Page 285 - Media Effects Advances in Theory and Research
P. 285
274 SPARKS AND SPARKS
success in the game. However, if they pressed the “hurt” button, they
were told that the handle would become too hot to touch and, thus, would
result in hurting the child’s progress in the game. Those children who
watched the violent film clip prior to being placed in this situation were
more likely to press the “hurt” button and more likely to keep the button
pressed for a long duration than were the children who watched the
sports program. Stein and Friedrich (1972) conducted another experiment
with children that randomly assigned subjects to view Batman and Super-
man cartoons (violent condition) or episodes of Mister Rogers Neighborhood
(prosocial condition). During the two weeks of observation following this
manipulation, the children who viewed the violent cartoons were more
likely to be aggressive in their interactions with other children than were
the children who viewed the prosocial programming. Both of these early
experiments, along with the ones by Bandura mentioned earlier, helped to
attract attention to the potential problem of media violence as a facilitator
of aggression.
In contrast to these early experiments that used children as the research
subjects, Leonard Berkowitz conducted a series of experiments that used
college students as subjects (Berkowitz & Alioto, 1973; Berkowitz & Geen,
1966, 1967; Berkowitz & LePage, 1967; Berkowitz & Powers, 1979;
Berkowitz & Rawlings, 1963). The typical paradigm employed in these
investigations was to expose subjects who were either provoked or
unprovoked by an experimenter to either violent media or nonviolent
media. Following exposure, Berkowitz discovered that provoked subjects
behaved more aggressively to the experimenter after viewing violence
than after viewing nonviolence.
Laboratory experiments, although capable of providing unequivocal
evidence for cause-effect relationships, are more equivocal in their appli-
cation to various contexts that exist outside the laboratory. Scholars and
critics who offer a dissenting view from the strong consensus that exists
among social scientists on the effects of media violence usually feature
some version of the argument that laboratory experiments lack ecological
validity. As Zillmann and Weaver (1999) have recently noted, “It seems
that critics of media-violence research could only be satisfied with longi-
tudinal experimental studies in which, within gender and a multitude of
personality variables, random assignment is honored and exposure to
violent fare is rigorously controlled—that is, with research that in a free
society simply cannot be conducted” (p. 147). In addition, it also seems
that critics demand that researchers be able to set up real-world opportu-
nities for aggression in order to settle the controversy about the generaliz-
ability of laboratory findings to settings outside the lab. Of course, even if
it were possible to do so, researchers would never want to set up such
opportunities for ethical reasons. Despite the limitations of experimenta-