Issue Involvement As a Moderator of the Effects on Attitude of Advertising Content and Context

Richard E. Petty, University of Missouri-Columbia
John T. Cacioppo, University of Iowa
ABSTRACT - Two persuasion experiments were conducted manipulating the personal relevance of the message, the quality of the arguments employed, and the characteristics of the message source. The results suggested that message content factors are more influential than source characteristics under high involvement conditions. The reverse tends to occur under low involvement.
[ to cite ]:
Richard E. Petty and John T. Cacioppo (1981) ,"Issue Involvement As a Moderator of the Effects on Attitude of Advertising Content and Context", in NA - Advances in Consumer Research Volume 08, eds. Kent B. Monroe, Ann Abor, MI : Association for Consumer Research, Pages: 20-24.

Advances in Consumer Research Volume 8, 1981      Pages 20-24

ISSUE INVOLVEMENT AS A MODERATOR OF THE EFFECTS ON ATTITUDE OF ADVERTISING CONTENT AND CONTEXT

Richard E. Petty, University of Missouri-Columbia

John T. Cacioppo, University of Iowa

ABSTRACT -

Two persuasion experiments were conducted manipulating the personal relevance of the message, the quality of the arguments employed, and the characteristics of the message source. The results suggested that message content factors are more influential than source characteristics under high involvement conditions. The reverse tends to occur under low involvement.

THE INVOLVEMENT CONCEPT

Persuasion researchers within both social and consumer psychology have recently emphasized the need to distinguish between high and low involvement situations (cf., Petty, Cacioppo, & Heesacker 1981, Pay 1976). Although there are many specific definitions of "involvement" within both disciplines, there is considerable agreement that in high involvement situations, the persuasive message under consideration has a high degree of personal relevance to the recipient, whereas in low involvement situations, the personal relevance of the message is rather trivial. In social psychology, this view is best represented by the work of the Sherifs who have argued that high involvement occurs when the message has "intrinsic importance" (Sherif & Hovland 1961, p. 197) or "personal meaning" (Sherif et al. 1973, p. 311), and in consumer psychology this view is well-represented by Krugman (1965) who defines involvement as the number of "personal references" or connections that recipients make between the message and their own lives (p. 355).

Despite the widespread agreement that involvement has something to do with the personal relevance of a message, there is little agreement on the research operations employed in studying involvement. Some of the research on involvement conducted by both social (e.g., Hovland, Harvey, & Sherif 1957) and consumer (e.g., Newman & Dolich 1979) psychologists has investigated existing groups that differed in the extent to which an issue or product was personally important, or has employed designs allowing subjects to assign themselves to high and low involvement groups. These methods, which are correlational in nature, confound involvement with all other existing differences between the high and low involvement groups. Other social (e.g., Rhine & Severance 1970) and consumer psychologists (e.g., Lastovicka & Gardner 1979) have defined involvement in terms of the specific issue or product under consideration. This procedure, of course, confounds involvement with aspects of the issue or product that are immaterial to their personal relevance. Finally, some researchers have studied involvement by varying the medium of message presentation. Interestingly, however, some investigators have argued that television is a more involving medium than is print (Worchel et al. 1975), whereas others have argued Just the opposite (Krugman 1967). A preferable procedure that keeps recipient, message, and medium characteristics constant for high and low involvement conditions was introduced by Apsler & Sears (1968) and is the method employed in the studies to be reported here. In this procedure subjects are randomly assigned to high and low involvement conditions and receive the same message via the same medium, but high involvement subjects are led to believe that the issue or product has some personal relevance whereas low involvement subjects are not.

In addition to the methodological differences that have plagued the involvement concept, another area of disagreement concerns the effects on persuasion that involvement is expected to have. The Sherifs have argued that increased involvement is associated with increased resistance to persuasion (cf., Sherif, Sherif & Nebergall 1965). The notion is that on any given issue, highly involved persons exhibit more negative evaluations of a communication because high involvement is associated with an extended "latitude of rejection." Thus, incoming messages on high involvement issues are thought to have an enhanced probability of being rejected because they are more likely to fall within the unacceptable range of a person's implicit attitude continuum. This view has received considerable acceptance within social psychology (e.g., Eagly & Manis 1966, Greenwald 1980). Krugman (1965) has proposed an alternative view that has achieved considerable acceptance within consumer psychology (e.g., Ray 1974, Rothschild 1979). Under this second view, increasing involvement does not invariably increase resistance to persuasion, but instead shifts the sequence of communication impact. Krugman argues that under high involvement, a communication affects cognitions, then attitudes, then behavior, whereas under low involvement a communication affects cognitions, then behavior, and then attitudes. The focal goal of the present paper is to present and test a third view of how involvement affects persuasion.

