Personality and Individual Differences 51 (2011) 117–122

Contents lists available at ScienceDirect

Personality and Individual Differences

journal homepage: www.elsevier.com/locate/paid

Habitual computer game playing at night is related to depressive symptoms

Sakari Lemola a,b,⇑, Serge Brand c, Nicole Vogler a, Nadine Perkinson-Gloor a, Mathias Allemand d,
Alexander Grob a
a
University of Basel, Department of Psychology, Basel, Switzerland
b
University of Helsinki, Faculty of Behavioural Sciences, Helsinki, Finland
c
Psychiatric Hospital of the University of Basel, Basel, Switzerland
d
University of Zürich, Department of Psychology, Zürich, Switzerland

a r t i c l e i n f o a b s t r a c t

Article history: This study investigated whether the amount and circadian time of habitual computer game playing were
Received 20 September 2010 related to depressive symptoms in adolescents and young adults. We expected that habitual late playing
Received in revised form 9 March 2011 relates to more depressive symptoms beyond the effect of the total time of computer game playing as
Accepted 16 March 2011
playing at night may involve short, irregular, and disturbed sleep as well as misalignment of the circadian
Available online 8 April 2011
rhythm. 646 adolescents and young adults (ages 13–30; 90.9% males) who play the internet role-playing
game World of Warcraft completed an online questionnaire. Habitual computer game playing between
Keywords:
10 pm and 6 am was related to an increased risk of high depression scores independent of the total
Depression
Computer game playing
amount of playing. Adolescents (ages 13–17 years) were most vulnerable when habitually playing during
Circadian rhythms early night (i.e., 10–12 pm), while emergent adults (ages 18–22 years) showed more vulnerability when
Sleep duration habitually playing late at night (i.e., after 2 am). The effect was partly mediated by daytime sleepiness but
Sleep problems not by sleep loss or insomnia problems.
Daytime sleepiness ! 2011 Elsevier Ltd. All rights reserved.
Adolescence
Young adulthood

1. Introduction improve when the circadian rhythm gets normally entrained to

Short, irregular, and disturbed sleep, as well as problems of cir-
cadian rhythmicity are related to poor adjustment particularly dur-
ing adolescence and young adulthood (Roberts, Roberts, & Duong,
2008; Shirayama et al., 2003; Wolfson & Carskadon, 1998). High
school students with short sleep on school-nights reported in-
creased levels of depressed mood and daytime sleepiness in com-
parison to individuals who sleep longer (Wolfson & Carskadon,
1998). Further, late bedtimes were associated with higher inci-
dence of depressive symptoms and suicidal ideation (Gangwisch
et al., 2010). In a related vein, individuals with delayed sleep phase
syndrome are at increased risk for mental health problems (Okawa
& Uchiyama, 2007; Shirayama et al., 2003). These sequels may con-
stitute a vicious circle and perpetuate sleeping difficulties and
desynchronization of the circadian rhythm. Indicative of a causal
mechanism, studies show mental health and well-being to
Abbreviations: WoW, World of Warcraft; MMORPG, Massively Multiplayer
Online Role-Playing Game; ADS-K, short version of the ‘Allgemeine
Depressionsskala’.
⇑ Corresponding author. Address: Missionsstrasse 62a, 4055 Basel, Switzerland.
Tel.: +41 61 267 0638; fax: +41 61 267 0661.
E-mail address: sakari.lemola@unibas.ch (S. Lemola).
the exogenous day–night cycle by oral melatonin treatment (Rah-
man, Kayumov, & Shapiro, 2010).
The present study aims at examining the interplay of computer
game playing during the night, sleep patterns, and depressive
symptoms. According to every day experience of parents and
teachers excessive computer game playing and in particular habit-
ual playing late at night have a detrimental impact on psychologic
functioning. Moreover, psychologists and pediatricians are nowa-
days often confronted with children and adolescents suffering from
addictive computer game playing. In a representative study of
more than 15,000 9th graders in Germany, 15.8% of the adolescent
boys and 4.3% of the girls play computer games for more than 4.5 h
per day and 3% of the boys and 0.3% of the girls can be considered
as addicted to computer games according to ICD-10 derived crite-
ria (Rehbein, Kleimann, & Mössle, 2010). The computer game with
the highest addiction risk was the Massively Multiplayer Online
Role-Playing Game (MMORPG) World of Warcraft (WoW).
MMORPGs are competitive role-playing games normally played
in cooperative groups. These groups gather for online-playing
appointments and, as the success of the whole group depends on
the collaboration of all group members, individual players are
pressured to join the gatherings, which may explain the addic-
tive potential (Bessière, Seay, & Kiesler, 2007). MMORPGs are

