\ , TELEVISION AND GROWING UP: THE IMPACT OF TELEVISED VIOLENCE REPORT10 THE SURGEoN GENERAL UNITED STATES PUBLlC HEALTH SERVICE FROM THE SURGEON GENERAL'S SCIENTIFIC ADVISORY COMMITTEE ON TELEVISION AND SOCIAL BEHAVIOR U.S. DEPARTMENT OF HEALTH, EDUCATION, AND WELFARE Health Services and Mental Health Administration National Institute of Mental Health 5600 Fishers Lane Rockville, Maryland Staff Members Eli A. Rubinstein George A. Comstock John P. Murray Michael Adler Eileen Marchak Susan S. Lloyd-Jones Joseph D. Reckley Margaret D. Salladay Laura A. De Lisi Vice Chairman. Surgeon General's Scientific Advisory Committee Senior Research Coordinator Research Coordinator Staff Assistant Research Assistant Editor Administrative Officer Secretary Secretary Former Staff Members Douglas A. Fuchs John P. Robinson Harold Leigh Thomas Brubeck Deborah Cutler Jan W. Lipkin Senior Research Coordinator (through 6/7(J) Research Coordinator (through 9/70) Administrative Officer (through IO/701 Information Officer (through 5/7 1) Research Assistant (through 8/70) Secretary (through 4/70) Richard A. Moore HEW Secretary's liaison with Advisory Committee (through spring 1970) Advisory Committee Members Ira H. Cisin George Washington University Thomas E. Coffin National Broadcasting Company Irving L. Janis Yale University Joseph T. Klapper Columbia Broadcasting System, Inc. Harold Mendelsohn University of Denver Eveline Omwake Connecticut College Charles A. Pinderhughes Boston University Ithiel de Sola Pool Massachusetts Institute of Technology Alberta E. Siegel Stanford University Anthony F. C. Wallace University of Pennsylvania Andrew S. Watson University of Michigan Gerhart D. Wiebe Boston University . . . 111 Letter of Transmittal December 31, 1971 Dr. Jesse L. Steinfeld Surgeon General Department of Health, Education, and Welfare Washington, D.C. 20201 Dear Dr. Steinfeld: We are pleased to transmit our report on the research available in our study of television and social behavior. We have been careful to keep in mind that this committee was estab- fished as a scientific body. Our major concern has been to assess the research carefully and come to conclusions justified by the data. As the report shows, this has been a very complex issue, for which there are no simple answers. We trust that this report will help to advance the understanding of these complexities. Respectfully submitted. Alberta E. Sleael. Ph.% iv Foreword This report is the result of over two years of effort by a distinguished committee of behavioral scientists. Their task has been difficult. The impact of televised violence on the viewer, as a reading of the report will show, is embedded in a complicated set of related variables. The conscientious effort by the committee to avoid an oversimplifi- cation of the problem has produced a document which may seem, at times, too technical. However, I believe that this report and the five vol- umes of research reports, which serve as a basis for the committee con- clusions, make a major contribution to an understanding of the role of television in influencing the social behavior of children and young peo- ple. The conclusions reached by the committee are carefully worded and merit the serious attention of all persons and groups concerned about the effects of viewing television. As the committee notes, these conclu- sions are based on substantially more knowledge than was available when the committee began its deliberations. But the research still leaves many questions unanswered. Without detracting from the importance of its conclusions, the committee specifies some of these unanswered ques- tions and urges that they be addressed in the future. This report will undoubtedly be scrutinized carefully by people who will be looking for support for their own prior point of view. Individuals with strong convictions on either side of the question about the effects of televised violence may not be satisfied. What these individuals will fail to recognize is that this set of conclusions, for the first time in this field of inquiry, sets a solid and extensive base of evidence in an appro- priate perspective. In that sense, the report and the research on which it is based represent a major contribution. V The committee is to be congratulated for the work it has done. The successful conclusion of the task is even more significant because of the explicit consensus among so broadly representative a group of scien- tists. I wish to commend the committee, the researchers, and the staff for a job well done. vi PREFACE All the available statistics confirm the pervasive role television plays jn the United States, if not throughout the world. More people own tele- vision sets and more people watch television than make use of any other single mode of mass communication. It is no wonder then that television is the subject of much attention, both directly as it serves its purpose and indirectly as a source of con- cern to examine how well it serves its purpose. All manner of inquiry &out the input of television on the lives of the American public has been and is being made. The issues about public television, cable televi- sion, and the role of television in election campaigns are all in the news today. The question of violence on television has been one issue that was raised almost immediately after television became a major contender for the leisure time and attention of the public. There have been a number of prior public examinations of this issue, and a number of statements and conclusions have been made. The committee has taken into account these earlier studies in reaching its own conclusions. We have also had the benefit of an extensive body of new data which we have carefully examined. A great deal of work is reflected in the pages of this report and in the concurrently published five volumes of technical reports, which have served as the major source of new information. We believe this work makes a major contribution to this area of scientific inquiry, and we wish here to acknowledge our indebtedness to the researchers and staff who brought that research to a successful conclusion. Our task has not been easy. We have tried to come to as carefully objective a conclusion as the data warranted. We suspect the debate will not end here. We are dealing with a complex and changing set of phe- nomena. Reassessment is inevitable as new evidence becomes available and as changes occur in what television presents and how it is presented. Our report consists of two parts: a Summary of Findings and Conclu- sions and a detailed report. vii Table of Contents Foreword ............................................................................. V preface ................................................................................ Summary Chapter: Findings and Conclusions ............................ Vi;' Chapter 1: Chapter 2: Chapter 3: Chapter 4: Chapter 5: Chapter 6: Chapter 7: Chapter 8: Chapter 9: References Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 Violence in Society and in the Television Medium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 Some Problems of Research on the Impact of Television . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Television Content . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . i: Changing Patterns of Television Use . . . . . . . . . . . . . . . . . . . . . . . . . 51 Television and Violence in the World of Children . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61 Television and Adolescent Aggressiveness . . . . . . . . . . . . . . . . 77 Current Knowledge and Questions for Future Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...*.*. . . . . . . . . . . . . . . 111 The Unfinished Agenda . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117 ,.......,.................................................................., 129 Appendices A: Surgeon General's Scientific Advisory Committee on Television and Social Behavior- Initial Operations, June-October, 1969 . . . . . . . . . . . . . . . . . . . . . . . . . . . 139 B: Television and Social Behavior Program Reports and Papers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147 C: Experiments on Children's Imitation of Aggressive Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..I 159 D: Experiments on Disinhibition of Aggressive Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161 E: The Interpretation of Correlation Coefficients . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165 The Report 1x Summary of Findings and Conclusions The work of this committee was initiated by a request from Senator John 0. Pastore to Health, Education, and Welfare Secretary Robert H. Finch in which Senator Pastore said: 1 am exceedingly troubled by the lack of any definitive information which would help resolve the question of whether there is a causal connection between tele- vised crime and violence and antisocial behavior by individuals, especially chil- dren. 1 am respectfully requesting that you direct the Surgeon General to appoint a committee comprised of distinguished men and women from whatever professions and disciplines deemed appropriate to devise techniques and to conduct a study under his supervision using those techniques which will estab- lish scientifically insofar as possible what harmful effects, if any, these pro- grams have on children. The question raised by this request has been this committee's central concern. However, the research program that was undertaken has at- tempted to place this question within a larger context. For this reason, the committee's title deliberately emphasizes more than the issue of tel- evised violence and aggressiveness and more than the question of televi- sion's harmful effects during childhood and youth. At the same time the committee was explicitly enjoined from drawing policy conclusions. Our task has been to state the present scientific knowledge about the effects of entertainment television on children's behavior, in the hope that this knowledge may be of use to both citizens and oficials concerned with policy. The findings we will summarize represent the issues and questions treated in the body of the report. They derive primarily from the re- search conducted under this program but take account also of past re- search and other current research. THE TELEVISION EXPERIENCE It would be difficult to overstate the pervasiveness of television in the United States. Census data indicate that 96 percent of American homes 2 TELEVISION AND GROWING UP have one or more television sets. The average home set is on more than six hours a day. Most adults report watching at least two hours daily. Most children also watch at least two hours daily. For most people, whatever their age, television viewing is a daily experience. Although not everyone watches every day, many watch for much longer than two hours. Television viewing stands in sharp contrast to the theater, movies, and other entertainment presented outside the home in that it does not usually involve such exclusive or focused attention. Viewers of all ages regularly engage in a wide range of activities while the set is on. The extent to which this discontinuity of attention alters what would be perceived and understood from television were attention undivided is a moot question. Young children before the age of six usually cannot successfully divide their attention. As a result, what they get from tele- vision is probably generally restricted to what is taken in while viewing with full attention and is perceived bereft of a larger context. As the child grows older, he becomes more able to follow at least the rough continuity of what is taking place on television while he is simultaneous- ly doing other things. The casual acceptance of viewing, however, does not equal indiffer- ence to television. By the first grade, a majority of boys and girls exhibit individual taste in program selection and preference for characters. Among younger children, situation comedies and cartoons are most popular. Sixth graders like family situation comedies and adventure programs. Tenth graders prefer adventure programs and music and vari- ety programs. Children and adolescents are attracted to programs fea- turing characters their own age. The propensity to view television changes as the individual goes through the major stages of maturation. Frequent viewing usually begins at about age three and remains relatively high until about age 12. Then viewing typically begins to decline, reaching its low point during the teen years. When young people marry and have families, the time they spend viewing tends to increase and then remain stable through the middle adult years. After middle age, when grown children leave home, it rises again. Many questions about television are presently unanswerable. Three basic ones concern the future character oftelevision,the influences and dynamics involved in the choosing of programs by individual viewers, and the underlying needs served by television that lead to its present extensive use. It would appear that television, like other media, is progressing through a series of stages from intriguing novelty to accepted common- place to possible differentiation as a servant of varied tastes. New devel- opments-UHF, public television, cable, cassettes, portable minisets- suggest that in the future the programming available may become in- creasingly varied and that the mass audience may become a diversity of FINDINGS AND CONCLUSIONS 3 smaller segments, each with its special interests. Newspapers, maga- zines, and radio provide examples of similar evolution. Why people choose to view what they do, and why they view so much, remain open questions after 20 years of commercial broadcast- ing. From the various rating services it is easy to deE?mine what audi- ences choose to view from among what if offered. The process by which choices are made, and the basic appeal that leads to persistent viewing at all ages, remain obscure. VIOLENCE ON TELEVISION Studies of media content show that violence is and has been a promi- nent component of all mass media in the United States. Television is no exception, and there can be no doubt that violence figures prominently in television entertainment. People are probably exposed to violence by television entertainment more than they are exposed by other media because they use television so much more. In regard to dramatic entertainment on television, and with violence defined as "the overt expression of physical force against others or self, or the compelling of action against one's will on pain of being hurt or killed," an extensive analysis of content has found that: -The genera! prevalence of violence did not change markedly be- tween 1967 and 1969. The rate of violent episodes remained constant at about eight per hour. -The nature of violence did change. Fatalities declined, and the proportion of leading characters engaged in violence or killing declined. The former dropped from 73 to 64 percent; the latter, from 19 to five percent. The consequence is that as many violent incidents occurred in 1969 as in 1967, but a smaller proportion of characters were involved, and the violence was far less lethal. -Violence increased from 1967 to 1969 in cartoons and in come- dies, a category that included cartoons. --Cartoons were the most violent type of program in these years. Another study concluded that in 1971 Saturday morning program- ming, which includes both cartoons and material prepared for adults, approximately three out of ten dramatic segments were "saturated" with violence and that 71 percent involved at least one instance of hu- man violence with or without the use of weapons. There is also evidence that years high in violence also tend to be years high in overall ratings, and that the frequency of violent programs in a year is related to the popularity of this type of program the previous year. This suggests that televised violence fluctuates partly as a function of the efforts of commercial broadcasters to present what will be maxi- mally popular. 