of the experience of PTSD.
However, while there are other emotional reactivity measures, they do not include
the numbing aspect which has been found to affect PTSDs “development, main-
tenance, and treatment” (p. 830; Orsillo et al., 2007). This scale was normed on U.S.
military veterans and thus was most directly relevant to the current inquiry. Items
are clustered into five subscales: positive subscale, sad subscale, general subscale,
anger subscale, and fear subscale. Orsillo et al. report that these “demonstrated
good to excellent internal consistencies with the following Cronbach’s alpha levels:
positive subscale, ␣⫽.91; sadness subscale, ␣⫽.88; anger subscale, ␣⫽.87; fear
subscale, ␣⫽.81; general subscale, ␣⫽.81” (p. 3). This scale was also reported as
having good test–retest reliability as well as convergent and discriminate validity.
Trauma Inventory. This 38-item scale was adapted from the Loss/Trauma
Questionnaire of Eng, Kuiken, Temme, and Sharma (2005). The part used herein
inquired about the incidence and intensity of nine types of trauma:
●Physical Assault: assault, incest, sexual assault, mutilation, physical abuse
●Recurrent Physical Assault: repeated physical assault (as defined above)
●Recurrent Emotional Abuse: repeated verbal aggression, humiliation, ne-
glect, or isolation
●Criminal Victimization: armed robbery, burglary, kidnapping, drive-by
shooting
●Negligent Injury: drunk driving resulting in physical harm, a serious car
accident resulting in physical harm, inappropriate or negligent medical treat-
ment resulting in physical harm
Video Games as Nightmare Protection 225
●Civil, Domestic, or Industrial Disaster: serious fires, collapse of a struc-
ture (e.g., bridge), crash of a transportation system (e.g., plane, train),
technological accident (e.g., exposure to radiation), work-related accident
(e.g., explosion)
●Cultural Violence: war, genocide, terrorism, torture
●Natural Disasters: earthquake, hurricane, tornado, avalanche, forest fire,
flood
●Other trauma: for example, life threatening illness, animal attacks, freak
accidents
Each trauma was briefly defined and then followed by a yes/no question as to
whether it had been experienced and at what age it was experienced. Two addi-
tional questions were asked about intensity of the effects of the experience and
nightmare occurrence after the experience. Unfortunately, due to a problem with
the software question formatting, these were not usable data.
Recent and Impactful Dream Recordings Scale. This questionnaire is
adapted from Lee (2009). Respondents are asked first to supply the most recent
dream they can recall and then an impactful dream which they believe deals
with their military experience. An impactful dream is asked for rather than a
nightmare in order to allow for a wider range of replies and not to assume that
there were nightmares. In each case subjects are instructed to tell the dream in
as much detail as they can. The dream recording instructions were the following:
Please pick the most recent dream that you recall, preferably from last night. Please describe
this dream as exactly and as fully as you can remember it. Try to tell the dream story, from
beginning to end, as if it were happening again (and without any interpretation or explana-
tion). Your report should contain, if possible, a description of:
●all the objects, places, characters, and events in your dream;
●the entire sequence of actions and events, from the beginning to the end of your dream;
●your moment-to-moment thoughts and feelings, from the beginning to the end of your
dream; and
●any unusual, incongruous, or implausible dream thoughts, feelings, objects, places, char-
acters, or events.
After giving the time when the recent dream occurred, subjects were then
asked to provide a “Military Experience Impactful Dream”. These instructions
were as follows:
An impactful dream is one that continues to influence your thoughts and feelings even after
you have awakened. Please choose one such dream that is potentially related, either directly
or indirectly, to your military experience.
The rest of the military impactful dream instructions were the same as for the
recent dream instructions.
Impactful Dreams Questionnaire (IDQ). This scale is adapted from Zadra,
Pilon, and Donderi, (2006) and Busink and Kuiken (1996). The first part asked
about 15 emotions and how intense the dreamer had experienced them during the
impactful dream, as per Zadra et al. (2006). However, there was a problem with the
item structure due to software issues; thus this information was lost. These emo-
tional evaluations of the dream were followed by a list of 19 questions asking about
226 Gackenbach, Ellerman, and Hall
the military impactful dream, as per Busink and Kuiken originally and most
recently restructured by Kuiken (2009), in order to classify the dream as nightmare,
existential, or transcendental.
Procedure
Prior to the data collection phase of the study, the research assistant partici-
pated in a private crisis intervention paraprofessional training. This student worked
on the project as part of her undergraduate coursework at a western Canadian
distance education university. This training was designed to sensitize her to poten-
tial problems in the subject correspondence phase of the study.
The sampling technique was convenience sampling. The next stage of the
project was an exploration of possible local, regional, and online places where
current and former military personnel might be contacted. Calls for research
participation were posted to several websites including military sites (e.g., www
.military.com, www.usmessageboard.com, www.army.ca, www.americasarmy.com),
gaming sites (e.g., www.militarygamers.net, www.gamers-forum.com), and comic
book sites (e.g., www.the-master-list.com). A major psychology research website
was also used to solicit potential subjects. A Facebook page was created supporting
the research project with cross posts to two Canadian Forces pages, a U.S. Army
support page, a U.S. Veteran, and a World of Warcraft fan site. Brochures and
bulletins were handed out and posted at military bases in western Canada and in
southern California. Support was sought from Military Family Resource Centres as
well as some social gathering spots for the military around bases in these two
regions. Finally, 55 possible participants wrote to the first author after a series of
media articles appeared about our research laboratory wanting more information
on this the research topic and/or offering to help as a research participant.
Some notices of the research included the web address of the questionnaire,
that is, the psychology research website, while other possible participants were
contacted by the research assistant and asked if they would be interested in
participating in a research study. Queries came to the assistant from the posters and
brochures as well as from media attention to the research program. Participants
were advised that related questionnaires to the study’s purpose of examining
gaming in the military would also be administered. At this juncture the explanation
of these related surveys was kept broad, including personality scales, life experi-
ences, and dreams questions. Those who agreed to participate were sent the
questionnaire web address. While IP addresses and referring URLs were gathered
automatically by the survey software, 130 of those who entered the survey did not
have a referring URL. This means that they entered from an e-mail response or
typed in the web address to their browser. The largest referring URL was the
United States psychology research site (n⫽143).
If potential participants answered the first question about playing video games
in the affirmative, they were presented with an electronic informed consent for the
remainder of the prescreening questions. Only those who fulfilled the criteria were
moved on to the second informed consent which described the entire study. If ever
in doubt about a potential participant, the researchers did not move forward with
that person. For instance, there were a few cases where a soldier wrote the first
Video Games as Nightmare Protection 227
author and indicated they he or she suffered from PTSD. In these cases our
response was an appreciation for their interest in our work but no invitation to
participate was sent. Those who went through the brief prescreening questions and
did not meet the criteria were taken to a debriefing statement explaining the
hypothesis of the study and the background research findings. The e-mails that were
received due to the media attention had been sent to the principle researcher and,
where appropriate, they were forwarded to the research assistant who contacted
these interested individuals and gave them the web address of the study. Thus, for
some subjects who found out about the research through media, there was another
initial screening by the principle researcher.
