There is and has been much debate on how much, if any effect media violence has on children and their development. There are many studies suggesting that it does have an effect, and even demonstrating these effects. The NIMH report (1982) went as far as to say, "In magnitude, television violence is as strongly correlated with aggressive behaviour as any other variable that has been measured" (pp. ? ? ? ? Hickey, 1995). Meanwhile, the makers of video games will argue that the statistics against media violence are dubious, and that the only aggression shown is good-natured fun (Reichhardt, 2003). In this essay I will look at the many ways of studying the effects of media violence, broadly encompassing films, video games and television violence, and the methodological issues surrounding each.
The main methods of studying the effects of media violence on children are observations (in a laboratory and in naturalistic settings), case studies, psycho-physiological measurements and longitudinal studies. Each method has its own advantages in obtaining the information being looked for, and each has inherent problems, whether it be reliability, validity or ethics. Bushman (1995) conducted an observational study in a laboratory setting where participants watched either violent or non-violent videotapes. After each viewing they competed against an opponent on a reaction time task where the loser was blasted with an unpleasant noise (Bushman, 1995). Because the environment was controlled and structured, it was possible to make direct comparisons and draw the conclusion that participants would blast their opponents with significantly louder noise than when non-violent videotapes had been watched.
The structured environment can also provide an equal opportunity for each individual participant to complete the same tasks, making comparison and interpretation of behaviour easier. It would also be possible for this experiment to be replicated and the results compared. There are a number of problems associated with this, though. To begin with, it is difficult to make general, real-life applicable statements and conclusions about research conducted in such an unnatural setting, where natural behaviour is less likely due to unfamiliar surroundings, and a situation that isolates just one aspect of a possible real-life experience completely out of context. Another issue is that the link between watching of violent and non-violent videotapes and aggressive behaviour is only a correlation, so causality cannot be established. Observations also take place in more naturalistic settings, which tend to reflect the participants' normal behaviour more accurately than a laboratory setting.
The disadvantage with observing in a naturalistic setting is that the environment cannot be controlled, and therefore it may be more difficult to draw clear conclusions. There are problems inherent in using observation to study the effects of media violence on children. Observation studies often involve sampling by looking for a particular behaviour, for example aggression, to occur in a given time frame, or by observing participants for specific target behaviour for as long as it takes for this behaviour to occur. So if a participant watches a violent or non-violent video, or plays a violent video game, he or she may be observed for a given time period. If the participant there may be no conclusion or an incorrect conclusion made about the effects of media violence on that particular participant displays no aggressive behaviour, however the participant may display aggressive behaviour out of that time; it is not representative. Another factor that may affect a child's aggressiveness is whether or not the child identifies with the aggressor.
The more the child identifies with the aggressor, the more likely it is that aggressive behaviour will follow by the child. Similarly, aggressive behaviour is only likely to follow if the media violence shown is justified in the film / game /television show (Turner and Berkowitz, 1972). In one experiment, children's television viewing was controlled; one group watching violent cartoons and the other watching non-violent cartoons for a specified time each day. The amount of aggression displayed by the children in their daily activities was "carefully recorded" (Steuer, Applefield & Smith, 1971). But who was to say what was to be classified as aggressive behaviour? And were there levels of aggression or was throwing a toy as aggressive as kicking somebody? Durkin (1999) observed participants playing video games. His interpretation of the behaviour seen was "kids having a good time." He described the aggressive behaviour being displayed as being "generally good-natured, and accompanied by laughter." The main type of aggression observed was "robust treatment of the equipment." This could be concerning for a number of reasons.
It shows the effect of observer bias. Observation is an action, and the observer carries his / her own ideas about what they are expecting or want to see. In this case, Durkin may well have been in receipt of funding for his study by the entertainment industry, in whose interests it would be not to find any strong link between video game violence and aggressive behaviour (Reichhardt, 2003). As seen in this example, interpretation of observed behaviour is subjective, and therefore another may ignore what is deemed as significant by one observer, so we cannot be sure that we are always getting the full picture as all behaviour is unlikely to be recorded.
