Personal Relevance As Moderator of the Effect of Public Service Advertising on Behavior

William K. Darley, University of Toledo
Jeen-Su Lim, University of Toledo
ABSTRACT - This paper examines the interaction of personal relevance and feelings toward a public service advertisement (Aad) and feelings toward the "drunk driving" issue (Aissue) on behavioral intention (BI). The results show that personal relevance significantly moderates the Aad-BI relationship as well as the Aissue-BI relationship. Directions for future research and public policy implications are discussed.
[ to cite ]:
William K. Darley and Jeen-Su Lim (1991) ,"Personal Relevance As Moderator of the Effect of Public Service Advertising on Behavior", in NA - Advances in Consumer Research Volume 18, eds. Rebecca H. Holman and Michael R. Solomon, Provo, UT : Association for Consumer Research, Pages: 303-309.

Advances in Consumer Research Volume 18, 1991      Pages 303-309

PERSONAL RELEVANCE AS MODERATOR OF THE EFFECT OF PUBLIC SERVICE ADVERTISING ON BEHAVIOR

William K. Darley, University of Toledo

Jeen-Su Lim, University of Toledo

ABSTRACT -

This paper examines the interaction of personal relevance and feelings toward a public service advertisement (Aad) and feelings toward the "drunk driving" issue (Aissue) on behavioral intention (BI). The results show that personal relevance significantly moderates the Aad-BI relationship as well as the Aissue-BI relationship. Directions for future research and public policy implications are discussed.

INTRODUCTION

Despite the recent increase in research efforts aimed at a greater understanding of attitude toward the advertisement and its role in the persuasion process (Shimp 1981, Burton and Lichtenstein 1988; MacKenzie, Lutz, and Belch 1986; Gorn 1982), many aspects of this complex phenomenon remain that have yet to receive proper theoretical and empirical consideration (Lutz 1985). The bulk of advertising research has been oriented toward what is said or felt about the advertisement that affects choice processes (Cushing and Douglas-Tate 1985). Previous studies have focused on the attitude toward the advertisement (Aad)-behavioral intention (BI) relationship with little interest in the possible moderating influence of other factors.

It is widely believed that ad effectiveness is moderated by audience involvement (Greenwald and Leavitt 1984). Although the data base regarding the behavioral consequences of persuasion has been growing, many questions remain unanswered. While past studies have focused on commercial or product advertising, the relevance to an applied and important realm of public service advertising has been ignored.

This study extends our understanding of individuals' affective reactions to a persuasive message. It examines the overall moderating effect of personal relevance upon the relationship between feelings toward a public service advertisement (Aad) and behavioral intention as well as between feelings toward an issue (Aissue) and behavioral intentions.

Attitude-behavior consistency has been the subject of much study. However, the affective reactions to an advertisement as they relate to intentions have received little attention, particularly in the context of the moderating role of personal relevance. In this study, consumers' reactions to a public service announcement introduce a new dimension in the subject matter as compared with previous studies.

HYPOTHESIS

Two types of involvement can affect message acceptance. One type concerns the extent to which the attitude issue under consideration is personally important to the recipient. The other type concerns the extent to which the particular attitudinal response adopted is of personal importance to the individual (Petty and Cacioppo 1979). This study focuses on the extent to which the attitude issue under consideration is of personal importance.

Personal relevance of an issue is one determinant of the route to persuasion that an individual is liable to follow while selecting between the alternatives of "central" versus "peripheral" routes to persuasion. Petty, Cacioppo, and Schumann (1983) contend that under the central route, attitude change is based on the careful and diligent consideration of the message claims or content. In contrast, under the peripheral route, attitude change is based on peripheral cues such as source attractiveness, quantity of arguments presented, and musical background. High personal relevance is expected to enhance thoughtful evaluation of message content. Low personal relevance is expected to lead to less thoughtful evaluation of the message content (Petty and Cacioppo 1979, 1981; Petty, Cacioppo, and Goldman 1981). Subjects involved with an issue are more likely to process message arguments in a systematic manner, employing the central route, than are less-involved subjects (Chaiken 1980). Borgida and Harvard-Pitney (1983) also found that more-involved subjects are more responsive to the message content, whereas less-involved subjects are more susceptible to perceptual salience effects.