No individual identities were gathered in the questionnaire at any phase and
there was no way to know which of the people who received e-mails actually went
on to participate. Thus all specific responses to the online questionnaires were
completely anonymous.
Participants was told in the informed consent prior to the full questionnaire
that there were six questionnaires to be filled out and that it should take between 1
and 1.5 hours. They were told that they would receive a debriefing about the study
upon completion of the surveys and that they could put their name on a list to
receive either the executive summary and/or the full report of the study’s findings.
Upon entering the full questionnaire part of the website, each participant
agreed electronically to participate after reading and agreeing to the informed
consent. Upon completion of the survey they were shown a debriefing statement.
This statement was placed at the bottom of each electronic page of the survey, “We
are aware of the sensitive nature of the information that we just requested. Please feel
free to discontinue your participation or disallow the use of your responses at any
time.” If subjects decided to stop their participation before finishing the survey, they
were encouraged in the informed consent to write the researcher and receive the
debriefing statement.
RESULTS
Those who made it through the prescreening, N⫽98, were divided into
gaming groups as a function of the frequency of their self-reported current game
play. The high-end group (N⫽64) were those who reported playing daily or weekly
while the low-end group (N⫽22) were those who reported playing video games
but less often (monthly, yearly, or rarely). By way of verification, the high-end
group was compared to the low-end group in terms of several other game play
variables and were scored significantly higher in each case. These included length
of play, t(83) ⫽4.17, p⬍.0001, number of games played, t(84) ⫽3.195, p⬍.002,
and age began play, t(84) ⫽2.33, p⬍.02. High-end gamers play longer, played
more games, and started earlier in life than the low-end gamers.
Further supporting these group definitions was the information about types of
games. Ninety-six percent of the high-end group preferred game genres that were
typical of the classic hard core gamer: for example, first person shooter, massively
multiplayer, action/adventure, simulation, fighting, and strategy. While the majority
of the low-end group also preferred these same classic genres (67%), a significant
228 Gackenbach, Ellerman, and Hall
minority of the low-end gamers (29%) preferred casual games.
1
A small percentage
of each group preferred sport and driving games (3% of the highs and 5% of the
lows). Another question regarding game preferences was gathered where partici-
pants were asked if they had just played a video game prior to filling out the
questionnaire. While the high-end gamers were equally likely to say they had (N⫽
33) or had not (N⫽30) been playing a video game just prior to filling out the
questionnaire, the low-end gamers were more likely to report not having played
(N⫽19) than having just played (N⫽1),
2
(1) ⫽14.092, p⬍.0001. These
respondents listed a variety of games as played the most just prior to filling out the
questionnaire. Role playing games, such as Mass Effect, were identified 19 times;
Action Adventure games, such as Dead Rising, identified seven times; Shooter
games, such as Call of Duty Modern Warfare 2, identified 23 times; Strategy games,
such as Civilization 4, identified 11 times; Simulations, such as Madden 2011,
identified six times; and Casual games, such as Yohoho, identified three times. As
in the examples just provided, most of the games (67%) just played were war or
battle type games across genre.
Video game groups were then compared in terms of general and military
demographics. Gender and education differed across groups for general demo-
graphics, with more women and higher educated individuals falling in the low-end
gamer group (Gender:
2
(1) ⫽15.21, p⬍.0001); Education:
2
(3) ⫽17.07, p⬍
.001). There was no difference in marital status or racial/ethnic background. As for
military demographics there were relatively fewer enlisted soldiers in the low-end
gamer group,
2
(2) ⫽6.45, p⬍.04; but none of the other military demographics
evidenced a gamer group difference (i.e., country of military service, service com-
ponent, branch of service, occupational category while in the military). Of partic-
ular interest to the present study were any group differences in military deployment
or combat. None of these seven questions [i.e., sum of deployments, t(84) ⫽.06;
deployed or not (t(84) ⫽⫺.524; combat experience, t(84) ⫽⫺.15; witnessed others
being wounded or killed, t(84) ⫽1.49; discharged a weapon, t(84) ⫽.09; danger of
being wounded, t(84) ⫽.85, was wounded, t(84) ⫽1.62; or felt in great danger
of being killed, t(84) ⫽.10] showed gamer group differences. Both groups reported
on average 1.5 deployments in their military history and answered affirmatively to
an average of two of the six combat questions.
In order to determine the possible effects of game play on dreams that might
be nightmarish, covariates predictive of nightmares needed to be controlled. These
were assessed in the present study. Affect distress was assessed using the Emotional
Reactivity and Numbing Scale (Orsillo et al., 2007; ERNS), with five subscale
scores. Affect load was determined in terms of personal history, sum of lifetime
traumas, and in terms of military experience, sum of combat experiences, and
whether they were deployed. Gamer group ttests on each of these variables were
computed and are portrayed in Table 1. It can be seen that three of the subscales
of the affect distress measure (ERNS) showed group differences while none of the
measures of affect load resulted in gamer group differences. Specifically, the
low-end gamer group were sadder, angrier, and more fearful than the high-end
1
Casual games take less time to play and are cognitive and emotionally less engaging, that is, lower
in presence, than the traditional “hard” core games (Gackenbach & Bown, 2011b).
Video Games as Nightmare Protection 229
gamer group soldiers. ERNS subscale norms for both PTSD and none-PTSD
soldiers are also listed in Table 1. It can be seen that the current set of soldiers
had some differences and some similarities to the veterans used in the ERNS
samples. Specifically, both gamer groups fell between the PTSD and none-PTSD
positive subscale scores. While for the sadness ERNS subscale, both gamer groups
fell below the two norm groups. The general, or emotional numbing, subscale
scores for the gamer groups in this study were at or below those of the norm groups.
Anger subscale scores were higher for the low-end gamer group than the norms.
Finally, the fear subscale scores for this sample were below the ERNS norm groups.
In order to determine the effects of gaming on nightmarish content in dreams, all
affect distress and affect load variables were used as covariates in all analysis of
dream content.
