Escolar Documentos
Profissional Documentos
Cultura Documentos
Objectives: To test whether the results of the accumu- Results: Effect size estimates were combined using meta-
lated studies on media violence and aggressive behavior analytic procedures. As expected, the short-term effects of
are consistent with the theories that have evolved to ex- violent media were greater for adults than for children
plain the effects. We tested for the existence of both short- whereas the long-term effects were greater for children than
term and long-term effects for aggressive behavior. We for adults. The results also showed that there were overall
also tested the theory-driven hypothesis that short-term modest but significant effect sizes for exposure to media
effects should be greater for adults and long-term ef- violence on aggressive behaviors, aggressive thoughts, an-
fects should be greater for children. gry feelings, arousal levels, and helping behavior.
Design: Meta-analysis. Conclusions: The results are consistent with the theory
that short-term effects are mostly due to the priming of
Participants: Children younger than 18 years and adults.
existing well-encoded scripts, schemas, or beliefs, which
Main Exposures: Violent media, including TV, mov- adults have had more time to encode. In contrast, long-
ies, video games, music, and comic books. term effects require the learning (encoding) of scripts,
schemas, or beliefs. Children can encode new scripts, sche-
Main Outcome Measures: Measures of aggressive be-
mas, and beliefs via observational learning with less in-
havior, aggressive thoughts, angry feelings, physiologi-
terference and effort than adults.
cal arousal (eg, heart rate, blood pressure), and helping
behavior. Arch Pediatr Adolesc Med. 2006;160:348-352
T
HE BODY OF EMPIRICAL RE- violence and aggression that this exposure
search linking children’s ex- produces.Wethenusemeta-analysestoshow
posure to media violence that,onthewhole,theavailableempiricaldata
with subsequent increases in showthepatternsonewouldexpectfromthis
their aggressive and violent theory. Although the focus of this article is
behaviorwasalreadysubstantialbythe1970s. on exposure to media violence, the theoreti-
The 1972 Surgeon General’s Scientific Ad- cal premise is that the same processes oper-
visory Committee on Television and Social ate when children are exposed to media vio-
Behavior report1 and the National Institute lence as when they are exposed to violence
of Mental Health 10-year follow-up report2 on the street, in the home, or among their
providedwidelyaccessiblesummariesofthis peers.
growingbodyofresearch.Bythe1980s,most The psychological processes that link
childdevelopmentscholarshadacceptedthe children’s exposure to violence with sub-
theory that exposure to media violence, at sequent increases in children’s aggres-
least during some periods of a child’s devel- sive behaviors can be divided into those
opment, increased their risk for aggression.3 that produce more immediate but tran-
However, it has only been in the 1990s and sient short-term changes in behavior and
more recently that meta-analyses have pro- those that produce more delayed but en-
videdsystematicsummariesoftheentirebody during long-term changes in behavior.
of research.4,5 Similarly, it has been during Long-term increases in children’s aggres-
this time that the attention of researchers has sive behavior are now generally agreed to
turned more toward investigating the pro- be a consequence of the child’s learning
cesses producing the effects and developing scripts for aggressive behavior, cogni-
a more coherent theoretical understanding tions supporting aggression, and aggres-
Author Affiliations: Vrije
of the effects.6 sion-promoting emotions through the ob-
Universiteit, Amsterdam, the
Netherlands (Dr Bushman); and
Inthisarticle,wedescribewhatwebelieve servation of others behaving violently. This
Institute for Social Research, are the major psychological processes that observational learning generally requires
University of Michigan, account for the relationship between chil- the repeated observation of violence. On
Ann Arbor (Drs Bushman dren’s exposure to different kinds of violence the other hand, short-term increases in
and Huesmann). in their world and the increase in the risk for children’s aggressive behavior following
(REPRINTED) ARCH PEDIATR ADOLESC MED/ VOL 160, APR 2006 WWW.ARCHPEDIATRICS.COM
348
(REPRINTED) ARCH PEDIATR ADOLESC MED/ VOL 160, APR 2006 WWW.ARCHPEDIATRICS.COM
349
Measure Average Correlation (95% CI) Studies, No. Participants, No. Fail-Safe N
Aggressive behavior 0.19 (0.19 to 0.20) 262 48 430 151 984
Helpful behavior −0.08 (−0.11 to −0.04) 59 3243 157
Aggressive thoughts 0.18 (0.17 to 0.19) 140 22 967 34 447
Angry feelings 0.27 (0.24 to 0.30) 50 4838 8654
Physiological arousal 0.26 (0.20 to 0.31) 27 1356 927
existence of a well-encoded network of aggressive scripts, TV programs, films, video games, music, and comic books). Sec-
beliefs, and schemas. The development and elaboration of ond, the study needed to include a measure of aggressive behav-
such networks require time and repeated learning experi- ior, aggressive cognition, aggressive affect, physiological arousal,
ences, so adults are more likely to possess well-elaborated, or helping behavior. A total of 431 studies involving 68 463 par-
ticipants met these criteria. There were 264 studies involving
rich networks of associations involving aggressive scripts,
50 312 children and 167 studies involving 18 151 adults.
beliefs,andschemas.Consequently,observationsofviolence
byadultswillprimeasetofrelatedaggressiveconstructslarger
CODED CHARACTERISTICS
than that for children and will prime them more rapidly.
