We get a clearer picture about the extent of TV violence effects when we know more about the way children watch televised violence. For example, Ekman and his associates (Ekman et al., 1972) found that those children whose facial expressions, while viewing televised violence, depicted the positive emotions of happiness, pleasure, interest or involvement were more likely to hurt another child than were those children whose facial expressions indicated disinterest or displeasure.

The long-term influence of television has not been extensively investigated but we do have indications from several major studies. In an initial longitudinal study Lefkowitz and his colleagues (Lefkowitz et al., 1972) were able to demonstrate long-term effects in a group of children followed-up over a ten-year period. In this instance, Eron (1963) had previously demonstrated a relationship between preference for violent media and the aggressive behaviour of these children at the age of eight. One question now posed was, would this relationship hold at later ages? To answer this question, the investigators obtained peer-rated measures of aggressive behaviour and preferences for various kinds of television, radio and comic books when the children were eight years old. Ten years later, when the members of the group were eighteen years old, the investigators again obtained measures of aggressive behaviour and television program preferences. The results for boys indicated that preference for television violence at age eight was significantly related to aggression at age eight (r = .21), but that preference for television violence at age eighteen was not related to aggression at age eighteen (r = .05). A second question posed was, could this adolescent aggressiveness be predicted from our knowledge of their viewing habits in early childhood? And, the answer seems to be yes. The important finding here is the significant relationship, for boys, between preference for violent media at age eight and aggressive behaviour at age eighteen (r = .31). Equally important is the lack of relationship in the reverse direction; that is, preference for violent television programs at age eighteen was not produced by their aggressive behaviour in early childhood (r = .01). The most plausible interpretation of this pattern of correlations is, that early preference for violent television programming and other media is one factor in the production of aggressive and antisocial behaviour when the young boy becomes a young man.

In more recent, short- term, longitudinal studies conducted by Lefkowitz and Eron and by their colleagues (Eron, 1982; Huesmann, Langerspetz & Eron, 1984; Sheehan, 1983), they found some short-term effects of viewing violence on aggressive behaviour of children in the United States, Australia and Finland.

Finally, the 22-year longitudinal study (Huesmann, Eron, Lefkowitz & Walder, 1984)--a follow-up to the earlier Lefkowitz et al. (1972) study--has found significant causal-correlations (r = .41) between violence viewing at age eight and serious interpersonal criminal behaviour at age 30.

In a different approach, a study by Belson (1978) has substantiated other long-term effects and has helped pin down which types of programs have the most influence. Belson interviewed 1565 youths who were a representative sample of thirteen to seventeen-year-old boys living in London. These boys were interviewed on several occasions concerning the extent of their exposure to a selection of violent television programs broadcast during the period 1959-71. The level and type of violence in these programs were rated by members of the BBC viewing panel. It was thus possible to obtain, for each boy, a measure of both the magnitude and type of exposure to televised violence (e.g. realistic, fictional, etc.). Furthermore, each boy's level of violent behaviour was determined by his own report of how often he had been involved in any of 53 categories of violence over the previous six months. The degree of seriousness of the acts reported by the boys ranged from only slightly violent aggravation such as taunting, to more serious and very violent behaviour such as: 'I tried to force a girl to have sexual intercourse with me; I bashed a boy's head against a wall; I threatened to kill my father; I burned a boy on the chest with a cigarette while my mates held him down'. Approximately 50 per cent of the 1565 boys were not involved in any violent acts during the six-month period. However, of those who were involved in violence, 188 (12 per cent) were involved in ten or more acts during the six-month period. When Belson compared the behaviour of boys who had higher exposure to televised violence to those who had lower exposure (and had been matched on a wide variety of possible contributing factors), he found that the high- violence viewers were more involved in serious violent behaviour. Moreover, he found that serious interpersonal violence is increased by long-term exposure to (in descending order of importance):

In summarising the extent of the effects, we agree with Comstock (Comstock & Paik, 1991) that there are multiple ways in which television and film violence influence the viewer. Comstock suggests four dimensions: Efficacy relates to whether the violence on the screen is rewarded or punished; Normativeness refers to whether the screen violence is justified or lacks any consequences; Pertinence describes the extent to which the screen violence has some similarity to the viewer's social context; and Suggestibility concerns the predisposing factors of arousal or frustration. Drawing on these four dimensions, Comstock suggests (Comstock & Paik, 1991, pp. 254-255) situations for which we have experimental evidence of the effects of film or television violence: