The Psychology of Entertainment Media

bundet

The Psychology of Entertainment Media - Blurring the Lines Between Entertainment and Persuasion TOC II. The Programs Between the Ads: The Persuasive Power of Entertainment Fiction and Narrative Chapter Eleven - Paths From Television Violence to Aggression: Reinterpreting the Evidence Chapter Twelve - Between the Ads: Effects of Nonadvertising TV Messages on Consumption Behavior 10. CONSUMER CULTIVATION 183 be objectively compared with the same aspects of the real world. Examples include asking respondents to estimate the percentage of Americans who are involved in a violent crime; the percentage of the American workforce that consists of lawyers, doctors, or police ofﬁcers; the percentage of marriages that end in divorce; and so forth. Value-system or second-order measures pertain to the values, attitudes, and beliefs that might be cultivated from television content. Examples include asking respondents if they are afraid to walk alone at night, measuring their level of mistrust, their acceptance of violence, their belief that their spouse would be unfaithful, or their level of materialism. What makes this distinction interesting is that, according to Hawkins and Pingree’s (1982) review of the literature up to that point, the size and reliabil- ity of the cultivation effect tends to vary as a function of the type of judgments. The cultivation effect tends to be observed more strongly and more often for ﬁrst- order (demographic) than for second-order (value-system) beliefs. There are at least two nonmutually exclusive explanations for this pattern of ﬁndings. One ex- planation is that only one type of judgment—judgments related to the prevalence of particular constructs that occur often on television—is inﬂuenced by television viewing. Judgments related to values and attitudes that might be developed from the lessons of television, however, are simply not affected by amount of viewing. A second explanation is that the judgments differ in terms of the processes involved in constructing them. This possibility implies at least two important things: Tele- vision may inﬂuence each of the two types of judgments in different ways, and different factors may mediate or moderate the relation between television viewing and the two types of judgments. Cultivation Judgments As Psychological Judgments. If one scrutinizes the types of judgments that have been classiﬁed as ﬁrst- or second-order judgments in terms of how psychologists categorize them, it is apparent that they differ in fundamental ways. First-order judgments for the most part consist of judgments of probability or set size (Shrum, 1995). Examples include estimating risk (e.g., risk of crime) and estimating the number or percentage of instances in which a particular category (e.g., millionaire) occurs within a larger, superordinate category (e.g., Americans). Second-order judgments typically consist of attitude, value, or belief judgments. Examples include beliefs related to trust, the extent to which the world is a mean and violent place, and whether achievement is reﬂected in product ownership, just to name a few. Viewed this way, it is quite possible that ﬁrst- and second-order judgments differ in terms of the way in which they are constructed. Moreover, decades of research in social and cognitive psychology have detailed the processes involved in constructing these judgments, which is useful in determining the role that certain types of inputs (e.g., television information) may play in this process. 184 SHRUM, BURROUGHS, RINDFLEISCH A Process Model of First-Order Cultivation Judgments How Are the Judgments Constructed? In attempting to construct a process model for cultivation effects, it seemed reasonable to start with ﬁrst-order (set-size and probability) judgments, given that those are the types of judgments for which cultivation effects have tended to be more robust and consistent. So how are those types of judgments constructed? Research by Tversky and Kahneman (1973; see also Kahneman & Tversky, 1982) suggests that these types of judgments are often made through the application of particular heuristics, or rules of thumb. Speciﬁ- cally, judgments of set size and probability tend to be based on the application of the availability heuristic (Tversky & Kahneman, 1973) or the simulation heuristic (Kahneman & Tversky, 1982). In using the availability heuristic , people base their judgments of set size or probability on how easy a relevant example comes to mind: The easier it is to recall, the higher the estimate. Thus, people tend to estimate that words in the English language that start with the letter K occur more often than words that have K as the third letter (Tversky & Kahneman, 1973, Study 3), even though the opposite is in fact the case. This result is presumably because words tend to be organized in memory according to their ﬁrst letter, and thus words that start with K are more easily recalled. Similarly, 80% of people tend to estimate that accidents account for more deaths than do strokes, even though strokes account for about 85% more deaths than do accidents (Lichtenstein, Slovic, Fischhoff, Layman, & Combs, 1978). Again, this is presumably because accidents are easier to recall or imagine than strokes. When judging set size or probability, a relevant example may not be available in memory (i.e., present in memory) or, if available, not particularly accessible (i.e., not easily retrieved). Thus, the availability heuristic cannot be applied. In these instances, people may resort to basing their estimates on the ease with which a relevant example can be imagined. This is an example of the simulation heuristic . Supporting this notion, research has shown that when people are induced to imagine a particular event such as winning a contest (Gregory, Cialdini, & Carpenter, 1982) or contracting a disease (Sherman, Cialdini, Schwartzman, & Reynolds, 1985), they provide higher estimates of the probability that they will experience these events compared with people who are not induced to imagine such events, and these relations are mediated by ease of imagining (Sherman et al., 1985). Relation to Media Consumption. The studies just noted, along with numerous others, clearly document that accessibility of relevant examples or ease of con- struction of a scenario inﬂuences estimates of set size and probability. Those with more accessible examples or greater ease of construction provide higher estimates. This has been shown to occur in both experimental studies and ﬁeld studies. But what inﬂuences this accessibility? Clearly, in the experimental studies, accessibil- ity is manipulated. But what of the ﬁeld studies of Lichtenstein et al. (1978)? Why did people tend to greatly overestimate the number of deaths caused by accidents but greatly underestimate deaths caused by strokes? Lichtenstein et al. speculated 10. CONSUMER CULTIVATION 185 that accessibility is inﬂuenced by media coverage, suggesting that media public- ity of such dramatic events as accidents and homicides increases accessibility of these examples relative to less dramatic and publicized causes of death such as strokes. This speculation was supported by a content analysis of newspaper articles showing just such differences in coverage (Combs & Slovic, 1979). These studies suggest that media consumption may inﬂuence the accessibility of constructs that are commonly portrayed. It follows, then, that differences in media consumption (all other things being equal) may inﬂuence levels of accessibility of relevant constructs. If so, then for judgments of set size and probability, if the availability or simulation heuristic is used, then heavier media consumers should provide higher estimates of set size or probability than lighter media consumers. In fact, this is exactly what cultivation theory predicts. Constructing the Process Model. From this point, it is a fairly simple leap to the development and testing of a process model of cultivation. Such a model predicts that heavier television viewing will make relevant examples more ac- cessible in memory than lighter viewing (Proposition 1) and that this enhanced accessibility will result in higher estimates, indicating a mediating role of acces- sibility (Proposition 2). Note, however, that the notion that television examples would be used in the construction of these judgments is not necessarily intuitive. When estimating the prevalence of lawyers or police ofﬁcers in the workforce, it is unlikely that people would consciously use an example of a TV lawyer or police ofﬁcer to construct this judgment. Thus, if such TV examples are indeed used in constructing real-world judgments, then people are likely unaware of the source of these examples (because they are made fairly quickly) and thus do not discount the television examples as an invalid source (Proposition 3). Finally, con- ditions that facilitate or inhibit the use of judgmental heuristics such as availability and simulation should correspondingly increase or decrease the cultivation effect. These conditions include motivation to process information, which should inhibit the cultivation effect (Proposition 4) and lack of ability to process information, which should facilitate the cultivation effect (Proposition 5). Approximately a dozen studies have validated and replicated each of these key propositions (for a review, see Shrum, 2002). Addressing Second-Order Cultivation Judgments As noted elsewhere (Shrum, 1995, 2002), the psychological process model shown in Figure 10.1 is mute with respect to second-order judgments such as attitudes and values. This is unfortunate because, as Gerbner et al. (2002) have noted, it is the extrapolation or symbolic transformation of the television message into more general perspectives and ideologies that is perhaps more interesting and better captures the concept of cultivation theory than a focus on simple perceptions and beliefs that are captured by ﬁrst-order measures. For this reason, it seems useful 186 SHRUM, BURROUGHS, RINDFLEISCH to extend the process model to include second-order judgments such as attitudes and values. In developing the model for ﬁrst-order judgments, we ﬁrst started with the question of how such judgments are made, working backward to understand how television information might inﬂuence these judgments. In applying this approach to second-order judgments, it quickly becomes apparent that ﬁrst- and second- order judgments are made quite differently. For one, judgments of set size and probability are virtually always memory based (Hastie & Park, 1986). That is, when asked to form a judgment about the probability of being involved in a violent crime or the incidence of millionaires in the United States, people would not be likely to have such answers stored in memory. Rather, they would construct them by recalling relevant examples or scenarios. Thus, ﬁrst-order judgments are likely constructed at the time the judgment is required (e.g., responding to a research query; playing Trivial Pursuit). For this reason, we would expect that memory for these examples would correlate with both the independent (TV viewing) and dependent ( judgments) variables (Hastie & Park, 1986). In fact, that is what the studies have consistently shown (Busselle & Shrum, in press; Shrum, 1996). This process also implies that conditions operating at the time of judgment would be more likely to impact the TV–judgment relation than would conditions operating at the time of encoding or viewing. Consistent with this notion, judgment conditions such as time pressure (Shrum, 1999a), task involvement (Shrum, 2001), and source discounting (Shrum et al., 1998) have been shown to moderate the cultivation effect. Conversely, conditions or variables operating at the time of viewing (e.g., attention while viewing, intention to view, perceived reality of television, need for cognition) have been shown to have virtually no effect on either the magnitude of ﬁrst-order judgments or memory for TV information (Busselle & Shrum, in press; O’Guinn & Shrum, 1997; Shrum, 1996, 2001; Shrum et al., 1998). In contrast, the construction of values, attitudes, and beliefs is likely made in a different manner. It is of course possible that attitudes and beliefs could be con- structed in a memory-based fashion. When asked to provide an attitude toward a particular object, people may attempt to recall relevant information (both cogni- tions and affect) and then construct their attitude in real time. This would likely occur when people do not have a readily accessible attitude or belief to provide when a request for attitude expression is made. If they did have an attitude or be- lief readily accessible, they would instead simply retrieve their prior constructed attitude or belief and report it (Hastie & Park, 1986; see also Carlston, 1980; Lichtenstein & Srull, 1985, 1987; Lingle & Ostrom, 1979). But consider the types of attitudes and beliefs that are typically measured in cultivation research. These measures assess the extent to which people believe the world is a violent place, are afraid to walk alone at night, approve of violence by police, believe crime is the most important political issue, do not trust others, and so forth. These types of beliefs are commonly used in everyday life. Thus, they are likely to already exist for most people, having been formed long ago 10. CONSUMER CULTIVATION 187 and reinforced or updated on a consistent basis. This is even more so the case for constructs such as personal values. By deﬁnition, values are stable and enduring beliefs that everyone possesses and that serve as guides to behavior over the course of a lifetime (Rokeach, 1973). Thus, they are formed at a relatively early age and then changed (either made stronger or altered) as new information comes in. This process of constructing judgments on the basis of incoming information (as opposed to retrieved information) is what Hastie and Park (1986) term an online judgment. If it is true that these types of judgments are formed in an online fashion, important implications are made for the role that television viewing might play in the formation of those judgments. In particular, it suggests that these types of judgments are likely made (developed, reinforced, or altered) during the viewing process. If so, it also implies that conditions operating during viewing may affect the inﬂuence of television information rather than conditions operating at the time the judgment is required by some external situation (e.g., being asked by a researcher). Note that this is essentially the opposite of the process involved with ﬁrst-order (memory-based) cultivation judgments, which depend on recall at the time the judgment is required and thus should be affected by conditions present at that time rather than at the time of viewing. Supportive Evidence. The notion that second-order cultivation judgments might be inﬂuenced by conditions at the time of viewing is a proposition that was addressed in a recent study (Burroughs et al., 2002). In that study, we investigated the relation between television viewing and the consumer value of materialism. Materialism is commonly viewed as the value placed on the acquisition of material objects, such as expensive cars, homes, and clothes (Richins & Dawson, 1992). Because content analyses have consistently shown that the world portrayed on tele- vision tends to be more afﬂuent and materialistic than the real world (Hirschman, 1988; Lichter et al., 1994; O’Guinn & Shrum, 1997), we expected that, consis- tent with cultivation theory, these materialistic messages would be internalized by viewers, resulting in higher levels of materialism for those who viewed relatively more television in general. However, we expected this positive relation between television viewing and materialism to be moderated by certain factors that might affect the processing of the message during viewing. These factors included the degree to which viewers tend to be attentive while viewing and the extent to which viewers tend to elaborate on the television message while viewing. We expected that those who pay more attention while viewing would be more persuaded by the television message than those who pay less attention, and we also expected that those who elaborate more on the message (those higher in need for cognition; Cacioppo & Petty, 1982) would also be more affected than those who elaborate less. The results were as expected. We found that level of television viewing was related to materialism: The more people viewed television, the more materialistic they were. However, also as expected, this relation was moderated by the two 188 SHRUM, BURROUGHS, RINDFLEISCH process variables. Speciﬁcally, the positive relation between television viewing and materialism was stronger for those who paid more attention to the program while viewing than for those who paid less attention and stronger for those higher in need for cognition than for those lower in need for cognition. Implications for Model Development and Causality The moderating role of attention while viewing and need for cognition is consistent with our theorizing that the process of cultivation for second-order cultivation judgments tends to occur during viewing. The variables of attention and need for cognition were intended to capture processes that were taking place during the viewing process. As such, it is highly unlikely that such variables would moderate the cultivation effect if in fact the judgments were memory based, at least in the pattern we observed. This last phrase is an important qualiﬁer. It is possible that greater elaboration and greater attention could indeed have an effect on the extent to which television information is used in a memory-based judgment. However, as has been shown in previous studies (Shrum, 2001; Shrum et al., 1998) the effect should be just the opposite of the one we observed: Greater attention and elaboration should lead to more source discounting (discounting the television information because it is not veridical) and thus reduce rather than inﬂate the cultivation effect. The pattern of results we observed also has important implications for causality. As we noted earlier, correlational results are always open to alternative explana- tions of causal paths. However, the pattern of moderating effects that we observed are difﬁcult to explain in reverse causality or third-variable terms. In particular, attention and elaboration are process variables that necessarily occur during rather than prior to viewing, making a reverse causality explanation untenable. And, al- though it is still possible that some third variable is driving the TV–materialism relation, that variable would also have to account for the two moderating effects we observed. Given that constraint, it is unclear what that third variable might be. Limitations of the Study There is one important limitation to the results of Burroughs et al. (2002). That limitation pertains to the use of need for cognition as a surrogate measure of elaboration during viewing. It could certainly be argued that those higher in need for cognition would not enjoy such a cognitively easy task as television viewing. Moreover, if they did view, they might be more prone to counterarguing than sup- port arguing. In fact, Burroughs et al. found that need for cognition was indeed negatively correlated with television viewing. However, we would argue that for those who are high in need for cognition who do decide to view frequently, a continual counterarguing of the television message would be a particularly miser- able experience. Rather, we expect that those