There should be positive relationships for both adults The following characteristics were coded for each study: (1)
and children between their degree of exposure to media type of study (ie, longitudinal study, laboratory experiment,
violence earlier in life and their subsequent long-term dis- field experiment, correlational study); (2) participant age (ie,
plays of aggressive scripts, beliefs, and schemas. Obser- children aged ⬍18 years, adults); and (3) type of dependent
vational learning of scripts, schemas, and beliefs and the variable (ie, aggressive behavior, aggressive thoughts, angry feel-
ings, physiological arousal, helping behavior).
desensitization of negative emotional reactions to vio- Each article was coded by one of us (B.J.B.). To assess cod-
lence are the 2 processes that contribute to these effects. ing reliability, an additional 3 judges each coded a random
These long-term effects should be more pronounced for sample of 50 articles. A different random sample (with replace-
children than for adults because observational learning and ment) was selected for each judge. There was perfect agree-
desensitization both depend on the observer having a mind ment among judges on coded variables.
that is readily modifiable. The younger the children are, the
moresusceptibletheyshouldbetoencodingaggressivescripts, META-ANALYTIC APPROACH
schemas, or beliefs through observational learning; also, the
The effect size index was the correlation coefficient. Because the
younger the children are, the more quickly their emotional
distribution of the correlation coefficient is not normal unless the
responsestoobservingviolenceshouldhabituate.Foradults, population correlation coefficient equals 0, we applied Fisher z
learning new scripts, schemas, and beliefs requires replac- transformation to each correlation coefficient before pooling them.
ing old ones, and the process is likely to take longer and re- Each z score was weighted by the inverse of its variance (ie, N−3).
quire the observation of more powerful scenes. In the rest Thus, larger studies had more weight computing the average cor-
of this article, we examine whether meta-analyses of the ac- relation. Then, we transformed z scores back to correlations. We
cumulated empirical data on media violence are consistent also computed a 95% confidence interval for the average corre-
with these predictions. lation. If the 95% confidence interval excludes the value 0, the
correlation is significantly different from 0.
We used both fixed- and random-effects analyses, al-
METHODS though the 2 approaches yielded the same pattern of results.
Thus, we report only the fixed-effects analyses in this article.
DATA SOURCES In a meta-analysis, it is difficult to find unpublished studies.
Studies with nonsignificant effects are often not published; they
To retrieve relevant studies, we searched the PsychINFO da- end up in file drawers rather than in peer-reviewed journals. If
tabase from 1887 (starting date) to 2000 using the terms vi- studies in file drawers had been published, the average correla-
olen* or aggress* as well as the terms TV, televis*, film, movie, tion would be smaller. “Fail-safe N” is the number of studies with
screen, music, radio, video, video game, computer game, cartoon, nonsignificant correlations that are needed to reduce the aver-
comic, pornograph*, erotic*, news, book, magazine, or sport. The age correlation to 0.36 If the fail-safe N is large, then the results
asterisk is a wild card that gives all of the possible forms of the are probably not affected by publication bias.
word (eg, aggress* searches for the terms aggress, aggressed,
aggression, aggressive, aggressively, and aggressor). We re-
stricted the search to empirical studies involving human par- RESULTS
ticipants. We also searched the reference sections of previous
meta-analyses of violent media and aggression.4,33-35 OVERALL EFFECTS OF MEDIA VIOLENCE
STUDY SELECTION As shown in the Table, there were significant, if modest,
Two inclusion criteria were used. First, the study needed to in- overall positive effect sizes showing that exposure to me-
clude a measure or manipulation of violent media exposure (eg, dia violence was positively related to subsequent aggres-
(REPRINTED) ARCH PEDIATR ADOLESC MED/ VOL 160, APR 2006 WWW.ARCHPEDIATRICS.COM
350
Correlation
drawer problem is not that serious.
(REPRINTED) ARCH PEDIATR ADOLESC MED/ VOL 160, APR 2006 WWW.ARCHPEDIATRICS.COM
351
Announcement
(REPRINTED) ARCH PEDIATR ADOLESC MED/ VOL 160, APR 2006 WWW.ARCHPEDIATRICS.COM
352