Major trends in research over three decades are examined. The paper also indicates some possible connections between research on effects of media violence and recent developments in cognitive psychology. The evidence is organized according to the theoretical explanations that have been given for the commonly reported correlation between the viewing of violence and aggressive behavior. The paper is divided into three major sections. In the first, evidence from laboratory experiments is reviewed. In the second section, studies carried out in field-experimental settings are summarized. In the third section, evidence from nonexperimental studies is analyzed. In the conclusion, the authors suggest that the cognitive-neoassociationist hypothesis serves as the best explanatory model for the overall findings, and they indicate how some of the other hypotheses concerning media effects on aggression may be subsumed by it. (Author abstract)