It is not only the behaviour that is in danger of being viewed subjectively, but also the media. For example, is Pac Man gobbling a ghost a violent event? So would a case study method be more useful? We would be more likely to get a full, detailed picture of the child, and it would be easier to establish the role of individual factors on the child. By being able to observe a child's behaviour in a variety of everyday situations, we could see what effect watching violence on television, video or video games, has on the child and their behaviour. The type of child being observed in case study, however, is likely to have been selected for his / her unique characteristics.
For example the killers of two year-old Jamie Bulger, Robert Thompson and Jon Venables, received much attention and interest due to the relative uniqueness of the case. Detective-Sergeant Phil Roberts, who had been present at the interviews of Robert Thompson, said "These two were freaks who just found each other. You should not compare these two boys with other boys - they were evil" (The Independent, 25 November, 1993). These two boys were not specifically case studies before their terrible crime was committed, however they were studied in similar way with retrospect. Their lives were analysed in great detail, from the lack of nurture from their parents, to the violent videos the boys had been watching, and used as inspiration for their cruel attack on Jamie Bulger (Newson, 1994).
The problem here however, is that such specific studies and findings cannot be generalised and compared to other children. Particiapnts may be volunteers, which in itself may provide a biased sample of people with unique characteristics not because they have been specifically selected, but because of the characteristics of a person that would lead them to volunteer for psychological research in the first place. Volunteers have been found to be more easily influenced, anxious for approval, neurotic, moody and aggressive than non-volunteers (Ora, 1965). Another reason that participants may not necessarily be representative is that they may have been selected due to ease of availability.
Many experiments for example have been conducted on students. One study was conducted on boys in private boarding schools and care homes for neglected children where half the teenagers were allowed to watch violent television and the other half were not (Fesbach and Singer (1971). This study is ethically dubious. Knowing that there may be negative effects and consequences on and for the participants, the study still went ahead, and it is not clear whether or not the participants were aware that they were being observed in this way, in their own homes.
In the likely case that they were not, in order to eliminate as far as possible demand characteristics, this would consist a breach of and encroachment upon personal privacy. A more easily comparable method of studying the effects of exposure to media violence is psycho-physiological measurement. These measurements can show links between parts of the central nervous system and are much easier to generalise and compare between individuals. Bushman (1995) conducted a number of studies on participants watching violent videotapes and measured their cardiovascular arousal. He found that there was indeed a difference between participants who had watched violent videotapes and those who had watched non-violent videotapes, however the physiological effects are the same for sexual arousal and aggression, so it is difficult to determine the exact cause of the effects observed.
One big limitation of this method of studying the effects of exposure to media violence is the amount of time and the cost of carrying out such research. Lefkowitz et al (1997) observed a positive correlation between the amount of violence viewed by boys and their later behaviour; the greater the boys' preference for watching violent television at age 8, the more aggressiveness was displayed at this age and ten years later, at age 18. (There was no significant relation for girls). We may draw the conclusion from this, then that preference for and watching violent television can lead to more aggressive behaviour in boys, but what caused the boys' preference for watching aggressive television in the first place? The cause of this preference could also be a predisposing factor of aggressive behaviour. Another problem with longitudinal studies is that the most interesting participants often drop out before the end of the study! Time and cost are also major limitations of longitudinal studies. The advantages are that we can observe the consequent effects of events in the lives of the same individuals and groups of people.
We can reliably trace their development in regard to these events, however we cannot definitely say that these events are the cause of the observed behaviour. Bushman (1995) looked at the relationship between trait aggressiveness and the effects of media violence. Having read film descriptions, high trait aggressive participants were more likely to choose a violent film to watch than were low trait aggressive participants. In a later part of the study (discussed earlier in this essay) they also displayed more aggressive behaviour when competing against an opponent on a reaction time task. This relationship between trait aggressiveness and desire to view violent media is particularly interesting when we consider the validity of such studies.
