Processes and Effects in the Construction of Normative Consumer Beliefs: the Role of Television

L. J. Shrum, University of Illinois
Thomas C. O'Guinn, University of Illinois
Richard J. Semenik, University of Utah
Ronald J. Faber, University of Minnesota
ABSTRACT - The knowledge that individuals use to make decisions and judgments comes from both direct and indirect experience. Such information may result from participation, from social networks, or via more indirect routes such as mass media. With respect to mass media, consumer researchers have been concerned predominantly with the effects of advertising. It is the contention of this paper that the programs between the ads also convey a wealth of information with respect to consumption, and may very well influence the decision making process. This paper presents studies which attempt to assess this influence. Survey and reaction time data serve to address both process and effects in discerning how individuals go about constructing their concepts of the social reality of consumption. The results support the notion that heavy television viewing influences consumption perceptions such that they more closely resemble the "reality" of the television world.
[ to cite ]:
L. J. Shrum, Thomas C. O'Guinn, Richard J. Semenik, and Ronald J. Faber (1991) ,"Processes and Effects in the Construction of Normative Consumer Beliefs: the Role of Television", in NA - Advances in Consumer Research Volume 18, eds. Rebecca H. Holman and Michael R. Solomon, Provo, UT : Association for Consumer Research, Pages: 755-763.

Advances in Consumer Research Volume 18, 1991      Pages 755-763

PROCESSES AND EFFECTS IN THE CONSTRUCTION OF NORMATIVE CONSUMER BELIEFS: THE ROLE OF TELEVISION

L. J. Shrum, University of Illinois

Thomas C. O'Guinn, University of Illinois

Richard J. Semenik, University of Utah

Ronald J. Faber, University of Minnesota

ABSTRACT -

The knowledge that individuals use to make decisions and judgments comes from both direct and indirect experience. Such information may result from participation, from social networks, or via more indirect routes such as mass media. With respect to mass media, consumer researchers have been concerned predominantly with the effects of advertising. It is the contention of this paper that the programs between the ads also convey a wealth of information with respect to consumption, and may very well influence the decision making process. This paper presents studies which attempt to assess this influence. Survey and reaction time data serve to address both process and effects in discerning how individuals go about constructing their concepts of the social reality of consumption. The results support the notion that heavy television viewing influences consumption perceptions such that they more closely resemble the "reality" of the television world.

INTRODUCTION

We know that consumers are influenced by perceptions of what others have and do. Some of these perceptions are determined via direct experience, others through less direct representations. Unfortunately, models of social influence in the field of consumer behavior have not taken account of how or how much the institutionalization of television in American homes has impacted this process. When most extant social influence models were forged, television was still a relative novelty. Today the average household watches over seven hours of it every day (Nielsen Television Index 1990). Further exacerbating this situation is that as a field, consumer behavior tends to think of mass communication narrowly as advertising. The things between the ads (a.k.a. programs) are rarely if ever studied as having any consumption related influence. This is unfortunate, because these programs provide overt and implied information about consumption norms and behavior. The manner in which this material is acquired, processed and utilized may yield significant theoretical and methodological implications for both consumer socialization and social cognition.

This paper argues and presents data in support of the belief that television programming is a significant and overlooked source of consumption related perceptions. This effect should be stronger among those who watch more television than those who watch comparatively less. In other words, we believe that people who watch more television have consumption related normative beliefs which more resemble the world as portrayed on television than those who watch less television. Furthermore, we believe that these beliefs are more accessible in memory among those who view more frequently.

THEORY

Antecedent to many attitudes and behaviors are normative beliefs, or beliefs about what is average or modal. The influence of normative beliefs are thought to extend to a wide range of processes and outcomes, from basic perception to overt behavior. Further, we have no reason to suspect that this is any less the case in the domain of consumer behavior. What we believe others possess, use, desire and value are subject to the same processes of social construction as any other normative belief. Yet, there has been little research in the pursuit of a more complete understanding of the acquisition and maintenance of normative economic and consumption related social beliefs. This seems particularly surprising in the field of consumer behavior, since so many ads, designed to attract attention and influence behavior, work on the basis of creating a positive perception for the targeted reference group member (i.e., housewife, business executive, teenager, happy family). From the 1920's on, this has been the essential premise of "reason why" and "slice of life" advertising (Fox 1984; Liess, Kline and Jhally 1988). These ads depended on the creation of the perception of the average person with an average problem, (then halitosis, now gingivitis), that only the advertised good or service could remedy.