INVOLVEMENT AS A DETERMINANT OF CONTENT-BASED PERSUASION

Elsewhere we have proposed that the level of involvement directs the focus of a subject's thoughts about a persuasive communication (Petty & Cacioppo 1979). Specifically, we have suggested that under high involvement conditions, the focus of thought is on the content of the persuasive message, whereas under low involvement conditions, the focus of thought is on non-content cues. Thus, under high involvement, if the communication presents arguments that are subjectively cogent and compelling, the recipient's thoughts will be primarily favorable and persuasion will result. If the communication presents arguments that are subjectively specious and subject to counterargumentation, resistance to persuasion (and perhaps boomerang) will occur. Thus, contrary to the Sherifs' view, increasing involvement can lead to either enhanced or reduced persuasion depending upon the quality of the arguments presented in the message.

In contrast to this focus on the content of a message under high involvement conditions, we have suggested that subjects who are not involved are more likely to focus on such non-content cues as the rewards available for adopting a certain attitude, the attractiveness, credibility, or power of the communication's source, and the number of others who advocate a certain position. Focusing on each of the latter aspects of a communication allows a person to evaluate a message or decide what attitudinal position to adopt without engaging in any extensive cognitive work relevant to the issue or product under consideration. As Miller et al. (1976) noted: "It may be irrational to scrutinize the plethora of counter-attitudinal messages received daily. To the extent that one possesses only a limited amount of information processing time and capacity, such scrutiny would disengage the thought processes from the exigencies of daily life" (p. 623). Thus, when a person is not highly involved with a persuasive message (i.e., when the message has no personal consequences), we propose that the person relies on a short-cut means of evaluation. Although, like Krugman, we are proposing that there are separate processes governing persuasion under high and low involvement, unlike Krugman we believe that the sequence of communication impact is the same--cognitions, attitudes, then behaviors. The difference between the two processes lies in what cognitions are affected--cognitions dealing with issue-relevant argumentation (high involvement), or cognitions dealing with non-content features of the influence situation (low involvement).

EMPIRICAL TESTS OF THE INVOLVEMENT CONCEPT

In an initial test of our two-process model of involvement (Petty & Cacioppo 1979), subjects heard a counterattitudinal message containing either strong or week arguments under conditions of either high or low involvement. As expected, increasing involvement enhanced the production of counter-arguments to the weak arguments and increased the production of favorable thoughts to the strong arguments. Consistent with this finding, increasing involvement increased the persuasiveness of the strong arguments, but decreased the persuasiveness of the weak arguments. Although the results of this initial study did support the view that subjects do more thinking about the content of a message under high involvement than under low, it did not directly address whether subjects are more attentive to content-irrelevant cues under low involvement then under high.

Next, we report two experiments designed to test the full two-process model of involvement. In each study, subjects were exposed to a persuasive communication. In Experiment 1 (conducted in collaboration with Rachel Goldman) the message was presented on audio tape and concerned a change in a campus regulation. In Experiment 2, the message was presented in print form and concerned a new consumer product. The following variables were manipulated in each study: (a) the personal relevance of the message (high and low involvement), (b) the quality of the arguments which subjects heard or read in support of the advocated conclusion (strong or weak arguments), and (c) a characteristic of the source presenting the message (high or low expert source in Experiment 1, high or low attractive source in Experiment 2). The two-process model of involvement would expect that under high involvement conditions, persuasion would be affected most by the quality of the message arguments employed, but that under low involvement conditions, persuasion would be tied most strongly to the credibility or attractiveness of the message source.

EXPERIMENT 1

One hundred forty-five male and female undergraduates at the University of Missouri participated in order to earn extra credit in an introductory psychology course. The design was a 2 (Issue involvement: high or low) X 2 (Argument quality: strong or weak) X 2 (Source credibility: high or low) factorial. Subjects were run in groups of 3 to 16 in cubicles designed so that no subject could have visual or verbal contact with any other subject. Upon arrival at the laboratory, subjects read that they would be rating tapes for their sound quality. After reading these instructions, subjects heard one of four tapes over headphones. After listening to the appropriate communication, subjects completed the dependent variable booklets, and were then debriefed, thanked, and dismissed.

