J Youth Adolescence (2017) 46:884–897
DOI 10.1007/s10964-017-0646-z
EMPIRICAL RESEARCH
Video Gaming and Children’s Psychosocial Wellbeing: A
Longitudinal Study
Adam Lobel1 Rutger C. M. E. Engels2 Lisanne L. Stone3 William J. Burk4
Isabela Granic4
●
●
●
●
Received: 1 November 2016 / Accepted: 9 February 2017 / Published online: 21 February 2017
© The Author(s) 2017; This article is published with open access at Springerlink.com
Abstract The effects of video games on children’s psychosocial development remain the focus of debate. At two
timepoints, 1 year apart, 194 children (7.27–11.43 years
old; male = 98) reported their gaming frequency, and their
tendencies to play violent video games, and to game (a)
cooperatively and (b) competitively; likewise, parents
reported their children’s psychosocial health. Gaming at
time one was associated with increases in emotion problems. Violent gaming was not associated with psychosocial changes. Cooperative gaming was not associated with
changes in prosocial behavior. Finally, competitive gaming
was associated with decreases in prosocial behavior, but
only among children who played video games with high
frequency. Thus, gaming frequency was related to increases
in internalizing but not externalizing, attention, or peer
problems, violent gaming was not associated with increases
in externalizing problems, and for children playing
approximately 8 h or more per week, frequent competitive
gaming may be a risk factor for decreasing prosocial
behavior. We argue that replication is needed and that future
research should better distinguish between different forms
of gaming for more nuanced and generalizable insight.
* Adam Lobel
Adam@AdamLobel.com
1
Swiss Center for Affective Sciences, University of Geneva,
Chemin des Mines 9, Geneva 1202, Switzerland
2
Trimbos Institute, Da Costakade 45, Utrecht 3521VS, Netherlands
3
Overwaal, Centre for Anxiety Disorders, Pro Persona Tarweweg,
Nijmegen 6534AM, Netherlands
4
Behavioural Science Institute, Radboud University,
Montessorilaan, Nijmegen 6525HR, Netherlands
Keywords Psychosocial development Video games
Prosocial behavior Longitudinal
●
●
●
Introduction
Video games have rapidly become a universal aspect of
child development (Lenhart et al. 2008), and their quick rise
to prominence has stimulated scientific inquiry and public
concern (Ferguson 2013). With researchers stressing that
children may be particularly susceptible to the influence of
video game playing (Bushman and Huesmann 2006; Lobel
et al. 2014a), the effects of video games on children’s
psychosocial development remains highly debated. Video
games have thus been widely studied as a potential cause
for aggressive cognitions and behavior (Anderson et al.
2010; Carnagey and Anderson 2004), emotional problems
such as depression (Tortolero et al. 2014), and hyperactivity
and inattention (Gentile et al. 2012). In these lines of
research, video games are seen as a compelling entertainment medium whose clever use of feedback loops and
positive reinforcement schedules train unhealthy habits of
mind (Gentile and Gentile 2008a, b).
Conversely, researchers have recently begun to look at
video games as a domain for training healthy habits of mind
(Adachi and Willoughby 2012; Granic et al. 2014). From
this perspective, many video games reward communication
and cooperation as well as resolving negative emotions such
as frustration. Moreover, video games seem to provide a
context for the fulfillment of self-deterministic needs,
thereby positively contributing to psychological well-being
(Ryan et al. 2006). The current paper adds to the discussion
on gaming’s positive and negative consequences with data
J Youth Adolescence (2017) 46:884–897
from a longitudinal study that could address the relations
between different forms of video game playing and the
psychosocial development of children. Here, psychosocial
development refers broadly to the psychological and social
changes children undergo during development, including
changes in patterns of internalizing and externalizing problems, attention, and how children relate to peers.
Psychosocial Development and Gaming
In a recent review we argued for the potential of video
gaming to afford psychosocial benefits (Granic et al. 2014).
This perspective focuses on gaming as a modern and
meaningful form of play, and therefore as a context where
children’s developmental needs can be met (Fisher 1992;
Verenikina et al. 2003). Just as traditional forms of play
provide positive contexts for children’s psychosocial development (Erikson 1977; Piaget 1962; Vygotsky 1978), so too
video games seem to afford promise (Adachi and Willoughby 2012; Granic et al. 2014). This promise is in part
due to the ubiquity of gaming; with between 90 and 97% of
children playing video games (Lenhart et al. 2008), it seems
that social development has partly migrated from physical
playgrounds to digital ones. Moreover, video games have
become—particularly in the past decade—a more social and
emotionally rich entertainment medium. Thus, modern video
games may provide a context for children to bond with others
and learn the benefits of cooperation.