high in need for cognition who view frequently are the ones who enjoy watching television and thus would be more 10. CONSUMER CULTIVATION 189 likely to suspend disbelief and elaborate extensively than those high in need for cognition who are lighter viewers. Our pattern of results is consistent with that notion: It is the people who are both heavy viewers and high in need for cognition that exhibit the highest levels of materialism. CONCLUSION The arguments that we have presented in this chapter for the processes involved in the construction of second-order cultivation judgments are just that—arguments. Although we have discussed some empirical ﬁndings that support our reasoning, there is still quite a bit of work to do in terms of ﬂeshing out the entire process. Ideally, one would end up with a process model for second-order judgments that is similar to the one shown in Figure 10.1 for ﬁrst-order judgments, one that speciﬁes testable propositions that address the processes that mediate and moderate the effect of television viewing on judgments. Doing so would provide a major step in establishing the causal impact that television viewing has on the gamut of human judgment and behavior. REFERENCES Bargh, J. A. (1997). The automaticity of everyday life. In. R. S. Wyer (Ed.), The automaticity of everyday life: Advances in social cognition (Vol. 10, pp. 1–61). Mahwah, NJ: Lawrence Erlbaum Associates. Bryant, J., Carveth, R. A., & Brown, D. (1981). Television viewing and anxiety: An experimental investigation. Journal of Communication, 31 (1), 106–119. Bryant, J., Zillmann, D. (Eds.). (2002). Media effects: Advances in theory and research (2nd ed., p. 87). Mahwah, NJ: Lawrence Erlbaum Associates. Burroughs, J. E., Shrum, L. J., & Rindﬂeisch, A. (2002). Does television viewing promote materialism? Cultivating American perceptions of the good life. In S. Broniarczyk & K. Nakamoto (Eds.), Advances in consumer research (Vol. 28, pp. 442–443). Provo, UT: Association for Consumer Research. Bushman, B. J., & Anderson, C. A. (2001). Media violence and the American public: Scientiﬁc facts versus media misinformation. American Psychologist, 56 (6–7), 477–489. Busselle, R. W., & Shrum, L. J. (in press). Media exposure and the accessibility of social information. Media Psychology . Cacioppo, J. T., & Petty, R. E. (1982). The need for cognition. Journal of Personality and Social Psychology, 42 , 116–131. Carlston, D. E. (1980). The recall and use of traits and events in social inference processes. Journal of Experimental Social Psychology, 16 , 303–328. Combs, B., & Slovic, P. (1979). Newspaper coverage of causes of death. Journalism Quarterly, 56, 837–849. DeFleur, M. L. (1964). Occupational roles as portrayed on television. Public Opinion Quarterly, 28 , 54–74. Dixon, T., & Linz, D. (2000). Overrepresentation and underrepresentation of African Americans and Latinos as lawbreakers on television news. Journal of Communication, 50, 131–154. 190 SHRUM, BURROUGHS, RINDFLEISCH Gerbner, G., & Gross, L. (1976). Living with television: The violence proﬁle. Journal of Communica- tion, 26 (2), 173–199. Gerbner, G., Gross, L., Morgan, M., & Signorielli, N. (1980). The “mainstreaming” of America: Violence proﬁle no. 11. Journal of Communication, 30 (3), 10–29. Gerbner, G., Gross, L., Morgan, M., & Signorielli, N. (1986). Living with television: The dynam- ics of the cultivation process. In J. Bryant & D. Zillmann (Eds.), Perspectives on media effects (pp. 17–40). Hillsdale, NJ: Lawrence Erlbaum Associates. Gerbner, G., Gross, L., Morgan, M., & Signorielli, N. (1994). Growing up with television: The cultiva- tion perspective. In J. Bryant & D. Zillmann (Eds.), Media effects: Advances in theory and research (pp. 17–41). Hillsdale, NJ: Lawrence Erlbaum Associates. Gerbner, G., Gross, L., Morgan, M., Signorielli, N., & Shanahan, J. (2002). Growing up with television: Cultivation processes. In J. Bryant & D. Zillmann (Eds.), Media effects: Advances in theory and research (2nd ed., pp. 43–67). Mahwah, NJ: Lawrence Erlbaum Associates. Gregory, L. G., Cialdini, R. B., & Carpenter, K. M. (1982). Self-relevant scenarios as mediators of likelihood estimates and compliance: Does imagining make it so? Journal of Personality and Social Psychology, 43 , 89–99. Hastie, R., & Park, B. (1986). The relationship between memory and judgment depends on whether the judgment task is memory-based or on-line. Psychological Review, 93 , 258–268. Hawkins, R. P., & Pingree, S. (1982). Television’s inﬂuence on constructions of social reality. In D. Pearl, L. Bouthilet, & J. Lazar (Eds.), Television and behavior: Ten years of scientiﬁc progress and implications for the eighties (Vol. 2, pp. 224–247). Washington, DC: Government Printing Ofﬁce. Hawkins, R. B., & Pingree, S. (1990). Divergent psychological processes in constructing social re- ality from mass media content. In N. Signorielli & M. Morgan (Eds.), Cultivation analysis: New directions in media effects research (pp. 33–50). Newbury Park, CA: Sage. Head, S. W. (1954). Content analysis of television drama programs. Quarterly of Film, Radio, and Television, 9 , 175–194. Hirsch, P. (1980). The scary world of the nonviewer and other anomalies: A reanalysis of Gerbner et al.’s ﬁndings on cultivation analysis. Communication Research, 7 , 403–456. Hirschman, E. C. (1988). The ideology of consumption: A structural–syntactical analysis of “Dallas” and “Dynasty.” Journal of Consumer Research, 15 , 344–359. Hughes, M. (1980). The fruits of cultivation analysis: A reexamination of some effects of television watching. Public Opinion Quarterly, 44 , 287–302. Kahneman, D., & Tversky, A. (1982). The simulationheuristic. In D. Kahneman, P. Slovic, & A. Tversky (Eds.), Judgment under uncertainty: Heuristics and biases . New York: Cambridge University Press. Lichtenstein, M., & Srull, T. K. (1985). Conceptual and methodological issues in examining the re- lationship between consumer memory and judgment. In L. F. Alwitt & A. A. Mitchell (Eds.), Psychological processes and advertising effects: Theory, research and application (pp. 113–128). Hillsdale, NJ: Lawrence Erlbaum Associates. Lichtenstein, M., & Srull, T. K. (1987). Processing objectives as a determinant of the relationship between recall and judgment. Journal of Experimental Social Psychology, 23 , 93–118. Lichtenstein, S., Slovic, P., Fischhoff, G., Layman, M., & Combs, B. (1978). Judged frequency of lethal events. Journal of Experimental Psychology: Human Learning and Memory, 6 , 551–578. Lichter, S. R., Lichter, L. S., & Rothman, S. (1994). Prime time: How TV portrays American culture . Washington, DC: Regnery Publishing. Lingle, J. H., & Ostrom, T. M. (1979). Retrieval selectivity in memory-based impression judgments. Journal of Personality and Social Psychology, 37 , 180–194. Morgan, M., & Shanahan, J. (1996). Two decades of cultivation research: An appraisal and meta- analysis. In B. R. Burleson (Ed.), Communication yearbook 20 (pp. 1–45). Newbury Park, CA: Sage. 10. CONSUMER CULTIVATION 191 Newcomb, H. (1978). Assessing the violence proﬁles of Gerbner and Gross: A humanistic critique and suggestion. Communication Research, 5 , 264–282. Ogles, R. M. (1987). Cultivation analysis: Theory, methodology and current research on television- inﬂuenced constructions of social reality. Mass Comm Review 14 (1, 2), 43–53. O’Guinn, T. C., & Shrum, L. J. (1997). The role of television in the construction of consumer reality. Journal of Consumer Research, 23 , 278–294. Oliver, M. B. (1994). Portrayals of crime, race, and aggression in “reality based” police shows: A content analysis . Journal of Broadcasting & Electronic Media, 38 , 179–192. Richins, M. L., & Dawson, S. (1992). A consumer values orientation for materialism and its measure- ment: Scale development and validation. Journal of Consumer Research, 19 , 303–316. Rokeach, M. (1973). The nature of human values . New York: Free Press. Shanahan, J., & Morgan, M. (1999). Television and its viewers: Cultivation theory and research . New York: Cambridge University Press. Sherman, S. J., Cialdini, R. B., Schwartzman, D. F., & Reynolds, K. D. (1985). Imagining can heighten or lower the perceived likelihood of contracting a disease: The mediating effect of ease of imagery. Personality and Social Psychology Bulletin, 11 , 118–127. Shrum, L. J. (1995). Assessing the social inﬂuence of television: A social cognition perspective on cultivation effects. Communication Research, 22 , 402–429. Shrum, L. J. (1996). Psychological processes underlying cultivation effects: Further tests of construct accessibility. Human Communication Research, 22 , 482–509. Shrum, L. J. (1999a). The effect of data-collection methods on the cultivation effect: Implications for the heuristic processing model of cultivation effects. Paper presented at the meeting of the International Communication Association, San Francisco, CA. Shrum, L. J. (1999b). The relationship of television viewing with attitude strength and extremity: Implications for the cultivation effect. Media Psychology, 1 , 3–25. Shrum, L. J. (1999c). Television and persuasion: Effects of the programs between the ads. Psychology & Marketing, 16 , 119–140. Shrum, L. J. (2001). Processing strategy moderates the cultivation effect. Human Communication Research, 27 , 94–120. Shrum, L. J. (2002). Media consumption and perceptions of social reality: Effects and underlying processes. In J. Bryant & D. Zillmann (Eds.), Media effects: Advances in theory and research (2nd ed., pp. 69–95). Mahwah, NJ: Lawrence Erlbaum Associates. Shrum, L. J., & Bischak, V. D. (2001). Mainstreaming, resonance, and impersonal impact: Testing moderators of the cultivation effect for estimates of crime risk. Human Communication Research, 27 (2), 187–215. Shrum, L. J., Wyer, R. S., & O’Guinn, T. C. (1998). The effects of television consumption on social perceptions: The use of priming procedures to investigate psychological processes. Journal of Consumer Research, 24 , 447–458. Smythe, D. W. (1954). Reality as presented by television. Public Opinion Quarterly, 18 , 143–156. Tversky, A., & Kahneman, D. (1973). Availability: A heuristic for judging frequency and probability. Cognitive Psychology, 5 , 207–232. Van den Bulck, J. (2003). Is the mainstreaming effect of cultivation an artifact of regression to the mean? Journal of Broadcasting & Electronic Media, 47 (2), 289–295. Volgy, T., & Schwarz, J. (1980). Television entertainment programming and sociopolitical attitudes. Journalism Quarterly, 57 , 150–155. Zillmann, D. (1980). Anatomy of suspense. In P. H. Tannenbaum (Ed.), The entertainment functions of television (pp. 133–163). Hillsdale, NJ: Erlbaum. CHAPTER ELEVEN Paths From Television Violence to Aggression: Reinterpreting the Evidence George Comstock Syracuse University Entertainment provides at least three gratiﬁcations: a respite from the anxi- eties and pressures of everyday life; the opportunity to compare oneself with the demeanor, possessions, and behavior of others; and a means of keeping up with what is transpiring in the world. These gratiﬁcations have been particu- larly well documented in the case of television (Comstock & Scharrer, 1999), where a variety of measures of stress and interpersonal conﬂict predict greater afﬁnity for or consumption of television; where viewers tend to watch those on the screen like themselves in gender, age, or ethnicity more attentively (which would make comparisons more meaningful); and where there is a common be- lief that something can be learned from all types of programming and keeping abreast embraces both events in the world at large and the varied portrayals and depictions offered by the medium in news, sports, and entertainment program- ming. There are also less-positive consequences of attending to entertainment, with one of the most investigated by the social and behavioral sciences the fa- cilitation of aggressive and antisocial behavior by violent television and ﬁlm portrayals. The empirical evidence on the inﬂuence of violent television entertainment on aggression, bolstered by an extraordinary seven meta-analyses with varying emphases, often has been interpreted as supporting the view that attitudes serve as the link between exposure to violent programs and behavior. The evidence, in 193 194 COMSTOCK fact, gives equal support to the hypothesis that the link is the availability of genres of behavior in the minds of viewers—that is, the readiness by which they may be retrieved—rather than dispositions toward behavior. From this latter perspective, television violence is processed by young viewers in a manner similar to advertising within the elaboration likelihood model (ELM; Cacioppo & Petty, 1985; Petty & Cacioppo, 1990). The factors that seemingly promote favorable evaluations of aggressive behavior instead govern the salience of such behavior. Violent portrayals operate like television commercials, which most of the time, because they are low in persuasive argumentation, affect consumers by maintaining the salience of particular brands among purchase options (Comstock & Scharrer, 1999). The research on television violence and aggressive and antisocial behavior has been voluminous but has been generally assigned to a sphere of scientiﬁc endeavor quite distinct from consumer psychology. This is more a product of myopia and an inclination to honor the apparent boundaries of paradigms than it is one of distinc- tiveness among theories or any interest in decidedly different types of outcomes. In fact, it is a quick and comfortable journey that leads face-to-face with consumer psychology if one travels the backward loop from Bandura’s (1986) social cogni- tive theory to the health belief model (Becker, 1974). The health belief model was derived from social learning theory, an earlier version of what Bandura now calls social cognitive theory. The health belief model essentially argues that behavior related to health—such as food preferences, cigarette and alcohol consumption, seat belt use, and cancer examinations—can be changed by manipulating beliefs about personal risk, the availability of effective means of risk reduction, and the ease of access to and social acceptability of these means of risk reduction. It has been the conceptual model for such large-scale federally ﬁnanced interventions as the Stanford cardiovascular ﬁeld experiments (Farquar et al., 1977, 1990), in which persuasive campaigns, using the mass media and counseling by physicians, were employed to encourage individuals to behave in ways that would reduce the risk of cardiovascular mishap. Social cognitive theory, of course, is the un- derpinning for much of the experimental research on young children’s increased aggressiveness as a consequence of exposure to violent portrayals (e.g., Bandura, Ross, & Ross, 1963a, 1963b). The health belief model makes it clear that social cognition is a theory of persuasion, although one that emphasizes behavior, mod- eling of that behavior by attractive persons, relevance achieved through attributes of the portrayal that encourage identiﬁcation, and, in its application to the area of television violence and aggression, persuasion that is unintended. Thus, the con- nection between consumer psychology and television violence is not limited to the frequent consumption of media content that is a vehicle for commercials but extends to the fundamentals of theory, speciﬁcally, in the present case, the analogy provided by the peripheral or heuristic path of dual-processing theories such as the ELM. 11. FROM TELEVISION VIOLENCE TO AGGRESSION 195 THE MOST USEFUL GATEWAY At this point in time, the most useful gateway to the evidence on television violence and aggressive and antisocial behavior is through meta-analyses (Comstock & Scharrer, 2003). Meta-analysis estimates the magnitude of the relationships among variables. These estimates are more reliable and more valid than those produced by a single study (Hunt, 1997). It thereby enhances the quality and veracity of evidence. The role of meta-analysis, and its appeal, is aptly conveyed by the title of Morton Hunt’s (1997) account of its development and uses, How Science Takes Stock . It has become an important tool for drawing inferences when several and often many empirical studies exist on the same topic. The essence of meta-analysis is the estimating of the size of the relationship between two variables. Typically, there is an exhaustive search for unpublished as well as published studies. This is directed at the ﬁle drawer problem—published studies alone may result in an overestimate of the magnitude of a relationship because studies showing larger, statistically signiﬁcant outcomes are more likely to be published (Rosenthal, 1979). In turn, studies can be scored on various dimensions for quality and for ecological validity (conditions that parallel everyday life) so that the analysis eventually can focus on the studies that have the strongest claims for generalizing beyond their particular sets of data. In meta-analysis jargon, the goal is to estimate the effect size —the magnitude of the association between two variables. Despite this language, there is nothing about meta-analysis that ensures that the independent variable contributes to or in any way is causally related to the dependent variable. This is an inference that is left to the analyst. The beneﬁts of meta-analysis are the use of all retrievable data to estimate the magnitude of a relationship, the greater reliability and validity that result from the enlarged N , and the ability to code studies for any and every conceivable attribute so that the analyst can focus on those with particular characteristics or qualities. For example, in the case of violent television portrayals and aggressive or antisocial behavior, the analyst could isolate those experiments where a portrayal of justiﬁed aggression was the treatment and compare the effect size for the dependent variable with effect sizes for other treatments. The estimate based on several experiments would have greater reliability and validity than the estimate from the original 1963 experiment by Berkowitz and Rawlings (i.e., if the additional experiments were equivalent as a group in quality and ecological validity to the original). In effect, studies are treated analogously to respondents in a survey. The quality and validity of interpretation depends on the representativeness of the sample (in the case of meta-analysis, a census) and the selection of questions addressed to the sample (in the case of meta-analysis, the studies). The basic calculation is as follows: Effect Size = Mean t − Mean c ESD 196 COMSTOCK Where t = treatment, c = control, and ESD = estimate of the standard devia- tion. Because the standard deviation represents the same proportion of cases for all normal distributions, this produces an effect size that can be pooled with others for an aggregate estimate. The formula obviously ﬁts only experimental designs, but procedures have been developed to estimate the equivalent for surveys and other designs. In addition, procedures also have been developed to estimate the statistical signiﬁcance of any effect size and fail-safe numbers—the number of additional outcomes with null results that would be necessary to reduce the obtained estimate to null. Thus, meta-analysis in its current form makes it possible for the conﬁdence of the analyst to rest not only on the magnitude of the effect size but also on the estimated probability that an outcome is attributable to sampling variability and the estimated quantity of null results that would have to be uncovered to overturn an observed effect size. Nevertheless, meta-analysis is not a substitute for interpretation. The fretful analyst still must decide whether the outcomes of experiments are generalizable to everyday circumstances and whether positive correlations between two variables in survey data represent the operation of a third variable or causation, as well as the direction of causation. In the case of television and ﬁlm violence and aggressive and antisocial be- havior, interpretation is confronted with a large array of data—seven quantitative aggregations of study outcomes. The most comprehensive (Paik & Comstock, 1994) presents almost 1,150 instances in which an effect size was calculated. Fur- thermore, the outcomes of individual studies sometimes may supply information that clearly adds to, extends, and qualitatively enriches rather than is merely rep- resented in a meta-analysis. As a result, meta-analysis should not be thought of as a means by which all scientiﬁc questions are answered but as a source of particu- larly reliable and valid estimates of the relationships among variables that may be importantly augmented or qualiﬁed by the outcomes of particular studies. In the present instance, the meta-analytic data lead to a number of conclusions unambiguously thoroughly supported by empirical evidence. In addition, the ex- amination of the outcomes of a particularly compelling study leads to a surprise that suggests a revision in theory. Andison (1977), in a pioneering effort, simply categorized the outcomes of 67 experiments and surveys as to the direction and size of the relationship between vi- olence viewing and aggressive or antisocial behavior. Hearold (1986) was the ﬁrst to apply the now widely accepted meta-analytic paradigm in which the standard deviation is the criterion for assessing the magnitude of the relationship between variables to the literature on media and behavior. A student of Eugene Glass at the University of Colorado (who is credited with developing meta-analysis in the 1970s in an attempt to quantitatively discredit H. J. Eysenck’s claims that psychother- apy was ineffective) examined more than 1,000 relationships (drawn from 168 separate studies) between exposure to anti- and prosocial portrayals and anti- and 11. FROM TELEVISION VIOLENCE TO AGGRESSION 197 prosocial behavior. 