The impact of mass media violence on aggression has almost always been studied in the laboratory; this paper examines the effect of mass media violence in the real world. The paper presents the first systematic evidence indicating that a type of mass media violence triggers a brief, sharp increase in U.S. homicides. Immediately after heavyweight championship prize fights, 1973-1978, U.S. homicides increased by 12.46 percent. The increase is greatest after heavily publicized prize fights. The findings persist after one corrects for secular trends, seasonal, and other extraneous variables. Four alternative explanations for the findings are tested. The evidence suggests that heavyweight prize fights stimulate fatal, aggressive behavior in some Americans. (abstract Adapted from Source: American Sociological Review, 1983. Copyright © 1983 by the American Sociological Association) Media Violence Effects Homicide Causes Violence Causes Televised Sports Television Viewing Television Violence Sports Viewing-Related Violence 1970s Adult Violence 07-02

This paper shows that suicides increase immediately after a suicide story has been publicized in the newspapers in Britain and in the United States, 1947-1968. The more publicity devoted to a suicide story, the larger the rise in suicides thereafter. The rise in suicides after a story is restricted mainly to the area in which the story was publicized. Alternative explanations of these findings are examined; the evidence indicates that the rise in suicides is due to the influence of suggestion on suicide, an influence not previously demonstrated on the national level of suicides. The substantive, theoretical, and methodological implications of these findings are examined.

David P. Phillips, Kenneth A. Feldman, A Dip in Deaths Before Ceremonial Occasions: Some New Relationships Between Social Integration and Mortality, American Sociological Review, Vol. 38, No. 6 (Dec., 1973), pp. 678-696

This paper tests the artifnct explanation for the findings in Phillips (1983) and Bollen and Phillips (1982). (I) First test: gdeaths rise qFer real stories because ofthe timing of these stones, then deaths should also rise dter bogus stories (which fall a year ahead of the real ones) since they exhibit the same timing as the real stories. Using small samples, Baron and Reiss found a rise in deaths afier bogus stories. Using larger samples we found no such rise. Baron and Reiss restricted themselves to bogus year-ahead ston-es. but we also studied bogus year-behind stories (which fall a year behind the real ones). Because bogus behind stories exhibit the same timing as bogus ahead stories, they too should display a rise in deaths. However. no such rise can be found. (2) Second test: If deaths rise after stories because of heteroscedasticity. then these rises should disappear once heteroscedasticity has been removed. Honlever, all findings in Phillips (1983) and Bollen and Phillips (1982) persist after removal of heteroscedasticity. The findings also persist afier holiday eflects are eliminated. The above tests show that the artif4ct hypothesis is extremely implausible. Baron and Reiss suggest that the results in Phillips (1983) and Bollen and Phillips (1982) are an of heteroscedasticity and the timing of publicized stories. We will call this the artifact hypothesis. In contrast, Phillips and Bollen claim their results are due to of the publicized stories; this will be called the imitation hypothesis. This paper will use 'three methods for assessing the artifact and imitation hypotheses, starting with Baron and Reiss's own BOGUS-story approach.

A 1974 ASR paper by Phillips showed that suicides increase in the month of highly publicized suicide stories. Several subsequent papers have built on this finding and have suggested that publicized suicide stories trigger increases in covert suicides disguised as motor vehicle or airplane accidents. If the original ASR findings cannot be replicated, then doubt is cast on all these studies. In this paper we demonstrate that U.S. daily suicides increase significantly after highly publicized suicide stories appear on television evening news programs. We provide the first evidence that the increase in suicides occurs only after and not before the suicide story. In addition, we provide the first systematic study of the length of time a suicide story affects suicides; the effect probably does not extend beyond ten days. These findings support and extend the literature on imitative suicides.