4 TELEVISION AND GROWING UP TELEVISION'S EFFECTS Television's popularity raises important questions about its social effects. There is interest and concern in regard to many segments of the population-ethnic minorities, religious groups, the old, the unwell, the poor. This committee has been principally concerned with one segment, children and youth, and in particular with the effects of televised vio- lence on their tendencies toward aggressive behavior. People ask behavioral scientists various questions about television and violence. In our opinion the questions are often far too narrowly drawn. For example: (1) It is sometimes asked if watching violent fare on television can cause a young person to act aggressively. The answer is that, of course, under some circumstances it can. We did not need massive research to know that at least an occasional unstable individual might get sufficiently worked up by some show to act in an impetuous way. The question is faulty, for the real issue is how often it happens, what predispositional conditions have to be there, and what different undesirable, as well as benign, forms the aggressive reaction takes when it occurs. (2) It is sometimes asked if the fact that children watch a steady fare of violent material on television many hours a day from early childhood through adolescence causes our society to be more violent. Presumably the answer is, to some degree, "yes," but we consider the question mis- leading. We know that children imitate and learn from everything they see-parents, fellow children, schools, the media; it would be extraordi- nary, indeed, if they did not imitate and learn from what they see on tel- evision. We have some limited data that conform to our presumption. We have noted in the studies at hand a modest association between viewing of violence and aggression among at least some children, and we have noted some data which are consonant with the interpretation that violence viewing produces the aggression; this evidence is not con- clusive, however, and some of the data are also consonant with other interpretations. Yet, as we have said, the real issue is once again quantitative: how much contribution to the violence of our society is made by extensive violent television viewing by our youth? The evidence (or more accu- rately, the difficulty of finding evidence) suggests that the effect is small compared with many other possible causes, such as parental attitudes or knowledge of and experience with the real violence of our society. The sheer amount of television violence may be unimportant tom- pared with such subtle matters as what the medium says about it: is it approved or disapproved, committed by sympathetic or unsympathetic characters, shown to be effective or not, punished or unpunished? So- cial science today cannot say which aspects of the portrayal of violence make a major difference or in what way. It is entirely possible that some lNDINGSANDCONCLL'SIONS 5 ypes of extensive portrayals of violence could reduce the propensity to jiolence in society and that some types might increase it. In our present ;tate of knowledge, we are not able to specify what kinds of violence portrayal will have what net result on society. What are the alternatives? If broadcasters simply changed the quanti- tative balance between violent and other kinds of shows, it is not clear what the net effect would be. People hunt and choose the kinds of stimu- lus material they want. Violent material is popular. If our society changed in no other way than changing the balance of television offer- ings, people, to some degree, would still seek out violent material. How much effect a modest quantitative change in television schedules would have is now quite unanswerable. More drastic changes, such as general censorship, would clearly have wide effects, but of many kinds, and some of them distinctly undesirable. In our judgment, the key question that we should be asked is thus a complicated one concerning alternatives. The proper question is, "What kinds of changes, if any, in television content and practices could have a significant net effect in reducing the propensity to undesirable aggres- sion among the audience, and what other effects, desirable and undesira- ble, would each such change have?" The state of our knowledge, unfortunately, is not such as to permit confident conclusions in answer to such a question. The readers of this report will find in it evidence relevant to answering such questions, but far short of an answer. The state of present knowledge does not permit an agreed answer. EFFECTS ON AGGRESSIVENESS Television is only one of the many factors which in time may precede aggressive behavior. It is exceedingly difficult to disentangle from other elements of an individual's life history. Violence and aggressiveness are also not concepts on which there is unvarying consensus. This applies equally to events observed in real life or through the media and to behavior in which an individual may engage. Violence is a vague term. What seems violent to one may not seem so to another. Aggressiveness is similarly ambiguous, and its designation as antisocial depends not only on the act but also on the circumstances and the participants. For scientific investigation, terms must be defined precisely and un- ambiguously. Although various investigators have used somewhat dif- ferent definitions, generally both televised violence and individual ag- gressiveness have been defined as involving the inflicting of harm, inju- ry, or discomfort on persons, or of damage to property. The translation of such a conception into measurement procedures has varied very 6 TELEVISION AKD GROWING up widely, and whether antisocial activity is involved or implied is a matter for judgment in the specific instance. Effects on aggressiveness: evidence from experiments Experiments have the advantage of allowing causal inference because various influences can be controlled so that the effects. if any. of one or more variables can be assessed. To varying degrees. depending on de- sign and procedures. they have the disadvantages of artificiality and constricted time span. The generalizability of results to everyday life is a question often not easily resolvable. Experiments concerned with the effects of violence or aggressiveness portrayed on film or television have focused principally on two different kinds of effects: imitation and insGgation. Imitation occurs when what is seen is mimicked or copied. Instigation occurs when what is seen is followed by increased aggressiveness. * Imitatjon. One way in which a child may learn a new behavior is through observation and imitation. Some 20 published experiments doc- ument that children are capable of imitating filmed aggression shown on a movie or television screen. Capacity to imitate, however, does not imply performance. Whether or not what is observed actually will be imitated depends on a variety of situational and personal factors. No research in this program was concerned with imitation, because the fact that aggressive or violent behavior presented on film or televi- sion can be imitated by, children is already thoroughly documented. Instigatjon. Some 30 published experiments have been widely inter- preted as indicating that the viewing of violence on film or television by children or adults increases the likelihood of aggressive behavior. This interpretation has also been widely challenged, principally on the ground that results cannot be generalized beyond the experimental situation. Critics hold that in the experimental situation socially inhibiting factors. such as the influence of social norms and the risk of disapproval or retal- iation, are absent, and that the behavior after viewing, though labeled "aggressive," is so unlike what is generally understood by the term as to raise serious questions about the applicability of these laboratory find- ings to real-life behavior. The research conducted in this program attempted to provide more precise and extensive evidence on the capacity of televised violence to instigate aggressive behavior in children. The studies variously involve whole television programs, rather than brief excerpts; the possibility of making constructive or helping, as well as aggressive, responses after viewing: and the measurement of effects in the real-life environment of a nursery school. Taken as a group, they represent an effort to take into F~ND~NGSANDCONCLUSIONS 7 account more of the circumstances that pertain in real life. and for that reason they have considerable cogency. In sum. The experimental studies bearing on the effects of aggressive television entertainment content on children support certain conclu- sions, First, violence depicted on television can immediately or shortly thereafter induce mimicking or copying by children. Second. under cer- tain circumstances television violence can instigate an increase in ag- gressive acts. The accumulated evidence, however, does not warrant the conclusion that televised violence has a uniformly adverse effect nor the conclusion that it has an adverse effect on the majority of children. It cannot even be said that the majority of the children in the various stud- ies we have reviewed showed an increase in aggressive behavior in re- sponse to the violent fare to which they were exposed. The evidence does indicate that televised violence may lead to increased aggressive behavior in certain subgroups of children, who might constitute a small portion or a substantial proportion of the total population of young tele- vision viewers. We cannot estimate the size of the fraction, however. since the available evidence does not come from cross-section samples of the entire American population of children. The experimental studies we have reviewed tell us something about the characteristics of those children who are most likely to display an increase in aggressive behavior after exposure to televised violence. There is evidence that among young children (ages four to six) those most responsive to television violence are those who are highly aggres- sive to start with-who are prone to engage in spontaneous aggressive actions against their playmates and, in the case of boys, who display pleasure in viewing violence being inflicted upon others. The very young have difficulty comprehending the contextual setting in which violent acts are depicted and do not grasp the meaning of cues or labels con- cerning the make-believe character of violence episodes in fictional pro- grams. For older children, one study has found that labeling violence on a television program as make-believe rather than as real reduces the in- cidence of induced aggressive behavior. Contextual cues to the motiva- tion of the aggressor and to the consequences of acts of violence might also modify the impact of televised violence, but evidence on this topic is inconclusive. Since a considerable number of experimental studies on the effects of televised violence have now been carried out, it seems improbable that the next generation of studies will bring many great surprises, particular- 1~ with regard to broad generalizations not supported by the evidence currently at hand. It does not seem worthwhile to continue to carry out studies designed primarily to test the broad generalization that most or all children react to televised violence in a uniform way. The lack of uni- formity in the extensive data now at hand is much too impressive to war- rant the expectation that better measures of aggression or other metho- dological refinements will suddenly allow us to see a uniform effect. 8 `TELEVISION AND GROWING UP Effects on aggressiveness: survey evidence A number of surveys have inquired into the violence viewing of young people and their tendencies toward aggressive behavior. Measures of exposure to television violence included time spent viewing, preference for violent programming, and amount of viewing of violent programs. Measures of aggressive tendencies variously involved self and others' reports of actual behavior, projected behavior, and attitudes. The be- havior involved varied from acts generally regarded as heinous (e.g., arson) to acts which many would applaud (e.g., hitting a man who is at- tacking a woman). All of the studies inquired into the relationship between exposure to television violence and aggressive tendencies. Most of the relationships observed were positive, but most were also of low magnitude, ranging from null relationships to correlation coefficients of about .20. A few of the observed correlation coefficients, however, reached .30 or just above. On the basis of these findings, and taking into account their variety and their inconsistencies, we can tentatively conclude that there is a modest relationship between exposure to television violence and aggres- sive behavior or tendencies, as the latter are defined in the studies at hand. Two questions which follow are: (1) what is indicated by a corre- lation coeflicient of about .30, and (2) since correlation is not in itself a demonstration of causation, what can be deduced from the data regard- ing causation? Correlation coefficients of "middle range," like .30, may result from various sorts of relationships, which in turn may or may not be manifest- ed among the majority of the individuals studied. While the magnitude of such a correlation is not particularly high, it betokens a relationship which merits further inquiry. Correlation indicates that two variables-in this case violence viewing and aggressive tendencies-are related to each other. It does not indi- cate which of the two, if either, is the cause and which the effect. In this instance the correlation could manifest any of three causal sequences: -that violence viewing leads to aggression; -that aggression leads to violence viewing; -that both violence viewing and aggression are products of a third condition or set of conditions. The data from these studies are in various ways consonant with both the first and the third of these interpretations, but do not conclusively support either of the two. Findings consonant with the interpretation that violence viewing leads to aggression include the fact that two of the correlation coefficients at the .30 level are between earlier viewing and later measured aggression. However, certain technical questions exist regarding the measures em- ployed, and the findings can be regarded as equally consonant with the FINDINGSANDCONCLUSIONS 9 view that both violence viewing and aggression are common products of some antecedent condition or conditions. Various candidates for such a preceding condition can be identified in the data. These include preexisting levels of aggression, underlying per- sonality factors, and a number of aspects of parental attitudes and be- havior, among them parental affection, parental punishment, parental emphasis on nonaggression, and habitual types of parent-child commu- nication patterns. Several of these variables failed to operate statistical- ly in a manner consonant with common origin interpretations. At least two, "parental emphasis on nonaggression" and "family communica- tion patterns,*' operated