Yet despite the potential benefits of gaming for children’s
psychosocial development, scant empirical work has
explored these options (Hromek and Roffey 2009; Przybylski and Wang 2016). Instead, there has been a predominant focus on the potential psychosocial dangers of
gaming. A recent meta-analysis identified 101 studies that
investigated the effects of playing (violent) video games on
children’s and adolescents’ psychosocial health. Of these
studies, nearly 70 of them assessed whether (violent) video
games were related to externalizing problems (such as
aggression). In contrast, prosocial behavior (e.g. Gentile
et al. 2009) and internalizing problems (such as depression)
were each assessed in about 20 studies (e.g. Parkes et al.
2013). Just 9 studies assessed the relation between gaming
and attention problems (e.g. Bioulac et al. 2008) and even
fewer investigated the relation between gaming and children’s peer relationships (e.g. Przybylski 2014).
Several methodological shortcomings are also important
to highlight. First, gaming research among children has
predominantly been cross-sectional in nature—64 of the
101 studies identified in Ferguson (2015) were correlational.
The major limitation of these studies is that they do not allow
inferences about order. Moreover, many of these studies have
not controlled for relevant background variables such as
socio-economic status (SES) and gender. On the other hand,
885
while experimental studies allow researchers to draw causal
inferences, the real-world generalizability of such gaming
studies remain debated. Regarding studies on externalizing
problems in particular, researchers have questioned the ecological validity of the outcome measures used (see Anderson
and Bushman 1997; Ritter and Eslea 2005) and the extent to
which these studies used well-matched control conditions
(see Przybylski et al. 2014). Beyond these issues, as most of
these experimental studies were run in a single lab session,
these experiments do not give enough insight into the longterm consequences of playing video games.
Regarding internalizing problems, studies that examine
the link between gaming and emotional problems have
predominantly focused on “problematic gamers.” These are
individuals who habitually play for very many hours and
show other signs of dependency, such as avoiding social
interactions or obligations in favor of gaming (van Rooij
et al. 2011). Among adolescents, such gamers seem to have
elevated depression symptoms compared to their peers
(Messias et al. 2011). A recent, large scale, cross-sectional
study among Canadian adolescents also indicated that video
game play was positively associated with symptoms of
depression and anxiety (Maras et al. 2015). These findings
are consistent with the conclusions made in a review by
Kuss and Griffiths (2012). These problems seem to emerge
as a result of escapism; that is that problematic gamers seem
drawn to gaming as an escape from real world problems. As
a means of escape, gaming may offer temporary distraction,
but without alleviating real world distress, excessive gaming may only exacerbate said problems. Still, the crosssectional nature of past studies leaves open whether individuals with internalizing problems retreat to video games
as an escape, or whether gaming acts as a precursor to these
issues. Moreover, little is known about the relationship
between gaming and internalizing problems in children due
to the scarcity of research among this cohort.
Finally, hyperactivity and inattention has been investigated as a detrimental psychosocial outcome of gaming.
This research is premised on the perception that video
games are fast-paced and offer frequent rewards, thus
potentially habituating children to a steady stream of novel,
pleasurable stimuli. On the one hand, children with Attentional Deficit Hyperactivity Disorder have been shown to
play more video games than their peers (Mazurek and
Engelhardt 2013) and Gentile and colleagues (2012) argue
that there may be a bidirectional effect between attentional
problems and gaming. On the other hand, studies among
adults show that action video games may confer cognitive
benefits, including improvements in executive functioning
(Green and Bavelier 2012). Due to these conflicting findings, and a lack of longitudinal research among children, the
extent to which gaming may influence children’s attention
remains largely unknown.