From a-social science perspective, it is also essential to understand that life on television significantly differs from life directly experienced. Many things are over-represented (i.e., wealthy people), others under-represented (i.e., minorities). Television producers have over time decided what best fits the medium, and have structured portrayals accordingly. In the McLuhan (1964) sense of the medium being the message, this is reflected in things from the purely fictional portrayals of family life in the average situation comedy, to what is selected as news and how it is structured to best fit the demands and constraints of the medium. This disparity does, however, provide the researcher with a unique opportunity, the context for differentiating things learned via television versus more direct experience.

Cultivation

One of the more enduring, provocative and controversial contributions of mass communication research to social science has been in the area of social reality effects. Collectively, this research has demonstrated modest yet consistent associations between exposure to television and individuals' beliefs about various aspects of their social environment. Most commonly associated with Gerbner's cultivation the theory (Gerbner et al. 1977, 1980), research in this area has now extended beyond the bounds of that particular theory and constitutes a somewhat broader area of inquiry in mass communication known as social reality research (Hawkins and Pingree 1982). Still, both the general theoretical and methodological orientation of contemporary work in this area remains largely consistent with Gerbner's original conceptualizations. While there are some legitimate and well documented methodological criticisms of this research (Hirsch 1980, 1981), its intuitive appeal has kept it alive. It has-most recently been bolstered by supportive work in social psychology (Tyler 1980, 1984), and consumer behavior (O'Guinn et al. 1989; O'Guinn and Shrum 1990; Shrum, O'Guinn and Faber 1990).

The theory of cultivation holds that television viewing significantly assists in creating or cultivating a view of reality which is biased toward the highly formulaic and stylized narrative content of television. The more one views, the greater the effect. Cultivation research examines not the conscious acquisition and rational utilization of information, but rather the mere absorption of it. Summary beliefs about our social environment are built or "constructed" a la Berger and Luckman (1967), with bits of information from a number of sources, and these sources differ significantly in their properties. Television is a largely narrative, dramatic medium in which individual viewers willfully "suspend their disbelief," often in a very passive cognitive state (Ray 1973), for an average of four hours per day (NTI 1990). It therefore seems quite reasonable to cultivation theorists that television figures prominently in the individual's construction or summary beliefs about the nature of social things.

Psychologically, cultivation is a black box theory. It posits no explicit psychological dynamic. It offers instead a somewhat vague sort of socialization theory based on notions of frequency and passivity of viewing. Other theories may, however, contribute to understanding the cultivation effect, and assist in modifying and extending related social reality research. One is the work of decision scientists Kahneman and Tversky (1982; Tversky and Kahneman 1973) on the availability heuristic. The basic premise of this work is that the things more easily retrieved from memory are more accessible or more "available", and are thus disproportionately represented in judgments regarding the occurrence of events or the frequency of things. An heuristic is said to be most typically employed in low involvement situations or in simplifying complex tasks (see Folkes 1988). It could be that when television viewers are asked to estimate the incidence of something like the likelihood of being the victim of a violent crime, they rely on an heuristic in which mass media portrayals are simply more accessible. A finding supportive of this notion has been reported in at least one study. Lichtenstein et al. (1978) found that mass media exposure was positively correlated with overestimating the frequency of certain lethal events prominently displayed in the media.