1 Wood, Wong, and Cachere (1991) examined 23 experiments in and out of the laboratory in which the dependent variable was “unconstrained interpersonal aggression” among children and teenagers. Allen, D’Alessio, and Brezgel (1995) aggregated the data from 33 laboratory-type experiments in which the independent variable was exposure to sexually explicit video or ﬁlm portrayals and the dependent variable was aggression. Hogben (1998) conﬁned himself to 56 coefﬁcients drawn from studies measuring everyday viewing but included a wide array of aggression-related responses, including, in addition to aggressive or an- tisocial behavior, hostile attitudes, personality variables, and, in one case (Cairns, Hunter, & Herring, 1980), the content of invented news stories. Bushman and Anderson (2001) tracked the correlation between exposure to violent television or ﬁlm portrayals in experimental and nonexperimental designs in 5-year intervals over a 25-year period. Paik and Comstock (1994), in a comprehensive updating of Hearold’s assessment of the relationship between exposure to television vio- lence and aggressive and antisocial behavior, included 82 new studies for a total of 217 that produced 1,142 coefﬁcients between the independent and dependent variables. 2 These data represent the behavior of many thousands of persons ranging in age from preschool to adulthood, an assortment of independent and dependent variables, and a variety of research methods. For example, Paik and Comstock (1994) included the full range of ages, a variety of types of programming and nine categories of behavior, and laboratory-type experiments, ﬁeld experiments, time series, and surveys. These are all typically positive features of meta-analyses, which, as surveys of a literature, take on the characteristics of the area of inquiry in contrast to the more limited (and limiting) attributes of a single study. 1 The meta-analytic paradigm made its ﬁrst public appearance in the 1976 presidential address by Glass (1976) at the annual meeting of the American Educational Research Association in San Fransisco. This justiﬁably places him foremost among the pioneers of the method. However, at about the same time Robert Rosenthal at Harvard was at work on a similar scheme (Rosenthal & Rubin, 1978). Neither knew of what the other was doing (Hunt, 1997), an example of the frequent independence of innovations from the ventures of speciﬁc innovators. 2 Paik and Comstock (1994) focused on the most substantial portion of the Hearold (1986) meta- analysis. Hearold examined all possible combinations of antisocial and prosocial portrayals and be- havior, including the effect size between exposure to prosocial portrayals and antisocial behavior and the effect size between exposure to antisocial portrayals and prosocial behavior (although the data on these pairings expectedly was not voluminous). However, she did not include erotica or pornographic portrayals among the independent variables, and she included among her antisocial dependent vari- ables such outcomes as stereotyping, passivity, and feelings of powerlessness. Aggressive portrayals and behavior nevertheless made up the largest number of independent and dependent variables. Paik and Comstock conﬁned themselves to aggressive portrayals and outcomes, except for including erotica and pornographic portrayals among their independent variables. Thus, they updated the major portion of Hearold’s analysis and extended it in terms of coverage by including erotica and pornographic por- trayals and in terms of method by including estimates of statistical signiﬁcance and fail-safe numbers. 198 COMSTOCK These analyses irrefutably conﬁrm that there is a positive correlation between exposure to violent television and movie portrayals and engagement in aggressive or antisocial behavior. This holds for the data from experiments and for the data from surveys. All seven of the quantitative aggregations of data can be invoked on behalf of this outcome including the four not conﬁned to a narrow focus on partic- ular measures of exposure or quite speciﬁc outcomes (such as unconstrained in- terpersonal aggression)—the initial pioneering effort by Andison (1977), Hearold (1986), Bushman and Anderson (2001), and Paik and Comstock (1994). How- ever, interpretation quickly strides to center stage because this datum is not very informative about the processes responsible for the relationship. Experiments un- ambiguously document causation within their circumstances, but it is a matter of judgment whether their outcomes can be generalized to other circumstances, and positive correlation coefﬁcients within surveys document association in everyday life but by themselves are insufﬁcient to infer causation (Cook & Campbell, 1979). INTERPRETING THE EVIDENCE The case for causation is quite strong. It rests on two factors readily observable in the meta-analyses. The ﬁrst is the consistency of the outcomes for the experi- mental designs, where the positive effect sizes are quite robust ( r . = . 40 in Paik & Comstock 1994), statistical signiﬁcance is readily achieved ( p exceeds four digits in Paik & Comstock), and fail-safe numbers are huge (over 700,000 in Paik & Comstock). The second is the conﬁrmation by the survey designs—where the ef- fect sizes are more modest but respectable ( r . = . 19 in Paik & Comstock), while statistical signiﬁcance and fail-safe numbers (the quantity of null ﬁndings neces- sary to reduce a coefﬁcient to statistical insigniﬁcance) remain impressive—that the condition necessary to infer causation in everyday life exists. This condition is quite explicit: a positive correlation between everyday violence viewing and ev- eryday aggressive or antisocial behavior (i.e., a correlation outside the laboratory) that stoutly resists explanation other than by some causal contribution of viewing to behavior. The meta-analyses supply the correlation. Other studies supply the additional evidence. Thus, the experiments begin with evidence of causation, and the surveys conﬁrm that this outcome is generalizable to everyday life. The reverse hypothesis (Belson, 1978), though superﬁcially promising, is par- ticularly disappointing. This is the proposition that the positive correlation is ex- plained by the seeking out of violent entertainment by those prone to aggressive- ness. Admittedly, there are some sets of data in which aggressiveness predicts subsequent violence viewing, although the literature is quite mixed as to the reg- ularity or pervasiveness of such on outcome (Comstock & Scharrer, 1999). In fact, the analysis that supplies some of the strongest evidence against the reverse hypothesis also provides some data consistent with it. The difﬁculty is that the reverse hypothesis fails woefully as a complete explanation for the correlation. 11. FROM TELEVISION VIOLENCE TO AGGRESSION 199 The most convincing evidence comes from the reanalysis by Kang (1990) of the NBC panel data on elementary school children (Milavsky, Kessler, Stipp, & Rubens, 1982). 3 This repeated measures surveying of the same sample over a 3.5-year period permitted the calculation of coefﬁcients representing earlier view- ing and later behavior or earlier behavior and later viewing, permitting inferences about which was affecting the other. Kang found eight statistically signiﬁcant co- efﬁcients for a viewing-to-behavior effect but only four for a behavior-to-viewing effect, out of a total of 15 pairings of earlier and later measurement. Moreover, there was only one instance of reciprocal association (viewing predicted behavior and behavior predicted viewing within the same span of time). These data are wholly inconsistent with the reverse hypothesis as a comprehensive explanation. The be- havior variable in the NBC data was interpersonal aggression—hitting, ﬁghting, stealing, name calling. Belson (1978) in his survey of about 1,600 London teenage males extends the dismissal of the reverse hypothesis to seriously harmful behav- ior. He statistically manipulated his data so that he could compare directly the plausibility of the direct and the reverse hypothesis. He concluded that there was no support for the reverse hypothesis and strong support for the direct hypothesis that violence viewing increased the committing of seriously harmful acts (such as attempted rape, false report of a bomb threat, and use of a tire iron, razor, knife, or gun in a ﬁght) among a subsample with a high propensity for delinquent behavior (p. 390, Table 12.13). The impression given by Bandura and colleagues (Bandura et al., 1963a, 1963b), and often repeated in textbooks, after early experiments with nursery school chil- dren is that males are more likely to be affected. Meta-analysis demurs. Girls as well as boys appear to be affected. In Paik and Comstock (1994), effect sizes are similar for boys and girls in the data from the method that represents everyday events—surveys. Only in experiments are effect sizes clearly greater for males. Whether this false impression is attributable to entertainment history (aggressive females on the screen are now more frequent), social change (childrearing is prob- ably now somewhat more accepting in regard to the expression of aggression by girls), norms (the Stanford nursery school surely was a campground of gender role convention), or the preponderance of males as subjects in the experiments (they outnumber females by about 6 to 1) is a matter of speculation. What is not is that meta-analysis has supplied a corrective. Meta-analysis also leads to the conclusion that effects do not diminish with age, although they are largest among the very young. In Paik and Comstock, effect sizes displayed an upward shift among those 3 The NBC panel study involved the collection of data from elementary and high school samples over a 3.5-year period in two cities, one in the Midwest and one in the Southwest. There were six points of data collection, which led to 15 wave pairs of earlier and later measurement (I-II, I-III, I-IV, etc.). The focus was on earlier television violence exposure and later aggressive behavior. The lag time between measurements varied from 3 months to more than 3 years. The elementary school N for the shortest period was 497 and for the longest period, 112, with attribution reducing the N as the time lags became longer. 200 COMSTOCK of college age, although otherwise there was a modest decline from preschool to adulthood with increasing age. Thus, it is not justiﬁed to conclude that the effects of television violence on aggressive and antisocial behavior are limited to the very young, or that they decline as cognitive ability increases to understand what is transpiring on the screen, to separate fantasy from reality, and to make distinctions between stirring acts and devious motives. Again, this is a pattern that becomes more readily recognizable in meta-analyses because single studies typically are quite restrictive in the range of ages included. The experiments that began to accumulate in 1963 with the work of Bandura, Berkowitz, and their colleagues (Bandura et al., 1963a, 1963b; Berkowitz & Rawlings, 1963) drew attention to the possibility that exposure to violent portray- als increased aggressive and antisocial behavior. However, they also were rather starkly limited to immediate effects. Surveys began to address the possibility of effects over time by attempting to represent earlier viewing in the measure of ex- posure (McLeod, Atkin, & Chaffee, 1972a, 1972b), and Eron and Lefkowitz and colleagues (Lefkowitz, Eron, Walder, & Huesmann, 1972, 1977) reported a statis- tically signiﬁcant correlation between violence viewing at about age 8 years and aggressive behavior a decade later after controlling for aggression in the earlier time period (and, thus, its correlates, including any otherwise unmeasured causes of the behavior) as part of the 1972 U. S. Surgeon General’s inquiry into televi- sion violence. This was the beginning of the production of data sets that seemingly show an effect over time. The most recent is from Johnson and colleagues (Johnson, Cohen, Smailes, Kasen, & Brook, 2002), who report statistically signiﬁcant cor- relations between television viewing at age 14 years and aggressive and antisocial behavior at ages 16 and 22 years with the reverse hypothesis encountering the usual inhospitality among the data with the pattern about the same for those both high and low initially in aggressive and antisocial behavior. However, the clear- est documentation of effects over time occurs in the NBC panel data, where it is possible to distinguish between longitudinal effects and cumulative effects. NBC’s original analysis by Milavsky et al. (1982), found that there were two instances in which effects increased with the passage of time. In one case, the coefﬁcients among the elementary school sample became larger as the number of months between measurements increased, with the largest clustered among the ﬁve coefﬁcients representing the longest time spans. In the other, the coefﬁcients among the same elementary school sample became somewhat larger when there were no statistical controls for the inﬂuence of earlier viewing, implying that there was an aggregative inﬂuence of violence viewing. Two reanalyses of the NBC panel data provide further evidence of effects over time. Cook, Kendzierski, and Thomas (1983) concluded that there was evidence of increasing coefﬁcients with the passage of time among both the elementary school sample and the teenage sample and for several different measures of aggressive and antisocial behavior. Kang (1990) found that ﬁve of his statistically signiﬁcant eight viewing-to-behavior coefﬁcients were clustered among the longest time spans, whereas three of his four 11. FROM TELEVISION VIOLENCE TO AGGRESSION 201 behavior-to-viewing coefﬁcients were clustered among the shortest time spans. These analyses suggest that there are longitudinal effects and cumulative effects. The former probably represents the inﬂuence of earlier viewing on behavior that has become newly relevant or newly within the range of the individual. The latter indicates that inﬂuence accumulates. Thus, the data supports the developmental interpretation (Eron & Huesmann, 1987) that earlier viewing establishes traits that will persist and perhaps become even more pronounced while conﬁning the reverse hypothesis largely to the short term. Milavsky and colleagues (Milavsky et al., 1982) seized on socioeconomic sta- tus as a possible explanation for positive coefﬁcients in their panel data. They argued that the substantial representation of young people from households of low socioeconomic status (SES) led to samples where television viewing—and thus exposure to violent portrayals—and aggressive behavior would be correlated. Alas, SES proves inadequate to producing the required artifact. Socioeconomic status consistently has been judged as unable to fully explain positive associations between violence viewing and aggressive and antisocial behavior (Belson, 1978; Chaffee, 1972; Comstock & Scharrer, 1999). Furthermore, in the case of the NBC data, Cook et al. (1983) teased out a truly embarrassing rejoinder—the pattern that held for the males of increasing coefﬁcients as time spans increased, alleged by Milavsky and colleagues as an artifact of socioeconomic status, was duplicated with middle-class girls. The recorded effects are not trivial in seriousness or size. The effect sizes in Paik and Comstock (1994) for interpersonal aggression, the category of behavior for which the evidence is the strongest because it has been examined more often than any other type of antisocial behavior, are in the medium range by Cohen’s (1988) frequently employed criteria (Rosenthal, Rosnow, & Rubin, 2000). In- terpersonal aggression includes hitting, ﬁghting, name calling, and stealing and ordinarily would constitute an experience that most victims would prefer to avoid. Other seriously harmful or criminal outcomes have much smaller effect sizes, but they are statistically signiﬁcant and represent the inﬂiction of greater harm than merely hitting, ﬁghting, name calling, or stealing. The Belson data (Belson, 1978) provide a particularly striking example of a nontrivial outcome. He found in his London sample of teenage males that the viewing of violent television en- tertainment predicted the committing of signiﬁcantly more seriously hurtful (and decidedly criminal) acts than were committed by those like them in every other measured respect (other than the greater viewing of violence), and there was no evidence that this could be attributed to the reverse hypothesis. Similarly, Johnson and colleagues (2002) found that greater television viewing (which would imply greater exposure to violent portrayals) at an earlier age predicted more frequent assaults or physical ﬁghts resulting in injury among 16- and 22-year-old males with the plausibility of the reverse hypothesis diminished (as pointed out earlier) by the occurrence of the same pattern for those scoring lower as well as higher in aggressiveness at the earlier time period. 202 COMSTOCK Taking advantage of the Federal Communication Commission’s (FCC) televi- sion station license freeze in the late 1940s and early 1950s to conduct a quasi- experimental time series with switching replications (Cook & Campbell, 1979), a group led by the methodologist Thomas Cook (Hennigan et al., 1982) found con- sistent evidence across two different samples (cities and states) and at two points in time (early and late introduction of television) that television’s introduction was followed by a signiﬁcant rise in larceny theft. This is an outcome that has enjoyed two distinct interpretations: that it should be attributed to relative deprivation ac- celerated by the materialistic emphases of the medium (Hennigan et al., 1982) and that it represents emulation of television’s antisocial emphases in a manner that apprehension would be unlikely and sanctions modest (Comstock, 1991). In the Paik and Comstock (1994) meta-analysis, coefﬁcients were mostly in the medium range (by Cohen’s criteria) for simulated aggressive behavior (such as aggressive inclination measured by questionnaire or performance on an aggression machine) and minor aggression (such as violence done to an object or physical aggression against a person that would fall under the law’s radar), although some- times they were in the large range. 4 They were smaller when illegal activities were the dependent variable and became progressively smaller as the seriousness of the offense increased. Even so, criminal violence against a person achieved statistical signiﬁcance and scored a fail-safe number just shy of 3,000. The issue of the possi- ble triviality of the effect sizes for television violence viewing and aggressive and antisocial behavior is addressed somewhat differently by Bushman and Anderson (2001). They compiled a selection to compare with the effect size for all observa- tions in the Paik and Comstock (1994) m