Table 2. Means, Standard Deviations, N’s and F Values For Dream Type ⫻Gamer Group
ANCOVA With Affect Distress and Affect Load Controlled For Sum of All War Content
Dream
type
Video game
group Mean
Std.
deviation NF-values
Recent High 1.63 2.016 35 Main effect for dream type: F(1, 75) ⫽14.73,
p⬍.0001, partial eta
2
⫽.164
Main effect for gamer group: F(1, 75) ⫽4.96,
p⬍.03, partial eta
2
⫽.06
Low 1.27 1.272 11
Military High 2.41 2.408 32 Interaction dream ⫻gamer: F(1, 75) ⫽5.31,
p⬍.02, partial eta
2
⫽.07
Low 4.22 2.774 9
Table 1. Affect Distress (ERNS Subscales With Norms) and Affect Load (Combat, Deployed, and
Life Traumas) N’s, Means and Standard Deviations
Control variable
Video
game
groups N
Subscale
means
Std.
deviation
Norms
PTSD
soldiers
Norms No
PTSD
soldiers
ERNS positive subscale mean
(#items ⫽26)
High 69 93.29 13.63 90.33
‡
97.98
Low 22 96.772 18.02
ERNS sad subscale mean
(#items ⫽11)
ⴱ
High 69 32.6667 8.51 43.21 39.35
Low 22 36.5002 8.61
ERNS general (numbing)
subscale mean (#items ⫽7)
High 69 23.6523 4.74 24.62 28.25
Low 22 24.9088 6.21
ERNS anger subscale mean
(#items ⫽11)
ⴱⴱ
High 69 37.4198 7.45 38.10 34.61
Low 22 42.0453 6.92
ERNS fear subscale mean
(#items ⫽6)
ⴱⴱ
High 69 15.3912 5.10 20.54 19.40
Low 22 18.3636 3.58
sum of experiences in combat High 69 2.07 2.23158
Low 20 2.00 2.05196
deployed (1 ⫽no; 2 ⫽yes) High 69 1.64 .48419
Low 20 1.75 .44426
sum of all traumas High 69 10.62 3.18
Low 22 11.45 2.22
ⴱ
p⬍.07.
ⴱⴱ
p⬍.01.
‡
Norms differed between PTSD and none-PTSD subjects for all ERNS subscales but fear.
230 Gackenbach, Ellerman, and Hall
Dream Content Analysis
Dreams were content analyzed using three systems. First was Revonsuo’s
(2000, 2006; Revonsuo & Valli, 2000) threat simulation content analysis system. A
threatening event in a dream is one which meets at least one of the following two
criteria:
Objective threat: An event in a dream where, if the event was real, the physical
or mental well-being of any person would be endangered or where any person’s
physical resources or territory would be jeopardized (i.e., any event that would be
considered threatening if it should really occur in waking life). Such an event may
be directly witnessed by the dreamer reporting the event or only indirectly heard
about in the dream.
Subjective threat: An event in a dream that is interpreted or emotionally
experienced by the dreamer (i.e., the dream Self) to be somehow dangerous. Any
event in which the subject reports the feeling of danger or threat even if no
objective threat (as defined above) is reported to accompany this feeling.
Then Wilmer’s (1996) system of classifying dreams of Vietnam veterans was
used to content analyze these dreams. In this system dreams are classified as:
Category I is the characteristic terrifying nightmare of the actual event as if it were recorded
by cinema . . . Category II or the “variable” nightmares contain plausible war sequences that
conceivably could have happened but did not actually occur . . . Category III dreams are like
ordinary nightmares, but their identification with the specific trauma or place of the trauma,
that is, the Vietnam War, is always present. (p. 87)
Along with Hartmann (1984), Wilmer holds that movement away from a literal
replay of the dream indicates healing and integration with other autobiographical
memories. Finally, the degree of lucidity/control and game play types of activity in
the dreams was assessed.
Before the content analysis is discussed, the time when dreams occurred must
be examined. An ANOVA of Dream Type (recent/military) ⫻Gamer Group
(high/low) was computed on time of dream. A main effect for dream type emerged
(F(1, 72) ⫽19.177, p⬍.0001, partial eta
2
⫽.210). Military dreams were older than
recent dreams. The military dream mean was “within the last 6 months”, while
recent dream mean was “last week sometime”. Despite this difference, time of
dream was not used as a covariate because the time difference was built into the
questions about which dreams to report. Not all soldiers who participated were
currently in the military, so a military dream would be expected to have happened
Table 3. Wilmer Classification of Type of Nightmare as a Function of Dream Type and Gamer
Group and Norms
Dream
category
Count and
column
percentage
Recent
dream
Military
dream
High gamer
group
Low gamer
group
Wilmer vietnam
veterans norms
Actual Event
Nightmare
Count 0 4 3 1 189
Column % .0% 12.5% 6.1% 5.6% 53%
Plausible Event
Nightmare
Count 19 23 27 15 76
Column % 55.9% 71.9% 55.1% 83.3% 21%
Ordinary
Nightmare
Count 15 5 19 2 94
Column % 44.1% 15.6% 38.8% 11.1% 26%
Video Games as Nightmare Protection 231
awhile ago. “Within the last six months” is actually rather recent, relatively speak-
ing, as an indicator of how long ago it could have been. Likely the elimination of
individuals who had never gamed resulted in this finding: that is, subjects would
likely have been older and thus their military dreams had occurred longer ago.
Threat Simulation Dream Content Analysis
One judge was trained to rate the morning after dream reports using the
“Dream Threat Rating Scale” (Revonsuo, 2000). Using this system, dream analysis
is carried out in two phases. To begin, the researcher must identify and isolate the
description of any threatening events that may occur in a dream report. A threat-
ening event is one that meets at least one of the following two criteria as noted
earlier, “objective threat” or “subjective threat.”
Next, the identified threatening events are rated on the following eight subscales:
nature of the threatening event, target of the threat, severity of the threatening event
for the self, participation of the self in the threatening event, reaction of the self to the
threatening event, consequences of the threatening event to self, resolution of the
threatening event, and the source of the threatening event. Each of the subscales
allows for further classification within them. For example, an event being
analyzed using the “nature of the threatening event” subscale allows the re-
Figure 1. Aggressive nature of threat as a function of dream type and gamer group.
232 Gackenbach, Ellerman, and Hall
searcher to further break the event down by classifying it as one of a variety of
threatening events including escapes, accidents, diseases, catastrophes, and so
forth.
To ensure an adequate level of training, the judge rated dreams from Gackenbach
and Kuruvilla (2008) until she reached an 80% agreement with the threat simulation
coding on that original set of 10 dreams. Gamer Group (high/low) ⫻Dream Type
(recent/military) ANCOVA’s with five affect distress, ERNS subscales, and three
affect load, trauma experiences, as covariates were computed on the continuous
variables for the Threat Simulation Scales. A main effect for nature of threat coded in
terms of aggressiveness (1 ⫽no harm,2⫽nonaggressive harm,3⫽aggressive harm)
was significant for dream type (F(1, 51) ⫽6.99, p⬍.011, partial eta
2
⫽.12) such that
more aggressive threat was coded in military dreams than in the recent dreams. There
was also a significant gamer group by dream type interaction (F(1, 51) ⫽4.596, p⬍
.037, partial eta
2
⫽.083), which is portrayed in Figure 1.