The work of Tyler (Tyler 1984; Tyler and Cook 1984) is also important in this context. He and his colleagues have offered support for the belief that direct and indirect experience yield very different outcomes. Essentially, he argues that direct experience influences personal risk assessment, but not estimates of societal risk, and that indirect experience does just the opposite, influences environmental estimates, but not personal risk. So, for example, if a man was mugged in his neighborhood, this direct experience would lead him to believe that he is more at risk of violence and thus personal risk would be adjusted. However, seeing violent television programming will lead him to believe that the world in general is more violent, but have no impact on personal risk assessments. This may help to explain some of the inconsistencies in other social reality researchers' works, and may help predict precisely what will be more accessible in memory given the question posed by the research. This effect would also be consistent with most modern models of memory. Perhaps simply due to the higher probability of recently encountering a televised portrayal, such portrayals are more frequently represented, making their retrieval more probable.

One particular criticism of social reality research is that people know that television is not real, and therefore tend not to use the information gleaned from television. Indeed, some social reality research looks specifically at those viewers who tend to indicate that they believe the world of television to be real (termed "perceived reality", Potter 1986). However, such an assumption is not totally consistent with the person memory literature. In fact, many models of social cognition assume that subjects do not make an exhaustive search of memory for information bearing on a particular proposition, but instead search memory and use the first relevant piece of information encountered, which may or may not be initially perceived as real (see Wyer and Srull 1989; Sherman and Corty 1984). Furthermore, Wyer and Hartwick (1980) demonstrated that amount of processing influences both recall and judgments, showing that subjects remembered implausible arguments better than plausible ones, and were more apt to use these more easily recalled arguments as a basis for judgment. Thus, more accessible information, even though it is not necessarily more believable, may be used as a basis for judgment.

Social Reality and Consumption

A growing number of researchers interested in consumer behavior have written about the role of advertisements in creating symbolic and cultural beliefs in consumers (Mick 1986; McCracken 1986; Levy 1959). However, advertising represents only a small portion of media content. Furthermore, symbolic meaning is only one aspect of social reality. Perhaps even more fundamental, exposure to the "world" as portrayed on television has the potential to influence our perceptions of the very existence or incidence of things. If unchallenged, these perceptions become part of enduring cognitive structures. We begin to believe the world, or at least part of it, exists as it is constituted on television. Again, it is important to remember that these beliefs may not exist in a logical, rational or elaborated sense. It may, in fact, be the uncritical and vague way in which they are encoded and stored that gives them much of their power.

Very few studies have been published which have directly investigated consumer cultivation. Fox and Philliber (1978) examined the impact of television viewing on perceptions of affluence in the U.S. They found a significant relationship between amount of viewing and perceptions of affluence, but this relationship disappeared when control variables were applied. (Weimann 1984) conducted a study of Israeli viewers of American programs in Israel, and found that heavy viewers overestimated the percentage of Americans owning various household items as well as the average earnings of American families, relative to light viewers. Additionally, by testing different causal models, Weimann found that the data were best explained by a model in which control variables influence amount of viewing, but not cultivation effects directly, thus indicating that amount of viewing does have a direct effect on cultivation. However, the control variables did not include any measure of direct experience with American households. Lee (1988) and Lee and O'Guinn (1990) have studied cultivation among Taiwanese immigrants to the United States. By taking measures in Taiwan and at several levels of years lived within the United States, these researchers were able to use a cross-sectional design to study the process by which immigrants learned about the role and value of consumption objects. In Taiwan, respondents greatly overestimated material abundance in the United States, and this relationship was significantly explained by exposure to U.S. television programming. They also found that exposure to television remained a very strong predictor of beliefs concerning consumption behavior in Taiwanese immigrants, even after several years in the U.S.

It may also be important to consider the way consumer decision making is portrayed on television. Research has shown that decisions on television tend to infrequently consider important purchase decisions such as available finances and alternative choices (Faber 1978; Way 1982). Additionally, few sources of information are considered and decisions are made within a very short time frame (typically within one day) on television. Therefore, we may find that heavy viewers believe the average person to be less deliberate and thorough in their purchase decision making.

