A handful of content analysis studies of violence on television have been carriedout in Britain since 1970.
Much of this work has been funded by the broadcastingindustry. Halloran and Croll (1972) analysed broadcast material on BBC1 andthe ITV midlands region during one week of April 1971. They coded violentincidents occurring in fictional drama programmes and news, current affairs anddocumentary programmes. The coding instruments for violence were based onthose used by Gerbner and programmes were coded for the amount and type ofviolence they contained, the kinds of characters involved in violence and certainthemes or contexts in which violent events happened. At around the same time the BBC’s Audience Research Department undertooka more extensive analysis of television over a six-month period and analysedcontent from dramatic fiction, news and current affairs, documentaries, and lightentertainment programmes (Shaw and Newell, 1972). Unlike Halloran and Croll,the BBC researchers did not follow the Gerbner method precisely, although ingeneral terms their analysis was conceptually of the same type. Violence wastreated essentially as an explicit form of interaction between human beings.
However, because the BBC’s study sampled materials from a broader spectrumof programme types than did Gerbner’s, it was considered necessary to expand the definition of violence to include unintentional injurious acts and acts ofvandalism against property to code incidents of the sort often reported ordepicted in non-fictional programmes. Each of these British studies indicated that although a common feature ofprogramming, violence was not as prevalent on British television as on Americantelevision. In fictional drama, the only category of programming in which directcomparisons between American and British findings were possible, Halloran andCroll found that nearly 56% of programmes coded contained violence—compared with a reported 80% of fictional programmes on American television(Gerbner, 1972). In nearly 48% of fictional programmes, a major character was involved inviolence either as aggressor or victim. The rate of violence on British televisiondrama was 2.8 incidents per programme or just over four incidents an hour, againsubstantially less than an average rate of over seven incidents per hour reportedby Gerbner (1972) for American prime-time television drama.
In their more extensive analysis of British television content, Shaw andNewell found a somewhat greater prevalence of violence in fictional drama interms of the number of programmes containing at least one major violent incident(63%), but somewhat less violence in terms of the rate at which violent incidentsoccurred; average numbers of just under two violent incidents per programme orjust over two per hour were recorded. The prevalence and rate of violent incidents was also found in both studies tovary considerably across different forms of drama programme. Halloran andCroll noted that the most violent type of programme in the fiction category wasthe cartoon.
All cartoons coded by them contained some violence and the rate ofincidents per hour in such shows was nearly 34. None of the cartoon incidentswas fatal, however. Violence was also prevalent in feature films; 80% contained some violence,and incidents averaged over four per hour. This contrasted with plays, of which50% were violent, but even then, with an average incident rate of less than oneper programme, were unlikely to be very violent programme by programme.After cartoons, the most violent programmes coded by Halloran and Croll werecrime, western, and action-adventure shows. All programmes in these categoriescontained violence with an average rate of nearly eight incidents per hour. Turning to the BBC’s analysis of the prevalence of violence for differentprogramme forms or ‘themes’, once again, agreement with Halloran and Crollwas not complete.
Feature films contained at least one major violent incident in86% of cases which compares quite well with Halloran and Croil’s 80% figure,whereas the category of programmes labelled as ‘TV series’ by the BBC’sresearchers, which consisted largely of crime-detective, western and actiondramas, contained violence in only 75% of cases, as contrasted with all suchprogrammes monitored by Halloran and Croll. Inconsistencies across the violence measures produced by these two studiesbecome even more pronounced when consideration is given to mean rates of INTRODUCTION 29 30 INTRODUCTION violent incidents per programme type. The BBC team coded an average of justover three incidents per hour for feature films, one per hour less than Halloranand Croll. Even more significantly, they coded 2.2 incidents per hour for ‘TVseries’, a little over 25% of the average rate of incidents per hour recorded for thecrime-drama, western and action-drama categories by Halloran and Croll. Evenwhen examining particular themes or particular series, the BBC’s violence ratestatistics remained lower than those reported by Halloran and Croll. Thus, theBBC reported that westerns contained a higher rate of violent incidents than anyother type of television series, but at 3.9 incidents per programme, this was stillonly about half the rate reported by Halloran and Croll for crime-drama, westernand action-drama categories.
Cumberbatch et al. (1987) found an overall rate for dramatic fiction of 3.60violent acts per hour which reduced to 2.51 in prime time (8.00 pm to 11.
00 pm).These findings compared with Shaw and Newell’s (1972) result of 2.74 violentacts per hour for dramatic fiction, and the parallel analysis by Halloran and Croll(1972) who found 4.04 acts per hour. Cumberbatch et al. (1988) suggested thatthere may have been a degree of undercoding in the original BBC study.