bundet

eta-analysis (p. 481, Figure 2). The r . = .31 for media violence and aggression compares favorably in terms of magnitude with those for such pairings as passive smoking and lung cancer, calcium intake and bone mass, and homework and academic achievement. In their own cumulative meta-analysis, Bushman and Anderson found stable, statistically signiﬁcant corre- lations for experimental designs from 1975 to 2000, and for the ecologically most valid data, the nonexperimental designs representing real-life aggression or anti- social behavior, Bushman and Anderson found statistically signiﬁcant correlations that have increased in magnitude over the past 25 years. In contrast, in a parallel content analysis of major news media, they uncovered a decline in the frequency 4 Paik and Comstock (1994) divided outcomes into three categories: simulated aggression, minor aggressive behavior, and illegal activities. In effect, they created a scale of increasing social consequence in terms of the validity of the dependent variable. The ﬁrst included use of an aggression machine and self-report of aggressive inclination, play with aggressive toys, and miscellaneous simulated aggressive behavior. The second included physical aggression against an object, verbal aggression, and physical aggression against a person below the threshold of serious harm. The third included burglary, grand theft, and seriously harmful violence against a person. Self-report of aggressive inclination was largely represented by responses indicating what the individual would do in hypothetical situations. Thus, it was analogous to the expression of a norm, value, or attitude—a disposition—rather than an intention to behave in the future in a particular way in a speciﬁc situation. 11. FROM TELEVISION VIOLENCE TO AGGRESSION 203 that reports linked television and ﬁlm violence with an undesirable behavioral outcome (they record a sizable negative correlation of r =− . 68 between average effect sizes and the average ratings of news reports for their fealty to the scien- tiﬁc evidence over six data points between 1975 and 2000). They conclude, with considerable exasperation at the performance of the news media, that “regardless of preference for experimental or non-experimental methods, it has been decades since one could reasonable claim that there is little reason for concern about media violence effects” (p. 485). Finally, based on Paik and Comstock (1994), the U. S. Surgeon General’s recently released report on youth violence (U. S. Department of Health and Human Services, 2001) identiﬁes greater exposure to television violence between the ages of 6 and 11 years as an early risk factor for the committing of criminal violence equivalent to a felony between the ages of 15 and 18 years (p. 58, Box 4–1). The effect size ( r . = . 13) was categorized as falling into the small range, and the report concluded that the effect sizes for both aggression in general and physical aggression—apparently using slightly more liberal criteria than Cohen—achieved the large range. However, about three fourths of the near 20 factors identiﬁed as posing early risks for later violence also fell into the small category. Young people are particularly vulnerable to the inﬂuence of television and ﬁlm violence when they have attributes that predict greater exposure to violent portrayals or greater likelihood of engaging in aggressive or antisocial behavior (Comstock & Scharrer, 1999; U. S. Department of Health and Human Services, 2001). Five such attributes are low socioeconomic status, rigid or indifferent par- enting, unsatisfactory social relationships, low psychological well-being, and a predisposition for antisocial behavior. Although there are no data that directly document such interactions, there is data clearly linking these factors to either the independent or dependent variables, and thus logically they would appear to in- crease vulnerability. Exposure to television and to television violence is inversely associated with socioeconomic status (Comstock & Scharrer, 1999; Thornton & Voigt, 1984). When parent-child communication is open and constructive (in con- trast with families where communication is rigid, many topics are avoided, differ- ences in opinion are discouraged, and orders rather than explanations by parents are the rule), there typically is less television viewing by children and less exposure to violent television entertainment (Chaffee, McLeod, & Atkin, 1971; McLeod et al., 1972b). In addition, a variety of delinquent behavior is inversely associated with parental interest in the whereabouts of children and teenagers (Thornton & Voigt, 1984). Psychological discomfort and social conﬂicts predict greater exposure to television—those under stress, lonely, anxious, in negative moods, or in conﬂict with others apparently ﬁnd the medium a satisfying escape and thus score higher in viewing or attraction to television (Anderson, Collins, Schmitt, & Jacobvitz, 1996; Canary & Spitzberg, 1993; Comstock & Scharrer, 1999; Kubey & Csik- szentmihalyi, 1990; Maccoby, 1954; Potts & Sanchez, 1994). It is clear that the predisposed—those scoring higher in antisocial behavior or possessing attributes 204 COMSTOCK that are correlates of antisocial behavior—are most likely to be affected. Sur- veys (Belson, 1978; Robinson & Bachman, 1972), experiments (Josephson, 1987; Celozzi, Kazelskis, & Gutsch, 1981), and a meta-analysis (Paik, 1991) all record associations that are greater among or limited to the predisposed. We doubt that it is a necessary condition because results are so consistent for general populations in surveys and experiments, but, if it is, then it is also very widespread. Thus, television violence is most likely to add to the burdens of those who already face considerable challenge in coping with everyday life, and this becomes particularly clear when it is acknowledged that the kind of behavior likely to be increased by violent portrayals is also the kind likely to lead to conﬂicts with others and clashes with the law. Two independent analyses concur that among the independent variables respon- sible for effects on aggression one of the most powerful is violent erotica. Both Paik and Comstock (1994), who examined erotic as well as violent portrayals, and Allen, D’Alessio, and Brezgel (1995) who examined only erotic portrayals, although proceeding somewhat differently (as is usually the case in these matters) and producing somewhat different effect sizes (as would be expected), report effect sizes for violent erotica that are among the highest recorded. The data also provide an important contribution to the debate over whether it is the sex (Weaver, 1991) or the violence (Donnerstein, Linz, & Penrod, 1987) that is responsible. Effect sizes for violent erotica are consistently higher than they are for erotica without violence. However, coefﬁcients for erotica without violence are consistently positive. When Allen and colleagues (Allen et al., 1995), following the schema developed by the Attorney General’s Commission on Pornography (1986), divided their independent variables into portrayals of nudity without sex, erotica (sex without violence), and violent erotica, they found for the ﬁrst an inverse effect size and for the latter two positive and increasing effect sizes. The answer, then, is that it is both the sex and the violence, with the two creating a powerful joint stimulus. The inverse effects for nudity and the weaker effects for erotica presumably derive from the absence of cues that would facilitate aggression, including de- piction of the participants in sexual behavior as meriting callousness, derision, or contempt. It is tempting to argue that the inverse and weaker effects represent the lower levels of arousal induced, but the ﬁnding of Allen and colleagues (Allen et al., 1995) that throughout self-reported arousal was inversely (if modestly) correlated with aggression precludes such a rash inference. The data bestow considerable conﬁdence that the patterns observed are not the products of methodologically inferior or ecologically questionable undertakings. In neither the early meta-analysis of Hearold (1986) nor the more recent one by Paik and Comstock (1994) does the introduction of measures of study quality and ecological validity alter conclusions with one noteworthy exception. Hearold found that when she conﬁned her analysis to the studies scoring high in methodologi- cal quality (which would give the data particular credibility) the outcomes were 11. FROM TELEVISION VIOLENCE TO AGGRESSION 205 symmetrical for antisocial and prosocial portrayals. Antisocial portrayals were as- sociated with a positive effect size for antisocial behavior and a negative effect size for prosocial behavior. Prosocial portrayals were associated with a positive effect size for prosocial behavior and a negative effect size for antisocial behavior. One clear implication is the imposition of a double jeopardy by violent children’s programming: greater likelihood of antisocial behavior and lesser likelihood of prosocial behavior by the young viewer. Hearold (1986) also examined another design element—the matching in ex- periments of treatment and outcome variables. These experiments involved pure modeling, with the outcomes undeterred by the requirement that the subjects gen- eralize from one situation to another. She found that this design element doubled effect sizes for both antisocial and prosocial behavior. There are two implications. One is that commercial entertainment has less effect than might otherwise be the case because it often fails to match the situations in which viewers ﬁnd themselves. The other is that experiments with these characteristics should not be taken as of- fering an effect size likely to be duplicated in everyday life, except when portrayals and behavioral options exactly match. DISAPPOINTMENT AND SURPRISE There is also an element that is both disappointing and surprising. Often, exposure to violent portrayals has been a predictor of an attitudinal disposition favorable to aggressive or antisocial behavior, and, often, these outcomes have occurred in experimental designs. As a consequence, exposure in these cases can be said to be the cause of the disposition. For example, in the Paik and Comstock (1994) meta-analysis, a substantial number of the outcomes categorized under the rubric “aggressive intent” ( r = . 33, medium magnitude by Cohen’s criteria) consisted of responses indicating what the individual would do in hypothetical situations. Such measures represent attitudes or dispositions rather than clear-cut intentions in regard to speciﬁc circumstances. Thus, meta-analysis (and the many individual studies encompassed) provides evidence of a causal link between violence viewing and dispositions. Nevertheless, there is little direct evidence that dispositions are a key and necessary link between the independent and dependent variables— exposure to violent portrayals and aggressive or antisocial behavior. There is a data set that permits an examination of this issue and possesses at- tributes that give it unusual credibility. Belson’s (1978) sample of London male teenagers was very large (about 1,600) and is the only probability sample in the literature (as a result, it has the inferential nicety of clearly representing a larger population). Belson’s measurement was meticulous and statistical analysis assidu- ous. The respondents were personally interviewed by a conservatively attired male in a clinical setting. Boys were given false names to emphasize the conﬁdentiality of the investigation, and the names were used throughout to encourage forthright 206 COMSTOCK responses. The interviews were unusually lengthy—about 3.5 hours. This is the equivalent of almost 1,000 days of interviewing, an apt reﬂection of the scale of this enterprise. When it came to aggressive and antisocial behavior, they were asked to indicate whether, within the past 6 months, they had committed an act printed on a card by placing the card in a location clearly designated as representing “Yes,” “No,” or “Not sure.” The intent was to reduce guile resulting from shame or em- barrassment by eliminating the need for the respondent to speak about what he had done. The interviewer then engaged in forceful probes in regard to a “Not sure” and the frequency with which a confessed act had been committed. The labeling of these acts as to their seriousness was based on the rating of adult judges. Thus, the aggressive and antisocial outcome measures were sensitive, probing, and rooted in normative social judgments and presumably minimized both guile and conformity to interviewer expectancies. Thus, the data have a strong claim to representing real-world patterns. Among the variables included in the Belson (1978) survey were a number of items representing attitudinal dispositions—norms, values, and beliefs. He in- cluded them as possible outcomes associated with exposure to violent television entertainment. This makes it possible to embark on the path of theory testing and possible revision. I begin with Belson’s ﬁndings and then turn to a brief exami- nation of the evidence supporting the popular view that there is an attitudinal or dispositional link between viewing and behavior. Finally, I attempt to resolve any conﬂicts between the two bodies of evidence—in this case, by positing an alterna- tive path connecting violence viewing to aggressive and antisocial behavior. Belson obtained direct measures of four dispositions—antisocial attitudes, approval of vi- olence, hostile personality traits, and willingness to commit violence—as well as a dimension of behavior with a dispositional aspect, social contagion (antisocial behavior in the company of others). Reliability was high for the ﬁrst four, which were multiitem scales with 10 or more items for each disposition, whereas the latter represented a variety of antisocial acts. The ﬁrst four directly represent atti- tudes, norms, and values. The ﬁfth has a dispositional aspect because it represents indirectly the attitudes, norms, and values implied by the expressed beliefs and behavior of peers. None of these outcomes was predicted by the viewing of vio- lent television entertainment. Thus, there was no evidence of a dispositional link between viewing and behavior. Much of the research that has been central in the investigation of the relationship between exposure to television and ﬁlm violence and aggressive and antisocial be- havior would seem to assign a central place to attitudes, norms, and values. Both Bandura’s (1986) social cognition and Berkowitz’s (1984, 1990) neoassociation- ism give considerable weight to the categorizing of acts according to their effec- tiveness, social approval, and appropriateness for the circumstances. Both theories hold that portrayals with these characteristics are more likely to inﬂuence behavior. Such categorizing would seemingly depend on cognitive processes whose terminal state would be a disposition. The conclusion of Eron and Huesmann (1987), in 11. FROM TELEVISION VIOLENCE TO AGGRESSION 207 interpreting a positive correlation in survey data between childhood violence view- ing and adult aggressiveness, is representative of many others writing on television violence and antisocial behavior: It is not claimed that the speciﬁc programs these adults viewed when they were 8 years old still had a direct effect on their behavior. However, the continued viewing of these programs probably contributed to the development of certain attitudes and norms of behavior and taught these youngsters ways of solving interpersonal problems which remained with them over the years. (p. 196) Interpretation thus faces a quandary. The Belson (1978) survey provides strong evidence that attitudes, norms, and values are not invariably a link in real life in the causal chain between exposure and behavior. Numerous experiments also provide convincing evidence that aggression and antisocial behavior are increased when certain factors are present that would seem to operate through cognitive processes—portrayals that depict behavior as effective, socially approved, and appropriate for the viewer, as well as circumstances that place the viewer in the market for some behavioral guidance. These are the contingencies that Comstock and Scharrer (1999) refer to as efﬁcacy, normativeness, pertinence, and suscep- tibility. They represent, respectively, the degree to which a portrayed behavior is perceived as effective, as evidenced by reward or, for those intrinsically plea- surable, lack of punishment; as socially accepted, approved, or conventional; as relevant to the viewer, such as a perpetrator of the same age or gender or a vic- tim resembling a potential real-life target; and the degree to which the viewer is motivated or rendered vulnerable to being affected by the portrayal. 5 The Belson (1978) data clearly falsify the hypothesis that certain attitudes and norms (to use the phrasing of Eron and Huesmann, 1987) are a necessary link in the causal chain. The challenge is equally clear. If we accept the Belson ﬁnding, we must explain the role that these four factors have other than constituting a link through dispositions. The alternatives—draw a line through the Belson data, impugn the experiments—are not scientiﬁcally permissible. The interpretation that conforms to both sets of evidence is that these factors promote the incorporation and accessibility of the portrayed behavior. Behavioral 5 Susceptibility is a factor that has been found often to play a role in the inﬂuence of communication. In the case of aggressive or antisocial behavior, it is typically represented by frustration or provocation and in experiments often induced by the rude or insulting behavior of the experimenter. However, it also has been found to operate in other contexts. For example, in the Stanford cardiovascular ﬁeld experiment (Farquar et al., 1977, 1990) those who scored higher in risk proneness were more inﬂuenced by persuasive appeals intended to reduce behavior that contributed to cardiovascular disorder, and in the area of agenda setting (Comstock & Scharrer, 1999) those scoring higher in the need for orientation are more likely to be inﬂuenced by the emphases of media coverage in regard to the issues and topics they perceive as important. In each of these three instances, the inﬂuence of messages is somewhat contingent on individual motives, needs, and interests. 