in manners consonant with such an interpreta- tion, but the pertinent data were too limited to validate common origin status for either one. The common origin interpretation remains viable, however. Improved measures might possibly change the picture, and there is need for fur- ther and more refined investigation of the role played by personality fac- tors and by family and peer attitudes and behaviors. GENERAL IMPLICATIONS The best predictor of later aggressive tendencies in some studies is the existence of earlier aggressive tendencies, whose origins may lie in fami- ly and other environmental influences. Patterns of communication with- in the family and patterns of punishment of young children seem to re- late in ways that are as yet poorly understood both to television viewing and to aggressive behavior. The possible role of mass media in very ear- ly acquisition of aggressive tendencies remains unknown. Future re- search should concentrate on the impact of media material on very young children. As we have noted, the data, while not wholly consistent or conclu- sive, do indicate that a modest relationship exists between the viewing of violence and aggressive behavior. The correlational evidence from surveys is amenable to either of two interpretations: that the viewing of violence causes the aggressive behavior, or that both the viewing and the aggression are joint products of some other common source. Several findings of survey studies can be cited to sustain the hypothesis that viewing of violent television has a causal relation to aggressive behav- ior, though neither individually nor collectively are the findings conclu- sive. They could also be explained by the operation of a "third variable" related to preexisting conditions. The experimental studies provide some additional evidence bearing on this issue. Those studies contain indications that, under certain limit- ed conditions, television viewing may lead to an increase in aggressive behavior. The evidence is clearest in highly controlled laboratory stud- ies and considerably weaker in studies conducted under more natural 10 TELEVISION AND GROWING UP conditions. Although some questions have been raised as to whether the behavior observed in the laboratory studies can be called "aggressive" in the consensual sense of the term, the studies point to two mechanisms by which children might be led from watching television to aggressive behavior: the mechanism of imitation, which is well established as part of the behavioral repertoire of children in general; and the mechanism of incitement, which may apply only to those children who are predisposed to be susceptible to this influence. There is some evidence that incite- ment may follow nonviolent as well as violent materials, and that this incitement may lead to either prosocial or aggressive behavior, as deter- mined by the opportunities offered in the experiment. However, the fact that some children behave more aggressively in experiments after seeing violent films is well established. The experimental evidence does not suffer from the ambiguities that characterize the correlational data with regard to third variables, since children in the experiments are assigned in ways that attempt to control such variables. The experimental findings are weak in various other ways and not wholly consistent from one study to another. Neverthe- less, they provide suggestive evidence in favor of the interpretation that viewing violence on television is conducive to an increase in aggressive behavior, although it must be emphasized that the causal sequence is very likely applicable only to some children who are predisposed in this direction. Thus, there is a convergence of the fairly substantial experimental evidence for short-run causation of aggression among some children by viewing violence on the screen and the much less certain evidence from field studies that extensive violence viewing precedes some long-run manifestations of aggressive behavior. This convergence of the two types of evidence constitutes some preliminary indication of a causal relationship, but a good deal of research remains to be done before one can have confidence in these conclusions. The field studies and the laboratory studies converge also on a number of further points. First, there is evidence that any sequence by which viewing television violence causes aggressive behavior is most likely applicable only to some children who are predisposed in that direction. While imitative behavior is shown by most children in experiments on that mechanism of behavior, the mechanism of being incited to aggressive behavior by seeing violent films shows up in the behavior only of some children who were found in several experimental studies to be previously high in ag- gression. Likewise, the correlations found in the field studies between extensive viewing of violent material and acting in aggressive ways seem generally to depend on the behavior of a small proportion of the respon- dents who were identified in some studies as previously high in aggres- sion. INDINGS AND CONCLUSIONS 11 Second, there are suggestions in both sets of studies that the way chil- lren respond to violent film material is affected by the context in which t is presented. Such elements as parental explanations, the favorable or +mfavorable outcome of the violence, and whether it is seen as fantasy or reality may make a difference. Generalizations about all violent con- tent are likely to be misleading. Thus, the two sets of findings converge in three respects: a prelimi- nary and tentative indication of a causal relation between viewing vio- lence on television and aggressive behavior; an indication that any such causal relation operates only on some children (who are predisposed to be aggressive); and an indication that it operates only in some environ- mental contexts. Such tentative qnd limited conclusions are not very sat- isfying. They represent substantially more knowledge than we had two years ago, but they leave many questions unanswered. Some of the areas on which future research should concentrate in- clude: (1) Television's effects in the context of the effects of other mass media. (2) The effects of mass media in the context of individual devel- opmental history and the totality of environmental influences, particu- larly that of the home environment. In regard to the relationship be- tween televised violence and aggression, specific topics in need of fur- ther attention include: predispositional characteristics of individuals; age differences; effects of labeling, contextual cues, and other program factors; and longitudinal influences of televisian. (3) The functional and dysfunctional aspects of aggressive behavior in successfully adapting to life's demands. (4) The modeling and imitation of prosocial behavior. (5) The role of environmental factors, including the mass media, in the teaching and learning of values about violence, and the effects of such learning. (6) The symbolic meanings of violent content in mass media fiction, and the function in our social life of such content. Chapter 1 Introduction Previous scientific efforts to assess evidence of television's effects on youthful viewers have come to a variety of conclusions. Much testimo- ny has been collected to support the various positions, and opinions have been strongly expressed. At the time the work of this committee began in 1969, the most widely accepted summary evaluation of the research findings was probably that which emerged from a well-known 1961 study: For some children, under some conditions, some television is harmful. For oth- erchildren under the same conditions, or for the same children under othercon- ditions. it may be beneficial. For most children, under most conditions, most television is probably neither harmful nor particularly beneficial (Schramm, Lyle. and Parker. 1961). Nevertheless, some scientific studies were finding more controversial evidence. A small body of research had concluded that "witnessing ag- gressive TV programs serves to reduce or control the acting out of ag- gressive tendencies rather than to facilitate or stimulate aggression" (Feshbach, 1969). Other investigators had concluded that "the observation of aggression is more likely to induce hostile behavior than to drain off aggressive in- clinations" (Berkowitz, 1964). Against this backdrop of conflicting expert opinion, the committee began its work. HISTORY OF THE COMMITTEE The work of this committee was initiated by a request from Senator John 0. Pastore, Chairman of the Senate Subcommittee on Communica- 13 14 TELEVISION AND GROWING UP tions of the Senate Commerce Committee, in aletter of March 5, 1969. to Health, Education and Welfare Secretary Robert Finch, in which Sena- tor Pastore said: I am exceedingly troubled by the lack of any definitive information which would help resolve the question of whether there is a causal connection between tele- vised crime and violence and antisocial behavior by individuals, especially chil- dren. . . .I am respectfully requesting that you direct the Surgeon General to appoint a committee comprised of distinguished men and women from whatever professions and disciplines deemed appropriate to devise techniques and to conduct a study under his supervision using those techniques which will estab- lish scientifically insofar as possible what harmful effects, if any, these pro- grams have on children. On March 12, 1969, in a statement to the Communications Subcom- mittee, Surgeon General William H. Stewart announced that he would appoint an Advisory Panel of experts in the behavioral sciences. the mental health dis- ciplines, and communications to study the effects of televised violence. Their task will be to review what is presently known, and to design and to recommend the long-range research studies which will help answer the specific questions now under discussion. The Panel members will be knowledgeable about televi- sion and violence, and, of equal importance, experts in such related areas as social psychology. communication and learning, and the etiology of emotional disturbance. Dr. Stewart told the subcommittee that he would direct the National Institute of Mental Health to assume responsibility for the functions of the Advisory Panel and to provide technical staff for the study. On April 16, 1969, HEW Secretary Finch issued a directive authorizing the for- mation of the Surgeon General's Scientific Advisory Committee on Tel- evision and Social Behavior. The Secretary said the committee would confine itself solely to s-cientific findings and make no policy recommen- dations. Its approach, he said, would be similar to that of the Surgeon General's 1962-63 Committee on Smoking and Health, which limited it- self to developing factual data and conclusions about the possible causal relationship between smoking and health. "As far as this department is concerned," Secretary Finch said, "we have no mandate and no power that relate to commercial broadcasting and we do not seek any, but we do have a clear responsibility in the area of public health including the important field of mental health." Selection of members In selecting the advisory panel, the Surgeon General noted that it would be a scientific group and that its credentials should be recognized by the scientific community, the broadcasting industry, and the general public. Letters from the Surgeon General went out to a variety of academic and professional associations- including the American Sociological INTRODUCTION 15 Association, the American Anthropological Association, the American psychiatric Association, and the American Psychological Association. In addition, letters went to the National Association of Broadcasters, the Columbia Broadcasting System (CBS), the National Broadcasting Company (NBC), and the American Broadcasting Company (ABC). All these groups were asked to recommend knowledgeable scientists for membership on the Advisory Committee. Other distinguished social sci- entists, government officials, and members of the broadcasting industry were also asked for nominations. From the dozens of names proposed by these groups and individuals, a list of 40 was drawn up by the Office of the Surgeon General. This list of "recognized experts in the behavioral sciences and mental health dis- ciplines" was sent by the Surgeon General on April 28, 1969, to the pres- idents of the National Association of Broadcasters and the three nation- al commercial broadcast networks. Dr. Stewart asked the broadcasters to indicate "which individuals, if any, you believe would nor be appro- priate for an impartial scientific investigation of this nature ." "I am taking this step," the Surgeon General said, "because the stud- ies initiated by this group may involve the active collaboration of the tel- evision industry. I want to insure that all members of the advisory com- mittee are acceptable to the major networks and broadcasters." The National Association of Broadcasters and two of the networks responded by supplying a total of seven names of individuals they thought inappropriate to serve on the committee. From the remaining 33 names, II members were chosen. One committee member was not on the original list but was added to strengthen representation in one of the scientific disciplines. We believe some comment on this manner of selection is in order. Most of us were unaware of the selection procedure at the time the committee was formed and we believe there was a serious error in this process. We agree that nominations should have been sought from aca- demic and professional organizations as well as from broadcasters and other groups with relevant expertise and knowledge. However, we do not agree that any group should have been allowed to cite individuals as unacceptable. Such a procedure in effect shared responsibility for com- mittee appointment. We do not believe such responsibility should be shared. Moreover, we feel that future government advisory committees concerned with matters of public interest should be selected in such a way that no legitimate criticism about the manner of selection can be leveled afterward, either by the public or by the committee itself. We began our work as a committee on June 16-17. 1969. The general outline of the mode of operation of the committee and its initial activi- ties were summarized in a brief progress report issued in October 1969 (see Appendix A). 