In sum, these studies support the basic cultivation hypothesis, but are also subject to the standard criticisms and limitations leveled at such work. With the above mentioned studies in mind, this paper presents the results of three studies which seek to extend mass-mediated social reality research into the domain of consumer behavior and to explore the psychological process which may account for these effects. In the first study a general population survey was administered. Given the variance in life experience in different demographic groups, it was deemed vital to collect data from widely divergent groups. Since we were also interested in process, a series of quasi-experiments were conducted. We will present the preliminary results of the survey and the results of the two quasi-experimental reaction time studies.

STUDY 1 (ILLINOIS GENERAL POPULATION SURVEY)

Method

Sample and Procedure. An extensive survey was prepared and mailed to approximately 2800 members of the adult Illinois population. The sample was selected via a standard systematic procedure using city directories. The sample was designed to best represent the adult population of Illinois. As of the writing of this paper, approximately 800 completed surveys had been returned.

The questionnaire attempted to assess the respondents' perceptions of the incidence of ownership, use or participation in consumption related behaviors, as well as their perceptions of income distributions. There were 42 such items, 3{} of which were purely consumption oriented, and another 12 which were broader societal level questions concerning handguns, violence, AIDS, etc. Following each question respondents were also asked about direct experience. An example of the format is:

What percentage of American households have a convertible automobile?

Do you currently have a convertible?   Yes    No

If not, have you had one in the past five years?   Yes     No

Other examples of incidence measures were jacuzzis, maids or servants, the percentage of adult Americans suffering from gingivitis or athlete's foot, the percent who are millionaires, comparison shop while making a purchase, suffer from high blood pressure and have dandruff. The items chosen were ones which were deemed to be over-represented on television (as compared to real life), either via program depictions or in commercials. Respondents were also asked to complete the 24 item BeLk materialism scale (BeLk 1985), as well as a full set of demographic measures.

TABLE 1

THIRD ORDER PARTIALS BETWEEN TV AND PERCEPTION MEASURES

(CONTROLLING FOR AGE, INCOME AND EDUCATION)

Television viewing was assessed in two ways. First, we had respondents estimate the number of hours they watched television in an average week during specific dayparts or time periods (6:00am noon), (noon - 7:00pm), (7:00pm - 10:00pm), (10:00pm - 6:00am). Secondly, we asked them to estimate the weekly hours spent watching specific types or genres of shows: soap operas, action/adventure, drama, etc.

Analysis

As of the writing of this paper, only 240 cases of the incoming data had been input and analyzed. Simple frequencies indicate sample statistics being very close to known population parameters in Illinois and the U.S, including the amount of television viewed. For example, the average person reported watching 3.9 hours of television per day in our survey compared to Nielsen's 4.1 (N"II 1990).

Three demographic measures were found to covary with many of the dependent measures and amount of television viewed. For example, TV viewing covaried with age (r = +.24, p < .001); education (r = -.27, p < .0001); and income (r= -.19, p < .01). Therefore, these demographic measures were used as control variables. Materialism did not covary with these variables and therefore was not used as a control variable. Zero through third order partial correlations were calculated between viewing hours for each program type and the normative perception measures. Analyses were also nested within gender. This procedure was considered necessary because of observed differences in viewing by gender (NTI 1990; Shrum, O'Guinn and Faber 1990) and the likelihood that direct experience with many dependent variables would also vary by gender.

Results

The general hypothesis, that heavier television viewers would have significantly higher estimates on the perception measures, was supported. Simple and partial correlations between the consumption related percentage estimates and the total hours of television viewing per week provided strong support for the general hypothesis. Most of the dependent measures were correlated in the .2 to .3 range (p < .05). However, stronger results were found when correlations were computed within genre or type of television show.

Length restrictions and the sheer volume of correlations preclude us reporting more results. Table l does, however, offer some illustrative findings. Shown are third order partial correlations for males who watch two different popular television genres: movies and soap operas. Higher levels of movie viewing among males were associated with higher estimates of the percentages of American households with private tennis courts (r= .36), convertible automobiles (1=.16), car telephones (r=.26), maids or servants (r=.22) and swimming pools (r=.23). Movie viewing among males was also positively associated with estimates of the number of people suffering from gingivitis (r=.31), having annual incomes of less than $15,000 (r=.23), belonging to a private spa or gym (r=.21), who are millionaires (r=.2 ,), who are currently addicted to crack or cocaine (r=.20), suffering from bladder control problems (r=.20), having dandruff problems (r=.30), who have had cosmetic surgery (r=.31), and who have called a 900 number dating or companion service (r=.16).