208 COMSTOCK classes or genres would be inﬂuenced because we would expect some general- ization, although the effect would be greatest when there was a match between the portrayal and behavioral options. The function of these factors is to govern the salience of classes or genres of behavior in the repertoire of the individual. In terms of the theory of a link between perception and behavior, offered by Dijksterhuis and Bargh (2001), these factors determine the signiﬁcance or validity as a guide to behavior extended to the behavior portrayed. Efﬁcacy, normativeness, pertinence, and susceptibility would act as gatekeepers, to use communication research jargon. From this perspective, observation and attitudes, norms, and values often act as competitors rather than the coconspirators they are usually taken for. Attitudes, norms, and values may remain stable while the likelihood of a change in behavior increases. Good people (at least, those with constructive thoughts) may behave badly without any sign of a change in expressed dispositions. This lowers the predictive power of attitudes, norms, and values, which have a notoriously weak resistance to the demands of situational circumstances (Eagly & Chaiken, 1993; Terry & Hogg, 2000) and heightens the predictive power of observation. Efﬁcacy, normativeness, pertinence, and susceptibility become conditions on which the latter, rather than the former, rest. This revision of theory addresses two nagging puzzles. One is the quite frequent occasion when attitudes, norms, and values, and behavior part ways. The other is the similarly frequent occasion when the enormity of an act seems beyond a contribution by the media. We expect people to be consistent in thought and behavior, and we expect heinous acts to be outside the inﬂuence of observation or perception. The revision asserts that there are routes to behavior other than through dispositions (although they sometimes may be crucial) and that role of the media may be modestly conﬁned to gatekeeping but on that account sometimes crucial. The data support the view that one route by which entertainment affects be- havior is analogous to peripheral processing in the ELM. Accessibility or salience is a key element. However, it would be premature to conclude that this is the sole route by which behavioral effects occur. Attitudes, norms, and values surely sometimes play a role. One would occur when these cognitive dispositions and accessibility coincide, which would increase the likelihood of an effect; this would be a special case of susceptibility in operation. Another is when thoughtful moti- vation enters, such as Bandura’s (1986) famous example of a well-contrived (and ﬁnancially successful) airline bomb extortion attempt following the televising of a movie depicting such a caper. This analysis obviously ﬁts quite well with the general aggression model (GAM) proposed by Anderson and Bushman (2002), where cognitive processes may have an inﬂuence but are not a necessary element. The present analysis, however, addresses the speciﬁc question of whether atti- tudes, norms, and values—that is, cognitive elements—are a necessary part of any causal chain where exposure to violent portrayals predicts aggressive or antisocial behavior (as in the Belson data), as much research suggests, and the conclusion is that they are not. 11. FROM TELEVISION VIOLENCE TO AGGRESSION 209 The Belson (1978) data do not support a role for direct effects of attitudes, norms, and values. However, they are consistent with a role for cognitive dispositions through effects on accessibility and salience. In addition, the Belson data have nothing conclusive to say about the possibility that attitudes, norms, and values held by others may create an environment more or less favorable to modes of behavior made accessible or salient by the media. Thus, attitudes, norms, and values remain key concepts in explaining the effects of violent portrayals on aggressive and antisocial behavior although they are not a necessary mediating link between exposure and behavior. REFERENCES Allen, M., D’Alessio, D., & Brezgel, K. (1995). A meta-analysis summarizing the effects of pornog- raphy II: Aggression after exposure. Human Communication Research, 22 (2), 258–283. Anderson, C. A., & Bushman, B. J. (2002). Human aggression. Annual Review of Psychology, 53 , 27–51. Anderson, D. R., Collins, P. A., Schmitt, K. L., & Jacobvitz, R. S. (1996). Stressful life events and television viewing. Communication Research, 23 (3), 243–260. Andison, F. S. (1977). TV violence and viewer aggression: A cumulation of study results. Public Opinion Quarterly, 41 (3), 314–331. Attorney General’s Commission on Pornography. (1986). Final report . Washington, DC: U. S. Government Printing Ofﬁce. Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory . Englewood Cliffs, NJ: Prentice Hall. Bandura, A., Ross, D., & Ross, S. A. (1963a). Imitation of ﬁlm-mediated aggressive models. Journal of Abnormal and Social Psychology, 66 (1), 3–11. Bandura, A., Ross, D., & Ross, S. A. (1963b). Vicarious reinforcement and imitative learning. Journal of Abnormal and Social Psychology, 67 (6), 601–607. Becker, M. H. (Ed.). (1974). The health belief model and personal health behavior. Health Education Monographs, 2 (4), 324–473. Belson, W. A. (1978). Television violence and the adolescent boy . Westmead, UK: Saxon House, Teakﬁeld. Berkowitz, L. (1984). Some effects of thoughts on anti- and prosocial inﬂuences of media events: A cognitive-neoassociationistic analysis. Psychological Bulletin, 95 (3), 410–427. Berkowitz, L. (1990). On the formation and regulation of anger and aggression: A cognitive- neoassociationistic analysis. American Psychologist, 45 (4), 494–503. Berkowitz, L., & Rawlings, E. (1963). Effects of ﬁlm violence on inhibitions against subsequent aggression. Journal of Abnormal and Social Psychology, 66 (3), 405–412. Bushman, B. J., & Anderson, C. A. (2001). Media violence and the American public: Scientiﬁc facts versus media misinformation. American Psychologist, 56 (6–7), 477–489. Cacioppo, J. T., & Petty, R. E. (1985). Central and peripheral routes to persuasion: The role of message repetition. In L. F. Alwitt & A. A. Mitchell (Eds.), Psychological processes and advertising ef- fects:Theory, research and application (pp. 91–111). Hillsdale, NJ: Lawrence Erlbaum Associates. Cairns, E., Hunter, D., & Herring, L. (1980). Young children’s awareness of violence in Northern Ireland: The inﬂuence of Northern Irish television in Scotland and Northern Ireland. British Journal of Social and Clinical Psychology, 19 , 3–6. Canary, D. J., & Spitzberg, B. H. (1993). Loneliness and media gratiﬁcation. Communication Research, 20 (6), 800–821. 210 COMSTOCK Celozzi, M. J., II, Kazelskis, R., & Gutsch, K. U. (1981). The relationship between viewing televised violence in ice hockey and subsequent levels of personal aggression. Journal of Sport Behavior, 4 (4), 157–162. Chaffee, S. H. (1972). Television and adolescent aggressiveness (overview). In G. A. Comstock & E. A. Rubinstein (Eds.), Television and social behavior: Television and adolescent aggressiveness (Vol. 3, pp. 1–34). Washington, DC: U. S. Government Printing Ofﬁce. Chaffee, S. H., McLeod, J. M., & Atkin, C. K. (1971). Parental inﬂuences on adolescent media use. American Behavioral Scientist, 14 , 323–340. Cohen, E. E. (1988). Children’s television commercialization survey . Washington, DC: National As- sociation of Broadcasters. Comstock, G. (1991). Television and the American child . San Diego, CA: Academic Press. Comstock, G., & Scharrer, E. (1999). Television: What’s on, who’s watching, and what it means . San Diego, CA: Academic Press. Comstock, G., & Scharrer, E. (2003). The contribution of meta-analysis to the controversy over tele- vision violence and aggression. In Gentile, D. A. (Ed.). Media Violence and Children . Westport, CT: Praeger. Cook, T. D., & Campbell, D. T. (1979). Quasi-experimentation: Design and analysis issues for ﬁeld settings . Chicago: Houghton Mifﬂin. Cook, T. D., Kendzierski, D. A., & Thomas, S. A. (1983). The implicit assumptions of television research: An analysis of the 1982 NIMH report on television and behavior. Public Opinion Quar- terly, 472 , 161–201. Dijksterhuis, A., & Bargh, J. A. (2001). The perception-behavior expressway: Automatic effects of social perception on social behavior. In M. P. Sanna (Ed.), Advances in Experimental Social Psy- chology (Vol. 33, pp. 1–40). San Diego, CA: Academic Press. Donnerstein, E., Linz, D., & Penrod, S. (1987). The question of pornography: Research ﬁndings and policy implications . New York: Free Press. Eagly, A. H., & Chaiken, S. (1993). The psychology of attitudes . Orlando, FL: Harcourt. Eron, L. D., & Huesmann, L. R. (1987). Television as a source of maltreatment of children. School Psychology Review, 16 (2), 195–202. Farquar, J. W., Fortmann, S. P., Flora, J. A., Taylor, C. B., Haskell, W. L., Williams, P. T., et al. (1990). Effects of communitywide education on cardiovascular disease risk factors: The Stanford Five-City Project. Journal of the American Medical Association, 264 (3), 359–365. Farquar, J. W., Maccoby, N., Wood, P. D., Alexander, J. K., Breitrose, H., Brown, B. W., Jr., et al. (1977) Community education for cardiovascular health. Lancet, 1 , 1192–1195. Glass, G. V. (1976). Primary, secondary, and meta-analysis of research. Educational Researcher, 5 , 3–8. Hearold, S. (1986). A synthesis of 1043 effects of television on social behavior. In G. Comstock (Ed), Public communication and behavior (Vol. 1, pp. 65–133). New York: Academic Press. Hennigan, K. M., Heath, L., Wharton, J. D., Del Rosario, M. L., Cook, T. D., & Calder, B. J. (1982). Impact of the introduction of television on crime in the United States: Empirical ﬁndings and theoretical implications. Journal of Personality and Social Psychology, 42 (3), 461–477. Hogben, M. (1998). Factors moderating the effect of television aggression on viewer behavior. Communication Research, 25 , 220–247. Hunt, M. (1997). How science takes stock . New York: Russell Sage. Johnson, J. G., Cohen, P., Smailes, E. M., Kasen, S., & Brook, J. S. (2002). Television viewing and aggressive behavior during adolescence and adulthood. Science, 295 , 2468–2471. Josephson, W. L. (1987). Television violence and children’s aggression: Testing the priming, social script, and disinhibition predictions. Journal of Personality and Social Psychology, 53 (5), 882–890. Kang, N. (1990). A critique and secondary analysis of the NBC study on television and aggression . Unpublished doctoral dissertation, Syracuse University, Syracuse, NY. 11. FROM TELEVISION VIOLENCE TO AGGRESSION 211 Kubey, R., & Csikszentmihalyi, M. (1990). Television and the quality of life: How viewing shapes everyday experience . Hillsdale, NJ: Lawrence Erlbaum Associates. Lefkowitz, M. M., Eron, L. D., Walder, L. O., & Huesmann, L. R. (1972). Television violence and child aggression: A followup study. In G. A. Comstock & E. A. Rubinstein (Eds.), Television and social behavior: Vol. 3. Television and adolescent aggressiveness (pp. 35–135). Washington, DC: U. S. Government Printing Ofﬁce. Lefkowitz, M. M., Eron, L. D., Walder, L. O., & Huesmann, L. R. (1977). Growing up to be violent: A longitudinal study of the development of aggression . Elmsford, NY: Pergamon. Maccoby, E. E. (1954). Why do children watch television? Public Opinion Quarterly, 18 (3), 239–244. McLeod, J. M., Atkin, C. K., & Chaffee, S. H. (1972a). Adolescents, parents, and television use: Adolescent self-report measures from Maryland and Wisconsin samples. In G. A. Comstock & E. A. Rubinstein (Eds.), Television and social behavior: Television and adolescent aggressiveness (Vol. 3, pp. 173–238). Washington, DC: U. S. Government Printing Ofﬁce. McLeod, J. M., Atkin, C. K., & Chaffee, S. H. (1972b). Adolescents, parents, and television use: Self-report and other-report measures from the Wisconsin sample. In G. A. Comstock & E. A. Rubinstein (Eds.), Television and social behavior: Television and adolescent aggressiveness (Vol. 3, pp. 239–313). Washington, DC: U. S. Government Printing Ofﬁce. Milavsky, J. R., Kessler, R., Stipp, H. H., & Rubens, W. S. (1982). Television and aggression: A panel study . New York: Academic Press. Paik, H. (1991). The effects of television violence on aggressive behavior: A meta-analysis . Unpublished doctoral dissertation, Syracuse University, Syracuse, NY. Paik, H., & Comstock, G. (1994). The effects of television violence on antisocial behavior: A meta- analysis. Communication Research, 21 (4), 516–546. Petty, R. E., & Cacioppo, J. T. (1990). Involvement and persuasion: Tradition vs. integration. Psycho- logical Bulletin, 107 (3), 367–374. Potts, R., & Sanchez, D. (1994). Television viewing and depression: No news is good news. Journal of Broadcasting and Electronic Media, 38 (1), 79–90. Robinson, J. P. & Bachman, J. G. (1972). Television viewing habits and aggression. In G. A. Comstock & E. A. Rubinstein (Eds.), Television and social behavior: Television and adolescent aggressiveness . (Vol. 3, pp. 372–382). Washington, DC: U. S. Government Printing Ofﬁce. Rosenthal, R. (1979). The “ﬁle drawer problem” and tolerance for null results. Psychological Bulletin, 86 , 638–641. Rosenthal, R., Rosnow, R. L., & Rubin, D. B. (2000). Contrasts and effect sizes in behavioral research: A correlational approach . New York: Cambridge University Press. Rosenthal, R., & Rubin, D. B. (1978). Interpersonal expectancy effects: The ﬁrst 345 studies. Behavioral and Brain Sciences, 3 , 377–415. Terry, D. J., & Hogg, M. A. (Eds.). (2000). Attitudes, behavior, and social context . Mahwah, NJ: Lawrence Erlbaum Associates. Thornton, W., & Voigt, L. (1984). Television and delinquency. Youth and Society, 15 (4), 445–468. U. S. Department of Health and Human Services. (2001). Youth violence: A report of the Surgeon General . Rockville, MD: U. S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Injury Prevention and Control, Substance Abuse and Mental Health Services Administration, Center for Mental Health Services, National Institutes of Health, National Institute of Mental Health. Weaver, J. (1991). Responding to erotica: Perceptual processes and dispositional implications. In J. Bryant & D. Zillmann (Eds.), Responding to the screen: Reception and reaction processes (pp. 329–354). Hillsdale, NJ: Lawrence Erlbaum Associates. Wood, W., Wong, F., & Cachere, J. (1991). Effects of media violence on viewers’ aggression in unconstrained social interaction. Psychological Bulletin, 109 (3), 371–383. CHAPTER TWELVE Between the Ads: Effects of Nonadvertising TV Messages on Consumption Behavi or Maria Kniazeva University of San Diego Research on the impact of nonadvertising television programs (e.g., series, soap operas, TV movies) on consumption behavior cannot boast the recognition gained by inquiries into the effects of television advertising. The latter research obviously has been built on the assumption that such effects exist and explores variables contributing to the effectiveness of this inﬂuence. The research on the inﬂuence of nonadvertising television exposure on consumer attitudes and behavior appears to be fragmented and overlooked in the marketing discipline. The major question, “Are the entertaining programs a powerful engine for change in daily consumption activity?” still seems to need further inquiry. If we turn to experts in the search for answers, we can ﬁnd strong advocates of the real effects of nonadvertising television programs on consumption behavior. For example, activists of the international organization Soap Operas for Social Change credit soap operas with generating demand for goods like sewing machines and condoms (Williams, 2001). Though the reported effects were witnessed in devel- oping countries, they support the TV producers’ general belief that nonadvertising television programs can encourage positive behavioral changes. The logical ques- tion is, to what extent have such claims and hopes been supported theoretically? This is the question that this chapter aims to address. This chapter focuses speciﬁcally on the effects of nonadvertising television programs on consumption behavior and discusses these effects in terms of the psychological processes of encoding, interpretation, and memory. The objective is to analyze how researchers in different ﬁelds have conceptually deﬁned and 213 214 KNIAZEVA empirically identiﬁed the effects of television viewing on consumers’ attitudes to- ward goods and services and consumer behavior in the marketplace. This research is a response to invitations to investigate the link between marketing and nonad- vertising media that has repeatedly sounded in the marketing discipline in recent years. Speciﬁcally, Lehmann (1999) urged researchers to broaden their current narrow focus on mass advertising by considering the cumulative effect of market- ing and media on consumer welfare. John (1999) welcomed efforts to understand the inﬂuence of subtle messages delivered by television. Hirschman and McGriff (1995), who represented an initial step in employing marketing skills toward the application of televised images to addressing social problems, expressed hope that studies like this will lessen destructive consumer behavior. Varey (1999) argued that studying the effects, and not just the effectiveness, of media should help in dealing with the consequences of the wholesale marketization of society. In light of these academic calls, it is timely to review the existent body of re- search on the impact of nonadvertising media on consumption behavior. To ﬁll this need, this chapter offers an analysis of reported ﬁndings on how and why nonadver- tising television has been found to inﬂuence consumption behavior. The presented literature review of the academic research conducted in the past 20 years examines the most popular and the most controversial media channel—television. Commer- cial advertising is excluded from the scope of this work due to the substantial attention that television advertising has already received in marketing literature. In this way, attention is diverted from the omnipresent and often annoying and intrusive commercial messages that prompt audiences to resist and neglect them to those equally ubiquitous but subtle and hidden messages to which customers deliberately expose themselves during the programs. Speciﬁcally, this chapter investigates how television messages are encoded by producers and interpreted by receivers and presents the ﬁndings under several head- ings. After providing a brief historical overview of the phenomenon of television, I illuminate theoretical explanations for the effects of TV on consumption behav- ior. Second, I summarize a set of empirical evidence of these effects, including both immediate and long-term effects. Third, I analyze research of cultural texts. Finally, I discuss public policy implications, identify some unanswered questions, and suggest directions for future research. In the overall discourse, two major streams of inquiry become visible: one exploring nonadvertising television’s ef- fects as emotional enhancers of commercials and the other examining these effects as cognitive enhancers of commercials. Both types of enhancers are seen as me- diating the impact