16 TELEVISION AND GROWING UP Observations on the general nature of advisory corn m ittees While this is not the place to offer elaborate commentary on the organ- izational and operational problems of committees and commissions formed to examine complex social problems, some discussion is appro- priate. More extended analyses have already been advanced by Lipsky (1971) and Wilson (1971). If the following elements are present, there will almost certainly be serious controversy: (1) Present the committee with a complex question about which there is both public and scientific controversy. This is al- most bound to be the case, or there would be no demand for the commit- tee in the first place. (2) Ask the committee to arrive at unequivocal con- clusions. Again, this is a Iikely circumstance. (3) Announce the commit- tee formation publicly, thus emphasizing its importance and stature. (4) Give the committee a severely limited time period in which to reach its conclusions. These four circumstances, of course, are almost inevitable attributes of the commission or committee approach to examining current social problems. They are cited, not to make excuses for the work done by such bodies, but rather to point out that these circumstances need to be recognized as another dimension of the difficulty of dealing with sub- stantive problems in this way. Our committee was not immune to these difficulties. The differences of opinion which have arisen during the life of this committee, about the meaning of scientific data on the issue of television and its relationship to social behavior, have.been the sort expected in any complex area of investigation. They reflect the lack of unanimity among scientists work- ing in this area. Comparing the task of this Advisory Committee with that of the Sur- geon General's Advisory Committee on Smoking and Health may be useful. In both instances the Surgeon General convened advisory groups to examine an issue of public health. The original request from Senator Pastore asking for the convening of this group was stimulated "because of the outstanding contribution made by [the Surgeon General's] Com- mittee through its report on smoking and health." The Committee on Smoking and Health reached its conclusions after a comprehensive reexamination and reevaluation of existing scientific evidence. The present committee, in contrast, has had available new research specifically sponsored to provide it with additional scientific data. The committee began its work immediately after a comprehensive examination of existing evidence in the area of televised violence had been made by the National Commission on the Causes and Prevention of Violence. Indeed, on September 23, 1969 (one day before our second INTRODUCTION 17 committee meeting), the National Commission issued its statement on violence in television entertainment programs. That statement, the work it represented, and the reaction it received underscored the original deci- sion to sponsor new research rather than to rely solely on reexamining preexisting material. THE RESEARCH PROGRAM One million dollars was made available for the support of new re- search, and a secretariat, the Television and Social Behavior Program, was organized within the National Institute of Mental Health to provide staff support for the work of the Advisory Committee. The committee worked closely with the staff throughout the life of this program. However, a committee composed of individuals with oth- er full-time responsibilities is not able to administer a large scale re- search program. The staff secretariat took major responsibility for find- ing competent investigators who were willing to undertake pertinent research within the time constraints. The staff also was responsible for selecting those proposals which seemed most likely to provide signifi- cant data and for monitoring the studies until their completion. Research strategy At the outset two alternative research strategies were considered: (a) attempt to develop a single, unified research project, or (b) seek out a series of individual studies which would address a variety of related questions and which would provide an interrelated set of findings. The former did not seem feasible, given the time limits and the present state of the art in this field. Between August 1969 and April 1970, 40 formal research proposals were submitted and reviewed for possible funding. A system of formal review, similar to that used to evaluate research contracts for the Na- tional Institutes of Health and the National Institute of Mental Health, was instituted to select the applications to receive financial support. For the Television and Social Behavior Program, groups of four to seven senior scientists in the researcher's field of expertise met on nine occa- sions to review proposals. Each review committee consisted largely of social scientists in the field who were not affiliated with the Television and Social Behavior program and senior staff members of the National Institute of Mental Health Intramural and Extramural Programs. In ad- dition, one or two members of the Scientific Advisory Committee, func- tioning individually as experts, were present at most meetings. The committee as a whole did not select the research projects. 18 TELEVISION AND GROWING UP Research projects In the end, 23 independent projects were funded which provided a multidimensional approach to the assessment of television's effects. These 23 projects- many of which involved more than one study and sometimes more than one report- and a number of specially commis- sioned papers form much of the basis for our inferences and conclu- sions. (For a list of all reports and papers, see Appendix B.) Although the projects vary widely in subject, scope, and approach, there were similarities among them in many instances, and the program staff and the investigators attempted to link them so that they could provide a coherent set of findings. This was done at both the investiga- tion and interpretation levels and resulted in the review and interpreta- tion as a group of sets of studies with common features, and in the inves- tigators' sharing of ideas, methods. measures, and in one instance. ex- perimental subjects.' The reports and papers were divided into five groups according to their common concerns and their theoretical and empirical orientations. `In one instance, two research teams (Liebert and Baron, 1971; Ekman et al.. 1971) col- laborated in an experimental study to conduct very different investigations using the same subjects (children), stimulus materials (violent and nonviolent television), and dependent variable (the choosing of a response that would either allegedly help or hurt an unseen- and actually nonexistent-other child playing a game). Liebert and Baron (1971) studied the relationship between exposure IO television violence and a tendency to aggress. Ekman et al. (1971) used subjects' facial expressions as they viewed to study their emotional reac- tions to violent and nonviolent television content, and related emotional reaction to subse- quent aggressive and helping behavior. In another cooperative endeavor. surveys of adolescents in a Maryland school system were conducted by three research teams (McIntyre and Teevan, 1971; McLeod, Atkin, and Chaffee, 1971a; Ward, 1971) who shared both subjects and data collection resources. In addition, one set of investigators used the Maryland data in conjunction with data on another sample IO aetter test the consistency of results (McLeod et al., 197la). To obtain a consistent criterion for assessing the amount of violence viewed by their subjects, many investigators used the violence ratings of television series arrived at by Greenberg and Gordon (1971b) in their study of television critics' and public perceptions of television violence (Baldwin and Lewis. 1971; Foulkes et al.. 1971; Friedman and John- son, 1971: Lefkowitzet al., 1971; LoSciuto, 1971; Lyle and Hoffman, 197la; Mclntyre and Teevan, 1971; McLeod et al., 197la. 197lb; Robinson and Bachman, 1971). Several inves- tigators made use of Gerbner's extensive content analysis (1971b) for a working definition of violence, and Clark and Blar#enburg (1971) modified this definition for their own pur- poses and used his data to validate their retrospective content analysis instruments. In a similar manner, Murray (1971) used Bechtel, Achelpohl, and Akers's (1971) tapes of sub- jects' viewing behavior in their own living rooms as a means of perfecting interobserver reliability. Murray (1971) also used the viewing diary developed by LoSciuto (1971) to measure behavior in regard to television. Another example of common methods concerns specific questionnaire items. Eight in- vestigators sought to measure television content in relation to violent or deviant behavior by asking subjects to name their four favorite television shows (Bechtel et al., 1971: Chaf- fee and McLeod. 1971b; Friedman and Johnson, 1971; Lefkowitz et al., 1971; LoSciuto, 1971; McIntyre and Teevan, 1971; Murray, 1971; Robinson and Bachman, 1971), and many used the same wording to query subjects about the amount of time they spent view- ing. The data provided by these common measures permitted the testing of patterns de- rived from the totality of results. [NTRODUCTION 19 One investigator in each of four groups then attempted to integrate the findings in an "overview" paper (Chaffee, 1971; Greenberg, 1971; Lie- bert, 1971; Lyle. 1971); an "overview" for the remaining group was prepared by the staff (Comstock, 1971). Each of these papers represents the individual author's perspective. Each of the five published volumes representing the work sponsored by the Television and Social Behavior program is introduced by the appropriate overview paper. NATURE OF THE REPORT The designation of this committee as one concerned with television and social behavior is especially significant. The committee's title em- phasizes more than just the issue of violence, and more than the ques- tion of the impact of televised violence on the behavior and attitudes of children and adolescents. While the latter remained a central concern, research conducted for this program also `studied such topics as the amount of time spent watching television, activities displaced or en- hanced by television viewing, television advertising and viewer reactions to it, learning of specific information and role expectations from televi- sion, and the comparative effects of black and white and color television on the information learned from a television program. The research pro- gram was both strengthened and made more difficult by the effort to place the problem in a larger context: nonetheless we cannot claim that this report or the work of this research program covers the entire subject of television and social behavior. We are aware of the difficulties of obtaining unequivocal answers to many questions about television's effects on viewers. Television is only one part of a complex web of elements that may influence people's atti- tudes and behavior. It is difficult to design studies which isolate the ef- fects of television content from these other variables. As a result, gener- alizing from laboratory experiments, surveys, or short-term studies to the long-term, real-time world can be risky. Television and special subgroups We also believe it important to note that other age groups and seg- ments of the population may be as responsive to the influence of televi- sion as are children. For example, elderly people, especially those in homes for the aged, as well as confined or institutionalized individuals for whom television is a major recreational activity and source of infor- mation, deserve special consideration in any assessment of the effects of television viewing. But little is known about this at present. Ultimately, of course, the needs and desires of the general viewing public will also have to be included in any attempt at a comprehensive analysis and eval- uation of television's influence. 20 TELEVISION AND GROWING UP The vicarious nature of television viewing Moreover. the vicarious nature of television viewing presents another difficulty in conceptualizing the effects of television. For example, view- ing televised violence is very different from being present at a violent encounter. The viewer may identify with the aggressor, but he does not himself deliver any blows or fire any weapons. He may identify with the victim, but he does not himself experience any pain, sustain any wounds, or shed any blood. There is no way he can intervene to prevent or terminate the aggressive exchange, no way he can retaliate against the aggressor, bring the criminal to justice, succor the victim, or comfort the bereaved. His involvement is remote, detached, vicarious, and thus only partial. The inactivity of the television viewer as a detached onlooker may it- self be the essence of the television viewing experience. His detachment may contribute to his own dehumanization. On the other hand, the con- scious experiencing of rich and even lurid fantasy without allowing it to spill over into unacceptable real-life behavior is generally acknowledged as characteristic of good mental health. More than a decade ago, Bauer and Bauer (1960) commented on this issue: For good or ill, experience via the mass media is predominantly vicarious. Looked at from the long-range point of view of the impact of the media on the population, this fact may in itself have more profound implications (which we cannot anticipate) upon the personality of future generations than the actual content of the communications conveyed by the mass media. Changing technology Equally important is the fact that we are examining television as it is today. Tomorrow's technological innovations will certainly bring changes in the medium and in the way it is used. With increased availa- bility of UHF stations, the growth of cable television, and the develop- ment of cassette systems, there will be greatly increased potential for viewer control in selection of programs. A CAVEAT AND A REQUEST The very existence of this Committee is perhaps testimony to a public tendency to expect quick and easy answers to difficult problems and to abdicate responsibility by "delegating" it to institutions rather than making individual decisions. Some people, moreover, seem inclined to be moralistic about the symbolic representation of violence on television and to blame televised violence for what happens in the real world. These tendencies may lead to attributing the phenomenon of violence to INTRODUCT ION 21 simple and easily correctible factors rather than to the more complex sources in our society. We wish to emphasize, however, that we are not concerned with blame or with making moral judgments. Our concern is with scientific evidence on television's effects. Throughout our deliberations we have been aware that television is one of the many influences which affect how people grow, learn, and behave toward their environment and toward one another. Our know- ledge of the human organism -to say nothing of the social organism-is far from definitive. We have attempted to take a small step toward great- er understanding of the medium of television and the implications it may have for society. We must urge that, in addition to this formal report to the Surgeon General, the serious student of television's effects examine the reports and papers on which we have drawn. They are being published concur- rently with this report to permit social scientists and others concerned with the issues involved to evaluate independently the work supported by the Television and Social Behavior Program and the validity of the conclusions reached by this committee. This committee can do no more than offer our own interpretation and evaluation of the findings. Chapter 2 Violence in Society and in the Television Med/um Individual children differ in the readiness with which they can learn to be aggressive or nonaggressive; genetic and other biological factors play a role in these differences (Berkowitz, 1962; Feshbach, 1970). Most small children are capable of learning to be aggressive and nonaggres- sive, cooperative and rebellious, trustful and suspicious, accommodat- ing and initiating, selfish and sharing, and constructive and destructive to varying degrees. Reinforcing and inhibiting life experiences deter- mine which patterns are more prominently developed. The frequency and intensity of activation, associated rewards or punishment, prevail- ing values, and available role models influence the character of these patterns. TELEVISION AND PERSONALITY DEVELOPMENT In infancy, neurophysiological patterns are immature, and behavioral responses are immediate. direct. generalized, and apt to be "all or none" in character, with considerable potential for change and reversal of response. In the course of early childhood development, the matura- tion of central nervous system tissues and the patterning of tissue func- tion by experience make available a wide range of direct and indirect, generalized and localized, complete and partial, immediate and delayed responses. Some patterns of response are reinforced and some are in- hibited. Patterns which are reinforced at one time may be inhibited at 23 24 TELEVISION AND GROWING UP another. In the course of training, education, and acculturation, patterns of varying intensity and complexity are developed and associated with one another, so that particular behavioral responses and roles are mani- fest in interactions with other persons. Most children over ten years of age show varying degrees of shame, guilt, and inhibition associated with crying, sucking, messiness, hitting, and other behaviors which they freely and comfortably displayed in ear- ly childhood. A stimulus which reinforces a response in early childhood may inhibit the same response in later childhood when inhibitory mecha- nisms are more highly developed. A specific response which has been learned may be employed at one time for constructive purposes and at another time for destructive purposes. The act of hitting which initiates an assault may at other times be employed for protection or for preven- tion of injustice. The physical, intellectual, and emotional resources of adolescents; their motivation toward independence from their families, toward au- tonomy and development of personal identity; and their proclivities for forming groups often render them capable of successful aggressive, anti- authority behavior for the first time. While most of this behavior repre- sents a phase in development and in this respect is prosocial in nature, it is often disquieting and disrupting to parents and other author'ities who are challenged. When these interactions are poorly handled by any of the parties involved, antisocial behavior may be one result. The precise impact televised content might have at particular points in the matura- tion process has yet to be determined. The complexities of developmental processes in childhood and adoles- cence and the variations from one individual to another make it difficult to predict the effects of any single carefully controlled stimulus upon behavior and impossible to predict fully the effects of the wide variety of visual and auditory stimuli offered in television programs. We need much more information in order to delineate the effects of televised vio- lence upon the behavior and development of children. To obtain it, it would be necessary to conduct both short-term and longitudinal re- search in controlled laboratory situations and in naturalistic settings; with young people at various stages of development, of differing charac- ter, from differing cultures, in varying emotional states; using a variety of stimuli arranged in varying sequences and with variable complexity. Many speculations are possible. but hypotheses have been tested only for very few circumstances and ages; these cannot be validly general- ized to apply to ages, states, and situations different from those which were investigated. TELEVISION AND SOCIALIZATION The socialization process is also a complex one. For a child discover- ing his inner and outer world and learning to respond to each, television VIOLENCE IN SOCIETY 25 may be an important source of models which demonstrate when, why, and how aggression can be appropriate. Each individual lives in a comparatively circumscribed context. Communication media offer opportunities for contact with a broader spectrum of experiences. Television, with its visual and auditory im- pact. is capable of providing vicarious experience with lifestyles and values from many different social contexts. It also provides a setting in which a young person might learn the strategies, tactics, and techniques of aggression. However, whether he puts to use what he learns and behaves aggres- sively will not depend only on what he sees or does not see on television. Nor will it depend only on what he sees or does not see in any other dis- crete experience in his own life. Although the causal antecedents of ag- gressive behavior are not fully understood, it is certain that they are di- verse. numerous, and complex in their relationship to each other and to aggressiveness. The impact of television viewing can only be fully understood when we know something about a young person's own-nature. his family, his neighborhood, his school. and other major circumstances and influences in his life. The strongly emotional experiences that occur in a child's re- lations with other members of the family and with peers are especially important. This is not to deny the potential importance of television. Rather, it is to say that other factors are also potentially important. These elements invariably contribute a context which influences the effects television has on the viewer. The family, the church, the legal system, and the military, among oth- er institutions. communicate codes, ethics, and guidelines for aggression and violence. The extent to which television reinforces or weakens these codes or guidelines is not presently known. Commercial television in the United States has not primarily attempted to be a teaching agent: its self-chosen primary role has been to errtertain. Entertainment, however-whether via television or not-may unobtru- sively convey ideas. information, sentiments, and values to the mem- bers of a society. Enculturating factors and his developing conscience provide criteria that nay help a young person to clarify which values and behaviors, presentea in entertainment, are to be emulated in reality and which are to be kept in the realm of fantasy. DIFFERENTIATING BETWEEN REALITY AND FANTASY Each person in the television audience is exposed to a broad variety of stimuli. These stimuli constitute a complex continuum ranging from what was conceived of as fantasy to mediated views of reality. Each person in the audience perceives and further interprets the stimuli through his own patterns of ideas, values, and responses. 26 TELEVISION AND GROWING UP Perceptions, interpretations, and responses to the same stimulus not only vary from individual to individual, but also vary from time to time within the same individual. The viewer watching a cartoon or a purely fictional drama may be aware of and acknowledge the fantasy nature of the stimuli, but through primitive unconscious identification processes he may respond psychologically and physiologically as if the stimuli are real and personally involve him. States of comfort or discomfort, plea- sure or pain, and even verbal communications or participating move- ments may be evoked. It is possible that stimuli from a television screen in a box occupying a small portion of a room arouse neurophysiological patterns similar to or different from those aroused in interpersonal experiences with real peo- ple. We do not yet know how the neurophysiological experience asso- ciated with witnessing a fight between two real people would compare with the neurophysiological experience associated with witnessing filmed images of that fight on a television screen. Responses of children and adults Genrally, infants and young children are less able than older persons to distinguish stimuli which are products of fantasy from those which are products of reality. Most children are more apt than older people to respond emotionally and physically, as well as ideationally, to their own fantasies and to the fantasies presented to them as if they were reality. In varying degrees adults, too, may experience reactivation of pat- terns which were more prominent during childhood. Many elements in the emotional experiences of adults are associated with emotional expe- riences from their childhood, and it is not uncommon for adults to enjoy relationships, interests, and activities of which they were fond during childhood. Indeed, much of the content communicated through the me- dia, including television. engages the "child part" of adults as well as their mature aspects. Parental influence In normal parent-child interaction, the differentiating of make-believe from real is a complex and extended process at best. In the television- child setting, the task is further complicated because the child is often left largely to his own devices. To him, the difference between film clips of actual combat or a real riot, and dramatic portrayals of similar con- flicts, may not always be clear. Commericals may further blur distinc- tions since they often consist of fantasy about real things. If fictional violence continues to appear in television entertainment, should special steps be taken to assist children in identifying it as fic- tion? Can fictional violence on television play a constructive role as a VIOLENCE IN SOCIETY 27 psychological safety valve which vents socially unacceptable hostility by offering vicarious experience to some persons? Can televised vio- lence stimulate psychological inhibitory mechanisms in some viewers which reduce their likelihood of imitating that behavior? Does televised violence instigate or facilitate for some viewers release of aggressive or violent impulses? Does a high concentration of violence in televised content convey impressions of permissiveness toward or expectations of violent behavior to some persons ? How do influences from family, school, religion, laws, neighborhood environment, peers, genetic, phy- siological and cultural factors interact with various television viewing experiences? Do the images on a television screen provide a "fantasy" stimulus quite unlike that provided by real people in the room? Which persons tend to differentiate and which tend to confuse fantasy and real- ity? Are these behavioral effects beneficial or detrimental, prosocial or antisocial, adaptive or maladaptive? These are some of the many questions which have motivated system- atic inquiry and scientific research on the effects of television on social behavior. WHAT THE CONTENT OF TELEVISION REFLECTS Television content inevitably reflects the values, the points of view, and the expectation of audience response held by those involved in the production process. Drama. light or serious, documentaries, "specials," variety and mu- sic programs, and news are quite different types of format and in many respects involve quite different considerations. All, however, require the making of decisions as to what will be presented from the voluminous amount of potential material. The values reflected in these decisions are no less relevant because they are generally unarticulated. The decisions made take on importance because all these varieties of television fare can structure the audience member's relationship to reality. To varying extents and in various ways, they can engage conscience, modify or mobilize opinion, and challenge or confirm beliefs. Audience response to news programs, for example, depends to a con- siderable degree upon the televised content, and this depends in part on the selection and editing process. Selection of an emotionally charged part of a speech and omission of the context in which it was given might increase the audience involvement but also might contribute to false be- liefs by offering an unbalanced view. Suggestible persons may be strongly influenced or even exploited by the ideas and advice offered through television and other media. Other viewers may be freed from restrictive ideas and false beliefs to which they have been bound. Media may be used to promote conflict or to re- solve it. The moderator of a panel show, for example, may help 28 TELEVISION AND GROWING UP representatives of different schools of thought to fight with one another or to find common interests, to collaborate, synchronize, and harmonize their contributions. Stereotypes In addition to violence, an area of major concern has been television's potentiality for perpetuating, reinforcing, or modifying social stereo- types about groups defined by such criteria as sex, ethnic background. and social class. Many children in the United States, especially those in big cities, have never met an American Indian. But American children have had endless hours of experience with "Indians" who ride horses across the plains, stalk wagon trains, and raid camps of white soldiers. Much of what American children "know" about American Indians may well have been derived from watching television dramas and movies rerun on televi- sion. For many years, blacks were seen usually as servants, slaves, or buf- foons, less often as athletes or fighters, almost never as clergymen, phy- sicians, teachers, attorneys, or policemen. Black Americans protested that such stereotypic portrayals conditioned other Americans to think of them as inferior to whites. This protest has now been heard, and vigor- ous efforts are now being made to present movie and television dramas in which black actors appear in a broad diversity of roles. Since television may play a role in shaping opinion and attitudes, it is important to pay attention to which persons, groups, and interests are presented in a favorable light and which are presented unfavorably. Tel- evised content can suggest who may be considered benign and who may be considered a threat to society. The responsibility of decision-making Decisions made by persons at various levels in the television industry determine what is broadcast, when it is broadcast, and how what is broadcast is treated-from point of view to camera angle. The media may offer an avenue of expression for a few or for many. Unfortunately, the powerful and the powerless, the wealthy and the poor, the influential elites and nonelites do not have equal access to the television cameras and microphones, and the impact of television may be differentially felt. In general, the powerful, influential. and elite have opportunity to initate and control the content and uses of television in ways that the powerless, the poor, and the nonelite do not. In these in- teractions one party's interests are often supported while the interests of other parties are sacrificed. This places an especially heavy responsibili- ty on those who determine which aspects of reality shall be given the special salience bestowed by television treatment. VIOLENCE IN SOCIETY 29 DEFINITIONS AND DIMENSIONS OF VIOLENCE The possible effect of televised violence on the behavior and attitudes of children is the major focus of this research program. The National Commission on the Causes and Prevention of Violence (1969) in examin- ing the history of American society made these points: America has always been a relatively violent nation. Considering the tumultuous historical forces that have shaped the United States, it would be astonishing were it otherwise. Since rapid social change in America has produced different forms of violence with widely varying patterns of motivation, aggression, and victimization, vio- lence in America has waxed and waned with the social tides. The decade just ending. for example. has been one of our most violent eras-although probably not the most violent. Exclusive emphasis in a society on law enforcement rather than on a sensible balance of remedial action and enforcement tends to lead to a decaying cycle in which resistance grows and becomes ever more violent. For remedial social change to be an effective moderator of violence, the changes must command a wide measure of support throughout the