No Game - University of Notre Dame

Do Prosocial Videogames have Stronger Effects than Violent Videogames?
Darcia Narvaez, PhD -- Bradley Mattan -- Carl MacMichael –
Abstract
Many previous studies have suggested that violent media increases aggressive thoughts and feelings in the viewer. Because videogames are a relatively new
form of media, the effects of videogame play have been the subject of fewer studies, though recent research has begun to shed light on the topic. This
study compared the cognitive effects of a prosocial videogame, in which the player was tasked with healing sick non-player characters (NPC’s), against
those of a violent game. Results indicate that participants were prosocially primed after playing the healing game, though there was no significant
aggressive priming after the violent game. A helping-oriented game may promote a prosocial expectancy bias.
Introduction
 Previous research indicates a tendency for violent
media to promote a “hostile expectation bias” in its
viewers (e.g. Bushman & Anderson, 2002;
Anderson, Gentile, & Buckley, 2007). A hostile
expectation bias is defined as “the tendency to
expect others to react to potential conflicts with
aggression.”
Violent Condition
Helping Condition
Violent
(n=31)
Helping
(n=28)
Collector
(n=19)
Game Expectant
(n=21)
No Game
(n=22)
Current Study
 The current study examines the possibility that
prosocial videogames can have a positive effect on
the cognitions of the player, just as aggressive
games can have a negative one.
 Since there are so few helping-oriented
videogames on the market, custom-designed
scenarios were created for all conditions, using the
level design toolset of the 2002 role-playing game
Neverwinter Nights.
Materials
 3 Videogame Conditions
 Violent (aggressive priming): Player kills
bandits with a melee weapon and collects their
gold. Bandits grow progressively tougher and
more numerous as player progresses, and
some must be killed quickly before they escape
 Helping (prosocial priming): Player heals sick
NPC’s, and collects a “note of gratitude” from
the healed. Sick characters require more
healing as player progresses, and some must
be healed quickly before they die of disease
 Collector (control condition): Player collects
bags of gems lying on the floor, in a race
against mice that aim to do the same. Bags and
mice increase in number as player progresses
 2 Additional Control Conditions
 Game Expectant: Expected to play game, but
did not until end, & No Game: No game played
 Story Stems
 All subjects write 20 possible responses for
each of 3 incomplete story stems (Bushman
and Anderson, 2001).
Table 1. Percentage of Response Type (Aggressive,
Prosocial, Neutral) by Condition (Violent, Helping,
Collector, Game Expectant, No Game)
Condition
 However, very little research has been conducted
on the effects of prosocial media or the existence
of a prosocial expectation bias, and there are no
previous studies investigating thoughts & feelings
after playing a prosocial videogame.
Research Question
Does a prosocial videogame increase
prosocial expectancy in the player?
Method & Procedure
 Participants were 116 undergraduates:
57 males, mean age = 19.53; 18 minority
59 females, mean age = 19.61; 15 minority
 Participants were randomly assigned to a
condition. They played the videogame for ten
minutes, then completed the three ambiguous
story stems. Afterwards, subjects provided
demographic information and indicated their
degree of experience with videogames (7-point
Likert-type scale). Players also rated their
arousal before playing, after playing, and after
finishing the stories stems.
Scoring
 Responses to the story stems were scored for
neutral, aggressive, or prosocial content
(thoughts and actions that the character would
experience next). Aggressive responses were
those that would do physical or psychological
harm to the other person in the story, and
prosocial responses were those that supported
the welfare of the other or the relationship.
University of Notre Dame
Department of Psychology
Results
 The MANOVA measured the effects of 4 factors on
responses to story stems: condition, experience in
videogames, sex, and age. It was significant only for
condition (Wilks’ lambda =.26, F(12,74.373) = 4.33,
p<.0001, eta2 = .37). The other variables were
nonsignificant, except that condition by age
approached significance, Wilks’ lambda =.25,
F(33,83.197) = 1.50, p=.07, eta2 squared = .37.
 As predicted, the helping condition generated a
significantly higher number of prosocial responses.
The univariate analysis was significant for prosocial
responses, F(4,82= 10.08, p<.0001, eta2 =.57, and
also for neutral responses, F (4,82)=3.80, p<.013,
eta2 =.34; it approached significance for aggressive
responses, F (4,82)=2.47, p<.07, eta2 =.25.
 The violent condition did not produce a significant
increase in aggressive responses in comparison to the
other conditions. Instead, there were more aggressive
responses than prosocial ones across all conditions.
This may indicate a baseline hostile expectancy bias
associated with conventional videogame experience.
Aggressive
Responses
.23 (.11)
Prosocial
Responses
.11 (.05)
Neutral
Responses
.66 (.10)
.17 (.09)
.15 (.07)
.65 (.09)
.21 (.10)
.11 (.06)
.69 (.08)
.17 (.08)
.07 (.05)
.76 (.08)
.17 (.08)
.05 (.04)
.77 (.08)
Discussion
 The results suggest that prosocially oriented
videogames have at least a short-term priming
effect for prosocial thoughts, feelings, and
attributed behaviors. These findings are
consistent with other research on the positive
effects of prosocial media (Hogan & Strasburger,
in press).
Future Directions
 Further research into this area may lead towards
the conception of a Generalized Prosocial Model
to mirror the General Aggression Model, the
predominant theory explaining the effects of
violent media. If constant exposure to violence
in videogames reinforces aggressive schemas
and attitudes, so too might repeated practice of
prosocial behavior in videogames increase
prosocial affect.
All correspondence regarding this poster can be addressed to Dr. Darcia Narvaez at: [email protected]