High personal relevance leads to high attitude-behavior consistency. Sivacek and Crano (1982) found that the degree to which an individual perceives an attitude as hedonistically relevant affects the relationship between attitude and behavior. Leippe and Elkin (1987) found that issue-involved subjects were more likely to engage in attitude-behavior consistency. Thus, high personal relevance leads to high attitude-behavior consistency.

Attitude change induced via the peripheral route tends to be weak compared to attitude change resulting from central route processing. Individuals who are deeply involved with a product are more likely to behave in accordance with their deeply held attitudes and feelings. Strong affect or high order affect will tend to lead to strong behavioral intention. Conversely, weak affect or low order affect will tend to lead to weak behavioral intention (Smith and Swinyard 1988). The foregoing discussions suggest the following hypotheses concerning feelings toward the advertisement (Aad) and behavioral intention (BI) relationship as well as feelings toward the drunk driving issue (Aissue) and behavior intention (BI).

H1(a): The relationship between Aad and BI will be moderated by the level of personal relevance.

H1(b): For high personal relevance condition, the relationship between Aad-BI will be stronger than for low personal relevance condition.

H2(a): The relationship between attitude toward the issue (Aissue) and BI will be moderated by the level of personal relevance.

H2(b): For high personal relevance condition, the relationship between Aissue-BI will be stronger than for low personal relevance condition.

METHODOLOGY

Subjects, Research Design and Exposure Conditions

Subjects were undergraduate business students enrolled at an urban midwestern university. They were all above 21 years of age, were juniors and seniors, were working part or full time, and held driver's licenses. Eighty-seven subjects participated in the study and were randomly assigned to one of the three conditions (high personal relevance, low personal relevance, and control group). Accordingly, there were twenty-nine subjects per cell or condition.

The stimulus for the present study was an actual public service announcement put together by a State Governor's Task Force on Drunk Driving. It was a professionally produced sixty-second commercial. None of the respondents had been previously exposed to this public service announcement. Appendix 1 presents the actual wording of the public service announcement.

To operationalize personal relevance, subjects in the high personal relevance condition were told that an identical "get-tough" drunk-driving campaign was to be initiated in their own state starting on a specified date (see Appendix 1). Thus, the message would affect all subjects personally. In the low personal relevance condition, subjects were told that the campaign was currently going on in another state. The message, therefore, would personally affect none of the subjects. The foregoing procedure is similar to the approach of Petty and Cacioppo (1981), and Petty, Cacioppo, and Schumann (1983).

Subjects listened to the radio ad. The same version was presented to the two personal relevance groups. Written instructions were provided before and after the radio ad. After listening to the radio ad, subjects responded to the primary or dependent variables. Responses were also elicited concerning manipulation check items and ancillary items such as demographic variables.

Dependent Variables

The dependent measures collected in this study were (a) feelings toward the public service advertisement, Aad, (b) feelings toward the drunk driving issue, Aissue, and (c) Behavioral Intention, BI. Following the presentation of the ad stimulus, subjects reported their reactions to the aforementioned measures.

(a) Feelings toward the advertisement (Aad)

This was measured by asking subjects to complete the statement: "I feel the advertisement I just heard is ...." Responses were to be made in terms of four evaluative semantic differentials. This part of the questionnaire was designed to assess relevant affect toward the commercial. The scales were bipolar (bad/good, irritating/non-irritating, uninteresting/interesting, unpleasant/pleasant) and were anchored -3 to +3. This measure was consistent with the current conceptualization and operationalization of this construct (see for examples Edell and Staelin 1983; Gardner 1983). A positive score indicates a positive feeling toward the advertisement.

(b) Feelings toward the drunk-driving issue (Aissue)

Aissue was measured using a 3-item semantic differential scale. Subjects were asked to indicate their general feelings about "drunk-driving" as a pressing issue by completing the statement: "I feel concern about drunk-driving is... (useless/useful, unimportant/important, meaningless/meaningful)." The scales were bipolar and anchored -3 to +3. A positive score in this scale indicates the subject perceives the drunk-driving issue as a pressing one.