community. Official efforts to impose change that is resisted by a dominant majority frequently prompt counterviolence. Finally. Americans have been. paradoxically, a turbulent people but have en- joyed a relatively stable republic. Our liberal and pluralistic system has histori- cally both generated and accommodated itself to a high level of unrest. and our turmoil has reflected far more demonstration and protest than conspiracy and revolution. Within these broad conclusions, the Commission examined the histo- ry of violence, with attention to both individual and group violence and to effects of television and other media upon these. At least two things are clear from reading the Violence Commission report, as well as the primary references on violence and aggression which the Commission used. The first is that violence has characterized our society throughout its history, and the second is that there is no simple or universal explana- tion of the causes of violence. In fact, there is not even a clear consen- sus about what constitutes violence. What is "violent?" The character of an act does not, by itself, define whether the act is violent. The effect, the social context, the moral framework, the degree of legitimization, and the amount and kinds of group endorsement of the act are very relevant to the definition of violence in the real world. For example, while many societies sanction parents' use of physical force to control and train their children, the same force, employed by other per- sons in a different context, might be defined as violence. Although their use of force is not so widely permitted, children often employ force in 30 TELEVISION AND GROWING UP their dealings with other persons-especially other children-and in their expression of feelings. Over time, most individuals will internalize their society's moral codes and mold their behavior accordingly. Whether or not the use of physical force will be defined as violence depends upon one's perspective and upon the context, as well as upon the nature of the act. The recipients of forceful action generally define such action as violent more readily than do initiators of the action. Thus: -The same act may be considered violent under some circum- stances and not under others. -The same act may be judged as violent by one person and not by another. -The same act may be generally accepted and labeled nonviolent when committed by one person but may be generally rejected as violent when committed by another. -The same violent act may be accepted at some ages but not all others. or may be accepted among males but not among females. -The same violent act may be rejected if one initiates it but may be approved as self-protection against another's attack. -Violence may be accepted if it is deemed necessary to protect a person, a property. or an important belief. -Destroying or hurting another by psychological or verbal means. which are generally more subtle than physical actions, will often not be considered as violence. -The ethics of violence may be blunt; line-of-duty violent acts of soldiers and police may be acceptable. -The ethics of violence may be more subtle. It may be acceptable to hit back. but not in the groin or in the eye. -An act by a person we like or idealize is less apt to be considered violent than the same act by a person we dislike or denigrate. -Violence to right a wrong may be acceptable by an acknowledged official but not by ordinary citizens, some of whom may even be expected to accommodate to injustice. Defining agcjression Throughout this report the terms "aggression" and "violence" are employed almost always in reference to antisocial behavior. We ac- knowledge that this usage is neither comprehensive nor precise. Howev- er, this usage is so common that its meaning is communicated easily. The word "aggression" has generally been associated with antisocial or destructive implications. Within psychoanalytic theory, on the other hand, aggression refers to the mobilization, organization, and applica- tion of energy to a task which may be constructive or destructive, proso- cial or antisocial. VIOLENCE IN SOClETY 31 In his review of literature on effects of media portrayals of violence, Weiss t 1969) noted the difficulty of arriving at a generally accepted con- ceptualization of aggression. A vast and varied array of behaviors may be considered aggressive, depending upon effects. upon intent. upon context. upon associated feelings and fantasies. and upon other factors. There is no aggreement either among lay persons or among scientists about how fantasized aggression. verbal aggression, and physical ag- gression may be compared. Nor is there agreement about what consti- tutes an aggressive act in real-life experiences or about the degree to which behavior measured in a laboratory is analogous to that in a natur- alistic setting. Aggression against an inanimate object is not always ac- cepted as the functional equivalent of aggression against an animate one. Would the inanimate object have been struck if it could hit back? Is aggressive behavior in play a functional equivalent of aggressive behav- ior with intent to harm? Sociopolitical aspects of.violence and aggression When we consider behavior within a societal context, the meaning of concepts such as "violence," "aggression," "order," and "disorder" is defined by sociopolitical processes. Similarly, decisions about the par- ticular manner in which "violent" acts are to be handled-for example, with a "show of force" or the actual use of "deadly force" by officials -are also essentially sociopolitical in nature. In a staff report to the Violence Commission, Skolnick (1969) dis- cussed the political and public policy aspects of defining, labeling, and handling violence. The kind of acts which are classified as "violent," as well as those which are not so classified, vary according to who provides the definition and who has the superior resources for disseminating and enforcing his definitions. The legislative process is involved in the for- mulation and enactment of criminal laws and of specific penalties for engaging in behavior so defined and officially prohibited. For example, the behavioral act of killing another person does not automatically nor even necessarily constitute murder. If the killing can officially be viewed as justified or in self-defense, for example, it will not be labeled as mur- der. Similarly, the young man setting fire to a Vietnamese hut may be considered a dutiful citizen and soldier: the same man burning a grocery store in New York or Chicago may be viewed as a dangerous criminal engaged in arson and related crimes. Almost every society, including primitive societies, legitimizes for the sake of its own maintenance some aggression and violence against inter- nal and external threats. Every society has inconsistent norms and mores. Every society talks a better, purer, more noble game than it plays. 32 TELEVISION AND GROWING UP Aggression and violence are always the legitimized privilege of authori- ty, whether it be within the setting of the family. within a tribe, or within a nation. Some aggression and violence have been an outcome of disagree- ments between individuals or groups over cherished values and beliefs which, in themselves, are conflicting at times. In a competitive society, strong motivations toward productivity and rewards may lead to high standards of living for some people and exploitation. suffering. and un- fairness for others. Those who focus their attention upon the productivi- ty and the high standard of living have a legitimate basis for their ap- proval of this process; those who focus attention upon the exploitation and unfairness have a legitimate basis for their disapproval. People often accommodate and adjust for long periods of time to damage, injury, or psychological trauma caused by such inequities as crippling discrimination on the basis of socioeconomic status or race. Severely destructive effects may be tolerated, but they are rarely de- fined as violence if they are brought about slowly enough, within a framework of accepted values and laws, and by group rather than indi- vidual action. Such legitimized and processed violence may have a large number of victims reflected in death rates, morbidity rates, vulnerability to exploitation, and other forms of human suffering. Neglect is not considered violence even if it results in death. Sudden damage to an individual or an object is generally recognized as violence while slow, erosive damage is apt to be perceived as violence only by the victim. In like manner, one who holds, envelops, or imprisons another against his will seldom perceives the violence experienced by the one who is held. Dimensions of violence and the television industry The television industry, in the production of programs with violent content, variously deals with or neglects these definitions and dimen- sions. The length of programs restricts the extent to which complexities can be developed. The beliefs, values, and definitions which exist in the minds of television decision-makers produce additional limitations in the conceptualization of violence on television. The economics of mass media lead to the presentation of violence in such a way and in such dimensions as suit the tastes of a highly heterogeneous audience. Addi- tionally, if content is presented which is not accepted to influential per sons and important public officials, problems of other kinds may devel- op. Thus, in many ways the practicalities of continually balancing rela- tionships with the audience, with public officials, with advertisers, and with numerous other interests foster limitations of various kinds on tele- vision content. Unless persuasive influences develop in new directions, VIOLENCE IN SOCIETY 33 the present patterns seem likely to continue, as a result of both con- scious and unconscious psychological and social pressures. DEFINING VIOLENCE FOR RESEARCH PURPOSES Any comprehensive consideration of the issue of violence in televi- sion content must take into account as many dimensions and complexi- ties of violence as possible, not confine itself to narrowly restricted as- pects. When violence must be defined for research purposes, however, it inevitably is stated in a restricted form. In his analysis of television con- tent in research sponsored in this program, Gerbner (197 1 b) points out: Violence connotes a great variety of physical and mental violations, emotions, injustices, and transgressions of social and moral norms. For this study violence was defined in its strictest physical sense as an arbiter of power. Analysts were instructed to record as violent only `the overt expression of physical force against others or self, or the compelling of action against one's will on pain of being hurt or killed.' The expression of injurious or lethal force had to be credi- ble and real in the symbolic terms of the drama. Humorous and even farcical violence can be credible and real, even if it has a presumable comic effect. But idle threats, verbal abuse or comic gestures with no real consequences were not to be considered violent. The agent of violence could be any sort of creature, and the act could appear to be accidental as well as intentional. All characters serve human purposes in the symbolic realm, and accidents or even `acts of na- ture' occur only on purpose in drama. An example of what investigators considered "violent" filmed materi- al is a specially assembled 45minute videotape used by Greenberg and Gordon (1971c), which the authors described as follows: This 4%minute tape contained 75 separate scenes of violence which varied in length from five to 120 seconds. All violent sequences were scenes in which characters physically harmed themselves or another person (e.g., hitting or shooting), overtly intended such harm (e.g., shooting but missing), or physically damaged some inanimate object (e.g., smashing furniture). Scenes of yelling or shouting were also recorded as examples of verbal aggression. Liebert and Baron (197 1) employed three-and-one-half-minute action sequences from the television series The Untouchables. Stein and Fried- rich (1971) used 12 20-minute episodes of Balman or Superman as an "aggressive" television film diet in their study of four-year-olds. This illustrates the principle that violence is operationally defined by the choice of specific stimulus material. One researcher, however, defined media violence in a very different and much broader way. Clark (1971) argues that violence can be almost imperceptible and slow as well as sudden, and that the media can be vio- lent as well as convey violence. In Clark's view, since television is a way of learning about the worth of one's self and others, the medium does 34 TELEVISION AND GROWING UP violence to blacks and other minorities by portraying them in ways that lower their self-esteem. Television violence, in his terms, is the "slow mental disintegration" that "the mass media commit by virtue of their effects on the black self-image. " As a result, Clark studies identification with television characters, because he believes that identification is the psychological process through which the violence he attributes to televi- sion is inflicted and is an index of the harmful effects of television and other influences on the wellbeing of minorities. While violence defined in this manner can produce destructive effects and many victims, these effects result from the use of psychological force rather than physical force. Operational definitions of violence and aggression generally emphasize specific physical actions which cause discomfort or injury to a person or damage to property. Chapter 3 Some Problems of Research on the Impact of Television A number of recurring questions arise in the process of reviewing what is known about the impact of television. Representatives of many diverse disciplines are trying to understand and formulate the effects of media experience upon human behavior. In each discipline there are diverse schools which rely upon different theories and different meth- ods. They exist in relatively separated and isolated compartments. In addition to these general problems, a number of specific research questions must be addressed before even tentative conclusions on the nature of television's effects can be advanced: What are the special problems associated with studying television's impact in childhood? What is the nature of the television stimulus? What are the strategies for investigating the impact of television? How much can these studies tell us about the viewer's behavior in response to television? BEHAVIOR IN EARLY CHILDI-iOOb A large number of studies conducted over the past two decades, con- cerned with the years of immaturity in humad beings and other species, have convinced specialists in child development that the early period of life is critically important. These studies support the age-old observation that "as the twig is bent, so the tree will grow." The child's learning dur- ing the first five or six years sets the foundation fbr lifelong patterns of behavior and for further learning. Attitudes and values, as well as habits 35 36 TELEVISION AND GROWING UP of thinking and reacting to other people, are set during this formative period. Child psychologists and child psychiatrists think of the young child as especially susceptible to influence (whether for good or for ill) during the years of his life when he is vitally dependent on other individ- uals for his very survival and growth. Young children are naturally curious and eager to learn all they can