The Cumberbatch et al. (1987) study found a decline in television violenceeven against the depressed rates of Shaw and Newell. Thus, overall, theproportion of programmes with violence in them declined considerably. The twomajor channels, BBC1 and ITV, showed decreases from 49% to 28% and from51% to 32%, respectively, while the minority channel, BBC2, showed a declinefrom 55% to 26%. The least decline was in dramatic fiction where 63% ofprogrammes contained violence in 1971 compared with 58% in 1986. Otherprogramme genres that could be directly compared showed a consistent pattern ofreduction in violence. Light entertainment reduced from 30% to 26%; currentaffairs (magazine) programmes reduced from 56% to 40%; and cartoons reducedfrom 89% to 63% in the proportion of programmes within the genre thatcontained any violence. During the early 1990s, a series of monitoring studies were published by theBroadcasting Standards Council in Britain.
These studies monitored not only howmuch violence there was on television, but also how much sex and bad language.The first of these studies monitored television output over four separate two- orthree-week periods in 1991 and 1992. The second study examined televisionouptut over a single two-week period in 1993, and the third study monitored twoseparate two-week spells in 1994. This exercise involved a survey of panels ofrespondents who were interviewed about the programmes they watched and kepta viewing diary. In 1992, a content analysis exercise was also run in parallel tothe viewer-based monitoring exercise for one of the sampled weeks, whichassessed the amount and distribution of violence on the four UK terrestrialchannels, BBC1, BBC2, ITV and Channel 4. In 1993 and 1994, the contentanalysis was repeated and extended to cover two separate weeks each year, oneof which also covered output from three satellite movie channels, Sky Movies, The Movie Channel and Sky Movies Gold (see Broadcasting Standards Council,1994, 1995).
The content analysis conducted in 1994 sampled a total of 648 programmesover two weeks of terrestrial television output, focusing on transmissionsbetween 5.30 pm and midnight every evening. A further 168 programmes werevideo-recorded from the three satellite channels during the second of these twoweeks. During the first week of analysis in 1994, a total of 603 violent scenes werefound in 165 programmes (51% of all programmes monitored on the fourterrestrial channels during that week). These figures compared with 573 violentscenes in 162 programmes (53% of all programmes monitored) for the sameweek in 1993. The violence in both years covered 3% of broadcast time, withfictional drama programmes and films contributing the greatest amount of anyprogramme genres to overall violence levels.
During the second week analysed in 1994 and 1993, which also coincided withthe single week content analysed in 1992, further year-on-year comparisons weremade. It was reported that 165 programmes monitored in 1994 (51% of total)contained 648 violent scenes, compared with 155 programmes (50% of total) in1993 which contained 666 violent scenes, and 143 programmes in 1992 (43%)which contained 478 violent scenes. The total amount of violence thereforeexhibited a steady increase during evening broadcast hours on the four terrestrialchannels. This increase was further reflected in changes to the rates ofoccurrence of violent scenes per hour, recorded as 4.0 per hour in 1994, 4.
0 in1993, and 2.9 in 1992. In 1992, violent scenes occupied 2% of broadcast timemonitored, increasing to 3% in 1993 and to 4% in 1994. Certain programmetypes appeared to contribute disproportionately to this rise in violence on thesmall screen. Between 1993 and 1994, for this particular week, rates of violentscenes per hour fell for all major categories of programmes (e.
g., national news,factual programmes, light entertainment, sport, religion, children’s programmesand fictional drama) except one—films. Films exhibited a marked increase, year-on-year, in levels of violence, from 4.
0 violent scenes per hour in 1993 to 7.1 in1994. Levels of violence on the three satellite movie channels were generally muchhigher than the level on terrestrial channels. In 1994, nearly nine out of ten(87%) programmes monitored contained some violence.
This represented anincrease on 1993 (73%). In 1994, however, the numbers of violent incidents (1,107) fell, year-on-year, from the 1993 level (1,139). The rates of occurrence ofviolent scenes per hour, however, remained fairly stable at 7.8 in 1993 and 7.6 in1994. BROADENING THE DEFINITION OF VIOLENCE The quantitative measurement of violence on television via content analysis hasbeen largely a process of defining the concept of violence clearly and applying INTRODUCTION 31 32 INTRODUCTION the definition in an accurate and consistent manner to television programmes.What the commonly assumed definitions overlook is that there are many formsof violence other than that which involves purely physical injury and harm.There is emotional and psychological violence, verbal violence, institutional andsymbolic violence.