0191-8869/$ - see front matter ! 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.paid.2011.03.024
118 S. Lemola et al. / Personality and Individual Differences 51 (2011) 117–122
exceptional from other computer games as many MMORPG-play-
ers typically develop a high commitment toward their virtual char-
acter, which they often keep for months or even years, and toward
their cooperative group (Bessière et al., 2007). WoW is played
world wide by more than 11 million people (Cavalli, 2008) and
was indicated as the favorite computer game by nearly 10% of male
and 1.3% of female students in Germany. On average, these stu-
dents indicated that they played WoW for 230 min per day
(Rehbein et al., 2010).
Scientific research on the impact of computer game playing on
well-being and adjustment is still scarce and mixed (cf. Cain &
Gradisar, 2010). Existing evidence suggests that excessive com-
puter game playing can be a risk factor for poor adjustment (Wea-
ver et al., 2009). One mechanism by which playing computer
games – and in particular MMORPGs – may relate to poor adjust-
ment, is that many people play the game during nights which
may lead to sleep loss, sleep disturbances, and delayed sleep phase.
Griffiths, Davies, and Chappell (2004) for instance, found that
nearly one in five players of the internet role-playing game Ever-
quest sacrificed sleep in order to play the game. Generally, time
spent playing computer games associates with later bed times
and fewer sleep hours in children and adolescents (Fuligni & Hard-
way, 2006; Van den Bulck, 2004). Moreover, playing computer
games before bedtime can lead to higher activity levels of the cen-
tral and autonomic nervous system and may result in increased
sleep latency and decreases in subjective sleepiness (Higuchi, Mot-
ohashi, Liu, & Maeda, 2005). A representative finish study indicated
that intensive computer usage had a negative impact on perceived
health only if it co-occurred with deteriorated sleep habits and
waking-time tiredness (Punamäki, Wallenius, Nygard, Saarni, &
Rimpela, 2007).
In the present study we tested whether habitual computer
game playing during the night is related to depressive symptoms
in adolescents and young adults who play WoW. We expected that
habitual late playing relates to more depressive symptoms beyond
the effect of the total time of computer game playing as playing at
night may involve short, irregular, and disturbed sleep as well as
misalignment of the circadian rhythm. Moreover, we tested
whether the relation between habitual late playing and depressive
symptoms is mediated by short sleep duration, sleep problems,
and/or daytime sleepiness. Thus, the present study is thought to fill
a gap in the literature as it addresses the interplay of excessive
computer game playing, sleep patterns, and mental health in the
already quite large group of adolescents and young adults who
are committed to internet role-playing games. As these hypothe-
sized relations may differ between individuals from different
developmental stages, the questions were also addressed sepa-
rately among adolescents (ages 13–17), emerging adults (ages
18–22), and young adults (ages 23–30).
2. Method
2.1. Procedure
Participants were recruited during winter 2009–2010 by con-
tacting them through the internet role-playing game WoW and
connected forums. Before giving consent, participants were in-
formed on the purpose of the study, assured that participation
was completely anonymous and voluntary, and that any data col-
lected would be treated in strictest confidence. Particularly, partic-
ipants were assured that no personal data were communicated to
any other player, nor were IP-addresses of participants accessible
to the researchers. Moreover, participants could stop or withdraw
from the study without giving any further explication. The entire
study has been carried out in accordance to the ethical standards
for research involving human subjects required by the University
of Basel and laid down in the Declaration of Helsinki (World Med-
ical Association, 2008, online at www.wma.net).
2.2. Participants
The sample consisted of 646 WoW players. Of these, 90.9% were
male (n = 587), 9.0% were female (n = 58); one participant did not
indicate the gender. Participants were aged between 13 and
30 years, with a mean age of 19.7 years (SD = 3.6 years). To address
the research questions within groups of the same developmental
stage, the sample was subdivided into a group of adolescents (ages
13–17, n = 190), emerging adults (ages 18–22, n = 347), and young
adults (ages 23–30, n = 129). Accordingly, participants outside
these age groups were excluded, i.e., participants younger than
13 years (n = 6), or older than 30 years (ages 31–45, n = 44), as well
as participants who did not indicate their age (n = 38). Of the par-
ticipants completing the questionnaire, 84.8% (n = 548) indicated
Germany as their country of residence, 6.5% (n = 42) were from
Austria, and 6.5% (n = 42) were from Switzerland; the remaining
2.2% (n = 14) were from other countries.
2.3. Measures
The online questionnaire was designed with the online survey
application software Survey Monkey (www.SurveyMonkey.com,
Portland, OR). The questionnaire was divided into three sections,
i.e., (1) demographic background, (2) playing habits such as
amount and the usual time of day of playing WoW, and (3) depres-
sive symptoms and sleep duration, sleeping problems, and daytime
sleepiness. On average participants completed the questionnaire in
16 min (SD = 6 min).
2.3.1. Computer game playing
Participants indicated how many hours in total they played
WoW during a typical week. Further, they indicated how often in
a typical week they played WoW during seven different time slots
(6–12 am, 12 am–6 pm, 6–8 pm, 8–10 pm, 10–12 pm, 12 pm–
02 am, and 02–06 am) on a 4-point scale with the categories ‘less
than once/week’, ‘on 1–2 days/week’, ‘on 3–4 days/week’, and ‘on
5–7 days/week’.
2.3.2. Sleep duration, sleep problems, and daytime sleepiness
To assess sleep duration and problems, items were taken and
modified from a widely used manual on coping with sleep distur-
bances (Backhaus & Riemann, 1996) and from the Insomnia Sever-
ity Index (ISI; Bastien, Vallières, & Morin, 2001). Participants were
asked how much sleep they usually get (a) on weekdays or work-
days, and (b) on weekends or non-workdays in hours. Sleep prob-
lems were measured with three items on sleep problems (‘‘Have
trouble falling asleep,’’ ‘‘Wake up during the night and have diffi-
culty going back to sleep,’’ and ‘‘Wake up too early in the morning
and to be unable to get back to sleep’’) using a 5-point scale rang-
ing from 1 (never) to 5 (almost always) and one item on the general
quality of sleep during the past week. The four items were highly
interrelated and were combined to build a sleep problems score
with a Cronbach’s alpha of 0.65. Higher mean scores reflect higher
levels of sleep problems. Daytime sleepiness was measured with
the Epworth Sleepiness Scale (Johns, 1991), which includes eight
items and assesses the general level of daytime sleepiness by ask-
ing participants the chance of falling asleep/dozing in different dai-
ly situations with a Cronbach’s alpha of 0.65. Higher mean scores
reflect an increased daytime sleepiness.
 S. Lemola et al. / Personality and Individual Differences 51 (2011) 117–122 119