Also significant was a dream type main effect (F(1, 51) ⫽10.186, p⬍.002,
partial eta
2
⫽.166) and interaction (F(1, 51) ⫽3.244, p⬍.078, partial eta
2
⫽
.06) for severity of threat recoded to be continuous (1 ⫽none,2⫽trivial,3⫽
social/psychological,4⫽life threatening). In both analyses the military dreams
had more severe threat than the recent dreams which was accounted for
primarily by the low-end gamers. None of the other three threat simulation
variables treated as continuous were significant for either main effects or
interactions: threat simulation present, sum of the number of threats from target
threat, and consequences of threat. For conceptual reasons it’s important to
point out that self-participation was marginally significant when only ENRS
were covariates (dream main effect: F(1, 55) ⫽3.35, p⬍.07, partial eta
2
⫽.06;
dream by gamer interaction: F(1, 55) ⫽2.97, p⬍.09, partial eta
2
⫽.05). The
dream ego was less likely to participate in the low end gamers military dreams
than in the high end gamers military dreams.
Two threat simulation variables were treated as repeated measures, source
of threat (i.e., personal, media, fiction and unidentified) and resolution of threat
(i.e., happy, unhappy, discontinuous, dream ends). Thus ANCOVA’s for gamer
group by dream type by each of the within subject threat simulation variables
were computed with affect distress and affect load as covariates. Significant
three way interactions were evident in each case (source of threat F(1,
50) ⫽4.64, p⬍.04, partial eta
2
⫽.085; and resolution of threat F(1, 48) ⫽3.90,
p⬍.05, partial eta
2
⫽.085). For the resolution of threat it can be seen in Figure
2 that a different pattern of resolutions to the threat was evident for the high-
versus the low-end gamer groups. For the low-end group, consistent with
previous findings on threat simulation, there were more abrupt awakenings in
the military dreams than in the recent ones while the opposite was the case for
the high-end gamer group. There was no difference between dream types for
happy resolutions for the high-end gamers, while military dreams were less
likely to end in happy resolutions for the low-end gamers. Another distinction
between the resolutions of the high- versus the low-end gamer groups, is that for
the lows there was not much difference in types of resolution for the recent
dream but big differences in their military dreams. For the high-end gamers, on
the other hand, there were different patterns of resolution for the two different
types of dream.
Video Games as Nightmare Protection 233
The source of threat three way interaction is portrayed in Figure 3, where
it can be seen that the largest source was personal but that the direction of
this threat differed as a function of type of dream for high- versus low-end
gamers. Specifically, high-end gamers evidenced more personal threat in their
military dreams while low-end gamers personal threat source did not differ as a
function of dream type. The other three possible sources of threat (i.e., media,
fiction, or unidentified) did not differ as a function of dream type or gamer
group.
Figure 2. Resolution of threat as a function of gamer group and type of dream.
234 Gackenbach, Ellerman, and Hall
War Content in Dreams
Wilmer’s scale is constructed to code for war type content, as well as
instructions for categorization of dreams as one of three types of nightmare. He
developed the scale during his work with Vietnam veterans who sought therapy
for PTSD. The war type content included act frequency counts for under attack,
war/battle, the dead, firefights, killing women and children, killing enemy,
killing buddies, captured, somewhere in war/battle, being wounded, chase and
Figure 3. Source of threat as a function of gamer group and type of dream.
Video Games as Nightmare Protection 235
running, home, being killed, animals, decapitation, looming danger, shot down,
atrocities/ mutilation, and return to war/battle. These were summed and a
Gamer Group (high/low) ⫻Dream Type (recent/military) ANCOVA with the
same affect distress and affect load covariates was computed. Both main effects
and the interaction were significant. These Fvalues and means are portrayed in
Table 2. Military dreams were higher in war type content overall as were low
gamers’ dreams, but the interaction makes this clearer. Specifically, it was the
military dreams of the low-end gamers that had the most war content, which is
consistent with the threat simulation findings.
Wilmer also instructs classification of dreams into one of three types of
nightmares: actual events, plausible events, and ordinary nightmares. A chi-square
on this classification as a function of dream type,
2
(2) ⫽9.33, p⬍.01 was
significant and as a function of gamer group,
2
(2) ⫽4.90, p⬍.08 was near
significant. The numbers and percentages along with Wilmer’s norms are portrayed
in Table 3. The military dreams were most likely to be classified as actual or
plausible event nightmares, while the ordinary nightmares were more likely to
come from the recent dream offerings. In terms of the gamer group effect, there
was no difference in terms of actual event nightmares but the plausible event
nightmares were considerably higher (83%) among the low-end gamers than
among the high-end gamers (55%). This was reversed for the ordinary nightmares
with high-end gamers dreams being classified more (39%) here than the low-end
gamers dreams (11%).
Lucid/Control and Gaming Dream Content Analysis
A dream coding system has been developed in this laboratory for coding
lucidity and related variables as well as game play in dreams (Gackenbach & Rosie,
2009; Gackenbach, Rosie, Bown, & Sample, 2011; Gackenbach et al., 2009). It was
further refined for the current inquiry. In this iteration there were two sections:
lucid/control type scales and game type scales. In the lucid/control scales coders
2
were asked to rate the lucidity, knowing subjects were dreaming while dreaming, as
well as prelucid content (i.e., talking about dream in the dream, false awakening,
wondering if it was a dream only to conclude it was not, and out-of-body experience
in dream). A version of Kahan and LaBerge’s (1994) MACE scale was adapted for
coders to code if each type of thinking was present in the dream. Dream control was
coded in terms of self, dream characters, and dream environment. Finally, the
dream ego’s stance, that is, first person to third person, was assessed. As the intent
is not to examine nightmarish content, the previous covariates of affect load and
distress were not used as covariates. Thus, there were no covariates on these
analyses.
In these ANOVA’s of dream type ⫻gamer group there was no significance:
Lucid & Prelucid Dream
3
; Dream:, F(1, 59) ⫽0.23, ns; Group:, F(1, 59) ⫽0.90, ns;
2
We’d like to thank Katherine Wisniewski and Mary-Lynn Ferguson for their help on this part of
the project and Grant MacEwan University for a grant to support this coding.
3
Sum of recoded judges’ evaluations of lucid and prelucid content.
236 Gackenbach, Ellerman, and Hall
Dream ⫻Group:, F(1, 59) ⫽1.45, ns; Dream Control (Self, character, & environ-
ment); Dream:, F(1, 59) ⫽0.64, ns; Group:, F(1, 59) ⫽0.14, ns; Dream ⫻Group:,
F(1, 59) ⫽1.08, ns; and Dream Ego Stance; Dream:, F(1, 59) ⫽0.38, ns; Group:,
F(1, 59) ⫽0.38, ns; Dream ⫻Group:, F(1, 59) ⫽.38, ns. The MACE subscale scores
were mean evaluations by judges of items that logically clustered. These clusters
were focus on attention, emotion, thinking, and task. There was a MACE by Dream
Type interaction which is portrayed in Table 4 with Fvalues. According to the
judges, there was more attention used in the military dreams overall than in the
recent dreams while there were no dream type differences in emotions evident,
thinking, and task focus with emotions being rated the highest.