Interesting and fairly substantial third order partials were also observed when looking at soap opera viewing among men. Heavier television viewing correlated positively with higher levels of estimates of the percentage of adult Americans who suffer from hemorrhoids (r=.17), percent of American high school students regularly using illegal drugs (r-.12), percent of adult Americans suffering from athlete's foot (r=.63), percent of Americans who have had cosmetic surgery (.29), and the percent of adult Americans who have visited a friend or relative in the hospital in the last month (r=.64). These activities and advertising related health problems are common soap opera fare.

Several things remain to be done with respect to the survey data. One of the more pressing is to analyze the data while controlling for direct experience. This may allow us to get at some of the theoretical issues raised by Tyler (1984) and others. Also, once more of the data is in, we will be able to perform more comprehensive and sophisticated multi-factor analyses to determine viewing patterns as well as perception factors or clusters.

STUDY 2A AND 2B (REACTION TIME STUDIES)

In order to address the issues of process in making the perceptual estimates, a reaction time methodology was employed. Two specific hypotheses were tested: 1) heavy television viewers would give estimates which more resembled television portrayals relative to light television viewers, and 2) heavy television viewers would make these inferences quicker, thus showing faster response latencies. It is our contention that those who tend to watch more television should have a greater number of informational bits stored in memory. Television moves much faster than real life, consequently leading to numerous impressions about world beliefs. When asked to make an inference, subjects tend to use the information most accessible. This may take the form of a previously constructed belief. If this is the case, then subjects will most likely retrieve that judgment and terminate the search. Alternatively, if no previously constructed belief exists, then subjects will search for relevant information to use in making a probabilistic inference. In either case, if increased television viewing results in greater quantities of relevant information being deposited in memory, then it is quite probable that these television-originated bits of information will be used to make a judgment or indicate the likelihood of the occurrence of a particular event. Furthermore, if such information is over-represented in memory, then it would be more accessible, resulting in shorter response latencies.

Method

Samples and Procedures. Study 2a and 2b consisted of two separate but methodologically similar quasi experiments conducted during the 19891990 academic year. Because a prior study (Shrum, O'Guinn and Faber 1990) had indicated gender differences on the perception measures, Study 2a consisted of 55 undergraduate males in an introductory advertising class at a large midwestern university. Study 2b used the same subject pool, but consisted of 130 subjects, all of whom were women. Subjects in both studies performed the exercise on a microcomputer. Following established procedures for most reaction time studies (see Fazio 1990 for a review), subjects were instructed to be both quick and accurate. Upon directions to start the exercise, subjects were instructed via computer screen to press the space bar in order to receive a question. When the space bat was pressed, the first question appeared. Because the particular reaction time program employed did not allow for anything other than single digit responses, subjects indicated their response by pressing a key labeled from 0 to 9. These integers corresponded to an intuitive scale encompassing percentages from 0% to 99% (i.e., 0 = 0 - 9%, 1 = 10 - 19%, 4 = 40 - 49%, and so on). As soon as a key was pressed, the question disappeared and the subject received a prompt to press the space- bar for the next question. An internal clock recorded the time between when the space bar was pressed (causing the question to appear) and when the response was entered.

The single digit response scale was somewhat problematic due to the subjects' constraints on differentiation. Consequently, Study 2b attempted to overcome this problem by having subjects record via pencil and paper the precise number they were thinking of when they entered the scaled value on the computer. After each answer was recorded on the microcomputer, subjects were instructed to immediately write down their answers onto a questionnaire. These measures, rather than the computer recorded values, are reported.