Independent Variables

Argument quality.  All subjects heard a communication that advocated that seniors be required to pass a comprehensive exam in their major area as a requirement for graduation. The strong version of the message provided persuasive evidence (statistics, data, etc.) in support of the exam (e.g. institution of the exams has led to a reversal in the declining scores on standardized achievement tests at other universities). In contrast, the weak version of the message relied more on quotations and personal opinion and examples to support its position (e.g. the author's major advisor took a comprehensive exam and now had a prestigious academic position). The strong arguments were taken from a pool that had elicited primarily favorable thoughts in a pretest, and the weak arguments were taken from a pool that had elicited primarily counterarguments.

Issue involvement.  Subjects in the high involvement conditions heard the speaker advocate that the comprehensive exams should begin in the 1979-1980 academic year at their university (in which case they would all be affected personally by the proposal). In the low involvement conditions, the speaker advocated that the exams be initiated in the 1989-1990 academic term.

Source credibility.  The high credible source was described as a professor of education from Princeton University who had conducted a study of comprehensive exams nationally. The low credible source was a junior at a local high school who had prepared a term paper on the topic.

Dependent Variables

After hearing the tape, subjects completed two measures of opinion about the topic. First, subjects rated the concept "senior comprehensive exams" on four 9-point semantic differential scales (harmful-beneficial, wise-foolish, good-bad, favorable-unfavorable) that were summed to form a general measure of evaluation. Next, subjects responded to an 11-point Likert-type scale concerning their extent of agreement with the speaker's proposal. The responses to these two attitude measures were converted to standard scores and averaged to form an index of communication acceptance.

Following the key attitude measures, subjects completed some manipulation check measures and other ancillary questions. Finally they were given 4 minutes to list as many of the arguments provided in the communication as they could remember. Two judges, blind to the experimental conditions rated each argument listed for accuracy (p. 92). Similar statements of the same argument were only counted once. Disagreements between judges were resolved by consulting a third judge.

Results

Analyses on the manipulation check measures indicated that the three independent variables were successfully varied. Subjects hearing the high credibility speaker rated him as more "qualified" (M = 6.4) than subjects who were exposed to the low credibility induction (M = 5.8), F(1,137) = 4.86, p < .03. Also, subjects hearing the strong arguments rated their quality as being significantly higher (M = 8.9) than subjects exposed to the weak arguments (M = 4.5), F(1,137) = 51.02, p < .001. Finally, subjects in the high involvement condition rated the likelihood that the University of Missouri would institute comprehensive exams during their stay as higher (M = 5.5) than subjects in the low involvement conditions (M = 2.7), F(1,137) = 5.12, p <.02.

The means for each cell on the measure of communication acceptance are presented in Table 1, A 2 X 2 X 2 analysis of variance on this measure yielded two main effects and two qualifying interactions. A main effect for credibility, F(1,137) = 6.06, p < .02, Indicated that the high credibility communicator induced more acceptance (M = .21) than the low credibility communicator (M = -.21). A main effect for argument quality revealed that the strong arguments produced more agreement with the position advocated (M = .36) than did the weak arguments (M = -.36), F(1,137) = 20.35, p <.001.

Of most interest are the two interactions, however. An Involvement X Arguments interaction, F(1,137) = 6.05, p < .02, demonstrated that the strong arguments produced significantly more persuasion than the weak only under conditions of high personal involvement. Also, an Involvement X Credibility interaction, F(1,137) = 3.92, p <.05, revealed" that the high credibility speaker produced significantly more persuasion than the low only under conditions of low personal involvement.

TABLE 1

EFFECT OF INVOLVEMENT, ARGUMENT QUALITY, AND SOURCE CREDIBILITY ON ATTITUDES TOWARD COMPREHENSIVE EXAMS

Finally, an analysis of the argument recall scores indicated that subjects were able to recall more of the strong (M = 4.2) than the weak (M = 3.2) arguments, F(1,137) = 14.93, p <.001, but involvement did not affect argument recall.