The game in dream coding was approached in a variety of ways. First, distinc-
tions were made between the ways that games might be portrayed in dreams
ranging from the dream being a game to an offhand mention of gaming. Thus these
variables were coded as present or absent and then summed with blanks converted
to zeros:
a. In the dream the dream ego is in the game world—the dream is the game
(leave blank if none of these)
1. stated in the dream transcript
2. implied in the dream transcript
b. In the dream the playing of a video game (leave blank if none of these)
1. Dream ego playing a video game
2. Watching others play a video game
3. Other (i.e., video game playing in the background as a movie etc.)
c. In the dream games are mentioned (leave blank if none of these)
1. Nonvideo game (i.e., sports watching on TV or watching live or playing
“real” sports; casino gambling)
2. Video game (i.e., shopping for a video game, dressed as a video game
character but clearly NOT playing it, winning a X-Box)
It can be seen in Table 4 that only in the case of the recent dreams for the
high-end gamers was there some indication of gaming in the dream. It should be
noted that the similarity of military themed content to military games played by
these soldiers may be confounding these results.
Gaming was also taken up in terms of the five subscales generally defined from
the Hall and Van de Castle content analysis of dreams (1966). These were coded as
present or absent and in terms of game content and included:
Characters: self is game character, other is game character, self changes into
game character, self controls game character
Activities: physical, movement, location change, verbal, visual, auditory, thinking
Emotions: positive emotions, negative emotions, neutral emotions
Settings: locations
Objects: architecture, household, food, travel, streets, regions, nature, body
parts, clothing, communication (including technological), money, miscella-
neous. Look especially for weapons.
A sum score of these coded items was significant for the dream by group
interaction, see Table 4. Again the interaction approached traditional levels of
significance, in the same direction as the previous analysis of game content. Finally,
the content of all dreams with or without game content was coded using a reduced
Video Games as Nightmare Protection 237
Table 4. Gamer Group (High/low) ⫻Dream Type (Recent/military) ANCOVA’s With no Covariates on Various Measures of Lucid/Control, Gaming Content
in Dreams and Self Evaluations of Dreams (IDQ)
Variable
Finding for dream type
(recent/military) ⫻gamer
group (high/low) with no
covariates
Mean/SD/N recent
dream high-end
gamers
Mean/SD/N recent
dream low-end
gamers
Mean/SD/N military
dream high-end
gamers
Mean/SD/N military
dream low-end gamers
Game Playing in Dream
§
Dream ⫻Group Interaction:
F(1, 55) ⫽3.91, p⬍.05,
partial eta
2
⫽.07
0.43/0.84/23 0/0/8 0/0/21 0.13/0.35/8
Sum of HVDC Subscales
for Game Elements
Dream ⫻Group Interaction:
F(1, 59) ⫽3.10, p⬍.08,
partial eta
2
⫽.05
0.96/1.65/25 0.00/0/8 0.14/0.64/22 0.38/1.06/8
MACE: Attend MACE: Emotion MACE: Think MACE: Task
Judges evaluation of
attention, emotion,
thinking, tasks in
dream (MACE)
Dream ⫻group ⫻MACE: 2
way interaction dream ⫻
MACE F(1, 59) ⫽3.30,
p⬍075, partial eta
2
⫽
.053
Recent: 1.42/.83/33
Military: 2.23/1.61;30
Recent: 2.76/1.77/33
Military: 2.87/1.81/30
Recent: 1.66/.80/33
Military: 1.56/.63/30
Recent: 2.56/1.35/33
Military: 2.48/1.62/30
§
Sum of game in dream includes dream is game (stated or implied), game being played in dream (playing, watching, mention), and games mentioned (video game
other than played, other games). In this case only number of words in dream was a covariate.
238 Gackenbach, Ellerman, and Hall
version of the ESRB
4
classification criteria. These variables included violence, sex,
drugs, language, humor, and gambling. The presence or absence of each was coded
and summed. There were no significant findings: Dream: F(1, 59) ⫽0.34, ns; Group:
F(1, 59) ⫽0.39, ns; Dream ⫻Group: F(1, 59) ⫽.28, ns.
IDQ Results
The final set of statistical analyses were for questions on the IDQ. This was
filled out by the respondent only in terms of their military dream. While not all
respondents filled it out, there were enough responses to warrant looking at the
results. In ttests on these items all but five of the 19 items resulted in no gamer
group difference. The five that did evidence a gamer group difference are summa-
rized in Table 5. It can be seen that the high-end gamers rated their dreams as
higher in all but avoiding harm which was rated higher by the low-end gamers and
is consistent with the judges, dream content analysis.
DISCUSSION
Does video game play inoculate soldiers against the negative effects of night-
mares associated with war trauma? The results of this inquiry suggest a qualified
yes, for those not currently suffering from PTSD symptoms. While there is much
work to be done, and prescriptive suggestions are premature at this point, these
results imply that video game play, especially of the war and battle type, may offer
help in terms of practice in fighting the enemy in imaginal realms, in this case
virtual, and in terms of the numbing toward violence that is oft cited in the
aggression modeling literature on gaming.
In this inquiry, individuals who play video games and are or have been in the
military were solicited to participate in an online survey. They were screened out
for recent evidence of PTSD. Two dreams were collected from each participant: a
4
The ESRB is the Entertainment Software Review Board that classifies games for consumer
information.
Table 5. Means, SD, and N’s for Significant Items From The IDQ as a Function of Gamer Group
Video game groups N Mean Std. deviation
IDQ spreading warmth High 33 1.52 1.176
Low 11 .91 .302
IDQ repeatedly avoid harm High 33 2.09 1.588
Low 11 3.27 1.489
IDQ movements vigorous and energetic
ⴱ
High 33 3.06 1.767
Low 11 2.00 1.549
IDQ movements well balanced and graceful
ⴱ
High 32 1.59 1.160
Low 11 1.18 .405
IDQ sudden shift in visual perception High 32 2.50 1.545
Low 11 1.45 1.036
ⴱ
Approached traditional significance levels at p⬍.1.