Subjects were taken through several practice questions. After beginning, the first three items they encountered were also treated as practice questions. In order to rule out the possibility that reaction time results may be confounded because high television watchers, for whatever reasons, may be faster at responding on the computer, two measures were used to establish baseline reaction time. These items were chosen because they represented questions which should take virtually no time to compute. The two items were "Half of 100 is what?" and "How old are you?", and were the fourth and fifth items encountered. The reaction times for these two items were averaged for each person to create a baseline index. All reaction times were then transformed by subtracting the baseline index from the raw reaction time for each response. Such a transformation should theoretically yield a reaction time measure which only measures the time needed to develop an answer, thus eliminating individual differences in reading speed and dexterity. Although the raw reaction times of the baseline questions were not faster for high television viewers, indicating that high television viewers do not respond faster for dexterity reasons, the transformation was still used in order to remove variance due to individual response times apart from the decision making process.

Subjects were asked to indicate hours of viewing by daypart and by genre. The results presented for both studies represent a combination of hours of prime time viewing and afternoon viewing. These time periods were considered to give robust portrayals of consumption related phenomena. The low viewership within each daypart precluded individual analyses of each daypart alone, so the categories were collapsed. Subjects were divided into either high or low television viewing based on a median split of total weekly viewing hours. The median for the collapsed category for both studies was 8 hours per week,

Dependent Measures. The questions were much the same as those used in Study 1, i.e., "What percentage of Americans drive a convertible?", "What percentage of American households have a Mercedes?", etc. Study 2a used 60 measures and Study 2b used 42 measures. Only a portion of the questions pertained to consumer behavior issues, and only these are reported.

Results

The results of the two studies are shown in Table 2, and lend partial support for our hypotheses. In the first study, the reaction times tended to be significantly faster for the high television viewers, while the perception measures were significant for only 3 of the 7 measures, though all in the right direction. The dependent measures which yielded the most promising results tended to be fairly specific questions about ownership: % American households which have a private plane was significant for both reaction time and perception; % American households with a sports car was significant for reaction time. Two more general questions showed some significant differences between high and low viewers: % Americans who cheat on their taxes was significant for reaction time and perception; Go of American households which live beyond their means was significant for reaction time. In the second study, the perception measures tended to show differences, while the reaction times showed significant differences in only 1 of 4 instances, but again, all in the predicted direction. % Americans who have at one time or another used the services of a prostitute was significant for reaction time and perception; % of Americans who comparison shop was significant for perception, but in this case high TV viewers underestimated relative to low viewers, a finding consistent with television portrayals; the % of American high school students who regularly use illegal drugs and % American households with an annual income of less than $15,000 were significant for perception, consistent with the findings in Study 1; % of adult Americans who have called a 900 number dating or companion service was only significant for reaction time.

Similar results were obtained on the nonconsumer behavior perception questions. The data in total showed that virtually all of the reaction time measures were in the predicted direction (56 of 60 in Study 2a, 37 of 42 in Study 2b), with high television viewers responding faster. Approximately 70% of the perception measures were in the predicted direction.

DISCUSSION

Both the Illinois general population survey in Study 1 and the reaction time studies in Study 2a and 2b yield results consistent with our general hypotheses and theory. The data indicate that the more people watch television, the higher they estimate the incidence of product ownership and wealth, relative to low television viewers. In addition, consistent with the notion that television portrays a world of extremes, high television viewers gave higher estimates for both the number of households making less than $15,000 as well as the number of millionaires. As expected, stronger associations were demonstrated in the general population- survey, since college students watch decidedly less television than the general population, and also have more similar experiences. Thus, one should not be surprised to find weaker effects among-students.

The perception measures also indicated enough differences to encourage further exploration into this phenomenon. However, more work needs to be done in refining both the measurement instrument and the measurement method. A closer analysis of the television message may elicit more thoughtful and meaningful questions. A more detailed look at viewing patterns, or analyses within genre, would seem to be promising, since studies have shown that different genres do indeed convey qualitatively different messages (Greenberg 1980; Potter and Ware 1987). It should also be noted that the analyses within genre are very basic. More sophisticated analyses of isolating effects from specific types of programs are planned.

