Violence is a common characteristic of many television programs. On British television, for example, an estimated average of one violent scene occurs every 16 min (Barlow & Hill, 1985). At schools, many educators have reported a marked increase in children's aggression both on the playground and in the classroom during the past decade (Barlow & Hill, 1985). It is not uncommon for these two observations to be juxtaposed and used as a basis for speculations concerning causal relationships.
Cashmore (1994) wrote that the focus of at least 1,000 research publications has been the possibility of a link between television violence and actual violence; in over three quarters of these publications, the authors have claimed that such a link does indeed exist and have provided various interpretations of its mechanisms. Nonetheless, others have managed to conclude that there is no connection at all. On balance, it seems likely that any relationship that may exist between watching television violence and perpetrating actual violence is likely to be a complex one, and a number of contributing factors must be considered.
Repeated exposure to television violence has been implicated as a major factor in the gradual desensitization of individuals to such scenes (Cline, Croft, & Courrier, 1973). It has been argued that this desensitization, in turn, may weaken some viewers' psychological restraints on violent behavior (such as guilt and fear of retaliation) and their fear of social disapproval (Doob & Wood, 1972). Turner, Hesse, and Peterson-Lewis (1986) concluded that the balance of evidence supports the notion that watching television results in a long-term increase in aggression in boys but not in girls, even though (a) boys and girls watch similar amounts of television (Van Evra, 1990) and (b) girls recall the content of the programs as well as, or better than, boys (Field & Anderson, 1985). A possible explanation for any gender differences may lie in the types of programs preferred by boys and girls. Boys typically prefer violent programs; girls prefer nonviolent ones. Girls are also reported to like violence less, approve of it less, and see it as less realistic; they are more frightened and distressed by television violence, respond to it more emotionally, and watch it in a more involved and less detached way than boys do (Van Evra, 1990).
Regardless of the types of television programs preferred, children who watch a lot of television tend to see more violent programs than children who watch relatively less television (Wiegman, Kuttschreuter, & Baarda, 1992). Children who spend more time watching television, especially violent programs, are also more likely to show later aggression, restlessness, and a belief in a "scary world" (Singer, Singer, & Rapacynski,1984).
A common and important point raised when the potentially harmful effects of violent television programs are discussed is that any significant negative effects tend to be specific and selective to those already predisposed to violence. A number of studies have shown that those young viewers who are more aggressive state that they watch or prefer to watch more violent programs (Wober, 1988). Cross-cultural research indicates that this finding appears to hold true in almost every country studied (Huesman & Eron, 1986).
Given the potential impact of television programs on children, the control of their viewing is of major concern. Such control is often regarded as the parents' responsibility, although evidence suggests that relying on parents is often a poor safeguard. Parental control over children's viewing habits appears to have been declining for several decades (Barlow & Hill, 1985). Rubinstein (1983) found that although there was much parental concern about the levels of sex and violence on television, parents exerted a relatively low level of control over what heir children watched.
In addition, when parents have placed restrictions on what children can watch, the children have found many ways of getting around these limitations. Typical strategies have included watching on another television in a bedroom, sneaking downstairs and watching through a door, or watching at a friend's house (Barlow & Hill, 1985).
Overall, existing research has shown a complex interrelationship of personal characteristics, patterns of media usage, and attitudes toward screen violence. Our aim in this study was to clarify this relationship. Using multivariate techniques, we assessed the attitudes of secondary school children toward violence on television and determined how those attitudes related to the children's personal characteristics and media usage.
The participants were 316 children (156 girls and 160 boys) attending four local, single-sex schools: a boys' grammar school, a boys' secondary modern school, a girls' grammar school, and a girls' secondary modern school. Entrance tests for the academically selective grammar schools are administered at age 11, and children who do not gain entrance to these institutions attend secondary modern schools. Three classes of pupils from each school participated, from years 7 (11- to 12-year-olds), 9 (13- to 14-year-olds), and 11 (15- to 16-year-olds). All members of each class participated. The majority (93.7%) of the children's parents had occupations that were classified, according to the Registrar General's classification of occupations (Reid, 1989), as professional or skilled (manual and nonmanual).
We constructed a 47-item questionnaire that consisted of two parts, each preceded by its own printed instructions. In Part 1, 17 questions collected basic personal details (age, school grade, sex, parents' occupations) and information about the respondents' television and video viewing facilities, preferences, and habits. The remaining 30 items concerned attitudes toward violence on television. These items were based on verbatim statements extracted from semi-structured interviews previously conducted with three 11- to 12-year-olds who did not participate in the current study. All attitude questions were responded to on 5-point rating scales.
Each class in each school was tested separately. Two research assistants conducted the testing sessions. The questionnaires were distributed, and the instructions for Part 1 were read aloud. Any questions were answered in a nondirective manner. The children then completed Part 1 of the questionnaire in silence, after which the instructions for Part 2 were read aloud, and any further questions were answered. Part 2 of the questionnaire was then completed in silence.