2.3.3. Depressive symptoms Table 2

Depressive symptoms were assessed using the short version of Descriptives of study variables stratified by age groups.

the ‘Allgemeine Depressionsskala’ (ADS-K; Hautzinger & Bailer, Ages 13–17 Ages 18–22 Ages 23–30 P (age
1993; English original: ‘Center for Epidemiological Studies Depres- (n = 190) (n = 327) (n = 129) trend)
sion Scale’; CES-D; Radloff, 1977) containing 15 items that address M (SD) M (SD) M (SD)
various aspects of depressive symptoms during the past 7 days, Age 15.9 (1.0) 19.6 (1.4) 25.4 (2.2) –
which were rated on a 4-point Likert-type scale ranging from 0 (oc- Sleep duration on 7.4 (1.1) 7.1 (1.2) 7.3 (1.2) 0.35
curred never or rarely) to 3 (occurred most of the time or always). weekdays (h)
For binary analyses the sum score was dichotomized: A total score Sleep duration on 9.5 (1.6) 9.1 (1.6) 8.7 (1.4) <0.001
weekends (h)
of 18 or more points reflected a clinically relevant depressive syn-
Sleep problems 4.3 (2.6) 4.3 (2.6) 4.4 (3.0) 0.80
drome (Hautzinger & Bailer, 1993). Cronbach’s alpha for the ADS-K Sleepiness 6.9 (3.5) 7.0 (3.7) 7.5 (3.3) 0.15
scale was 0.89. Depression 10.6 (8.2) 10.6 (8.1) 8.5 (6.6) 0.03
Playing duration 24.1 (20.4) 22.4 (15.3) 19.4 (12.8) 0.03
per week (h)
3. Results

3.1. Descriptive statistics

On average participants were playing WoW for 22.1 h per week
(SD = 15.4; MD = 20 h/week; range: 1–130 h; for parametric analy-
ses the values of 7 participants (1.1%) who indicated to play more
than 80 h per week were restricted to the value of 80 h) which
equals an average daily playing duration of 189 min. Comparisons
between the three age groups revealed that playing time slightly
decreased with age (linear trend: F(2/643) = 4.5, P = 0.03; post
hoc Dunnett’s test comparing adolescents with emergent adults:
P = 0.73; and adolescents with young adults: P = 0.05).
Table 1 indicates how often participants in three age groups
played during different time slots throughout the day. In each
age group the most preferred time for playing was between 6
and 8 pm and between 8 and 10 pm. Among adolescents 18.4%
played on five or more days per week during night time (i.e., be-
tween 10 pm and 6 am), while among emerging adults there were
29.4%, and among young adults there were 23.6% who played five
or more days per week at that time. In each group only few partic-
ipants indicated to habitually play after 2 am. Moreover, habitual
playing in mornings (between 6 and 12 am) was rare.
Table 2 indicates means and standard deviations of the study
variables stratified by age groups. Sleep duration during weekday
nights was 7.23 h (SD = 1.22) which was significantly shorter than
on weekend nights (M = 9.17 h, SD = 1.59); F(1/641) = 967.8,
P < 0.001. The participants had an average score of 10.2 on the
ADS-K depression scale (SD = 7.9) which is in a similar range as
in ADS-K population samples (e.g., Hautzinger & Bailer, 1993,
M = 10.7, SD = 8.0, Cohen’s d = !0.06; Lehr, Hillert, Schmitz, &
Sosnowsky, 2008, M = 9.4, SD = 7.8, Cohen’s d = 0.10).
3.2. Logistic regression analyses
Multiple logistic regression analyses adjusting for gender, age,
sleep duration on weekdays and on weekends, and playing dura-
tion per week were conducted to predict the risk of high scores
on the ADS-K depression scale associated with different habitual
circadian playing times. In analyses on the whole sample, in-
creased risk of depression scores was associated with habitually
playing at night (i.e., between 10 and 12 pm: OR = 1.32, 95%
CI = 1.05–1.67, P = 0.02; between 12 pm and 2 am: OR = 1.33, 95%
CI = 1.04–1.69, P = 0.02; and between 2 and 6 am: OR = 1.49, 95%
CI = 1.07–2.08, P = 0.02), while habitually playing during the day
(i.e., 06–12 am, 12 am–06 pm, 06–08 pm, 08–10 pm) was not sig-
nificantly associated with risk of increased depression scores (all

Table 1
Habitual circadian playing times in three age groups.

