LEONARD D. ERON MONROE M. LEFKOWITZ ROWELL HUESMANN LEOPOLD 0. WALDER AMERICAN PSYCHOLOGIST APRIL 1972 253 DOES TELEVISION VIOLENCE CAUSE AGGRESSION? '

LEONARD D. ERONMONROE M. LEFKOWITZ

ROWELL HUESMANNLEOPOLD 0. WALDER

AMERICAN PSYCHOLOGIST • APRIL 1972 • 253

DOES TELEVISION VIOLENCE CAUSE AGGRESSION? '

• Longitudinal data were collected on 427 teenagers of an

• original group of 875 children who had participated in a study of third-grade children in 1960

• The information collected about these subjects in both time periods falls into two classes:

• (a) measures of aggression and

• (b) potential predictors of aggression.

• During the third-grade interviews, four different data sources had been used: the subject, his peers, his mother, and his father.

• Ten years later, the data sources were the subject and his peers. For convenience, this later time period will be designated as the thirteenth grade.

• The variables used in the study are listed in Table 1.

• Two variables are of particular importance: peer-rated aggression and preference for violent television programs.

In the third grade, peer-rated aggression scores were obtainedby asking each child to nominate any of hisclassmates on 10 "guess who" items describing aggressivebehavior; for example, "Who pushes and shoves otherchildren?"

13thThe number of possible peers any subject couldnominate was widened beyond his own third-gradeclassroom to include his classmates through high school.

In the third grade, the children's preferences for violenttelevision were obtained by asking each mother for herchild's three favorite television programs. All programsmentioned were then categorized as violent or nonviolentby two independent raters with 94% agreement in theirratings.

Each subject received a score according to the number ofviolent television programs he was reported by his motheras favoring. Scores ranged from 1 (for no violent programs)to 4 (for three violent programs).

In the 10-year follow-up study, each subject himself wasasked for his four current favorite television programs.All programs were then categorized for presence or absenceof violence by two independent raters who wereonly a few years older than the subjects themselves.

The 427 subjects studied in both the third andthirteenth grades consisted of 211 males and 216females whose modal age at the time of thethirteenth-grade interview was 19.

The dropout rate from participation in the study between the third and thirteenth grade was considerably higher for subjects who displayed high aggression in the third grade than for those who displayed low aggression.

• The correlations between preference for television violence and aggression at both grade levels are presented separately for boys and girls in Table 2.

Because of the large number of potential correlations, only those with a probabilityless than .01 on a two-tailed test were consideredsignificant and only those with a probabilityless than .001 on a two-tailed test were calledhighly significant. On the other hand, it shouldbe pointed out that the unusual skewness andkurtosis of many of the variables may impose alimit on the size of their intercorrelations.

Table 2 six variables

Boys - significant correlations between TV and peer related agg.Aggression in 3rd grade and TV violence in 3rd

Aggression in 3rd and aggression in 13th

TV in 3rd and aggression in 13th

Self –ratings & peer-ratings Antisocial behavior 13th and aggression in 13th

Delinquency predisposition and aggression 13th and antisocial 13th

GirlsGirls no TV linkAggression in 3rd and aggression in 13th

Self –ratings & peer-ratings Antisocial behavior 13th and aggression in 13thDelinquency predisposition and aggression 13th and antisocial 13th

ANOVA - boys

Reveals – relationship between violent TV in 3rd grade and aggression in 13th

(either peer rated or self-rated)

Six possible relationships between aggression and TV viewing - Boys at 2 ages

a) TV viewing 3rd grade causes aggression 3rd grade and 13th

b) TV viewing 3rd grade causes aggression in 3rd grade and this causes aggression in 13th

(product of intermediates smaller than endpoint)

c) Aggression in 3rd grade causes child to watch violent TV and this causes aggression at 13th

(product of intermediates smaller than endpoint)

d) Aggression in 3rd grade causes both violent TV watching in 3rd grade and aggression in 13th

e) Aggression in 3rd grade causes aggression in 13th both but not violent TV watching in 13th grade

To help refute alternative hypotheses remove factor of ….and determine relationship

Multiple regression

The order in which the variables were entered reveals that a preference for watching television violence was the most "useful" third-grade variable in the prediction. It explained more of the variance than any other predictor.

The coefficients in Table 5 showthat a preference for violent television in thirdgrade is the -major contributor to thirteenth-gradeaggression among the third-grade variables.