Thirteen of the questionnaires contained unusable or incomplete responses. Because these problems appeared to be spread evenly across the groups, we omitted those questionnaires from subsequent analyses, leaving 303 questionnaires completed by 155 boys and 148 girls.
We used a principal-components analysis with varimax orthogonal rotation to reduce the number of attitude items on the questionnaire. With Kaiser's (1960) criterion, nine factors with eigenvalues greater than 1.0 were extracted. Those factors accounted for 53.7% of the variance on the attitude questions.
On the basis of their responses to the attitude items, we divided the participants into two groups. The groups were constructed by using the SPSS Quick Cluster program and the factor scores for each participant on each of the attitude factors. Group 1, representing liberal attitudes toward violence on television, consisted of 183 children. Group 2 (120 children) contained children with more extreme views about violence on television. Though generally positive toward the media, these children were aware of and wary of violence on television, and they were generally more in favor of social control of the media. To confirm the adequacy of the clustering procedure, we conducted a series of nine univariate analyses of variance to test the differences between the cluster centers for each attitude factor. The cluster centers for the nine factors were significantly different, F(1,301) [greater than] 3.81, p [less than] .05.
Finally, we used the 17 items in Part 1 of the questionnaire as predictor variables in a stepwise discriminant function analysis. We conducted the analysis to identify which of the children's personal characteristics predicted their attitudes toward violence on television. One discriminant function was extracted, [[Chi].sup.2](1, N = 303) = 4.30, p [less than] .05. The amount of television watched on school days was the largest predictor variable and was entered into the discriminant analysis at Step 1 (Wilks's lambda = .99, p [less than] .05). No other variable achieved significance; thus, the function was determined solely by this primary variable, which had a unitary correlation with the function.
The overall pattern of results showed that the children in Group 1, those with more liberal attitudes toward screen violence, reported watching less television on an average school day (an average of 2.7 hr) than their more concerned counterparts (Group 2), who watched an average of 3.1 hr per day.
Although we examined a substantial number of personal characteristics, we found that the only variable that predicted children's attitudes toward violence on television was the number of hours spent watching television on school days. However, a preliminary exploration of the relationships among the predictor variables had shown a number of intercorrelations indicating that the impact of the other eight variables may have been mediated by the number of hours of television viewed on school days.
If, as suggested by Wiegman et al. (1992), children who watch more television are being vicariously exposed to more violence, control could focus on reducing either the absolute quantity of exposure or the violent content of programs. It is not surprising that the children in Group 2, who viewed a great deal more television than those in Group 1, had a higher sense of personal risk and suspicion and were more likely to perceive the world as a scary place (Singer et al., 1984).
It is encouraging that a substantial number of those children recognize that television has an impact on them and that there may be a need for some form of social control of television programming or watching. Somewhat less encouraging are reports that parental control over children's viewing has decreased over the past several decades and that those children who are less subject to parental control (and more able to avoid it) may be the ones who need it most (Barlow & Hill, 1985).
Barlow, G., & Hill, A. (1985). Video violence and children. London: Hodder and Stoughton.
Cashmore, E. (1994). . . . and then there was television. London: Routledge.
Cline, V. B., Croft, R. G., & Courrier, S. (1973). Desensitization of children to television violence. Journal of Personality and Social Psychology, 27, 360-365.
Doob, A. N., & Wood, L. E. (1972). Catharsis and aggression: Effects of annoyance and retaliation on aggressive behavior. Journal of Personality and Social Psychology, 22, 156-162.
Field, D. E., & Anderson, D. R. (1985). Instruction and modality effects on children's television attention and comprehension. Journal of Educational Psychology, 77, 91-100.
Huesman, L. R., & Eron, L. D. (1986). Television and the aggressive child: A cross-national comparison. New York: Erlbaum.
Kaiser, H. F. (1960). The application of electronic computers to factor analysis. Educational and Psychological Measurement, 20, 141-151.
Reid, I. (1989). Social class differences in Britain. Glasgow: Fontana.
Rubinstein, E. A. (1983). Television and behavior. American Psychologist, 38, 820-825.
Singer, J. L., Singer, D. G., & Rapacynski, W. S. (1984). Family patterns and television viewing as predictors of children's beliefs and aggression. Journal of Communication, 34, 73-89.
Turner, C. W., Hesse, B. W., & Peterson-Lewis, S. (1986). Naturalistic studies of long-term effects of television violence. Journal of Social Issues, 42, 51-73.
Van Evra, J. (1990). Television and child development. Hillsdale, NJ: Erlbaum.
Wiegman, O., Kuttschreuter, M., & Baarda, B. (1992). A longitudinal study of the effects of television viewing on aggressive and prosocial behaviour. British Journal of Social Psychology, 31, 147-164.
Wober, M. (1988). The extent to which viewers watch violence-containing programs. Current Psychology, 7, 76-92.
Please direct inquiries about this page to Dale Vidmar.