n (%)
<1 day/week 1–2 days/week 3–4 days/week 5–7 days/week
Ages 13–17 (n = 190)
06–12 am 157 (82.6) 29 (15.3) 1 (0.5) 3 (1.6)
12 am-06 pm 47 (24.7) 44 (23.2) 49 (25.8) 50 (26.3)
06–08 pm 26 (13.7) 26 (13.7) 61 (32.1) 77 (40.5)
08–10 pm 17 (8.9) 46 (24.2) 52 (27.4) 75 (39.5)
10–12 pm 50 (26.3) 70 (36.8) 35 (18.4) 35 (18.4)
12 pm–02 am 104 (54.7) 69 (36.3) 9 (4.7) 8 (4.2)
02–06 am 156 (82.1) 28 (14.7) 2 (1.1) 4 (2.1)
Ages 18–22 (n = 327)
06–12 am 274 (83.8) 39 (11.9) 10 (3.1) 4 (1.2)
12 am-06 pm 109 (33.3) 101 (30.9) 71 (21.7) 46 (14.1)
06–08 pm 38 (11.6) 63 (19.3) 109 (33.3) 117 (35.8)
08–10 pm 29 (8.9) 52 (15.9) 113 (34.6) 133 (40.7)
10–12 pm 56 (17.1) 90 (27.5) 87 (26.6) 94 (28.7)
12 pm–02 am 157 (48.0) 109 (33.3) 26 (8.0) 35 (10.7)
02–06 am 265 (81.0) 45 (13.8) 11 (3.4) 6 (1.8)
Ages 23–30 (n = 129)
06–12 am 103 (79.8) 23 (17.8) 2 (1.6) 1 (0.8)
12 am-06 pm 56 (43.4) 46 (35.7) 16 (12.4) 11 (8.5)
06–08 pm 26 (20.2) 31 (24.0) 36 (27.9) 36 (27.9)
08–10 pm 7 (5.4) 33 (25.6) 42 (32.6) 47 (36.4)
10–12 pm 31 (24.0) 45 (34.9) 25 (19.4) 28 (21.7)
12 pm–02 am 71 (55.0) 38 (29.5) 13 (10.1) 7 (5.4)
02–06 am 102 (79.1) 24 (18.6) 1 (0.8) 2 (1.6)
120 S. Lemola et al. / Personality and Individual Differences 51 (2011) 117–122
P-values > 0.11). Among adolescents, increased risk of depression
scores was associated with habitually playing between 10 and
12 pm (OR = 1.65 (95% CI = 1.05–2.60) P = 0.03), while habitually
playing at other times was not significantly associated with risk
of increased depression scores (all P-values > 0.06). Among emerg-
ing adults, increased risk of depression scores was associated with
habitually playing between 2 and 6 am (OR = 1.60 (95% CI = 1.02–
2.50) P = 0.04), while habitually playing during earlier time slots
was not significantly associated with risk of increased depression
scores (all P-values >0.20). Among young adults multiple logistic
regression revealed no significant increase in the risk of high
depression scores by habitual late playing (for all time slots, P-val-
ues >0.16).
3.3. Analysis of circadian playing time stratified by total playing
duration
Figure 1 displays mean values of the ADS-K depression scale in
the total sample and separately for the three age groups stratified
by the median of total playing duration per week (20 h/week) and
by the habitual time of playing during the day. Participants who
habitually played after 10 pm (i.e., participants who played at least
in one of the time slots 10–12 pm, 12 pm–02 am, 02–06 am for 5 or
more days per week; n = 161, 24.9% of the total sample) showed
higher depression scores than their counterparts who played less
often after 10 pm (total sample: F(1/644) = 12.8; P < 0.001, partial
g2 = 0.02; adolescents: F(1/186) = 3.71; P = 0.06, partial g2 = 0.02;
emerging adults: F(1/323) = 3.70; P = 0.05, partial g2 = 0.01; young
adults: F(1/123) = 10.12; P = 0.002, partial g2 = 0.08) while there
was no main effect of the total playing time (i.e., playing less than
20 h/week vs. playing 20 h/week or more) and no interaction of the
Fig. 1. Mean scores (95% CI) of depressive symptoms (ADS-K) stratified by groups with different total amounts of playing per week (1–19 h/week vs P20 h/week) and
different habitual circadian playing time (playing on <5 days/week after 10 pm vs playing on P5 days/week after 10 pm).
total playing time and habitual playing during the night (all
Ps > 0.20). Analyses with log-transformed depressive symptom
scores accounting for non-normality revealed equal results in
terms of direction, size, and significance of the coefficients and
are not presented.
3.4. Structural model
Finally, structural equation modeling was used to test whether
the relation between habitual playing during the night and depres-
sive symptoms was mediated by sleep duration on weekday and
weekend nights, sleep problems, or daytime sleepiness controlling
for age, gender, and total playing time per week (Fig. 2). The model
fit was acceptable (v2(19) = 22.0, P = 0.29; v2/df = 1.16;
CFI = 0.996; RMSEA = 0.016). Multi-group comparisons revealed
that constraining measurement and structural weights to be
invariant across the three age groups led to only marginally re-
duced model fit (v2(26) = 38.9, P = 0.05), indicating only marginal
differences between the age groups. Habitual playing during the
night modeled as a latent variable with the indicators ‘playing be-
tween 10 and 12 pm’ and ‘playing between 12 pm and 2 am’ pre-
dicted depressive symptoms. Constraining only the structural
weight between habitual playing during nights and depressive
symptoms to be equal across the age groups did not lead to a sig-
nificant reduction in model fit (v2(2) = 0.5, P = 0.76). Habitual play-
ing during the night was also predictive of short sleep on weekday
and weekend nights specifically in the youngest age group, and re-
lated to more daytime sleepiness. In contrast to expectations, in all
age groups habitual playing during the night was not predictive of
more sleep problems. Sleep problems were a moderately strong
predictor of depressive symptoms, while sleep duration on
 S. Lemola et al. / Personality and Individual Differences 51 (2011) 117–122 121