from life. Television is one potentially important source of knowledge, and by age two or three most American children have begun to watch and listen to television regularly (Schramm, Lyle, and Parker, 1961). However, most research studying the effects of television on children has not captured children's earliest experiences with television; instead, studies have concentrated on television's influence on school-age chil- dren and on adolescents. This is unfortunate; the years before the fifth birthday, when the child is especially open to new learning and new ex- periences, should be a period when television viewing might be especial- ly influential. Earlier studies (e.g., Bandura, Ross, and Ross, 1963) have documented that three-, four-, and five-year-old children imitate specific acts (including aggressive acts) which they have observed on film, at least in experimental circumstances. In the present series of studies, which will be described below, Stein and Friedrich (1971) were again able to document indications of television's impact as early as age three. A young child's reaction to television is potentially quite different from that of an adult. A child has only a limited range of past experience and does not have a well-established set of conceptual categories for clarifying his perceptual experiences. Many adults assume that because children catch the fun of some adult humor, they regularly operate on a higher level of sophistication than they actually do. If the stories or scenes which appeal to each age group were explored, one would probably discover that the child relates to humor which has a concrete rather than an abstract theme. The thinking of the three- and four-year-old is not logic as the adult sees it. At that age children are still free-associating through the day. The evolution of their thinking processes has not yet reached the stage where they voluntarily or involuntarily classify, sort, select, and organize information except in `very concrete and immediate terms. Certain children of superior intelli- gence who have had help with language and thinking in the family context do sometimes indicate that they can at least follow simple logical argu- ments, and their conversation often appears to make good sense to adults. However, the conversation of the overwhelming number of three- and four-year-old children is not always sensible in adult con- texts. In the same vein, the young television viewer often is unable to follow the theme of even a simple story (Klapper, 1969; Leifer and Rob- erts, 1971). It is unlikely that young children will understand the relative- ly complex motivations for and consequences of the behavior demon- strated by the television actor. RESEARCH PROBLEMS 37 DEFINING THE TELEVISION STIMULUS In order to assess the impact of television, we must clearly understand the nature of the television stimulus. A number of questions about tele- vision content, which raise important issues for research in this field, have been raised in earlier writings (e.g., Siegel, 1969). To what degree is the symbolic language of television different from or similar to other "languages" such as those used in interpersonal communication, live drama, serious music, and such? Is the "language" of television entertainment fare taken seriously by audiences, or does it carry within itself a heavy discounting element because of the potential artificiality of its excesses of cordiality, good humor, sincerity, intima- cy, and violence? Do audiences carefully attend to the symbols of tele- vision entertainment, or do these symbols merely reflect on irrelevant dimensions of life and thus require nothing more than superficial or cas- ual attention? Is the language of television especially "vivid," as some observers suggest? While television may be more vivid than other media like news- papers, comic books, or radio, how does it compare to listening to one's father or to a live concert or to seeing a professional football game in a stadium? And if the language of television is indeed more "vivid," is it necessarily more "effective" than, let us say, reading a fairy tale or lis- tening to a stereo recording of Peter and the Wolf? Can distinctions between "pure" entertainment content and "pure" information content be made from content analyses alone? Much re- search has shown that what may be information content for some view- ers may serve as entertainment content for others. Consequently, it is not easy to separate entertainment content from other types of content simply on the basis of an a prioriclassification scheme. Typically, televi- sion viewers in American homes are exposed to a complex mix of news, information, educational materials, advertising, propaganda, and enter- tainment fare. Any concern about the totality of reactions by viewers to television fare must also be concern about the totality of the symbolic stimuli to which they are exposed. A good deal of the "violent" content found in selected televised en- tertainment programs refers to times, places, characters, and events that are far removed from the actual life-space of the viewers; the programs are, in truth, fantasies which have no direct explicit application to con- temporary life (e.g., the "western," "science fiction," "ghost and hor- ror stories," the "period/costume drama"), but may in fact be symbolic of contemporary life. An interesting question arises here-namely, how and to what degree do content variables like "time of action," "type of action," and "place of action" that are removed from the current scene relate to contemporary audience reactions to this fare? Does this "dis- tancing" of symbols serve as another discounting factor so that the view- 38 TELEVISION AND GROWING UP er dismisses the materials as reflecting "just another story?" Or do these variables "wash out" and allow viewers to develop personal ana- logs for themselves regardless? Perhaps even more important questions are whether the young viewer perceives this "distancing," and, if so, how this perception relates to the likelihood that the child will adopt the televised behavior as a guide for his actions. What precisely constitutes portrayals of violence on television? In one approach, mentioned in Chapter 2, violent content is described in terms of discrete manifestations of physical aggressive behavior units in television programs. The unit of measure recorded in these studies is a specific act of observable behavior (e.g., punching, kicking, shooting). Each manifest act is generally given equal weight; the acts are summed up to reflect "violent" content as such. In another approach, it is sug- gested that aggressive behavior in television portrayals consists of an event made up of overt or covert aggression within the context of other nonaggressive events, or of an interpersonal tactic wherein aggressive behavior of some sort (rather than a nonaggressive tactic} is used to gain a specific end. Consequently, this unit of measure is the totality of the event or situation which includes the specific "aggressive" tactic em- ployed. Cutting across these two approaches are considerations of (1) whether the events and interpersonal tactics are reasonably capable of being adopted by a viewer quite literally, or (2) whether the portrayed event or tactic is symbolic and can only be adopted in keeping with the viewer's individual mode of expression of aggressive behavior. RESEARCH STRATEGIES In order to explore the possible influences on subsequent social be- havior of exposure to portrayals of violence on television, most of the studies in this program used one of two modes of investigation. One method can be described as applying the concepts and data-gathering techniques of field social survey research; the other, as applying the concepts and data-gathering techniques of the controlled laboratory experiment. Because the techniques used in either data-gathering method-survey or laboratory experiment-have critical bearing on the outcome of re- search, both methods will be given detailed attention as this report prog- resses. At this point it suffices to note that the distinctions between these two methods lie fundamentally in the manner in which data are gathered, rather than in the way they are ultimately analyzed and interpreted. Essentially, the social survey seeks to determine the relationships among and between variables as they may be distributed in relatively large samples either of a universe or of specific subpopulations. In con- trast, the laboratory experimental approach calls for isolating one varia- ble and testing its influence on the behaviors of small selected groups. RESEARCH PROBLEMS 39 One of the least complex experimental designs usually is composed of (I) a group (i.e., experimental group) which is exposed to a stimulus and (2) a group matched for similarities with the experimental group (i.e., control group) which is not so exposed. Implications of research Understanding the relationship between research results and free- ranging human behavior has been a persistent difliculty in attempts to apply scientific findings to the problems of daily life. Surveys and other correlational studies are usually unable to clarify sequential or causal relationships; experiments, while elucidating causality, usually require a simulation of certain behaviors in an experimental setting. Thus, each research strategy has some limitations. In experimental studies of the impact of television in early childhood, the problem is even more acute, according.to some observers, because the most definitive evidence comes from experiments in special play- rooms which are somewhat strange to the child. When achild views tele- vision, he usually watches in his own home surrounded by his family; critics suggest that the things the child learns and the behavior he dem- onstrates in this setting are quite different from what he learns and how he behaves in a special playroom. ' Some specialists concerned with the growth and development of children, on the other hand, believe that there is no clear distinction among settings for studying a child's behav- ior. They maintain that, for young children, the playground, the nursery school, and the playroom with a television set are not artificial but rather are part of the child's natural daily environment. Therefore, they hold, the behavior demonstrated in these settings can indeed be considered representative of the child's free-ranging behavior. Suspension of norms for behavior. The attempt to study the social effects of viewing television drama might be restated as the attempt to study the real-life behavior consequences of vicarious'experience. This relationship between a "fantasy stimulus" and a "reality response" raises some important questions for research. Certain aspects of this issue were discussed in the preceding chapter; however, further aspects have implications for research methodology. In culture after culture, for example, societies have exhibited games, entertainments. and ceremonies during which established norms for `The playroom in which a child psychologist conducts his or her research with young children is usually a small private room furnished with a table and chairs, a rug on the floor. and various toys. When the research concerns television, the furnishings include a television receiver. Usually there is a one-way vision mirror on the wall through which observers in the adjoining room may watch the child and make records of his behavior without intruding on it. Any technical monitoring apparatus-e.g., a tape recorder-is housed in the adjacent observation room. The playroom itself is planned to be cheerful, uncomplicated. and inviting. to provide a comfortable setting for the child. 40 TELEVISION AND GROWING up behavior are suspended; special codes which permit encroachments on norms or taboos come into force for a limited period. The requirement of truthfulness is suspended while the storyteller relates tales of youth- ful adventure. The prohibition of physical violence is suspended during games of contact sports. Norms for behavior between the sexes are somewhat relaxed during the Mardi Gras. In all such cases, the specta- tor, for a well-defined time period, enters into a moratorium on norms during which vicarious experience of otherwise unacceptable behavior is not only permitted but encouraged. This pattern may be referred to as an "entertainment scenario," in contrast to a "reality scenario" in which a person is expected to order his behavior in compliance with approved norms for everyday living. While the entertainment scenario tends to indulge impulse, the reality scenario tends to inhibit it. The entertainment scenario involves the assumption that socialization is well enough established that those involved can agree that during their interval of vicarious experience, everyday norms are suspended, not abolished. For example, a father and son at a football game may join in shouting to their team to commit all manner of violence against the op- posing team (entertainment scenario), and the son may have a little trou- ble "settling down" immediately after the game. But they both know that, once they have returned home, the son's interactions with his sister must conform to a completely different set of ground rules (reality scen- ario) than those which were appropriate on the playing field. Everyday experience suggests, at the same time, that the mood estab- lished in the entertainment scenario tends to persist. The demands of the reality situation and individual personality characteristics probably in- fluence the speed with which one moves from the entertainment scena- rio back to the reality scenario. The strength of the stimulus may also be a factor. For measurement to be fully valid, these potential differences be- tween the reality scenario and the entertainment scenario need to be taken into consideration. Unfortunately, there is little information avail- able that bears directly on this issue. Limitations of research In some research instances, it is necessary to alter or modify some aspects of the behavior studied. In research dealing with the impact of televised violence on children's aggressive behavior, the requirement that aggressive behavior be simulated is particularly important. No in- vestigator would place a child in a setting where he could clearly harm either himself or another child. Instead, he might substitute inanimate objects like large dolls for live persons as the object of aggression. Thus, experiments on the impact of televised violence have generally focused RESEARCH PROBLEMS 41 on indicators of interpersonal aggression, such as the child's report of his feelings and attitudes about hurting another person or his behavior in striking inanimate objects. Moreover, as Weiss (1969) points out in his review, "the testing situation is designed to give the impression that ag- gression is permissible if not encouraged; in the shock studies,2 aggres- sion is required and only the degree of aggression can vary." These con- siderations, as Weiss indicates, raise questions about "the propriety of referring to the responses used in the research as aggressive behavior." There are, of course, other aspects of research which must be under- stood in attempting to translate the experimental findings to daily life. Where the study of children's television viewing behavior is concerned, one aspect which must be studied is the child's overall psychological state for the day as well as for the moment. If he has been getting into mischief all day long, or if his caretaker has been irritable, or if he has not been