Discussion

The present study provided evidence for the view that the level of issue involvement is one determinant of whether content or noncontent factors dominate in producing persuasion. When the issue was of high personal relevance to the subjects, the manipulation of argument quality had a significant impact on attitudes, whereas the effect of source credibility was small and not significant. Under low involvement conditions, however, the reverse occurred--source credibility had large effect on persuasion, but argument quality had little impact. This pattern of results provides strong support for the view that high issue relevance motivates diligent processing of the content of a message. When the issue is relatively uninvolving, however, subjects appear to be unwilling to engage in the effortful cognitive work necessary to evaluate the quality of the arguments.

Under low involvement conditions, subjects were content to form their attitude on the basis of who said it, rather than on the merits of what was said. In the second experiment, a conceptual replication of Experiment 1 was conducted employing advertising stimuli.

EXPERIMENT 2

Two hundred and forty undergraduates at the University of Missouri participated in s study entitled "Evaluations of Advertisements," and received extra credit in an introductory psychology course. The design of the study was a 2 (Product involvement: high or low) X 2 (Message content quality: strong or weak) X 2 (Source attractiveness: high or low) X 2 (Sex of subject). Fifteen subjects were run in each cell of the design in groups of 4 to 8.

Upon arrival at the laboratory, the subjects were told that they would be examining a variety of advertisements and providing their impressions of them. Each subject was handed a booklet containing six magazine ads. Five of the ads were for real but relatively unfamiliar products (e.g. Lux cigarettes, Rolex watches), and one of the ads was for

a fictitious product (Vilance shampoo). The fictitious ad was fifth in the booklet of six and was prepared specifically for the present experiment. The subjects were told to turn through the booklet at their own pace and give a signal to the experimenter when they had finished examining the ads. Following perusal of the ad booklet, subjects completed a dependent variable booklet, and were then debriefed, thanked, and dismissed.

Independent Variables

Product involvement.  Preceding each ad in the booklet was a brief description of the purpose of the ad. All subjects read the some descriptions for the real ads, but the description for the bogus Vilance shampoo ad differed for subjects in the high and low involvement conditions. Subjects in the low involvement conditions read:

The product you are about to see is being put out by the Europia Collection based in Vienna, Austria. Their main interest lies in introducing the product to the rest of Europe. Before launching a new European campaign, they have distributed their advertising materials to Journalism schools in the U.S. This is in order to determine if the ad has enough basic appeal to make it worth pursuing. The psychology department is assisting the journalism, school in this evaluation.

Subjects in the high involvement conditions read a similar description except that they were told that the company was interested in introducing the product to the United States.

The University of Missouri has been chosen for research purposes. It is for this reason that the product will soon be introduced in the Columbia area. Since you will soon be able to purchase this product in Columbia, the company is asking University students to evaluate their proposed advertisement.

Source attractiveness.  Four different advertisements for Vilance shampoo were created to vary the source and message variables. Each ad looked similar in that it presented a male and female in their early 20s giving the reasons why they liked Vilance shampoo. In the high attractive ads photograph of a couple that previously had been rated as "extremely attractive" was used, and in the low attractive ads a photograph of a couple that previously had been rated as "somewhat unattractive" was used. An "extremely unattractive" stimulus was not used because it did not prove plausible in pilot testing.

Message quality.  In the strong argument conditions, the text that accompanied the photograph presented arguments for the shampoo that previously had been rated as compelling and persuasive (e.g. Vilance contains minerals that strengthen each hair shaft so it helps prevent split-ends). In the weak argument conditions, the text presented arguments that previously had been rated as unpersuasive (e.g. Vilance has a down-to-earth brown color that makes us feel natural).

Dependent Variables

Subjects were asked to answer number of questions about each product for which they had seen an ad. Different questions about the advertisements were also posed to maintain the cover story. The crucial measure of attitude toward the product consisted of the sum of subjects' ratings of Vilance shampoo on four 9-point semantic differential-type scales (good/bad, satisfactory/unsatisfactory, favorable/unfavorable, high quality/low quality).

Results

The means for each cell on the measure of attitude toward Valance shampoo are presented in Table 2. A 2 X 2 X 2 X 2 analysis of variance revealed no main effects nor interactions involving the sex of subject variable so this will not be discussed further. A main effect for the argument quality manipulation, F(1,223) = 35.41, p <.0001, indicated that the strong arguments produced more favorable attitudes toward the product (M = 6.6) than did the weak arguments (M = 1.3). A main effect for the attractiveness variable, F(1,223) = 8.46, p <.004, showed that the high attractive models induced more acceptance of the product (M = 5.3) than did the low attractive models (M = 2.7).