Sleep
problems

.43***
Sleep duration .45***; .42***; .49***
-.07 weekday
-.04; -.07; -.11

-.02
Sleep duration -.02; .02; -.14†
10-12 00-02 -.17*** weekend
-.27***; -.16*; .05
pm am
.01
-.11* Daytime .04; .00; .00
.79 .73 -.27**; -.04; -.10
sleepiness
.74; .79; .78 .74; .70; .79
.09* .08*
.06; .11*; .00
.08; .05; .22*
Habitual Depressive
.20***
playing at .22*; .18*; .31** symptoms
night
.03
.01; .02; .15

Total playing
.57*** -.12***
.52***; .64***; .59*** duration/week
-.01; -.10†; -.01

Age
-.02
.02; -.05; -.03

Gender

Fig. 2. Prediction of depressive symptoms by total amount of computer game playing per week, habitual computer game playing at night, sleep problems, sleep duration,
daytime sleepiness, and background variables: Structural equation modeling (v2(19) = 22.0, P = 0.29 v2/df = 1.16; CFI = 0.996; RMSEA = 0.016). Indices in the second line
represent coefficients in the age-subgroups (i.e., adolescents, aged 13–17 years; emergent adults, aged 18–22 years; young adults, aged 23–30 years). !P < 0.10; ⁄P < 0.05;
⁄⁄
P < 0.01; ⁄⁄⁄P < 0.001.