The attempts to address psychological process through the reaction time measures seem especially promising. Given the exceedingly large number of results in the predicted direction (90% over the two studies), it would seem that this is something more than mere chance or an easily detected third variable. Still, methodological concerns remain. For example, the response scale is by most standards much too long (Fazio 1990). A "Yes-No" or 'True-False" format may help to reduce the variances in the response times. Not only would such a format make the physical act of responding easier, but the time needed to formulate an answer would be shorter.

TABLE 2

PERCEPTION AND REACTION TIME DIFFERENCES (STUDIES 2A AND 2B)

While the reaction time results from student samples were quite promising, the perception differences were less impressive. As mentioned above, collecting data from non-students, who typically have a far broader range of television viewing and life experiences, may show greater cultivation effects. We are currently in the process of analyzing reaction time data collected from a nonstudent sample. Hopefully, such refinements of measures and techniques will prove helpful in teasing out the subtleties involved.

The effects of television programming have been of interest for some time to those in the field of mass communication. It seems that consumer researchers would also be interested in the effects of such a ubiquitous source of consumption information. Furthermore, the availability of this naturally occurring, and passively viewed background source, seems particularly attractive in its implications for social cognition.

REFERENCES

Belk, Russell (1985), "Materialism: Trait Aspects of Living in the Material World," Journal of Consumer Research, 12, 265-280.

Berger, P. L. and T. Luckman (1967), The Social Construction of Reality: A Treatise in the Sociology of Knowledge, Garden City, NY: Doubleday.

Faber, Ronald J. (1978), "Decision Making on Television: A Content Analysis," unpublished manuscript, University of Wisconsin.

Fazio, Russell H. (1990), "A Practical Guide to the Use of Response Latency in Social Psychology Research," in Review of Personality and Social Psychology, Vol. 11, ed. Ladd Wheeler, Beverly Hills: Sage.

Folkes, Valerie (1988), "The Availability Heuristic and Perceived Risk," Journal of Consumer Research, 15, 1 (June), 113-123.

Fox, Stephen R. (1984), The Mirror Makers: A History of Advertising and its Creators, New York: Random House.

Fox, Stephen and William Philliber (1978), "Television Viewing and the Perception of Affluence," Sociological Quarterly, 19, 103-112.

Gerbner, George, Michael Eleey, Marilyn Jackson-Beeck, Suzanne Jeffries-Fox, and Nancy Signorielli (1977), "TV Violence Profile No. 8: the Highlights," Journal of Communication, Vol. 27, 171-180.

Gerbner, George, Larry Gross, Michael Morgan and Nancy Signorielli (1980), "The 'Mainstreaming' of America: Violence Profile No. 11," Journal of Communication, 30, 10-27.

Greenberg, Bradley S. (1980), Life on Television: Content Analyses of U.S. TV Drama, Norwood, NJ: Ablex.

Hawkins, Robert, and Suzanne Pingree (1982), "Television's Influence on Constructions of Social Reality," in Television and Behavior: Ten Years of Scientific Progress and Implications for the Eighties, eds. D. Pearl, L. Bouthilet, and J. Lazar, Technical Reports, Washington: U.S. Government Printing Office, 224-247.

Hirsch, Paul (1980), 'The Scary World of the Nonviewer and Other Anomalies: A Reanalysis of Gerbner et al.'s Findings on Cultivation Analysis," Communication Research, 7 (4), 403456.

Hirsch, Paul (1981), "On Not Learning From One's Own Mistakes: A Reanalysis of Gerbner et al.'s Findings on Cultivation Analysis," Communication Research, 8, 3-37.

Kahneman, D. and Amos Tversky (1982), Judgment Under Uncertainty. Heuristics and Biases, New York: Cambridge University Press.

Lee, Wei-Na (1988), Becoming an American Consumer: A Cross-Cultural Study of Consumer Acculturation Among Taiwanese, Taiwanese in the U.S. and Americans, unpublished doctoral dissertation, University of Illinois.

Lee, Wei-Na and Thomas C. O'Guinn (1990), "The Consumer Acculturation of Taiwanese Immigrants," Working Paper, University of Texas.