of television viewing on consumption behavior. “OLD” MEDIA UNDER NEW ATTACKS The mass arrival of television in 1946 is often termed revolutionary, and after the 1949 debut of the TV set in Sears Roebuck catalog, it took only 14 years for the American public to recognize that they received more of their news from television 12. NONADVERTISING AND CONSUMPTION BEHAVIOR 215 than from newspapers. Moreover, surveys indicate that since the 1980s, more than two thirds of Americans have regarded TV not only as their primary but also their preferred source of news, and more than half believe it is the most credible source (Dennis, 1989). On the other hand, there were just 3 years between the ﬁrst sounds of gunshots from televised westerns in the 1955–1956 season and the ﬁrst formal complaint about TV violence by a U.S. senator. Published in Reader’s Digest under the title “Let’s Get Rid of Tele-Violence” (Kefauver, 1958), this article presumed a direct link between the rising rates of juvenile crime and television viewing and virtually opened up the still ongoing debates about the controversial impact of television on its viewers and society as a whole. (For reviews, see Comstock [this volume] and Shrum [this volume].) Today, television is attacked by public health specialists for the potential health risks that media exposure presents to children and adolescents, by sociologists for being a time waster and social isolator (Tonn & Petrich, 1998), and by social scientists and public policy ofﬁcials for the prevalence of violent content. Many of the sins attributed to television’s inﬂuence have either direct or indirect links with consumption behavior. Materialism, compulsive buying, smoking, and the antisocial consumption of drugs and alcohol are the “sins” that are referred to most often. The overall situation seems to be so critical that the American Academy of Pediatrics (AAP) suggested that pediatricians eschew televisions in their waiting rooms, provide guidance about media use in the home to parents and children during pediatric visits, and take a media history from patients by asking them about their TV viewing habits (Hogan, 2000). These steps were included in a 5-year, nationwide education campaign, Media Matters, launched by the AAP in 1997 with the purpose to protect youth from the potential harm of media messages and images. Thus, regardless of the dominant character of television, societal attitudes regard this medium as both a threat and an opportunity, a conﬂict that is rooted in the recognition of TV as a powerful message sender. STUDYING TV: UNREALISTIC ATTEMPT? Despite (and perhaps because of ) the pervasiveness of television, studying its ef- fects has never been considered an easy task. Yet at the dawn of television, scholars predicted not only the overwhelming popularity of an audiovisual communication channel but also the difﬁculty of grasping the effects of TV on people’s lives. The word totality best explains the phenomenon, which resulted from the intervention of TV in personal, political, and social lives. Some scholars even expressed concern over the prospects of possible studies: “Since it [television] has affected the totality of our lives, … it would be quite unrealistic to attempt systematic or visual presen- tation of such inﬂuence” (McLuhan, 1964, p. 317). Instead, what seemed a more feasible way of studying these effects was suggested—“to present TV as a complex gestalt of data gathered almost at random” (McLuhan 1964, p. 317). Forty years later, however, it is evident that research on the impact of television viewing on 216 KNIAZEVA people’s lives, including their consumption behavior, has been conducted in two major directions: One stream of research uses quantitative methods and focuses on responses that are outwardly observable; the second approach is characterized by qualitative research. Both directions have the same goal: explaining the effects of TV on various aspects of life. The studies include careful analysis of what is of- fered on the screen and of how this translates into viewers’ daily activities beyond the screen. Research on the impact of nonadvertising televised messages on behavior of viewers has been conducted in several disciplines and is presented by scholars of communications (e.g., Gerbner, Gross, Morgan, Signorielli, & Shanahan, 2002), psychology (e.g., Bandura, 1994), sociology (e.g., Fox & Philliber, 1978), child development (e.g., Potts, Doppler, & Hernandez, 1994), and preventive medicine (e.g., Cooper, Roter, & Langlieb, 2000). The authors agree in deﬁning the place of television in the individual decision-making and judgmental process. The shared theoretical framework assumes two important factors employed by the viewers: direct and indirect experiences. Direct experience refers to past activities in which an individual was personally engaged, whereas indirect experience includes activ- ities experienced by other people but of which the consumer is aware through two possible sources—a social network, mass media, or both. Thus, indirect experi- ence can be gained through friends, relatives, and neighbors and through television, print media, and radio. But to what extent do nonadvertising messages conveyed by television really matter in an individual’s consumption behavior? In the search for the answer to this general question, a substantial body of knowledge has been collectively generated. Scholars have analyzed statistical descriptions of the audience, engaged in con- textual analysis of the television messages and images, offered theoretical ex- planations and empirical evidence of the mechanism of television’s inﬂuence, and claimed to ﬁnd immediate and long-term effects of television watching on the audience. Their ﬁndings suggest that after the shift in the public’s prefer- ence from a print to a visual culture occurred, the artiﬁcial reality perpetually portrayed on the TV screen started serving as a subliminal frame of reference for the viewers in their consumption activity while affecting both “good” con- sumption (e.g., making the right lifestyle choices, developing healthy eating patterns) and “bad” consumption (e.g., smoking, consuming alcohol, doing drugs, practicing overconsumption). These ﬁndings are analyzed in the rest of the chapter. THEORETICAL EXPLANATION FOR THE EFFECTS OF TV ON CONSUMPTION BEHAVIOR A determined attempt to offer a theoretical explanation for the effects of nonad- vertising TV on consumption behavior is what unites the majority of the reported studies. Current research pursuing this goal can be grouped into the examination 12. NONADVERTISING AND CONSUMPTION BEHAVIOR 217 of consumption associative and consumption nonassociative television viewing and their effects. Thus, research that explores the explicit consumption context of the television messages or the feelings produced by watching the portrayals of certain products, goods, and services directly associated with their use by the characters of a TV program can be termed under the rubric consumption associa- tive viewing. Examples include the works by Hirschman and Thompson (1997), O’Guinn and Shrum (1997), and Shrum, Wyer, and O’Guinn (1998). In turn, re- search that is consumption context independent and does not focus on the depicted consumer products and their perception by TV viewers can be termed as research into consumption nonassociative viewing. This stream of research explores over- all emotions elicited by the program and examines TV programs as emotional enhancers of the advertisements. Examples include the works by Goldberg and Gorn (1987), Murry, Lastovicka, and Singh (1992), and Murry and Dacin (1996). Consumption Nonassociative Viewing The main objective of the consumption nonassociative research is to deﬁne which factors facilitate the greater effectiveness of commercials embedded in the pro- gram. The scholars engaging in this research mostly have tested what is considered to be conventional advertising wisdom that asserts the following scheme: If the program elicits positive emotions, the viewer will like the program; liking the pro- gram, in turn, leads to a positive attitude toward the advertisement; and this attitude is followed by a positive attitude toward the brand advertised. In accordance with this theory, there is empirical evidence that commercials are more effective when placed in happy rather than in sad programs. Thus, building on previous mood research that suggested that the effects of the elicited mood may be evidenced for at least 15–20 min, Goldberg and Gorn (1987) exposed their participants to advertisements embedded in a happy episode about frogs trained to improve their self-image and in a sad episode about the killing of a young child. The authors then asked their participants (among other questions) about their likelihood of pur- chasing the products advertised in the embedded commercials and demonstrated the difference in responses. Those seeing the happy episode felt happier while watching the commercials than did those who saw the sad episode. The happy episode viewers also demonstrated a stronger intention to purchase the advertised products than did viewers of the sad episode. However, it has been also argued that these program-elicited feelings do not directly inﬂuence attitudes toward the advertisement and brands but rather that viewer attitudes toward the program are what matter (Murry et al., 1992). Thus, dramas that elicit sadness may result in a positive attitude toward the advertisement similar to that produced by the positive-emotion programs, provided that the viewer likes the drama. Going further in this line of research, Murry and Dacin (1996) proposed and proved the hypothesis that “negative emotions elicited by a program will diminish program liking only if viewers believe the program to be either too realistic or personally relevant” (p. 441). This leads to the conclusion that negative 218 KNIAZEVA moods elicited by TV viewing are not a threat to the positive attitude toward advertisements and advertised brands when the individual does not feel personally threatened. Consumption Associative Viewing Whereas the research into consumption nonassociative viewing mostly explores nonadvertising television programs as emotional enhancers, the study of consump- tion associative viewing mostly regards TV programs as cognitive enhancers. This stream of research, as well as the previous one, also refers to a relation between ad- vertising and media, which it characterizes as symbiotic. Some of the most sound recent theory-building work was offered by Hirschman and Thompson (1997), who argued that advertising, though certainly a powerful inﬂuence on consumption, is not a dominant voice and is being supplanted by mass media nonadvertising mes- sages and images that largely deﬁne consumers’ beliefs and behavior. The strength of these subtle messages lies in their informal, unobtrusive nature because such mass media texts are not viewed with the same cultivated skepticism as actual advertisements. By examining the mediating role of nonadvertising forms of media on advertis- ing, Hirschman and Thompson (1997) expanded prior multidisciplinary research that ascertained that media enhance the effectiveness of some advertising by por- traying certain products as more appealing than others. In that role, media sort reality into meaningful social categories that provide a frame of reference from which consumers interpret their daily lives (Miller, 1988; Schiller, 1989). This the- oretical framework was further developed when Hirschman and Thompson deﬁned three interpretive strategies employed by consumers of televised messages in their relationships with the mass media. First is the inspiring and aspiring mode that is present when a media image is interpreted as representing an ideal self to which the consumer can aspire. The second mode is rather oppositional to the ﬁrst: Labeled by the authors as deconstructing and rejecting, it refers to consumer relationships with the mass media that are characterized by overt criticism of the artiﬁcial and unrealistic quality of the media representation. The third mode is tailoring and personalizing; termed identifying and individualizing, it occurs when consumers negotiate their personal goals and self-perceptions in relation to idealized images presented in the media. Proposing that consumers’ relationships with nonadvertising forms of mass media are an essential aspect of the perceived meaning they derive from advertisements, the authors implied that the process of decoding messages por- trayed on the screen is not a completely random and unpredictable affair. Moreover, they emphasized the interactive nature of television, which requires viewers who consume the TV images to actively construct the meanings associated with de- picted goods and services. In line with this conclusion, a number of studies have been designed to test whether television consumption can be argued to alter the 12. NONADVERTISING AND CONSUMPTION BEHAVIOR 219 perception of reality and subsequently affect consumer behavior (e.g., O’Guinn & Shrum, 1997; Shrum et al., 1998). Television and the Perception of Reality Several authors have claimed a causal relationship between viewing television programs and the formation of beliefs about social reality. According to their ﬁndings, heavy viewers of television do differ from light viewers in their perception of the real world, which in their mental construction closely reﬂects that portrayed on the screen. Thus, a positive relation between TV viewing and materialism was found (see Shrum, Burroughs, & Rindﬂeisch, this volume). Speciﬁcally, heavy viewers were found to overestimate the ownership of expensive products to a greater extent than did lighter viewers (O’Guinn & Shrum, 1997). Heavy viewers also overestimate the rate of crime (Gerbner, Gross, Morgan, & Signorielli, 1980a) and drug and alcohol consumption in society (Shrum & O’Guinn, 1993). They negatively view the elderly as unhealthy and ﬁnancially poor and perceive women as aging faster than men (Gerbner, Gross, Morgan, & Signorielli, 1980b). These works have employed Gerbner’s cultivation theory (Gerbner et al., 2002), which begins with an analysis of the aggregated messages embedded in television as a system and compares them with answers received in response to surveys designed to measure the correlation between televised images and viewers’ perceptions of the real world. Cultivation theory suggests that heavy exposure to television results in largely viewing the world the way it is depicted on the screen. Thus, the overrepresented possession of certain goods on television (luxury cars, cell phones, nice houses) leads to a perception of a similar level of possession in the real world and subsequently to a distorted picture of reality. In several attempts at exploring the mechanism by which concepts of the social reality of consumption are constructed, scholars examined the effects of televi- sion on perceptions of afﬂuence (Fox & Philliber, 1978; O’Guinn & Shrum, 1997; Shrum, 2001; Shrum, O’Guinn, Semenik, & Faber, 1991) and reported a signiﬁcant relationship between the two variables. Their ﬁndings supported the general hy- pothesis that heavier viewers of television perceive greater afﬂuence in America than do lighter viewers. This included perceptions related to the ownership of swimming pools, convertible automobiles, use of maids, and the percentage of millionaires. Cultivation theory, though widely accepted and tested by scholars, elicits some criticism for not providing a thorough explanation of its mechanism and not considering other variables (e.g., people’s socioeconomic situation or education) that also could lead to the social judgments and beliefs expressed by participants of studies (Fox & Philliber, 1978; McGuire, 1986). In response to this criticism, some research implicates the availability heuristic as a psychological mechanism responsible for the effects of television viewing on judgmental processes (Shrum, 2001). The availability heuristic refers to a mode of information processing that 220 KNIAZEVA links judgments of frequency or probability of occurrence with ease of recall. The availability heuristic was developed by decision scientists Tversky and Kahneman (1973), who suggest that people estimate the prevalence of patterns and events based on how easy it is for them to retrieve the related information from memory. According to the availability heuristic, heavy viewers of TV should have more in- formation received from television accessible in memory because they are exposed to a higher number of televised messages than are light viewers. These messages are stored in the memory over time, and when it comes to making judgments and decisions, heavy viewers of TV retrieve the accumulated information (Shrum et al., 1998). For this reason, heavy viewers should have relevant images (those frequently sent through the media channel) more accessible in memory. Thus, the application of the availability heuristic is seen as responsible for viewers’ estimates of the real frequency and probability of events portrayed on the screen, which may explain the psychological mechanism of the cultivation effect. In addition to the role television plays in consumer socialization, there has been some empirical evidence that it can play a role in consumer acculturation as well. The idea that the televised world is perceived as a close model of the real one and is turned to as a reference point for understanding society’s values and beliefs was evident in Reece and Palmgreen’s (2000) inquiry into the motives for watching television by foreigners. In a study of Asian Indian graduate students enrolled at American universities, the authors distinguished the motives of acculturation and reﬂection on values from among eight motives. Having conducted a survey of the convenience sample of 99 students, scholars argued that those subjects who wanted to learn more about current issues in their host country, understand the ways in which American people behave and think, and make American friends spent more time in front of the TV. The respondents also reported substantial exposure to host media back home. This led to assumptions that they were aware of what to expect on the screen and were prepared to decode the images encoded in news and sports coverage and situation comedies and movies—programs deﬁned as preferred in the search for valuable cultural information. Furthermore, the participants of the study were found to deliberately turn to indirect televised experience in the search for answers about real-life experiences in the host country. Lifestyle-Related Product Choices The conceptual research into the effects of TV also suggests that consumers’ incor- poration of television-mediated images of reality into their perceptions of how oth- ers live and consume is used during the making of lifestyle-related product choices. Thus, Englis and Solomon (1997) noted that images of afﬂuence transmitted via mass media become objects of desire for many who aspire to this quasi-mythical lifestyle. Conversely, some symbolical images associated with lifestyles that con- sumers tend to avoid are being stored as a negative mental set of objects. Moreover, the authors strongly assert that these pervasive consumption-rich images replete in 12. NONADVERTISING AND CONSUMPTION BEHAVIOR 221 mass media are no mere shadows playing across the screen of popular culture but become equally or even more important than actual behavior observed when con- sumers subconsciously engage in the construction of lifestyle meanings. Englis and Solomon have argued that these intangible lifestyle meanings are powerful social constructions developed in the minds of consumers and should not be regarded by marketers as secondary. After conducting several studies on consumers’ cognitive processing of media and marketing information and their consumption behavior, the scholars suggested a model that assigned the central role of generating and modifying lifestyle imageries to channel intermediaries—those including com- mercial vehicles (e.g., advertising), popular culture (e.g., television), and media hybrids. Overall, Englis and Solomon stressed that lifestyles should not be under- stood solely in terms of the actual patterns of behavior that marketers typically measure and use to cluster consumers. The Promising Future of Hybrids The use of messages embedded in television programs and labeled hybrids seems to have raised more ethical concerns than conceptual attention. These messages are paid for by the sponsor but are not explicitly identiﬁed as being such. A rela- tively new genre of marketing communication is growing (Balasubramanian, 1994) and is represented by product placements, infomercials, and program tie-ins. It is believed that hybrid messages avoid key disadvantages inherent in commercial advertising (i.e., a sponsor is identiﬁed and consumers can view the message with skepticism) and beneﬁt from key advantages provided by commercial advertising, such as the sponsor’s total control over the message content and its format. As a result, consumers perceive hybrid messages without any of the resistance they may practice while dealing with explicitly purchased messages. Conceptually, product placement hybrids appear to play a role in inﬂuencing consumer behavior similar to that of any other embedded product messages. Thus, Balasubramanian iden- tiﬁed several theories on which the expected effects of impacting consumers are built: attribution theory (Mills & Jellison, 1967), the classical conditioning princi- ple (Gorn, 1982), and the modeling paradigm (Bandura, 1994). Attribution theory assumes that television viewers perceive portrayed products without any reporting bias that could be attributed to a paid message. The classical conditioning principle assumes a positive paired-association enhancing product perception when a favor- able endorser image (unconditional stimulus) and a product (conditional stimulus) wisely match each other. The modeling paradigm assumes that models demonstrat- ing positive consequences of product used reinforce the process of persuasion. Public policy ofﬁcials invite researchers to employ the powerful persuasive potential of hybrid messages for conveying noncommercial messages designed to promote socially desirable consumption behavior. They call for embedding into television programs not so much products as social concepts and argue that the positive effects of such practices have been empirically supported. 