weekday and weekend nights was not related to depressive symp-
toms. More daytime sleepiness was modestly but significantly re-
lated to higher depressive symptom scores, specifically among
emergent adults. Again, models with log-transformed depressive
symptom scores revealed equal results in terms of direction, size,
and significance of the coefficients and are not presented.
4. Discussion
We found that habitual computer game playing between 10 pm
and 6 am was related to an increased risk of depression scores in
adolescents and emergent adult players of the internet role-play-
ing game WoW. While adolescents appeared to be particularly vul-
nerable to playing between 10 and 12 pm, emergent adults were
particularly at risk for depression when habitually playing after
2 am. Importantly, the effect of the habitual playing time was inde-
pendent of the total time of computer game playing and thus ap-
pears to be rather an effect of the circadian timing of computer
game playing than an effect of the actual total amount of hours
which adolescents and emergent adults spend playing. Interest-
ingly, the average depressive symptom score of the sample in the
present study was not higher than in population samples (Hautzin-
ger & Bailer, 1993; Lehr et al., 2008), which is in contrast to the
commonly held assumption that adolescents and young adults
who excessively play computer games are rather maladjusted. In
sum, this is the first time that a relation between the circadian tim-
ing of excessive computer game playing and negative outcomes
such as depressive symptoms could be shown.
Our results are congruent with research showing that late bed
times are related to a higher risk of depression and maladjustment
(Gangwisch et al., 2010; Okawa & Uchiyama, 2007; Shirayama
et al., 2003). A possible explanation of the pattern that adolescents
were more vulnerable to playing during early night is that there
were only very few adolescents who habitually played later than
12 pm, as possibly for many of them parents still set their bedtimes
during school-weeks. Adolescents may also be more vulnerable for
the effect of playing between 10 and 12 pm than the older age
groups. Moreover, our finding of an indirect effect of habitual late
playing on depressive symptoms via increased daytime sleepiness
is in agreement with research indicating that the effect of com-
puter game playing on adjustment is mediated by daytime tired-
ness (Punamäki et al., 2007).
According to generally held assumptions, late night activities
bear the risk of sleep loss and development of sleep disturbances,
which lead to poor adjustment. However and contrary to our
expectations, we could not find mediation of the effect of habitual
late playing by short sleep duration or sleep problems. In fact, sleep
problems were unrelated to habitual late playing and sleep dura-
tion was not predictive of depressive symptoms in the multivariate
structural equation modeling approach. The observation that the
mere sleep duration was not associated with depression is in line
with findings of previous studies with adolescents (e.g., Brand,
Beck, Hatzinger, & Holsboer-Trachsler, 2009). However, the obser-
vation that habitual late playing was not associated with sleep
problems deviates from results of a study by Higuchi et al.
(2005) showing that playing computer games before bedtime
interferes with sleepiness and leads to difficulties falling asleep.
A possible reason for this incongruence is that in the Higuchi
et al. study a shooter game was used in the experimental proce-
dure, which possibly might lead to higher arousal than the internet
role-playing game WoW.
122 S. Lemola et al. / Personality and Individual Differences 51 (2011) 117–122
The relation between habitual late playing and depressive
symptoms which was not mediated by sleep loss or insomnia
problems might be due to misalignment of the individual circadian
rhythm with the exogenous day-night cycle. This interpretation is
in agreement with a line of research showing that delayed sleep
phase is associated with poor adjustment, depression, and lower
estimates of quality of life (Gangwisch et al., 2010; Okawa & Uchiy-
ama, 2007; Shirayama et al., 2003) which in turn can be improved
by entrainment of the individual circadian rhythm with the exog-
enous day-light cycle by oral administration of melatonin (Rahman
et al., 2010) or morning bright light exposure (Okawa & Uchiyama,
2007).
Due to the cross-sectional design of the study it is not possible
to draw causal inferences. Hence, also alternative interpretations
are consistent with the data. For instance, it is possible that depres-
sive symptoms lead to delay of the sleep phase, which in turn may
fuel the preference of playing computer games late at night. More-
over, also a reciprocal relationship between habitual late playing
and depressive symptoms is plausible. Future studies should test
these possibilities.
Further issues also warrant cautious interpretation. First, we
used an online questionnaire to assess sleep duration and sleep
problems while no objective sleep assessment via sleep-EEGs or
actigraphy was applied (although there is evidence that sleep data
from sleep logs and questionnaires do fit well with objective data,
cf. Brand et al., 2010; Werner, Molinari, Guyer, & Jenni, 2008). In a
related vein, the marginal internal consistency of the scales mea-
suring sleep problems and daytime sleepiness has to be noted,
which may have attenuated the reported relationships. Second,
the present study particularly focuses on the group of computer
game players who play WoW, an internet role-playing game which
is known for its high potential for addictive playing (Rehbein et al.,
2010). Thus, our results do not necessarily apply to other groups of