Levy, Sidney (1959), "Symbols for Sale," Harvard Business Review, 37, 117- 124.

Lichtenstein, Sarah, Paul Slovic, Garuch Fischhoff, Mark Layman and Barbara Combs (1978), "Judged Frequency of Lethal Events," Journal of Experimental Psychology: Human Learning and Memory, 6 (November), 551-578.

Liess, William, Stephen Kline and Sut Jhally (1988), 'The Modern Advertising Industry," in Social Communication in Advertising: Persons, Products and Images of Well Being, New York: Methuen, 127-145.

McCracken, Grant (1986), "Culture and Consumption: A Theoretical Account of the Structure and Movement of the Cultural Meaning of Consumer Goods," Journal of Consumer Research; 13, 71-84;

McLuhan, Marshall (1964), Understanding Media: The Extensions of Man, New York: McGraw Hill.

Mick, David G. (1986), "Consumer Research and Semiotics: Exploring the Morphology of Signs, Symbols and Significance," Journal of Consumer Research, 13, 196-213.

Nielsen Television Index (1990), Nielsen National TV Ratings, New York: A.C. Nielsen Company.

O'Guinn, Thomas C., Ronald J. Faber, Nadine J. J. Curias and Kay Schmitt (1989), "The Cultivation of Consumer Norms," in Advances in Consumer Research, Vol. 16, ed. Thomas K. Srull, Provo, Utah: Association for Consumer Research.

O'Guinn, Thomas and L. J. Shrum (1990), "The Psychology of Normative Economic Beliefs: Mass Mediated Processes and Effects in Consumer Socialization," in Applied Economic Psychology in the 1990's, eds. Stephen Lea, Paul Webley and Brian Young, Exeter, England: International Association for Research in Economic Psychology.

Potter, W. James (1986), "Perceived Reality and the Cultivation Hypothesis," Journal of Broadcasting and Electronic Media, 2 (Spring), 159-174.

Potter, W. James and W. Ware (1987), "An Analyses of the Contexts of Anti-social Acts on Prime Time Television," Communication Research, 14, 664-686.

Ray, Michael L. (1973), "Marketing Communication and the Hierarchy-of-Effects," in New Models for Mass Communication Research, ed. P. Clarke, Beverly Hills: Sage, 147-176.

Sherman, Steven J. and E. Corty (1984), "Cognitive Heuristics," in Handbook of Social Cognition, Vol. 1, eds. Robert S. Wyer and Thomas K. Srull, Hillsdale, NJ: Lawrence Erlbaum Associates.

Shrum, L. J., Thomas C. O'Guinn and Ronald J. Faber (1990), "Television and the Social Construction of Reality," paper presented at the International Communication Association conference, Dublin, Ireland.

Tversky, Amos and D. Kahneman (1973), "Availability: A Heuristic for Judging Frequency and Probability," Cognitive Psychology, 5 (September), 207-232.

Tyler, Tom R. (1980), "Impact of Directly and Indirectly Experienced Events: The Origin of Crime-Related Judgments and Behaviors," Journal of Personality and Social Psychology, 39 (1), 13-28.

Tyler, Tom 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, Tom R. and Fay L. Cook (1984), "The Mass Media and Judgments of Risk: Distinguishing Impact on Personal and Social Level Judgments," Journal of Personality and Social Psychology, 47 (4), 693-708.

Way, Wendy (1982), 'The Consumer Content of Prime-Time Television: Implications for Consumer Educators," paper presented to the American Council on Consumer Interests.

Weimann, Gabriel (1984), "Images of Life in America: The Impact of American T.V. In Israel," International Journal of Intercultural Relations, 8, 185-197.

Wyer, Robert S. and Jon Hartwick (1980), "The Role of Information Retrieval and Conditional Inference Processes in Belief Formation and Change," Advances in Experimental Social Psychology, Vol. 13, ed. Leonard Berkowitz, New York: Academic Press.

Tyer, Robert S. and Thomas K. Srull (1989), Memory and Cognition in its Social Context, Hillsdale, NJ: Lawrence Erlbaum.

----------------------------------------