Video Games as Nightmare Protection 239
recent dream and an impactful military dream. Two groups of gamer soldiers were
identified in terms of their self reports of current game play frequency: high-end
gamers who reported playing daily or weekly and low-end gamers, who reported
playing monthly, yearly or rarely. This group classification was verified in a number
of ways. The high-end group played more games, started younger, played more
often, and was more likely to have just been playing a game prior to filling out the
survey, than was the low-end group. Additionally, while 67% of the low-end gamers
preferred the classic hard core genre (i.e., first person shooters, Massively Multi-
player Online, strategy, action, adventure), 96% of the high-end gamers had the
same preferences. Casual, driving and sport games (i.e., Farmville, baseball) were
preferred by 34% of the low-end gaming group. One respondent reflected upon this
game preference:
I would see many soldiers, in combat, with PSPs or anything we could hook up to 220v
electricity. When soldiers weren’t on patrol, we often had violent war games on our systems.
It was weird. Like we didn’t get enough violence.
Indeed in a story about gaming on the frontlines, one informant wondered if
the “combat-centric titles” might be associated with wanting to be in the military
and went on to speculate that
The average military member will never actually live out that exciting, epic firefight, the
moment of a decisive battle, the heat of combat, or the thrill of knowing you just outwitted
your mortal enemy and crushed them beneath your strategic might and skill, so gaming
definitely offers a way of living out that fantasy. Call it a strange form of escapism. (Ashcraft,
2011)
In this sample, combat-centric titles were definitely preferred.
Because Levin and Nielsen (2007) concluded that nightmares are predicted by
affect distress and affect load, information was also gathered on these two variables
in order to control for these effects. Specifically, Orsillo, Theodore-Oklota,
Luterek, and Plumb’s (2007) Emotional Reactivity and Numbing Scale (ERNS)
was administered to ascertain affect distress. Affect load information was gathered
through Eng, Kuiken, Temme, and Sharma’s (2005) Trauma Inventory and through
questions adapted from Smith et al., (2007) and Hoge, Auchterlonie, and Milliken
(2006) examining behaviors associated with military deployment. The gaming
groups were compared on the ERNS subscales and on the affect load variables with
no group differences in the latter but some in the former. Specifically, the low-end
gamers scored higher on the sadness, anger, and fearfulness subscales but, inter-
estingly, there were no group difference in the general (emotional numbing) or
positive emotions subscales. There were no gamer group differences in any of the
trauma indices, with both groups reporting from two to three traumas experienced
in their life history out of the nine possible traumas presented to them in the study;
two out of six combat experiences; and one and one half deployments.
Dreams were content analyzed with three coding systems. Revonsuo’s (2000,
2006; Revonsuo & Valli, 2000) evolutionary based threat simulation content anal-
ysis system, Wilmer’s (1996) system of classifying combat in the dreams of Vietnam
veterans, and a lucid, control, gaming coding developed for this study but based
upon previous work (Gackenbach, Rosie, Bown, & Sample, 2011). The pattern of
results for the threat simulation scales and the combat experiences coding was that
240 Gackenbach, Ellerman, and Hall
the low-end gamer group reported more threat and combat in their military dreams
than did the high-end gaming group. This was particularly evident in the severity
and aggressiveness of threat. Here is an example of a high threat military dream
from a low-end gamer:
I couldn’t find my rifle and something was chasing me. I searched the entire forest until I did
find my weapon. As I turned around to shoot what was hunting me - the trigger felt like it
was a 1,000 lb trigger pull. The rounds I was shooting were delayed and where not hitting
where I was aiming. (Subject #21)
While here is one from a high-end gamer:
I was told by my old Sargent [sic] to load up on the humvv [sic] in my gunners spot. he said
we were going to roll out to fight some were in Baghdad. we drove down to the combat area
where there was a brutal fight me and quite a few men against the insergants [sic]. i
remember shooting and seeing men fall on both sides. i saw the faces of the dead eyes wide
and staring at the sky soulless faces of friends. i walked dazed back to the humvv [sic] and
woke up. (Subject #115)
Aggression and combat motifs are present in both dreams, but the high-end
gamer is fighting back, even if “dazed” in the dream. Such active engagement may
be indicative of practice in game play; this notion is supported by the marginal
significance of the self participating in reaction to threat finding in military dreams.
Resolution of the threat was more likely to be located in the dream ending for the
military dream among the low-end gamers. However, the dream ending was more
common in the recent dream for the high-end gamers.
The gaming effect is further illuminated in the finding by the ways in which the
dream was resolved. Four possibilities were coded: happy, unhappy, discontinuity,
and dream ends. Figure 2 demonstrates that, across dream types, high-end gamers
had more happy endings and fewer unhappy endings for their recent dreams.
Discontinuous endings did not differ for the high-end gaming group across dream
types, but it did for the low-end gamers. Discontinuous elements in a dream are
thought to be a type of dream bizarreness (Revonsuo, 2000). Previous research has
found that bizarreness in dreams is higher for high-end gamers (Gackenbach,
Kuruvilla, & Dopko, 2009; Gackenbach & Dopko, 2011). Finally, the dream ending
is the typical resolution, if not the definition, of a nightmare; this was most true of
the low-end gamers’ military dreams.
The low-end gamer group had significantly more combat content in their
military dreams than the high-end gamer group, which is consistent with the threat
simulation findings. Here is an illustration of a military dream, coded as high in
conflict, from a low-end gamer:
I am securing an airport trying to get forgien [sic] nationalist out when a group of militia
comes up and wants to get in. When they figure out they cant [sic] get by they get mad and
be head a civilian and all i can do is watch. (Subject #210)
By way of comparison, here is a military dream coded as low in conflict from
a high-end gamer:
Had a dream that i was back in Iraq. not doing anything special just being there walking
around. I was at our base camp walking around. don’t remember talking to or seeing anyone
there just the camp its self. i don“ remember having any feelings about it other then i kind
of missed it. (Subject #113)
Video Games as Nightmare Protection 241
Wilmer also suggests classification of combat dreams into one of three types of
nightmare: actual, plausible, and ordinary. Chi-squares by type of nightmare for
subjects’ dreams, not surprisingly resulted in more military dreams being coded as
the worst type of nightmare. Consistent with the rest of the data analysis for this
study, low-end gamers had far more plausible nightmares (thus more severe) than
the high-end gamers, although there was no difference in actual nightmares.
The final set of dream coding completed was for lucid and control type dream
content, as well as for gaming content. Previous research has indicated that gamers
are likely to report more lucid and control type dreams than nongamers (Gacken-
bach, 2006, 2009). Game content analysis was conducted to discover whether
self-reported game play differences in the groups of soldiers was reflected in the
actual content of their dreams. There were no gamer group or dream type differ-
ence for any of the lucid, or control type, judges’ assessments, with the exception of
the MACE. The MACE subscales interacted with dream type such that dream type
differences in attention accounted for the interaction. Not surprisingly, the military
dreams were coded as requiring more attention than the recent dreams, while there
was no difference between the dreams in terms of emotions, thinking, or task focus.