Of most interest in the analysis, however, was a significant Involvement X Arguments interaction, F(1,223) = 10.84, p <.001. This interaction revealed that increasing the personal involvement of the advertisement enhanced the importance of the message content in affecting attitudes. When the relevance of the ad was increased, subjects responded more favorably to the ad with strong arguments, but less favorably to the ad with weak arguments. The Involvement X Source interaction was not significant this time though the means were directionally consistent with the findings of Experiment 1.

TABLE 2

EFFECT OF INVOLVEMENT, ARGUMENT QUALITY, AND SOURCE ATTRACTIVENESS, ON ATTITUDES TOWARD AN ADVERTISED PRODUCT

Discussion

Experiment 2 replicated the finding from Experiment 1 that involvement (personal relevance) is an important determinant of the extent to which content-based persuasion will occur. When the advertisement was high in personal relevance to the subjects, the quality or cogency of the arguments presented in the ad had a much greater impact on attitudes Coward the advertised product than when the ad was of low relevance. The study did not find strong support for the view that source factors were more important when the ad was low rather than high in personal relevance, however. In retrospect, this effect may not have been strong in this study because how the models looked may have been viewed as a relevant persuasive argument for some subjects! In other words, for the specific product employed (shampoo), the attractiveness of the models (especially their hair) may have served as persuasive testimony for the effectiveness of the product.

THE ROLE OF INVOLVEMENT IN PERSUASION

Taken together, the results of the two studies strongly indicate that under high involvement, message content is the prepotent determinant of the amount of persuasion that occurs. Less strongly, the studies suggest that under low involvement, non-content factors such as the credibility or attractiveness of the message source are more important. Thus, the present studies provide some evidence that attitude change is determined by different factors under high and low involvement conditions.

In another paper (Petty & Cacioppo 1981) we have argued that a persuasive massage can induce attitude change via one of two routes. Under the first, or central route, thinking about issue-relevant information is the most direct determinant of the direction and amount of persuasion produced. Attitude changes induced via this route tend to be relatively permanent and predictive of subsequent behavior. Under the second, or peripheral route, attitude change is the result of non-content cues in the situation. Changes induced via this route tend to be relatively temporary and are not highly predictive of subsequent behavior (see review by Cialdini et al., 1981).

According to this framework, involvement is a prime determinant of whether or not an attitude change is induced via the central or the peripheral route. The experiments reported here as well as the results of other recent studies are consistent with the view that under low involvement conditions, persuasion may typically be governed by such peripheral cues as source characteristics, concerns about desirable self-presentation, and/or one's social role, whereas under high involvement, persuasion may be governed more by message content factors such as the number, quality, and/or accuracy of the message arguments presented (cf., Cacioppo & Petty 1980, Chaiken 1980, Cialdini et al. 1976).

The level of involvement is not the only determinant of the route to persuasion, however. In addition to having the necessary motivation to think about issue-relevant argumentation, the message recipient must also have the ability to process the message if change via the central route is to occur. Thus, if involvement is high (and the person is motivated to think about the message content), but the arguments are too complex for the person to understand, or if too many distractions prevent issue-relevant thought, then the central route cannot be followed.

Finally, we note that attitude change via the central route is a very difficult way to change a person's attitudes. First, the message must show some personal relevance to the recipient. Second, the person must have the ability to process the message content. Third. the message must present arguments that elicit primarily favorable thoughts. If the recipient is able to counterargue the message, then increasing involvement will not facilitate persuasion. If a change can be produced via the central route however, the benefits are clear--the attitude change will tend to persist and be predictive of subsequent behavior. An alternative strategy is to induce attitude change via the peripheral route. Since the peripheral route induces only a temporary change, it will be necessary to constantly remind the recipient of the cue (e.g., attractive source) upon which the new attitude is based. Lutz (1979) provides the example of a person who drives Hertz Rent-a-Cars, not because the person has thought about the attributes of the company (central route), but only because he has been constantly reminded that O. J. Simpson endorses the company (peripheral route). If the favorable attitude about Hertz had been based on a full consideration of the positive features of the company, the favorable attitude would likely persist on its own. Since the favorable attitude is based on a positive peripheral cue however, the favorable attitude persists only so long as the cue remains salient (accomplished through advertising repetition). Such continually repeated positive cues may be sufficient to get a person to try the advertised product. Interestingly, once the person has tried the product, it may become more personally involving and may make the person more likely to think about the content of future advertisements about the product. In this manner a peripheral change can lead to a central one.