(c) Behavioral Intention (BI)

To measure behavioral intention, each subject was asked to indicate on a 7-point scale personal likelihood of (a) driving under the influence of alcohol when a "get tough" with drivers who drive under such influence is introduced in the individual's state, and (b) trying to stop a friend who wanted to drive under the influence of alcohol when a "get tough" with drivers under the influence of alcohol is introduced in the individual's state. These conative measures were coded 1 to 7, anchored by "unlikely" and "likely." The second measure was reverse-scored to maintain consistency so that the higher the score on the BI, the more likely the subject would drive under the influence of alcohol.

ANALYSIS

Moderated Multiple Regression (MMR) analysis was used to assess the statistical significance of personal relevance as a moderator variable. MMR examines the hypothesized interaction and main effects by model comparisons of the full- and restricted-regression models (Pedhazur 1982; Saunders 1956; Zedeck 1971).

An F test was performed to determine whether the addition of interaction terms significantly increases the variance explained in the dependent variable. Differences in the R2 were tested using a procedure recommended by Saunders (1956). The significant F test on R2 difference indicates the slopes of the regression equation for the subgroups are statistically different. Subgroup analysis was employed to investigate the directionality of the significant moderating effects (see for examples, Cohen and Cohen 1975; Anderson 1986). Graphs of the regression lines for each subgroup were evaluated to determine the nature of the interactions.

TABLE 1

CELL MEANS AND RELIABILITY OF AAD, AISSUE AND BEHAVIOR INTENTION

RESULTS

To assess personal relevance manipulation effectiveness, subjects were asked to tell how well they thought each of three phrases ("important to me," "meaningful to me," and "worth remembering") described the "get tough on drunk-driving campaign." Each scale was anchored by "not well at all" (1) and "extremely well" (7). These three phrases were abstracted from the "personal relevance" factor in Wells, Leavitt, and McConville (1971). In addition, this scaling parallels that employed in the latter study.

The manipulation checks for the personal relevance condition indicated this manipulation had been successful. Subjects in the high personal relevance condition scored higher than the subjects in the low personal relevance condition. For "important to me," "meaningful to me," and "worth remembering," means were 3.41, 3.55 and 2.62 respectively for the low personal relevance condition and were 4.00, 4.57 and 4.40 for the high personal relevance conditions. The differences between the respective high and low condition means were in the right direction and significant at p < .05.

Primary Variables

Table 1 presents the reliability measures of and the summated scores for the 4-item attitude toward the advertisement (Aad) scale, the 3-item attitude toward the drunk-driving issue (Aissue) scale, and the 2-item behavioral intention (BI) scale. The reliability measures were .82, .88, and .76 for Aad, Aissue and BI respectively.

For both Aad and Aissue, the high personal relevance condition elicited stronger feelings than the low personal relevance condition. The high personal relevance condition revealed a tendency of a more positive Aad (3.14) than the low personal relevance condition (2.0). The high personal relevance group also tended to consider the drunk driving issue (Aissue) as a more pressing issue (6.52) than did the low personal relevance subjects (5.55). In addition, for behavioral intentions (BI), the high- personal relevance condition generated a lower level of intentions (3.07 versus 3.86). Thus, for the foregoing, subjects were less likely to drive under the influence of alcohol. Also, the control group means were in the right direction.

The prediction of Hypothesis la was that the relationship between Aad and BI would be moderated by personal relevance. Table 2 shows the results of the moderated regression analysis with personal relevance as moderator. The results show significant moderating effect of personal relevance on the relationship between Aad and behavioral intention.

The change in R2 from .074 of the restricted model to .187 of the full model for Aad is statistically significant at the 0.05 level. The foregoing provides support for Hypothesis la.