feeling well, the sight of people being attacked and punished on television could have quite a different effect on him than the same scene might on a day when he had been generally successful and when his cop- ing skills were strong. To some extent, these variations in background conditions can be taken into account by a research design which uses an adequate number of subjects and randomly assigns these subjects to the various treatment conditions. But other factors enter in when we try to extrapolate the results from experimental studies to real life. When a young child is feel- ing strong. confident, and cared-for, he is not so prone to confuse fanta- sy with reality and decide that the world is too dangerous for him to cope with. The two-, three-, or four-year-old child whose mother is in the house may watch punishment and aggression on television with more detachment or aplomb than when she is not present and when he is un- certain that he is being well cared for. `Weiss refers to experiments in which subjects are directed to administer ostensible electric shocks. Chapter 4 Television Content Studies of television program content leave no doubt that among en- tertainment programs, violence figures prominently. There is also much violence in news programs, but the research on television content has focused mainly on dramatized entertainment programs. This focus, in itself, precluded a complete examination of the full spectrum of televi- sion and social behavior. Television offers a remarkable variety of program content, including news, sports, music, politics, education, discussion programs, and wor- ship services. These types of programs are scarcely mentioned in our studies, nor is any attempt made to explore their constructive contribu- tions to American life. It is taken for granted that television program- ming is on the whole consonant with modal interests and values. Indeed, if it were not, it could not survive, since it is dependent on voluntary audiences. There are few places in the United States where people receive as few as two television channels, and there are probably few individuals who, if they review the weekly schedules, will fail to find programming to suit their tastes. If they or their children spend large amounts of time view- ing television, they are under no requirement to do so. The emergence of public television and of cable systems promises further extension of al- ternatives, further diversity of offerings. It is widely believed that television increases children's vocabulary and extends their horizons (Steiner, 1963; Witty, 1966; Lyle and Hoff- man, 1971a). At the same time, and precisely because of the enormous popularity of television programming, there is concern about the possi- bility of negative effects on children. This concern relates particularly to fictional violence in entertainment programs. It is primarily this concern that motivated government sponsorship of the present project, and our studies are almost exclusively addressed to its exploration. 43 44 TELEVISION AND GROWING UP VIOLENCE: SENT AND RECEIVED As we noted in a previous chapter, violence takes many forms. There is verbal violence, fist fighting, violence with weapons, and there is the slapstick violence among cartoon characters. There is the violence of nature in storms, in fires, in hunting by predatory animals. There is so- cially approved violence (when the sheriff defeats the criminal) and dis- approved violence (when the criminal holds up the storekeeper). For reasons that are not clear, it is customary, in studies of violence in enter- tainment programs, to exclude the violence of football, basketball, hockey, baseball, boxing, automobile racing, skating derbies, wrestling, rodeos. The portrayal of violence cannot be assumed to have a one-to-one re- lationship with the perception of violence nor with the response to it. Although we know of no studies that would justify generalizing on this point, there are reports that individual children may experience distress at the televised portrayal of a pet being wounded but apparently feel no such reaction to what many adults would consider more extreme forms of violence. To speak of violence in television programs, then, is to speak of many things. Nevertheless, a study by Greenberg and Gordon (1971b) indi- cates a high'degree of agreement among ratings by 303 adult audience members and 43 television critics as to which television programs are most violent. Particularly interesting is their finding that, though half of their audience sample was given a definition of violence and half was not, the rank ordering of the raiings by the two audience groups led to nearly identical lists of "most violent" programs. The definition was: "By violence, I mean how much fighting, shooting, yelling or kill- ing there usually is in the show ," The 43 television critics were provided with this same definition of violence. Their ratings corresponded closely with those of the sample of audience members. The critics and the public agreed as to the 20 series they considered most violent. VIOLENCE IN PROGRAMS The most thorough study of violent content in television entertain- ment programs, or segments of programs, "that tell a story" has been conducted by Gerbner (1971b). His definition of an instance of violence is "the overt expression of physical force against others or self, or the compelling of action against one's will on pain of being hurt or killed." In addition to such acts as fighting, shooting, or killing, Gerbner includ- ed humorous and farcical acts, accidents, and acts of nature, so long as they appeared to be "credible and real." TELEVISION CONTENT 4s Gerbner's most recent study includes findings from his two earlier studies of the same sort. thus providing comparisons between findings in 1967, 1968. and 1969. These studies are primarily devoted to the enumer- ation and classification of violent incidents by trained coders who watched and coded videotapes of selected network programs for one week in October for each of the three years. He points out that his study is an analysis of program content, not of effects. Because Gerbner's findings have been inaccurately cited in several instances as referring to all network programs during the week of each year he studied, clarification of his data base is appropriate. The hours studied in Philadelphia in 1967 are shown in the following table. The hours studied in 1968 and 1969 are similar but not identical: ABC CBS NBC Sunday 4:00- 5:OO p.m. 7:00- 8:00 p.m. 7:30-10:00 p.m. 7:00-\O:OO p.m. 9:00-10:00 p.m. Monday 7:30-10:00 p.m. 7:30-10:00 p.m. 7:30-IO:00 p.m. Tuesday 7:30-10:00 p.m. 7:30- 8:30 p.m. 7:30-10:00 p.m. 9:30-10:00 p.m. Wednesday 7:30-1O:OO p.m. 7:30-10:00 p.m. 7:30- 9:OO p.m. Thursday 7:30-10:00 p.m. 7:30-10:00 p.m. 7:3X&10:00 p.m. Friday 7:30-10:00 p.m. 7:30-10:00 p.m. 7:30-1O:OO p.m. Saturday 9:00-l 1:OO a.m. 9:00-l 1~00 a.m. 9:00-l 1~00 a.m. 9:30-10:00 p.m. 8:30-10:00 p.m. 7:30-10:00 p.m. News programs. variety shows, and network specials were excluded because they did not contain plots or story lines. Within the,se samples. Gerbner found that: -The general prevalence of violence did not change markedly be- tween 1967 and 1969. The rate of violent episodes remained constant at about eight per hour. -The nature of violence did change. Fatalities declined, and the proportion of leading characters engaged in violence or acting as killers declined. The former dropped from 73 percent to 64 percent: the latter from 19 to five percent. The consequence is that as many violent inci- dents occurred in 1969 as 1967, but a smaller proportion of characters were involved, and the violence was far less lethal. -Violence increased from 1967 to 1969 in cartoons and comedies. These two program types are not mutually exclusive in Gerbner's classi- fication system. Much of the increase in violence in comedies is attribut- able to the inclusion of cartoons in the comedy category. -Cartoons were the most violent type of program. The number of cartoon programs increased. from 37 in 1967 to 38 in 1969. The percen- tage of these programs containing some violence increased from 94 per- cent in 1967 to 97 percent in 1969. Although the percentage of leading 46 TELEVISION ANDGROWING UP characters involved in killing declined from 14 percent in 1967 to one percent in 1969, on the average 88 percent of leading characters in car- toons were involved in violence for the 1967-69 period. -Whereas in noncartoon shows in 1969 the agent of violence was a human being in 78 percent of the cases, in cartoons this role was depict- ed as human in only 23 percent of the cases. Nature, animals, and acci- dents are the agents of violence in more than three-quarters of the cases. Gerbner also tried to place the violence he observed into some social and moral context by looking at its time, place, and setting and by noting the kinds of people who engaged in violence and the kinds of people who were its victims. He found that: -In 1969, law enforcement agents appeared in four percent of the cartoon episodes and in 19 percent of the noncartoon. When they did play a role in noncartoon episodes, law enforcement agents were in- volved in violence in 79 percent of the cases. -Violence is more likely to take place in the past or the future (rather than in the present) and tends to be set in exotic, far-off, or uni- dentifiable places (rather than in surroundings familiar to viewers). -Violence is most frequently committed by white middle- and upper- class males, unmarried and in the young adult or middle years. -Most televised violence occurs between strangers or slight ac- quaintances. Gerbner's study combines Saturday morning programming with dra- matic programs in prime time evening hours. Barcus (1971) focused on Saturday morning programming in a content analysis using a sample of 19 hours broadcast in Boston by three network stations and one inde- pendent. He found: -In regard to broad program format categories, that commercial and promotional messages accounted for approximately 19 percent of the time; that when programs were roughly classified either as entertain- ment or as information, entertainment accounted for 89 percent of the time; and that 62 percent of total content consisted of animation. -In regard to violent content, that approximately three out of ten dramatic segments were "saturated" with violence; that 71 percent had at least one instance of human violence with or without the use of weap- ons; and that, although in 52 percent of the segments violence was di- rected at humans, in only four percent did this result in death or injury. Qualitative aspects of violence portrayals While these content analyses deal with the more readily quantifiable aspects of violence on television (e.g., How many acts? Who committed them? Where did the action take place?), they do not focus on the more qualitative aspects (e.g., Was the violent act related to character and plot development or was it gratuitous? How vivid or gory was the act -~ELEvISI~NCONTEN-I- 47 itself? What were the consequences?), which may well have a bearing on possible deleterious effects (see Heller and Polsky, 1971). In this connection it should be noted that the National Association of Broadcasters Television Code, the self-regulatory instrument of the industry, has definite strictures on these more qualitative aspects of the presentation of violence. For example, the code stipulates: "Such sub- jects as violence and sex shall be presented without undue emphasis and only as required by plot development or character delineation. Crime should not be presented as attractive or as the solution to human prob- lems and the inevitable retribution should be made clear." At another point the code states that "the detailed presentation of brutality or phys- ical agony by sight or by sound are not permissible." Unfortunately, Gerbner's study does not indicate the extent to which these industry guidelines for mitigating possible negative effects of violent content have actually been achieved in current television programming. POPULARITY OF VIOLENCE IN THE MEDIA Violence, of course, has been portrayed in entertainment since the earliest dramas were sung by traveling musicians. Clark and Blanken- burg's (1971) data on a variety of media-prime time television drama, movies, a family magazine, newspaper front pages, and television news -make it clear that violence appears regularly and frequently in all me- dia. It has been a major component of American mass media since their inception. Because of the crude measures used and the inherent differences be- tween media. direct comparisons among media as to violent content are not feasible. However, since people report using television much more than other media. they are presumably exposed to more fictional vio- lence on television than in any other medium. Clark and Blankenburg (1971), using TV Guide synopses from 1953 to 1969 as their source of information, observed some tendency for the frequency of violence in prime time evening programs to peak approxi- mately every four years. They found no evidence that such fluctuations were related either to national crime rates (a point to which we will re- turn) or to Congressional or other prominent criticism of violence in tel- evision. They did find evidence that is consistent with the interpretation that televised violence fluctuates as a function of the efforts of broad- casters to satisfy public taste and achieve as large an audience as possi- ble-a .53 correlation between percentage of programs classified as vio- lent and mean Nielsen ratings for all evening programs and a .49 correla- tion between the average Nielsen rating of programs classified as violent in one year and the number of such programs broadcast in the following year. Thus, the years that are high in violence also tend to be high in 48 TELEVISION AND GROWING UP overall ratings, and new season program formats are likely to vary ac- cording to what was popular with audiences the previous year. The in- vestigators report that the latest violence "peak" occurred in 1967. Heavy viewers of televised violence The remarkable popularity among the adult population of television drama that includes violence is a social reality that cannot be avoided. In order to study the audience size and some demographic characteristics of adult viewers of television violence, Israel and Robinson (1971) ana- lyzed marketing research data collected by W. R. Simmons and Asso- ciates. Using data from 1968, 1969, and 1970, and employing a nationally projectable sample of respondents who kept viewing diaries for two weeks, Israel and Robinson classified as heavy viewers of "vio- lent television" those who reported viewing 8.5 hours of programs clas- sified as violent during the two-week period in 1969-70. (Six hours was the cutoff point in 1967; in 1968 it was 7.5 hours.) Approximately 12 per- cent of the males and 11 percent of the females qualified as heavy vio- lence viewers on this criterion in 1969-70. These heavy viewers account for only about one-third of the total au- dience for the programs classified as "violent." These figures, projected nationally, mean that more than one-tenth of American adults watch more than four hours a week of television violence. The heavy viewers of violence are disproportionately clustered among males over 50 years old and among males with less than a full high school education. Crime statistics an.d televised violence Clark and Blankenburg (1971) tested the hypothesis that crime statis- tics in real life might vary with the frequency of fictional