This was true of both gaming groups. Furthermore, emotional focus items were
coded by judges as the highest across dream and group, which of course fits with
one of the major functions of all dreams and especially of nightmares (Levin &
Nielsen, 2007, 2009). As for gaming referents in these dreams, it was found only for
the high-end gamers in their recent dreams. This could be a reflection of the
similarity between military action and its dream incorporation and the games that
many of the soldiers were playing.
Finally, some of the respondents filled out the IDQ which followed the military
dream request. While most of the items showed no gamer group difference, those
that did show a difference, reflected the dream coding by judges discussed above.
That is, three of the five items that showed a gamer group difference reflected a
positive perspective on the dream (i.e., spreading warmth, vigorous and energetic
movements and balanced and graceful movements); these were favored by the
high-end gamers. One item was neutral (i.e., sudden shift in visual perception) and
again favored by the high-end gamers; while one item was negative (i.e., avoid
harm) which was favored by the low-end gamer group. This self-report indication
is important, as it confirms the judges’ evaluations. The high-end gamers seemed
less bothered by their military dreams than the low-end gamers. The self-reported
movement items reflect the previously noted active involvement in the dreams
unfolding by the high-end gamers.
Limitations
The major limitations of this study are the sampling method, sample size, and
sample characteristics. Because this study was undertaken with no formal military
sanction, it relied entirely on volunteers. Considerable effort was put into recruit-
ment of volunteers and 377 did enter the survey site. However, due to prescreening,
only 98 actually participated with at least one dream. About 100 volunteers were
lost because they were not in the military. Those who were left were high-
functioning soldiers. On the other hand, due to the length of the survey, there was
242 Gackenbach, Ellerman, and Hall
some survey fatigue; consequently, some items went unanswered. In addition, at the
bottom of each page of the survey, a statement was placed encouraging
the participants to stop if they felt any discomfort. Thus the length of the survey and
the permissions to stop participating likely added to incomplete responses.
CONCLUSION AND FUTURE DIRECTIONS
There is some preliminary indication that playing video games in the military
may help inoculate soldiers against the negative effects (i.e., severe nightmares) of
deployment. This result is supported by a report from the Offices of the Command
Surgeon and of the Surgeon General (OCS/OSG, 2009) of the U.S. Army regarding
troops deployed to Afghanistan. A survey of the off-duty activities of soldiers was
conducted in order to understand resilience while deployed. Included in the survey
were questions about surfing the Internet, listening to music, physical training,
reading, and about the number of hours spent playing video games. This report
indicates that moderate video game play was associated with fewer psychological
problems. In their study, moderate game play was defined by hours per day (three
to four hours) and did not allow for frequencies other than daily. Thus, the
investigators found that fewer hours per day (one hour) and more hours per day
(six hours) were associated with psychological problems. Our study, on the other
hand, asked about length and frequency of play as well as game play-related
behaviors. These high-end gamers played 2 to 4 hours at a time (daily and weekly
frequencies) while the low-end group played from less than an hour to 2 hours
(rarer frequencies). Thus the data from this study is supported by that of the
OCS/OSG study.
The next research question, if these results are duplicated, is what type of
games are best for nightmare inoculation. While we gathered game play prefer-
ences and recent games played, this data was either long term or just prior to filling
out the questionnaire. We did not have day before the dream game play informa-
tion. However, gamers typically have favorite genres and games they are currently
playing over several days or weeks. While there are big differences between specific
FPS and role playing battle type games, what is the same, and relevant to the
current inquiry, is the learned response of fighting back in the face of threat in
imaginal/virtual worlds. This is common to the traditional “hard core” genre but not
as much true of the rapidly growing casual genre (Gackenbach & Bown, 2011b).
Additionally, one informant from this study wrote the first author and is now
keeping a detailed dream/game diary in a follow-up study. Following a request from
the researcher to rate each dream as to its similarity to the game played the day
before, he responded, “I’m assuming you just want to measure how ‘game-like’ the
dream was, rather than how specifically like a single game it was. Since usually,
there is a mix of several games and I often play more than one game in a day”
(Research Participant, personal communication June 16, 2011). Indeed this dream
from him illustrates the multimedia nature of at least high-end gamer dreams:
Someone stole my pistol, so I went searching through the streets of Adamstown for them.
Found someone (bad guys) in a van. Took a photo of them but they stole my iPhone. Ended
up with my old Nokia phone and took a photo with that and then ran through all these
backyards while they chased me to try and get this phone too, but I threw them off (good
Video Games as Nightmare Protection 243
feelings here . . . like the old days: “no one can catch me”). Was going to tell the cops that
had a station in this huge concrete bunker (awesome bunker: had big protruding chunks of
concrete that stuck out so the it was a little like a modern ”star“ shaped fortress), but then
I realized I could just call NCIS and get them to help out. (Research Participant, personal
communication May 20, 2011)
Not only is a game element mentioned (Modern Star Shaped Fortress) but so
too his phone (iPhone) and a TV show (NCIS). In multiple studies our group has
found that high-end gamers are high media consumers (Gackenbach, in press).
Thus, narrowing down which game or game element to play in order to be
inoculated against nightmares will be a very complex process.
REFERENCES
Anderson, C. A., & Dill, K. E. (2000). Video games and aggressive thoughts, feelings, and behavior in
the laboratory and in life. Journal of Personality and Social Psychology, 78, 772–790. doi:10.1037/
0022-3514.78.4.772
Ashcraft, B. (2011, February, 13, 2011). Video games on the front lines. Message posted to http://
www.kotaku.com.au/2011/02/video-games-on-the-front-lines/
Barrett, D. (2001). Trauma and dreams. Cambridge, MA: Harvard University Press.
Bushman, B. J., & Anderson, C. A. (2009). Comfortably numb: Desensitizing effects of violent media on
helping others. Psychological Science, 20, 273–277. doi:10.1111/j.1467-9280.2009.02287.x
Busink, R., & Kuiken, D. (1996). Identifying types of impactful dreams: A replication. Dreaming, 6,
97–119.
Davis, J. L., Byrd, P., Rhudy, J. L., & Wright, D. C. (2007). Characteristics of chronic nightmares in a
trauma-exposed treatment-seeking sample. Dreaming, 17, 187–198. doi:10.1037/1053-0797.17.4.187
Eng, T. C., Kuiken, D., Temme, K., & Sharma, R. (2005). Navigating the emotional complexities of two
cultures: Bicultural competence, feeling expression, and feeling change in dreams. Journal of
Cultural and Evolutionary Psychology, 3, 267–285.
Ferguson, C. J. (2009). Violent video games: Dogma, fear and pseudoscience. Skeptical Inquirer, 33,
38–43.
Gackenbach, J. I. (2006). Video game play and lucid dreams: Implications for the development of
consciousness. Dreaming, 16, 96–110.