Of course, these can be as difficult to define and record as ageneral concept of ‘violence’ as physical force. As Hodge and Tripp (1986) pointed out, violence can be understood as amulti-faceted concept which does not represent a unitary process or a single set ofevents. Violence can vary in its severity, justification, consequences, and theintentions of the perpetrator. While physical violence may be the mostcommonly perceived form of violence on television, a more complete measuringprocedure also might include other expressions of violence such as verbalviolence and violent images. Whether or not the use of force or infliction of harm or injury is perceived asviolent clearly depends on a number of considerations associated with theparticular circumstances surrounding the action.
The total context in which theaction takes place exerts a significant influence as to how the viewer willinterpret the episode or image. Violence on television cannot be taken simply atface value. How violent actions are perceived is related to social norms, personalvalues and the particular form and context of violence itself. Context is known to be particularly important to viewers’ judgements aboutviolence on television. Whether violence occurs in a realistic or fantasy context,or in a serious or humorous context can make a significant difference to the wayit is regarded by television audiences.
The perceived realism of the situation inwhich violence on television is depicted tends to be significantly correlated withhow violent the behaviour is rated by viewers (Gunter, 1985). Humour can alsodilute the degree to which violent behaviour is upsetting to viewers. A recentopinion poll in the United States indicated that 72% of Americans felt thattelevision entertainment contained too much violence and 80% believed thattelevision violence was harmful to society. Most people were not upset,however, about violence that appeared in a humorous context (Galloway, 1993).
Research elsewhere has shown that people may not even see much of theviolence when it is presented in a humorous way. Most viewers, for example, donot regard cartoons as violent (Gunter, 1985; Howitt and Cumberbatch, 1974),despite the fact that content analyses have reported far higher rates of violence oncartoons than anywhere else on television (e.g., Halloran and Croll, 1972;Gerbner et al., 1978, 1979). Some researchers have even argued that violencedoes not exist in humorous contexts such as cartoons, and that content analysisstudies which report high levels of violence in cartoons must therefore beemploying dubious violence definitions and coding frames of doubtful veracity(Morrisson, 1993a).
Contextual characteristics of violence With violence, the literature suggests a number of contextual characteristics thatprovide cues that viewers use in interpreting the meaning of televisionportrayals. Among these are reward, intention, motive, remorse, consequencesand presentation style. One of the most important contextual characteristics of any portrayal iswhether the action is rewarded or punished. Social learning theory predicts thatwhen people watch a behaviour that is rewarded, they are much more likely tolearn it than if the behaviour was punished (Bandura, 1977). Reward has featuredas a component of analysis in quantitative studies of television violence. In thatcontext, it has been thought by some researchers to exhibit a potential to contributetowards the development of antisocial behaviour among viewers.
In a study ofviolence on Canadian television, Williams et al. (1982) found that violence wasoften portrayed as a successful way to solve conflicts. This observation wasreinforced by American research indicating that the great majority of antisocialacts depicted on television was shown as being rewarded (Potter and Ware,1987). Much content analysis research in the past has indicated that violence ontelevision is mostly intentional.
Williams et al. (1982) found that in 81 hours oftelevision output they examined, over 97% of the violent acts were intentional.While violence may be intentional, it is important to know what a character’sintentions were. If a character commits a violent act accidentally, viewersinterpret this differently from if the character was portrayed carefully planningthe act. In general, intentions to commit violence have been found to be dividedinto three types (Mees, 1990). Using a social norms approach to defineaggression, Mees found three modes of intention that underlie conceptions ofmotivation for aggressive acts: (1) thoughtlessness (the aggressor should havetaken possible dangers into consideration but did not); (2) selfishness (theaggressor knows that the action will cause distress or harm, but accepts this andplaces his/her own interests above those of others); and (3) malice (wickedness isaccepted and intended by the aggressor). Motive is closely linked to intention.
A television character’s motives arefiltered through the viewer’s legal or moral context of behaviour (Gunter, 1985).Defensive or altruistic aggression may be interpreted as milder than offensive,intentional, or sadistic aggression. Gunter (1985) argued that unusual forms ofaggression in which a strong motive (sadistic or sexual) is apparent are perceivedas more serious. Thus, the intention of the character can alter a viewer’sinterpretation of a violent episode. In a study of Finnish television, Mustonen and Pulkinen (1993) measuredmotivation by dividing aggression into defensive and offensive groups.
Withinoffensive aggression there were five values: instrumental, masochistic, reactive-expressive, sadistic and altruistic. They reported finding that spontaneous acts(57%) were more numerous than planned aggression (27%) and that first strike