222 KNIAZEVA GRASPING TELEVISION EFFECTS The empirical study of television effects is deﬁned by an approach that prompts or analyzes responses that are outwardly o

bundet

bservable. These responses may be grouped into those characterized by immediate effects and those deﬁned by long- term impacts. Whereas the studies of immediate effects are usually conducted in the laboratory setting where the researchers experimentally induce the expected effects, the research on long-term effects usually aims at identifying the correlation between television viewing and observable changes in attitude and consumption behavior. Both approaches seem to have had strong advocates. Thus, the proponents of long-term effects ﬁnd strong support in the cultivation theory proposed by Gerbner et al. (2002). They argue that despite the time, energy, and money invested, the conceptualization of television’s effect as a short-term individual change has not produced research that helps understand the distinctive features of television. To them, these features include massive, long-term, and common exposure of large and heterogeneous publics to centrally produced, mass distributed, and repetitive systems of stories (Gerbner et al., 2002). Another theoretical model of the mass media’s effects on individual decision- making can be found in the works of Tyler (Tyler, 1984; Tyler & Cook 1984), who has extensively tested the impersonal impact hypothesis proposed in prior research. According to this hypothesis, decisions are based on the judgments people make. These judgments, in turn, should be divided between those on a personal and those on a societal level. Moreover, the researchers should clearly separate them because the media messages may produce contradictory and independent societal and personal judgments. Thus, the individual, inﬂuenced by presumably overrepresented violence on the TV screen, may perceive a similar rate of crime in the surrounding society. However, when it comes to personal judgment, the viewer may not think there is an increasing danger and a high likelihood of being personally affected by crime. For this reason, the decision making and following behavior may not always be ruled by television exposure, even if the viewer is a heavy viewer. The impersonal impact hypothesis suggests that the impersonal nature of mass media is less effective in inﬂuencing personal concerns than in inﬂuencing societal concerns. Furthermore, compared with the second type of indirect experience—a social network—mass media is also the least effective in impacting personal judg- ments. Consequently, mass media is argued to primarily inﬂuence judgments only on a general level. However, recently this theory has been questioned because of the implications that it poses for the processes involved in the construction of the different judgments. Shrum and Bischak (2001) point to several deﬁni- tional and analytical ambiguities about the relation between type of risk judgment (personal/societal) and experience modality (direct/indirect). They argue that the deﬁnitions of personal risk and mass media representations are not clear and leave room for different interpretations. Shrum and Bischak also suggest that the types of 12. NONADVERTISING AND CONSUMPTION BEHAVIOR 223 analysis used to test the impersonal impact hypothesis ignore the assessment of in- teraction between direct experience and media information. Thus, further research on the impersonal impact hypothesis is invited. Long-Term Effects Evidence of long-term effects of nonadvertising television has been proclaimed by several authors. DeJong and Winsten (1990) credited 80 network television episodes broadcast during two seasons for more responsible public alcohol con- sumption and the tendency to perceive the concept of a designated driver as a social norm. Evans et al. (1981) reported that social modeling ﬁlms can help deter smoking in adolescents. An example of the probable positive effect of a media campaign on the population’s beliefs and healthy behavior in respect to cardiovas- cular diseases (CVD) was offered in Finnegan, Viswanath, and Hertog’s (1999) interpretive analysis proposing a causal relationship between the two. The authors referred to an overall decline in smoking in the country and a parallel increase in public opinion that smoking causes heart disease and cancer. They interpreted the positive outcome and evidence of the secular trend in CVD-related beliefs, knowledge, and behavior as due to the strong inﬂuence of national media that had offered numerous heart disease stories, including about 5–10 per month on three TV networks alone. Their analysis suggested the potential strength of media inﬂuence when the media are guided by organized activity geared not only toward building an individual’s awareness but also toward creating an encouraging en- vironment. This activity, known in the marketing discipline as social marketing and based on agenda-building models of communication, provides for a uniﬁed approach to inﬂuencing social, behavioral, and policy change, where television as a powerful media channel takes a leading role. Immediate Effects There also exists a strong feeling among communication theorists that televised inﬂuence is best deﬁned in terms of the content people watch rather than the amount of television viewing (Hawkins & Pingree, 1982). The studies in support of content’s predominance are usually conducted in controlled laboratory settings. A series of them have been designed to empirically test the immediate effects of television inﬂuence on behavior. An example is a study that investigated the impact of risk taking by characters in television programs on children’s willingness to take physical risk (Potts et al., 1994). This problem has not only medical aspects but also a number of social ones because risk-prone behavior, especially when prompted in immature children and vulnerable teenagers, may later have direct links to such antisocial consumption behavior as drug and alcohol use. Potts, Doppler, and Hernandez assigned a group of 50 children ages 6 to 9 years, to three experimental conditions—television stimulus programs with infrequent physical risk taking, programs with frequent risk taking, and no TV stimuli. Their results 224 KNIAZEVA indicated that children exposed to high-risk TV programs self-reported higher levels of willingness to take risk than did those exposed to low- or no-risk TV programs. The authors interpreted their results as “evidence of a small effect that may accumulate into a larger effect across the many hours of TV viewed routinely by most children” (p. 328). Moreover, they argued that even relatively innocuous and humorous animated cartoons may result in a previously unidentiﬁed impact on children via the observational learning process because risk-taking television characters are rarely punished and even glamorized. It is clear that for ethical reasons laboratory studies cannot be designed to elicit strong effects proving the negative inﬂuence of television watching. The body of research exploring negative impact is more inclined to apply survey and observational and interpretive methods. Consequently, the ﬁndings are usually suggestive. The typical example is a study by Distefan, Gilpin, Sergeant, and Pierce (1999) that explored whether movie stars who smoke on and off screen may encourage adolescents to start smoking. After running a multivariate statistical analysis of data gathered in the 1996 California Tobacco Survey of more than 6,000 teenagers, the authors claimed that the suggestive inﬂuence of movie star smoking on their fans’ tendency to start smoking was evident and only slightly weaker than the inﬂuence of smoking by friends and family members. Though the ﬁndings did not establish that smoking portrayed in ﬁlms directly leads adolescents to smoke, the study supported the idea of a reversed causal order in that adolescents accepting smoking tend to favor actors and actresses associated with smoking. The results prompted Distefan et al. to send the alarming message that stars have the potential to be much more powerful role models for youth smoking behavior than parents and teachers, a claim that certainly has public policy implications and supports the use of televised appeals by stars to prompt desirable consumption behavior. In this light, Hirschman and McGriff’s (1995) study provided a novel empirical attempt to examine possible therapeutic uses of motion pictures in drug reha- bilitation programs. The study, designed to apply marketing theory and method toward assisting the treatment of addiction as a social problem, was conducted in a real-life setting. Recovering alcoholics and drug addicts were exposed to ﬁlms that they viewed on a television monitor and were then asked to ﬁll out a ques- tionnaire while evaluating the accuracy of the portrayal of addiction on the screen and the ability of the ﬁlms to stimulate their recovery. Having analyzed the data, the authors concluded that two factors deﬁned certain ﬁlms’ stronger constructive effect: The stories were not overly graphic or violent, and the ﬁlms modeled not only addiction but also a path to recovery. In contrast, the more extreme portrayals of addiction tended to stimulate neutral or negative response in some viewers who did not identify themselves with the portrayed images and thus led the viewers to reject an intended positive message. With this experiment, for which four ﬁlms with addiction stories were carefully chosen, Hirschman and McGriff enhanced the trend of research into cultural texts, including those presented on television screens, which leads directly to the next stream of research. 12. NONADVERTISING AND CONSUMPTION BEHAVIOR 225 RESEARCH ON CULTURAL TEXTS The methodological practice of examining the content of television programs has been termed telethnography and applied to explorations of the televised presen- tation of products and their surrounding consumption environment (Sherry 1995; Wells & Anderson, 1996). It is argued that the method provides rich data for un- derstanding the mental images attached by customers to products and services. In addition, these meanings are vital to understanding the exchange value of the products in the marketplace. Consequently, thorough analysis of product classes (e.g., automobiles, credit cards, breakfast cereals) allows for developing effective promotional campaigns (Hirschman & Thompson, 1997). In one of the ﬁrst attempts to explore the domain of televised narratives under the marketing angle, Hirschman (1988) explored the ideology of consumption encoded within two, (at the time) immensely popular television series, Dallas and Dynasty. She investigated the messages the series transmitted about consumption, how prominent characters behaved as consumers, and what their consumption behavior signiﬁed. Through the hybrid structural-syntactical method, Hirschman explored the symbolic meanings of products and the ways in which they were linked to actions and outcomes in a televised series. Her ﬁndings proposed that popular television programs are strong vehicles of consumption ideology, which also serve as vehicles of consumers’ projective self-reﬂection. It is in a proposed model of the dynamic relationships between consumption practices and cultural texts that Hirschman, Scott, and Wells (1998) presented a method of interpreting the symbolism of consumption practice and its symbolic depiction on the screen. They used data on coffee drawn from advertising and television programming and incorporated a broad historical and sociological per- spective into their discursive model. Their main theoretical conclusion was not limited to highlighting the power of television programs that constantly inform consumers’ beliefs and affect their behavior. To Hirschman, Scott, and Wells, it is important to remember that these effects are not isolated because they cooccur in media time adjacent to television commercials and thus are worth being included in the set of variables inﬂuencing consumers’ perception of products and services. The model also encouraged cross-cultural research application and examination of how speciﬁc product categories are portrayed in other cultures. Exploration in this area has been implemented in some marketing studies. Thus, a content analysis of consumption imagery in music television shown in the United States (MTV) and Sweden (MTV-Europe) suggested that the American sample of videos contains more consumption imagery and presents a higher frequency of brand mentions than those videos to which Swedish viewers are exposed (Englis, Solomon, & Olofsson, 1993). Thus, only 14.9 percent of the Swedish sampled videos contained brand mentions versus 38.9 percent of the videos sampled in the United States, Moreover, the ﬁndings revealed several differences in consump- tion imagery as a function of musical genre. Thus, rap videos consisted of more 226 KNIAZEVA dark-side consumption images (alcohol, weapons, and drugs), heavy metal favored band-related products, and dance music conveyed more fashion messages. Con- sidering music television to be an important agent of adolescent socialization, the authors highlighted the importance of studying consumption-relevant content of music to better predict and understand the effects of music on consumers, partic- ularly young consumers. DISCUSSION AND PUBLIC POLICY IMPLICATIONS Though it is often claimed that most of the inquiry into media’s effects has so far been behavioral and focused on responses that are outwardly observable (Walsh, 2000), the careful review of research on the effects of nonadvertising television on consumption behavior has demonstrated that this topic also has been marked by the incorporation of analysis of cultural texts and strong attempts at theory building. All these areas of research provide strong evidence that television functions “to survey the environment for us, to socialize us into society and culture, and to entertain and sell us things” (Finnegan et al., 1999, p. 50). But, because of its ubiquity, TV makes it really difﬁcult to distinguish those effects caused solely by television viewing, and the functions of television can be indeed compared to the “invisible operation of bacteria” (McLuhan, 1964, p. 320). Methodologically, the main challenge researchers face today is that there is presumably no authentic control group—people not affected by the phenomenon of television. For this reason, studies are often criticized for employing a nonrep- resentative sample. In addition, there is no clear agreement among scholars on the deﬁnitions of light and heavy viewers. Most research refers to calculations of the number of hours that the respondents self-report TV watching. This way the boundary between heavy and light viewing remains arbitrary and vague, and an individual who may only watch sports programs can easily be referred to as a heavy viewer. Another challenging point is the choice of methods. Whereas some scholars favor purely empirical studies, believing that laboratory-run experiments can test the power of TV inﬂuence on its audience and thus arm the public with scientiﬁcally proven evidence, others question the limitedness of such studies. Thus, Brown and Cantor (2000) have argued that “any experimental manipulation may be just a drop in the bucket compared with the massive exposure to the same kind of content” (p. 3) in everyday lives and that often the focus of research interest lies in long-term attitude or behavioral change rather than short-term effects. Seen in this light, it seems especially suitable that marketing researchers study- ing television effects have experimented with methods themselves and have in- cluded in their operational set such methods as telethnograpy (Sherry, 1995), the hybrid structural-syntactical method (Hirschman, 1988), historical discourse (Hirschman et al., 1998), and the application of grounded theory (Hirschman & 12. NONADVERTISING AND CONSUMPTION BEHAVIOR 227 Thompson, 1997). This range of methods has helped explore not only immediate but also long-term effects of exposure to television on consumption behavior. The recognition by scholars of the importance of studying nonadvertising tele- vision is only the ﬁrst step. The next step should be the exploration of differential effects of TV messages and images on viewers. It is evident now that early models of mass media effects assumed a homogeneous audience and that invitations to conduct differential effects research and explore selective exposure more system- atically have become more pronounced lately (Brown & Cantor, 2000; Roe, 2000; Walsh, 2000). Scholars have called for an examination of the mediating effects of such variables as education, gender, socio economic status, family life, ethnicity, and