REFERENCES

Apsler, R., and Sears, D.O. (1968), "Warning, personal involvement, and attitude change," Journal of Personality and Social Psychology, 9, 162-166.

Cacioppo, J. T. , and Petty, R. E. (1980), "Sex differences in influenceability: Toward specifying the underlying process," Personality and Social Psychology Bulletin, in press.

Chaiken, S. (1980), "Heuristic versus systematic information processing and the use of source and message cues in persuasion," Journal of Personality and Social Psychology, in press.

Cialdini, R. B., Levy, A., Herman, P., Kozlowski, L., & Petty, R. (1976), "Elastic shifts of opinion: Determinants of direction and durability," Journal of Personality and Social Psychology, 34, 663-672.

Cialdini, R.B., Petty, R. E. , and Cacioppo, J. T. (1981), "Attitude and attitude change," Annual Review of Psychology, in press.

Eagly, A.H., and Mania, M. (1966), "Evaluation of message and communication as a function of involvement," Journal of Personality and Social Psychology, 3, 483-485.

Greenwald, A.G. (1980), "The totalitarian ego: Fabrication and revision of personal history," American Psychologist, 35, 603-618.

Hovland, C.I., Harvey, O.J., and Sherif, M. (1957), "Assimilation and contrast effects in reactions to communication and attitude change," Journal of Abnormal and Social Psychology, 55, 244-252.

Krugman, H.E. (1965), "The impact of television advertising: Learning without involvement," Public Opinion Quarterly, 29. 349-356.

Krugman, H.E. (1967), "The measurement of advertising involvement," Public Opinion Quarterly, 30, 583-596.

Lastovicka, J. L., and Gardner, D. M. (1979), "Components of involvement," in Attitude research plays for hash stakes, eds., J. Maloney and B. Silverman, Chicago: American Marketing Association, 53-73.

Lutz, R. J. (1979), "A functional theory framework for designing and pretesting advertising themes," in Attitude research plays for high stakes, eds., J. Maloney and B. Silverman, Chicago: American Marketing Association, 37-49.

Miller, N., Maruyama, G., Beaber, R., and Valone, K. (1976) "Speed of speech and persuasion," Journal of Personality and Social Psychology, 34, 615-624.

Newman, L. M., and Dolich, I.J. (1979), "An examination of ego-involvement as a modifier of attitude changes caused from product testing," Advances in Consumer Research, 6, 180-183.

Petty, R. E., and Cacioppo, J. T. (1979), "Issue involvement can increase or decrease persuasion by enhancing message-relevant cognitive responses," Journal of Personality and Social Psychology, 37, 1915-1926.

Petty, R. E., and Cacioppo, J. T. (1981), Attitudes and persuasion: Classic and contemporary approaches, Dubuque, Iowa: William C. Brown Co.

Petty, R. E., Cacioppo, J. T., and Heesacker, M. (1981), "The use of rhetorical questions in persuasion: A cognitive response analysis," Journal of Personality and Social Psychology, in press.

Ray, M. L. (1974), "Consumer initial processing: Definitions, issues, and applications," in Buyer/Consumer information processing, eds., G. Hughes and M. Ray, Chapel Hill: University of North Carolina Press, 145-156.

Ray, M. L. (1976), "Attitude as a communication response," in Attitude research at bay, eds., D. Johnson and W. Wells, Chicago: American Marketing Association, 87-111.

Rhine, R., and Severance, L. (1970), "Ego-involvement, discrepancy, source credibility, and attitude change," Journal of Personality and Social Psychology, 16, 175-190.

Rothschild, N. L. (1979), "Advertising strategies for high and low involvement situations," in Attitude research plays for high stakes, eds., J. Maloney and B. Silverman, Chicago: American Marketing Association, 74-93.

Sherif, M., and Hovland, C. (1961), Social judgment, New Haven: Yale University Press.

Sherif, C. W. et al. (1973), "Personal involvement, social judgment, and action," Journal of Personality and Social Psychology, 27, 311-328.

Sherif, C. W., Sherif, M., and Nebergall, R. E. (1965), Attitude and attitude change, Philadelphia: Saunders.

Worchel, S., Andreoli, V., and Eason, J. (1975), "Is the medium the message? A study of the effects of media, communicator, and message characteristics on attitude change," Journal of Applied Social Psychology, 5, 157-172.

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