computer game players. Third, the sample consisted mainly of
male WoW players. This is congruent with the generally much
higher prevalence of male MMORPG players in other German sam-
ples (Rehbein et al., 2010). However, the findings can therefore not
be generalized to women. It remains subject of future research to
study the relation of circadian playing time with depressive symp-
toms also in samples containing more females. At last, it has to be
noted that analyses in the different age subgroups are to some de-
gree challenged by differences in statistical power due to unequal
sample sizes.
In summary, we found that habitual playing of the internet role-
playing game WoW at night was related to an increased risk of
depression scores among adolescents and emergent adults inde-
pendent of the total time spent playing. Our results thus extend
previous research on the risks of computer game playing for
adjustment in adolescence and emergent adulthood by emphasiz-
ing the role of the circadian timing of computer game playing.
References
Backhaus, J., & Riemann, D. (1996). Schlafstörungen bewältigen [Coping with sleeping
disorders]. Weinheim: Beltz Psychologie Verlags Union.
Bastien, C. H., Vallières, A., & Morin, C. M. (2001). Validation of the Insomnia
Severity Index (ISI) as an outcome measure for insomnia research. Sleep
Medicine, 2, 297–307.
Bessière, K., Seay, A. F., & Kiesler, S. (2007). The ideal elf: Identity exploration in
world of warcraft. Cyberpsychology & Behavior, 10, 530–535.
Brand, S., Beck, J., Hatzinger, M., & Holsboer-Trachsler, E. (2009). Perceived
parenting styles, personality traits and sleep patterns in adolescents. Journal
of Adolescence, 32, 1189–1207.
Brand, S., Gerber, M., Beck, J., Hatzinger, M., Pühse, U., & Holsboer-Trachsler, E.
(2010). Exercising, sleep-EEG patterns, and psychological functioning are
related among adolescents. World Journal of Biological Psychiatry, 11, 129–140.
Cain, N., & Gradisar, M. (2010). Electronic media use and sleep in school-aged
children and adolescents: A review. Sleep Medicine, 11, 735–742.
Cavalli, E. (2008). World of Warcraft Hits 11.5 Million Users. Wired. <http://
www.wired.com/gamelife/2008/12/world-of-warc-1/> Retrieved 03.08.10.
Fuligni, A. J., & Hardway, C. (2006). Daily variation in adolescents’ sleep, activities,
and psychological well-being. Journal of Research on Adolescence, 16, 353–377.
Gangwisch, J. E., Babiss, L. A., Malaspina, D., Turner, J. B., Zammit, G. K., & Posner, K.
(2010). Earlier parental set bedtimes as a protective factor against depression
and suicidal ideation. Sleep, 33, 97–106.
Griffiths, M. D., Davies, M. N. O., & Chappell, D. (2004). Demographic factors and
playing variables in online computer gaming. Cyberpsychology & Behavior, 7,
479–487.
Hautzinger, M., & Bailer, M. (1993). Allgemeine depressions skala, manual. Göttingen:
Beltz Test GmbH.
Higuchi, S., Motohashi, Y., Liu, Y., & Maeda, A. (2005). Effects of playing a computer
game using a bright display on presleep physiological variables, sleep latency,
slow wave sleep and REM sleep. Journal of Sleep Research, 14, 267–273.
Johns, M. W. (1991). A new method for measuring daytime sleepiness: The Epworth
sleepiness scale. Sleep, 14, 540–545.
Lehr, D., Hillert, A., Schmitz, E., & Sosnowsky, N. (2008). Screening depressiver
störungen mittels allgemeiner depressions-skala (ADS-K) und state-trait
depression scales (STDS-T): Eine vergleichende evaluation von cut-off-werten.
Diagnostica, 54, 61–70.
Okawa, M., & Uchiyama, M. (2007). Circadian rhythm sleep disorders:
Characteristics and entrainment pathology in delayed sleep phase and non-24
sleep–wake syndrome. Sleep Medicine Reviews, 11, 485–496.
Punamäki, R. L., Wallenius, M., Nygard, C. H., Saarni, L., & Rimpela, A. (2007). Use of
information and communication technology (ICT) and perceived health in
adolescence. The role of sleeping habits and waking-time tiredness. Journal of
Adolescence, 30, 569–585.
Radloff, L. S. (1977). The CES-D scale: A self-report depression scale for research in
the general population. Applied Psychological Measurement, 1977(1), 385–401.
Rahman, S. A., Kayumov, L., & Shapiro, C. M. (2010). Antidepressant action of
melatonin in the treatment of Delayed Sleep Phase Syndrome. Sleep Medicine,
11, 131–136.
Rehbein, F., Kleimann, M., & Mössle, T. (2010). Prevalence and risk factors of video
game dependency in adolescence. Results of a German nationwide survey.
Cyberpsychology, Behavior, and Social Networking, 13, 1–9.
Roberts, R. E., Roberts, C. R., & Duong, H. T. (2008). Chronic insomnia and its negative
consequences for health and functioning of adolescents: A 12-month
prospective study. Journal of Adolescent Health, 42, 294–302.
Shirayama, M., Shirayama, Y., Iida, H., Kato, M., Kajimura, N., Watanabe, T., et al.
(2003). The psychological aspects of patients with delayed sleep phase
syndrome (DSPS). Sleep Medicine, 4, 427–433.
Van den Bulck, J. (2004). Television viewing, computer game playing, and internet
use and self-reported time to bed and time out of bed in secondary-school
children. Sleep, 27, 101–104.
Weaver, J. B., Mays, D., Weaver, S. S., Kannenberg, W., Hopkins, G. L., Eroĝlu, D., et al.
(2009). Health-risk correlates of video-game playing among adults. American
Journal of Preventive Medicine, 37, 299–305.
Werner, H., Molinari, L., Guyer, C., & Jenni, O.-G. (2008). Agreement rates between
actigraphy, diary, and questionnaire for children’s sleep patterns. Archives of
Pediatrics & Adolescent Medicine, 162, 350–358.
Wolfson, A. R., & Carskadon, M. A. (1998). Sleep schedules and daytime functioning
in adolescents. Child Development, 69, 875–887.
World Medical Association (2008). World Medical Association Declaration of
Helsinki: Ethical principles for medical research involving human subjects.
6th revision, pp. 1-5.