cognitive development and for exploration of which attributes make messages more powerful and inﬂuential for different groups. They propose that television’s impact is often disputed not because it is not always visible but because effects often accumulate over time and are obscured by the individual differences of the viewers. Previous research indicates that gender is one of the most fundamentally dif- ferentiating factors in media use in terms of both quantity and content preferences (Roe, 2000). Thus, boys are more likely than girls to have a television in their bed- rooms, they tend to spend more time in front of the TV, and they prefer watching action, crime, sports, science ﬁction, and war programming, whereas girls rate mu- sic, talk shows, and soaps higher. Only comedy seems to attract equal attention. It is also known that boys watch more animated cartoons (Huston, Wright, Kerkman, & Peters, 1990), that African American and Hispanic youngsters are more likely than White peers to have TV sets in their bedrooms, and that TV viewing is the most substantial during early childhood and declines in middle adolescence (Roberts, 2000). These data provide a valuable starting point for further inquiry. Many researchers have already empirically highlighted the importance of ex- ploring mediating variables. Granting media’s inﬂuence, including that of televi- sion, for positive change in the public’s attitude toward smoking and health disease prevention, several scholars from different countries have provided evidence that not all sociodemographic groups have beneﬁted equally from media information (Brannstrom & Lindblad, 1994; Dennis, 1989; Finnegan et al., 1999). According to these scholars’ ﬁndings, people with fewer years of formal education or of lower socioeconomic status demonstrated only modest improvements in risk knowledge, which suggests that the phenomenon of a knowledge gap in information does not seem to be reduced by media. Though age differences in information processing have long been reported, it is especially important to recall the studies with a consumption context. Thus, it has been claimed that young children demonstrate a better ability to process audiovisual televised information than solely audio transmissions (Peracchio, 1992), and the elderly better process consumer information learned from television than from print media (Cole & Houston, 1987). Hirschman and Thompson (1997), in their study of motivational relationships between viewers and televised messages, reported to have discovered age- and 228 KNIAZEVA gender-based differences in the interpretations of television’s appeal. According to their ﬁndings, older people seem to express a more inspirational relationship with media images and more often refer to some of them as an ideal self to which they can aspire. The scholars also discerned a major difference between men’s and women’s interpretations of media messages with regard to the perceived need to resist them on the basis of being unreal and artiﬁcial. Whereas many of their female participants described themselves as resisting the distorting inﬂuences of televised messages, the male participants reported their ability to stand outside these inﬂuences. Another potential area for future research was pointed out by Englis and Solomon (1997), who believe that the role of cultural gatekeepers, including tele- vision, is overlooked with regards to symbolic encoding. According to these re- searchers, the process of how media select and design consumption imagery that is passed on to consumers through media channels needs more examination. Sim- ilarly, the process of how consumers decode product messages also deserves more exploration. Shrum et al. (1998) found that the order of data collection might affect the results and that it matters what is measured ﬁrst—television viewing or the es- timates. Consequently, more research is needed to test how the order of data collection may affect the results. In terms of data collection, it could also be es- pecially appropriate, given the nature of media research, to respond to Lehmann’s (1999) challenge to marketers to explore nontraditional ways of gathering data and not to ignore facts reporting—data widely accessible to the public and offered in news reports. Such examples of potential data collection could be presented by a print media story about a new trend evident among singers. This particu- lar story goes that singers introduce lucrative clothing lines while singing lyrics that mention which brands are hot and recording music videos that show how to wear these clothes (Herman-Cohen, 2001). Another story describes barbers in Afghanistan who were jailed for providing local males with popular haircuts styled after the American actor Leonardo DiCaprio (Wallwork, 2001). These real- life television effects were not prompted in the laboratory setting, but scholars nevertheless could try to incorporate them in their research conducted in con- trolled settings. Since the advent of television in 1946, researchers have actively speculated on the role of television in inﬂuencing people’s behavior, attitudes, and knowledge. The phenomenon of television is still full of contradiction. It is depicted as a threat and as an opportunity, it can promote and deter behavior, it can encourage positive and negative attitudes, it can lead to socially desirable and avoidable effects, it can advocate consumption or abstention from consumption, and television viewing itself can be intimate and social, global and local, innocent and evil, passive and active, cheap and expensive. Because of this conﬂicting nature, television poses numerous public policy implications; after half a century of research scholars recognize they still know little about the extent or consequences of television’s 12. NONADVERTISING AND CONSUMPTION BEHAVIOR 229 inﬂuence (Brown & Cantor, 2000), and public health specialists are unclear about how best to advise parents and children about media use in the home (Hogan, 2000). Public policy ofﬁcials ask scholars to study the effects of TV portrayals of tobacco, alcohol, and illicit drug use, urge parents to be more involved in children’s media exposure, and encourage viewers to resist the potentially harmful impact of television. In a changing media environment that embraces new channels of communica- tion, the “old” media of television has not yet been relegated to secondary status. In light of the increasing tendency to simultaneously use multiple media, televi- sion still remains a powerful shaper and mirror of consumers’ minds, lives, and experiences. The disturbing prediction is that because of information overload we may witness the decline of credible fact (Braman, 1993), and for this reason it will become “harder to distinguish between what is on-screen and in reality” (Walsh, 2000, p. 72), what is experience and what is ﬁction. Thus, the characters in televi- sion programs and the products they will use will even more deﬁne the consumers’ perception of their surrounding environment. This review of the body of knowledge accumulated by research on the effects of television on consumption behavior suggests that both marketers and public policy ofﬁcials need a working knowledge of televised messages to be able to communicate their goods and ideas. Because television viewing is argued to have a strong impact on the associations consumers have with lifestyles, both aspired and avoided, the media gatekeepers’ choices revealed on the screen should be constantly monitored for information about current and future consumption trends. This task becomes particularly necessary in light of assertions that producers of TV programs “read” from present consumer behavior—a conclusion that leads to a two-tiered perspective on the encoding and decoding of lifestyle information from media sources (Englis & Solomon, 1997). In addition, this behavior also demands the investigation of whether the strength of nonadvertising television effects depends on viewers’ involvement with the program (Murry & Dacin, 1996). Many research questions that go unanswered have direct links to public policy issues. Thus, what implications appear at the societal level, and is it good or bad for the public when TV viewers perceive the real world as a more afﬂuent place than it is in reality (O’Guin & Shrum, 1997)? Does the heavy TV exposure of children and teenagers make them socially savvy or more materialistic and greedy? Does it teach them the value of money and how to spend it (Brown & Cantor, 2000)? Does TV promote compulsive and violent behavior by portraying risk-taking television characters (Potts et al., 1994)? Though television violence has probably received more scrutiny than any other area (see Comstock, this volume), hot debates about the possible contribution of television violence to antisocial behavior haven’t so far been explicitly linked to consumer behavior. Assuming there are causal relationships between the violent images abundant on TV and the aggressive behavior of their recipients, it is worth exploring how consumption behavior changes in a crime-ridden society, town, or neighborhood. There is a 230 KNIAZEVA need to examine both immediate effects (demand on guns, violent toys, or crime- preventive goods and services, including locks, insurance) and long-lasting effects (lifestyles, values, etc.). The ﬁnal and perhaps most socially important question is whether the reported ﬁndings can be used to reduce the antisocial consequences of televised messages and enhance their positive potential. Can academic studies, as some researchers optimistically hope (Hirschman, 1991), contribute to solving destructive social problems based on compulsive and addictive consumer behavior? ACKNOWLEDGMENTS The author thanks John Graham, Alladi Venkatesh, Judy Rosener, and Connie Pechmann for the strong support that she found during the work on the earlier versions of this chapter. She also greatly appreciates helpful guidance she received from L. J. Shrum during the work on this chapter. REFERENCES Balasubramanian, S. K. (1994). Beyond advertising and publicity: Hybrid messages and public policy issues. Journal of Advertising, 23 (4), 29–46. Bandura, A. (1994). Social cognitive theory of mass communication. In J. Bryant & D. Zillmann (Eds.), Media effects: Advances in theory and research (pp. 61–90). Hillsdale, NJ: Lawrence Erlbaum Associates. Braman, S. (1993). Harmonization of systems: The third stage of the information society. Journal of Communication, 43 (3), 133–40. Brannstrom I., & Lindblad I. (1994). Mass communication and health promotion: The power of the media and public opinion. Health Communication, 6 (1), 21–36. Brown, J. D., & Cantor J. (2000). An agenda for research on youth and the media. Journal of Adolescent Health, 27 (2), 2–7. Cole, C. A., & Houston, M. J. (1987). Encoding and media effects on consumer learning deﬁciencies in the elderly. Journal of Marketing Research, 24 , 55–63. Cooper, C. P., Roter, D. L., & Langlieb, A. M. (2000). Using entertainment television to build a context for prevention news stories. Preventive Medicine, 31 (3), 225–231. DeJong, W., & Winsten, J. A. (1990). The use of mass media in substance abuse prevention. Health Affairs, 9 (2), 30–46. Dennis, E. (1989). Reshaping the media: Mass communication in an information age. Newbury Park, CA: Sage. Distefan, J. M., Gilpin, E. A., Sargent, J. D., & Pierce, J. P. (1999). Do movie stars encourage adolescents to start smoking? Preventive Medicine, 28 , 1–11. Englis, B. G., & Solomon, M. R. (1997). Where perception meets reality: The social construction of lifestyles. In L. R. Kahle & L. Chiagouris (Ed.), Values, lifestyles and psychographics (pp. 25–44). Mahwah, NJ: Lawrence Erlbaum Associates. Englis, B. G., Solomon, M. R., & Olofsson, A. (1993). Consumption imagery in music television: A bi-cultural perspective. Journal of Advertising, 22 (4), 21–33. Evans, R. I., Rozelle, R. M., Maxwell, S. E., Raines, B. E., Dill, C. A., & Guthrie, T. J. (1981). Social modeling ﬁlms to deter smoking in adolescents: Results of a three-year ﬁeld investigation. Journal of Applied Psychology, 66 (4), 399–414. 12. NONADVERTISING AND CONSUMPTION BEHAVIOR 231 Finnegan, J. R., Jr., Viswanath, K., & Hertog, J. (1999). Mass media, secular trends, and the future of cardiovascular disease health promotion: An interpretive analysis. Preventive Medicine, 29 (6), 50–58. Fox, S., & Philliber, W. (1978). Television viewing and the perception of afﬂuence. Sociological Quarterly, 19 , 103–112. Gerbner, G., Gross, L., Morgan, M., & Signorielli, N. (1980a). Aging with television: Images on television drama and conceptions of social reality. Journal of Communication, 30 , 37–47. Gerbner, G., Gross, L., Morgan, M., & Signorielli, N. (1980b). The “mainstreaming of America”: Violence proﬁle no. 11. Journal of Communication, 30 , 10–29. Gerbner, G., Gross, L., Morgan, M., Signorielli, N., & Shanahan, J. (2002). Growing up with television: The cultivation perspective. In J. Bryant & D. Zillmann (Eds.), Media effects: Advances in theory and research (pp. 43–67). Mahwah, NJ: Lawrence Erlbaum Associates. Goldberg, M. E., & Gorn, G. J. (1987). Happy and sad TV programs: How they affect reactions to commercials. Journal of Consumer Research, 14 , 387–403. Gorn, G. J. (1982). The effects of music in advertising on choice behavior: A classical conditioning approach. Journal of Marketing, 46 , 94–101. Hawkins, R. P., & Pingree, S. (1982). Television’s inﬂuence on constructions of social reality. In D. Pearl (Ed.), Television and behavior: Ten years of scientiﬁc progress and implications for the eighties (pp. 224–247). Washington, DC: U. S. Government Printing Ofﬁce. Herman-Cohen, V. (2001, March 19). Whose music are you wearing? Los Angeles Times , pp. E1, E3. Hirschman, E. C. (1988). The ideology of consumption: A structural syntactical analysis of Dallas . Journal of Consumer Research, 15 , 344–359. Hirschman, E. C. (1991). Secular morality and the dark side of consumer behavior: Or how semiotics saved my life. In R. H. Holman & M. R. Solomon (Eds.), Advances in Consumer Research (pp. 1–4). Provo, UT: Association for Consumer Research. Hirschman, E. C., & McGriff, J. A. (1995). Recovering addicts’ responses to the cinematic portrayal of drug and alcohol addiction. Journal of Public Policy and Marketing, 14 (1), 95–107. Hirschman, E. C., Scott, L., & Wells, W. D. (1998). A model of product discourse: Linking consumer practice to cultural texts. Journal of Advertising, 27 , 33–50. Hirschman, E. C., & Thompson, C. J. (1997). Why media matter: Toward a richer understanding of consumers’ relationships with advertising and mass media. Journal of Advertising, 26 , 43–60. Hogan, M. (2000). Media matters for youth health. Journal of Adolescent Health, 27 (2), 73–76. Huston, A. C., Wright, M. C., Kerkman, D., & Peters, S. M. (1990). Development of television viewing patterns in early childhood: A longitudinal investigation. Developmental Psychology, 26 , 409–420. John, D. R. (1999). Consumer socialization of children: A retrospective look at twenty-ﬁve years of research. Journal of Consumer Research, 26 , 183–213. Kefauver, E. (1958). Let’s get rid of tele-violence. Reader’s Digest , 23–25. Lehmann, D. R. (1999). Consumer behavior and Y2K. Journal of Marketing, 63 , 14–18. McGuire, W. J. (1986). The myth of massive media impact: Savagings and salvagings. In G. Comstock (Ed.), Public communication and behavior (pp. 173–257). New York: Academic Press. McLuhan, M. (1964). Understanding media: The extensions of man. New York: McGraw Hill. Miller, M. C. (1988). Boxed in: The culture of TV . Evanston, IL: Northwestern University Press. Mills J., & Jellison, J. M. (1967). Effects on opinion change of how desirable the communication is to the audience the communicator addressed. Journal of Personality and Social Psychology, 6 , 98–101. Murry, J. P., Jr., & Dacin, P. A. (1996). Cognitive moderators of negative-emotion effects: Implications for understanding media context. Journal of Consumer Research, 22 , 439–447. Murry, J. P., Jr., Lastovicka, J. L., & Singh, S. N. (1992). Feeling and liking responses to television programs: An examination of two explanations for media-context effects. Journal of Consumer Research, 18 , 441–451. O’Guinn, T. C., & Shrum, L. J. (1997). The role of television in the construction of consumer reality. Journal of Consumer Research, 23 , 278–294. 232 KNIAZEVA Peracchio, L. A. (1992). How do young children learn to be consumers? A script processing approach. Journal of Consumer Research, 18 , 425–440. Potts, R., Doppler, M., & Hernandez, M. (1994). Effects of television content on physical risk taking in children. Journal of Experimental Child Psychology, 58 , 321–331. Reece, D., & Palmgreen, P. (2000). Coming to America: Need for acculturation and media use mo- tives among Indian sojourners in the US. International Journal of Intercultural Relations, 24 (6), 807–824. Roberts, D. F. (2000). Media and youth: Access, exposure, and privatization. Journal of Adolescent Health, 27 (2), 8–14. Roe, K. (2000). Adolescents’ media use: A European view. Journal of Adolescent Health, 27 (2), 15–21. Schiller, H. I. (1989). Culture Inc. New York: Oxford University Press. Sherry, J., Jr. (1995). Bottomless cup, plug-in drug: A telethnography of coffee. Visual Antropology, 7 , 351–370. Shrum, L. J. (2001). Processing strategy moderates the cultivation effect. Human Communication Research, 27 (1), 94–120. Shrum, L. J., & Bischak, V. D. (2001). Mainstreaming, resonance, and impersonal impact. Testing moderators of the cultivation effect for estimates of crime risk. Human Communication Research, 27 (2), 187–215. Shrum, L. J., & O’Guinn, T. C. (1993). Process and effects in the construction of social reality: Construct accessibility as an explanatory variable. Communication Research, 20 , 436–471. Shrum, L. J., O’Guinn, T. C., Semenik, R. J., & Faber, R. J. (1991). Process and effects in the construction of normative consumer beliefs: The role of television. In R. H. Holman & M. R. Solomon (Eds.), Advances in consumer research (pp. 755–763). Provo, UT: Association for Consumer Research. Shrum, L. J., Wyer, R. S., Jr., & O’Guinn, T. C. (1998). The effects of television consumption on social perceptions: The use of priming procedures to investigate psychological processes. Journal of Consumer Research, 24 , 447–458. Tonn, B. E., & Petrich, C. (1998). Everyday life’s constraints on citizenship in the United States. Futures, 30 (8), 783–813. Tversky, A., & Kahneman, D. (1973). Availability: A heuristic for judging frequency and probability. Cognitive Psychology, 5 , 207–232. Tyler, T. R. (1984). Assessing the risk of crime victimization: The integration of personal victimization experience and socially transmitted information. Journal of Social Issues, 40 (1), 27–38. Tyler, T. R., & Cook, F. L. (1984). The mass media and judgments of risk: Distinguishing impact on personal and societal level judgments. Journal of Personality and Social Psychology, 47 (4), 693–708. Varey, R. J. (1999). Marketing, media, and McLuhan: Rereading the prophet at century’s end. Journal of Marketing, 63 , 148–159. Wallwork, L. W. (2001, February 25). Lockup for long locks. Parade Magazine ,2. Walsh, D. A. (2000). The challenge of the evolving media environment. Journal of Adolescent Health, 27 (2), 69–72. Wells, W. D., & Anderson, C. L. (1996). Fictional materialism. In K. Corfman & J. Lynch (Eds.), Advances in consumer research 23 , Provo, UT: Association for Consumer Research, 120–126. Williams, C. J. (2001, April 24). Entertained into social change. Los Angeles Times , pp. A1, A6, A7.