886
J Youth Adolescence (2017) 46:884–897
The potential influence of video games on social behavior
seems particularly relevant. This is because, compared to the
video games of just two decades ago, contemporary video
games have become increasingly social in nature (Olson
2010). Researchers such as Greitemeyer and Ewoldsen have
noted that just as some games predicate in-game progress on
violence, other games predicate progress on prosocial behaviors (Ewoldsen et al. 2012; Greitemeyer and Osswald
2011). For instance, many games designed for multiple
players feature cooperative game modes where players are
encouraged to work together with others. A number of studies support the hypothesis that cooperative gaming may
promote prosocial behavior (Dolgov et al. 2014; Ewoldsen
et al. 2012) and may curb aggressive behaviors (Jerabeck and
Ferguson 2013; Velez et al. 2014) (although many of these
studies feature the sorts of methodological shortcomings
mentioned above). In contrast to cooperative gaming,
researchers have also investigated whether competitive
gaming promotes aggression and discourages prosocial
behavior (Eastin 2007). For instance, Adachi and colleagues
performed a series of studies to test the relative extent to
which violent content and competitive play each promote
aggression (Adachi 2015). Using experimental and longitudinal designs, these studies indicated that in both the shortand long-term, competitive gaming may be a greater predictor of aggressive outcomes than violence alone. However,
cooperative and competitive gaming have yet to be researched in the way these forms of play most commonly occur in
the real world—in tandem. Thus, while researchers have tried
to individually assess the effects of these forms of play, they
often co-occur in the real-world of gaming most children
participate. This is because many competitive video games
not only allow cooperative modes, but the competition in
these games is often team-based. However, no longitudinal
studies to date have simultaneously investigated the influence
of both cooperative and competitive video game playing; this
is important as many video games designed for competitive
play are also team-based, and therefore allow for cooperative
play as well.
Second, this study targeted an under-studied population,
namely children between the ages of seven and eleven.
Indeed, despite claims that children are especially vulnerable
to the effects of video game playing (Bushman and Huesmann 2006), scant longitudinal research has targeted children. Children seem particularly susceptible to being
influenced by video games because, unlike adults, they are
still in the process of forming patterns for how they deal with
social and emotional challenges. The behaviors and patterns
of mind that are therefore promoted during video game may
have a greater impact on them than on adults. Moreover, as
children near adolescence, their peer network and relationships become increasingly important (Davies 2010). As a
result, the social interactions they enact and rehearse during
video game play may be of greater relevance for how they
interact with their peers in the real world. Finally, our
longitudinal design allowed us to simultaneously test for both
gaming and selection effects; in the former, video game
playing may precipitate psychosocial changes, whereas in the
latter, children who already show psychosocial deficits may
select video games as an outlet. Thus, our longitudinal design
also allowed us to investigate the tandem development of
video game playing and psychosocial health.
Five domains of children’s psychosocial health were
assessed at two timepoints—externalizing problems, internalizing problems, hyperactivity and inattention, peer problems, and prosocial behavior. Given the psychosocial
benefits of play, we expected video game playing at the first
time point to predict decreases in children’s (H1) externalizing problems, (H2) internalizing problems, (H3) peer
problems, and (H4) overall psychosocial problems by the
second time point. Given the lack of consensus in the
research, no predictions were made for the influence of
gaming on hyperactivity and inattention, or on prosocial
behavior, although exploratory analyses were conducted.
We also explored the potential relationships between violent
video game content and both externalizing problems and
prosocial behavior. Finally, we also hypothesized that (H5a)
cooperative gaming at the first time point would be associated with increases in prosocial behavior, whereas (H5b)
competitive gaming at the first time point would be associated with decreases in prosocial behavior.
Design and Hypotheses
Method
The present longitudinal study was designed to address the
gaps in the literature described above. First, we focused on
the potential psychosocial benefits that playing video games
may have for children. Thus, in addition to assessing
negative outcomes such as externalizing problems, internalizing problems, and hyperactivity and inattention, we
also focused on peer relations, and prosocial behavior.
Participants
Prosocial Behavior and Cooperative and Competitive
Gaming
Data were collected during home visits 1 year apart (T1 and
T2; days between visits: range 265–510, M = 392.22, SD
= 59.05). Participants were recruited from a pool of 298
participants already participating in research which tracked
children’s psycho-social health (Stone et al. 2010). Parents
J Youth Adolescence (2017) 46:884–897
Table 1 Child and parent
demographics at T1 and T2
887
Children
Age (years)
Sex
Age (years)
Sex
T1 (n = 194)
Range
7.27–11.43
Male
98 (50.5%)
M
9.22
Female
96 (49.5%)
SD
1.14
T2 (n = 184)
Range
8.31–12.68
Male
90 (48.9%)
M
10.24
Female
94 (51.1%)
SD
1.14
were contacted via letters sent to their homes and follow-up
phone invitations. At T1, the children’s gender was evenly
split (boys n = 98); 86.6% of parent reporters were female
(n = 168); with the exception of three adopted mothers and
one adopted father, all parents were the child’s biological
parent; finally, 96.9% of parents were Dutch (n = 188), with
the others coming from Suriname (n = 1) or nearby European countries (n = 5). The study’s procedures were
approved by the Behavioural Science Institute’s Ethical
Review Board under the Radboud University, and informed
consent forms were obtained from parents at both timepoints. Descriptive statistics for the sample at T1 and T2 are
reported in Table 1. Ten participants from T1 declined to
participate at T2. Additionally, data from ten parent reports
were missing at T2 because their data was not properly
saved by the recording software. With the exception of five
parents, all parent reports were provided by the same parent
at T1 and T2. Among parents, education level was low for
6.7%, medium for 30.4%, and high for 60.3%1.