Gackenbach, J. I. (2009). Electronic media and lucid-control dreams: Morning after reports. Dream-
ing, 19, 1–6. doi:10.1037/a0013781
Gackenbach, J. I. (in press). Video game play and dreams. In D. Barrett & P. McNamara (Eds.),
Encyclopedia of sleep and dreams. Santa Barbara, CA: ABC-CLIO.
Gackenbach, J. I., & Bown, J. (2011a). Mindfulness and video game play: A preliminary inquiry. New
York, NY: Springer. doi:10.1007/s12671-011-0049-2
Gackenbach, J. I., & Bown, J. (2011b). Video game presence as a function of genre. Loading. 5.
Retrieved May 12, 2011 from http://journals.sfu.ca/loading/index.php/loading/index
Gackenbach, J. I., & Dopko, R. (2011). The relationship between video game play, dream bizarreness,
and creativity. Unpublished manuscript.
Gackenbach, J. I., & Kuruvilla, B. (2008). The relationship between video game play and threat
simulation dreams. Dreaming, 18, 236–256. doi:10.1037/a0013782
Gackenbach, J. I., & Kuruvilla, B. (2008). Video game play effects on dreams: Self-evaluation and
content analysis. Eludamos. Journal for Computer Game Culture, 2, 169–186. Retrieved from
http://www.eludamos.org/index.php/eludamos/article/view/53/87
Gackenbach, J. I., Kuruvilla, B., & Dopko, R. (2009). Video game play and dream bizarreness.
Dreaming, 19, 218–231. doi:10.1037/a0018145
Gackenbach, J. I., Kuruvilla, B., Dopko, R., & Le, H. (2010). Chapter 5: Dreams and video game play.
In A. Soria & J. Maldonado (Eds.), Computer games: Learning objectives, cognitive performance
and effects on development, (pp. 127–136). Hauppauge, NY: Nova Science Publishers.
Gackenbach, J. I., Matty, I., Kuruvilla, B., Samaha, A. N., Zederayko, A., Olischefski, J., & Von
Stackelberg, H. (2009). Video game play: Waking and dreaming consciousness. In S. Krippner
(Ed.), Perchance to dream (pp. 239–253). Hauppauge, NY: Nova Science Publishers.
Gackenbach, J. I., & Rosie, M. (2009, May). Cognitive evaluation of video games: Players perceptions.
Poster presented at Future Play 2009, Vancouver, BC.
Gackenbach, J. I., Rosie, M., Bown, J., & Sample, T. (2011). Dream incorporation of video game play
as a function of interactivity and fidelity. Dreaming, 21, 32⫺50. doi:10.1037/a0022868
244 Gackenbach, Ellerman, and Hall
Hartmann, E. (1984). The nightmare: The psychology and biology of terrifying dreams. New York, NY:
Basic Books.
Hoge, C. W., Auchterlonie, J. L., & Milliken, C. S. (2006). Mental health problems, use of mental health
services, and attrition from military service after returning from deployment to Iraq or Afghanistan.
Journal of the American Medical Association, 295, 1023–1032. doi:10.1001/jama.295.9.1023
Kahan, T. L., & LaBerge, S. (1996). Cognition and metacognition in dreaming and waking: Comparisons
of first and third-person ratings. Dreaming, 6, 235–249.
Kuiken, D. (2009, June). Criteria for the classification of impactful dreams: Existential, nightmares, and
transcendent dreams. Paper presented at the Annual Conference of the International Association
for the Study of Dreams, Chicago, IL.
Le, H., & Gackenbach, J. I. (2009, June). Nightmares of video game players: What do they look like?
Paper presented at the Annual Meeting of the International Association for the Study of Dreams,
Chicago, IL.
Lee, M. (2010). Reflective awareness in dreams following loss and trauma. Unpublished Ph. D., Univer-
sity of Alberta, Edmonton, AB.
Levin, R., & Nielsen, T. A. (2007). Disturbed dreaming, posttraumatic stress disorder, and affect
distress: A review and neurocognitive model. Psychological Bulletin, 133, 482–528.
Levin, R., & Nielsen, T. A. (2009). Nightmares, bad dreams, and emotional dysregulation: A review and
new neurocognitive model of dreaming. Current Directions in Psychological Science, 18, 84–88.
doi:10.1111/j.1467-8721.2009.01614.x
Orsillo, S. M., Theodore-Oklota, C., Luterek, J. A., & Plumb, J. (2007). The development and psycho-
metric evaluation of the emotional reactivity and numbing scale. The Journal of Nervous and Mental
Disease, 195, 830–836. doi:10.1097/NMD.0b013e318156816f
Phelps, A. J., Forbes, D., & Creamer, M. (2007). Understanding posttraumatic nightmares: An empirical
and conceptual review. Clinical Psychology Review, 28, 338–355. doi:10.1016/j.cpr.2007.06.001
Revonsuo, A. (2000). The reinterpretation of dreams: An evolutionary hypothesis of the function of
dreaming. Behavioural and Brain Sciences, 23, 793–1121. doi:10.1017/S0140525X00004015
Revonsuo, A. (2006). Inner presence: Consciousness as a biological phenomenon. Cambridge: MIT Press.
Revonsuo, A., & Valli, K. (2000). Dreaming and consciousness: Testing the threat simulation theory of
the function of dreaming. Psyche: An Interdisciplinary Journal of Research on Consciousness, 6.
Retrieved from http://ezproxy.macewan.ca/login?url⫽http://search.ebscohost.com/login.aspx?
direct⫽true&db⫽psyh&AN⫽2001–03396-001&site⫽ehost-live&scope⫽site
Smith, T. C., Ryan, M. A. K., Wingard, D. L., Slymen, D. J., Sallis, J. F., & Kritz-Silverstein, D. (2008).
New onset and persistent symptoms of post-traumatic stress disorder self reported after deployment
and combat exposures: Prospective population based US military cohort study. British Medical
Journal, 336, 366–371. doi:10.1136/bmj.39430.638241.AE
Wiederhold, B. K., & Wiederhold, M. D. (2008). Virtual reality for posttraumatic stress disorder and
stress inoculation training. Journal of CyberTherapy & Rehabilitation, 1, 23–35.
Wilmer, H. (1996). The healing nightmare: War dreams of Vietnam veterans. In D. Barrett (Ed.),
Trauma and dreams (pp. 85–99). Cambridge, MA: Harvard University Press.
Zadra, A., Pilon, M., & Donderi, D. C. (2006). Variety and intensity of emotions in nightmares and
bad dreams. Journal of Nervous and Mental Disease, 194, 249–254. doi:10.1097/01.nmd
.0000207359.46223.dc
Zadra, A. (2010, June). Dream Content, Waking States and Well-Being: Why Dreaming is Psycholog-
ically Meaningful, Keynote address at the annual meeting of the International Association for the
Study of Dreams, Ashville, NC.
Video Games as Nightmare Protection 245