As discussed earlier, voluntary participants have a higher tendency towards aggression than do non-volunteers, so this would need to be taken into account when considering whether or not we have a representative sample, and the effect of the sample bias on the validity of the experiment. Again, this link is only a correlation, but it does offer another angle from which to view the effects of media violence on children, and suggests that other factors are likely to be at work. There are serious ethical issues concerning studying the effects of media violence on children. If we are expecting that there is a high likelihood of negative effects on the child, how is it possible to conduct ethical research? In some studies, parents have been present, however the presence of parents has been seen to reinforce the effects by making the child think that if the parent is present during the viewing, there is nothing wrong with it, and this acts as a positive reinforcement. (Berk, 2003). It is unethical to conduct a study requiring children to watch material that is inappropriate for their age, as children often do not have the life experience necessary to view some scenes in context.
As has been shown by many studies, this may have the effect of increasing aggression in the child. Other ethical issues, as discussed in reference to Fesbach and Singer's study on boys in boarding schools and residential Hampshire and London homes (1971), are issues of participant knowledge of the study and breach of personal privacy. Laboratory based observational studies have the advantage of being able to record comparable information about participants, about which conclusions can be drawn, with the disadvantage that the behaviour may not be typical in real life. Observational studies in naturalistic settings have the advantage of recording more natural and typical behaviour, but it not so easily comparable.
Observations provide a good snapshot of behaviour at any one time, so a participant may behave aggressively after watching a violent movie at the time of the experiment, but may not at any other time. Longitudinal studies are advantageous in this respect, with the disadvantages of time, cost, drop out rate and dubious links between events and subsequent behaviour. Case studies have the advantage of providing an in-depth picture, but cannot be generalised to the general population. All the comparable results from these studies are correlations and therefore we cannot establish cause and effect. However there are so many studies of all types as to suggest that there is almost certainly a direct link between media violence and children's behaviour. So while each individual study has its methodological flaws, I think that collectively all the studies can suggest a direct link.
References Berk, L. E. , (2003) Child Development Allyn and Bacon: Boston Bushman, B. J. (1995) The moderating role of trait aggressiveness in the effects of violent media on aggression Journal of Personality and Social Psychology 69, 950-960 Durkin, K. & Ais bett, K.
(1999) Computer Games and Australians Today Office Film Literature Classification, Sydney cited in Nature 424, 6497, 367-368 Fesbach S. , and Singer, R. D. (1971) Television and aggression. San Francisco: Jossey Bass cited in Carlson, N.
R. , Bus kist, W. & Martin, G. N.
Psychology The Science of Behaviour (2000) Massachussetts: Allyn and Bacon Hickey, N (1995) How Much Violence Is There? Impact of Mass Media Eldon, R. editor New York: Longman Publishers Lefkowitz, M. M. , Er on, L. D. , Walker, L.
O. & Hues mann l. R. (1997) Growing up to be violent: A longitudinal Study of the development of aggression. New York: Pergamon Press, cited in Psychology: The Science of Behaviour (European adaptation) Ora, J. P.
(1965) Characteristics of the volunteer for psychological investigation. Office of Naval Research Contract 2149 (03), Technical Report 27 cited in Malim, T & Birch, A. (1998) Introductory Psychology. Hampshire and London: Macmillan Press Ltd Reichhardt, T. (2003) Video Violence: Playing With Fire? Nature 424, 6497, 367-368 Steuer, F. B.
, Applefield, J. M. & Smith, R. (1971) Televised aggression and the interpersonal aggression of pre-school children Journal of Experimental Child Psychology 11, 431, 422-447, cited in Atkinson, R. L. , Atkinson, R.
C. , Smith, E. E. , Bem, D. J. , Hillard, E.
R. (1990) Introduction to Psychology. Florida: Harcourt Brace Jovanovic h Turner, C. W. & Berkowitz, L (1972) Identification with film aggressor (covert role taking) and reactions to film violence.
Journal of Personality and Social Psychology 21, 256-64, cited in Malim, T & Birch, A. (1998) Introductory Psychology Hampshire and London: Macmillan Press Ltd.