Procedure
Children provided self-reports during a face-to-face interview with an experimenter. Parents provided their survey
responses via an online questionnaire. Families were
rewarded a 20 and 30 Euro voucher check (per child) for
their participation at T1 and T2 respectively.
1
Parents
Secondary and higher education in the Netherlands is stratified.
Here, low refers to individuals who completed the lowest level of
secondary school, a vocational school track until the age of 16;
medium refers to individuals who completed a more advanced
vocational track until the age of 17; and high refers to individuals
with a university-level education, having attained a Bachelor’s degree
or higher.
T1 (n = 194)
Range
29.95–51.47
Male
26 (13.4%)
M
41.88
Female
168 (86.6%)
SD
3.66
T2 (n = 174)
Range
30.68–52.42
Male
24 (13.8%)
M
42.83
Female
150 (86.2%)
SD
3.76
Measures
Psychosocial health
Psychosocial health was measured by parent’s reports on the
Dutch version of the Strengths and Difficulties Questionnaire (SDQ (Goodman 1997); Dutch version (van
Widenfelt et al. 2003)). The SDQ uses a 3-point Likert
scale (0–2 Not true to Very true) and is comprised of five
sub-scales: (a) internalizing problems, (b) externalizing
problems, (c) hyperactivity/inattention, (d) peer relationship
problems, and (e) prosocial behavior. Consistent with Stone
and colleagues (2010) reliability was calculated using ω;
this reliability index has repeatedly been shown to yield
more accurate estimates than α, particularly so when data
are skewed, as is the case with SDQ (Stone et al. 2015;
Zinbarg et al. 2005). All sub-scales showed acceptable to
good reliability at T1 and T2: (a) internalizing problems
(sample: Many worries, often seems worried; ωT1 = .83;
ωT2 = .81); (b) externalizing problems (sample: Often fights
with other children or bullies them; ωT1 = .75; ωT2 = .89);
(c) hyperactivity/inattention (sample: Restless, overactive,
cannot stay still for long; ωT1 = .88; ωT2 = .89); (d)
peer problems (sample: Rather solitary, tends to play alone;
ωT1 = .83; ωT2 = .68); and (e) prosocial behavior (sample:
Shares readily with other children; ωT1 = .84; ωT2 = .78).
All sub-scales consist of five items with sum scores being
calculated for each sub-scale. The SDQ also includes
a total difficulties score, calculated as the sum scores of all
scales except for prosocial behavior (ωT1 = .95; ωT2 = .95);
this reflects children’s general psychosocial health.
Descriptive statistics for the SDQ measures are presented in
Table 2.
Gaming frequency
Children’s frequency of video game playing was assessed
by: (1) Parental reports for the number of hours their child
888
J Youth Adolescence (2017) 46:884–897
Table 2 Change in SDQ from T1 to T2
Violent gaming
T1
T2
M
SD
M
SD
Internalizing problems
1.99
2.1
1.76
1.86
Externalizing problems
1.01
1.39
0.84
1.46
Hyperactivity
2.99
2.55
2.89
2.62
Peer problemsa
1.14
1.63
0.97
1.27
Prosocial behavior
6.77
1.49
6.9
1.31
7.15
5.33
6.47
4.99
Total difficulties
b
a
Peer problems decreased from T1 to T2, t(173) = 2.09, p = .038
b
Total difficulties decreased from T1 to T2, t(173) = 2.29, p = .023
plays on average per week; (2) child reports for the number
of hours they had played video games during the past week;
(3) children’s ability to recall their gaming hours across a
whole week was scaffolded by an additional measure of
gaming frequency: in interviews, children looked over a
calendar with the experimenter and indicated for each day
over the past full week whether or not they had played a
video game in the morning, afternoon, and evening. “Video
games” were explicitly described to parents and children as
any game that can be played on an electronic device, and
several example games were listed.
Descriptive statistics for the frequency measures are
presented in Table 3. Both parent’s and children’s reported
hours of gaming were Windorized with a cut-off at 3 SD
above the mean (T1: parent reports: M = 5.76, SD = 3.87,
outliers n = 4; child reports: M = 4.86, SD = 4.25, outliers
n = 6. T2: parent reports M = 6.83, SD = 5, outliers n = 2;
child reports: M = 5.92, SD = 5.9, outliers n = 2). Children
reported an average of 7.88 discrete play sessions per week
(SD = 4.15) at T1 and 8.11 (SD = 4.78) at T2.
Moderate correlations were observed across the three
frequency measures at each time point (T1: r ≥ .47, p
< .001; T2: r ≥ .41, p < .001). Moderate correlations were
also observed within reporters across T1 and T2 (parental
report: r = .566, p < .001; child report: hours r = .367, p
< .001, calendar r = .485, p < .001). Game frequency was
operationalized as child reports of hours gaming2. As psychosocial health was reported by parents, this means that
our analyses were performed across reporters. This is preferred to analyses using only a single reporter as such
analyses introduce a potential single source bias (Burk and
Laursen 2010; Lobel et al. 2014a).
2
Analyses using parent reported hours yielded the same pattern of
results.
Similar to the methods in Anderson and Dill (2000) and
Prot et al. (2014), children were asked to report their
favorite video game(s) from the past several weeks. At T1,
Minecraft, Super Mario Bros., and Subway Surfer were the
most popularly listed games/franchises, each being reported
by 13 children. At T2, the most popular titles were more
diverse with 46 children listing Minecraft, 21 listing a title
from the Fifa franchise, and 18 listing Mario Party and Hay
Day each. Violent gaming was computed as a dichotomous
variable; children who listed a violent video game among
their favorite games were assigned a 1, and those who did
not were assigned a 0. Video games were classified as being
violent when gameplay required players to harm other ingame characters.
Cooperative and competitive gaming
Whereas third-party review boards provide information about
whether a game contains violent content, the extent to which
games are played cooperatively or competitively is not.
Violent content is therefore objectively observable based on a
game’s content and design, whereas competitive and cooperative play is more difficult to observe. Following Przybylski and Mishkin (2016), cooperative and competitive
gaming were therefore individually assessed by children with
a single item using a Likert scale (5-point scale: Never to
Every time or almost every time). Experimenters clearly
informed children what was meant by “cooperative” and
“competitive” gaming: Children were asked to think about the
times that they play video games, and to rate the frequency
with which, when playing, they play a game where they have
to “work together with others; that the game is cooperative”
and “play against others; that the game is competitive”.
Planned Analyses
All analyses were performed in R (R Core Team 2013). For
preliminary analyses, paired-sample t-tests were used to
determine whether children’s psychosocial health and
gaming frequency changed from T1 to T2, independent ttests were used to determine whether there were gender
differences on all variables at both timepoints, and correlations were calculated. To investigate our hypotheses, three
sets of structural path models were estimated with the
lavaan package (Rosseel 2012). In all models, full information maximum likelihood was used to account for
missing values and the Hubert-White covariance adjustment
(MLR in lavaan) was applied to standard errors in order to
deal with the lack of normally distributed variables.
In the first models, cross-lagged panel models were
estimated to test H1-H4; that is, whether gaming at T1
J Youth Adolescence (2017) 46:884–897
Table 3 Gender differences in
gaming frequency at T1 and T2
889
Parent hours
T1
Child hours
T2
T1
Child calendar
T2
T1
T2
M
SD
M
SD
M
SD
M
SD
M
SD
M
SD
Total
5.67
3.63
6.80
4.90
4.90
4.07
5.81
5.48
7.88
4.15
8.11
4.76
Boys
6.75
3.94
8.26
5.16
5.93
4.21
7.55
5.85
9.23
3.99
9.81
2.63
Girls
4.62
2.95
5.47
4.25
3.85
3.67
4.16
4.54
6.47
3.83
6.49
4.48
t
p
t
p
t
p
t
p
t
p
t
p
4.24
Purchase answer to see full
attachment