The partial regression coefficients are reported in Table 2. The partial regression coefficient for Aad by personal relevance interaction is significant at the .05 level(t = -2.12). The partial regression coefficient for the interaction term is .26, indicating a negative moderating effect of personal relevance on behavioral intention (BI). To further examine the directionality of this interaction, subgroup analysis was performed. In this procedure, subgroup regression lines were graphed by plotting scores of BI for the high (mean plus 1 standard deviation) and the low (mean minus 1 standard deviation) Aad and Aissue. For example, the BI score for the high Aad is calculated from the regression equation by substituting "Aad" with mean plus one standard deviation for the Aad value. For a detailed discussion of the procedure, see Pedhazur (1982), Cohen and Cohen (1975), and Anderson (1986). The slope of the subgroup regression line for the high PR group is steeper than that for the low PR condition (see Figure 1). The high PR group showed a significantly stronger Aad-BI relationship than did the low PR group. The results in Table 2 and Figure 2 provide support for Hypothesis lb. Therefore, the hypothesis that the relationship between Aad-BI would be stronger for the high personal relevance condition than for the low personal relevance condition is supported.

TABLE 2

MODERATED REGRESSION ANALYSIS ON BEHAVIORAL INTENTION WITH PERSONAL RELEVANCE (PR) AS MODERATOR

Hypothesis 2a stated that the relationship between attitude toward drunk-driving issue (Aissue) and behavioral intentions is moderated by personal relevance. The results presented in Table 2 show that the R2 change from the restricted (.141) to the full model (.286) for Aissue is statistically significant at the .01 level. This provides support for Hypothesis 2a.

The partial regression coefficient for Aissue by personal relevance interaction is significant at the .0} level with a t value of -3.31 (see Table 2). The partial regression coefficient for the interaction term is -.50 which suggests a negative moderating effect of personal relevance on behavioral intention. Subgroup analysis was performed to further examine the directionality of this interaction. The slope is steeper for the high personal relevance condition than for the low personal relevance condition (see Figure 1). Thus, the hypothesis that the relationship between Aissue-BI would be stronger for the high personal relevance condition than for the low personal relevance condition is supported (see Table 2 and Figure 1).

FIGURE 1

INTERACTIONS BETWEEN PERSONAL RELEVANCE AND INDEPENDENT VARIABLES

CONCLUSION AND DISCUSSION

The focus of this paper was to test the moderating role of personal relevance. The moderating hypothesis of personal relevance in the Aad-BI relationship (Hypothesis 1) was supported. Also, the evidence showed that the strength of the relationship between Aad-BI was significantly stronger for the high personal relevance group than for the low personal relevance group. The subgroup regression slope for the Aad was steeper for the high personal relevance group than for the low personal relevance group.

The moderating hypothesis of personal relevance in the Aissue-BI relationship was supported (Hypothesis 2). The evidence provided in Table 2 and Figure 1 showed that the strength of the relationship between Aissue-BI was significantly stronger for the high personal relevance condition than for the low personal relevance condition.

The foregoing results show that high personal relevance enhances the relationship between attitude toward the ad and intentions as well as between attitude toward the issue and intentions. Thus, the findings are consistent with the Elaboration Likelihood Model (Petty, Cacioppo, and Schumann 1983). Past studies have found that attitude change via the central route tends to be persistent and predictive of subsequent behavior for the central route. Consequently, for the high personal relevance condition, stronger relationships were found for the two types of feelings and behavioral intentions.

The results suggest some public policy implications, at least for the design of advertisements concerning social causes or issues. Public service advertising has generally focused on the content and source characteristics and has ignored the potential consequences of subjects' reactions toward the message and issue involvement. These results would suggest that elicited feelings toward the public service advertisement need to be carefully considered.

The personal relevance of an advertisement, with some creative ingenuity, is manipulatable in real world situations. Thus, public service announcements should be presented in such a way as to make them personally relevant to consumers. It is also important in public service advertising to stress a link between the issue in question and the self. Such personal connection should add to the effectiveness of public service advertising.

While this study adds to the growing but still uncertain research data base regarding the behavioral consequences of persuasion under differing involvement conditions, further research is needed. Investigations of the reactions in the context of actual product advertisements involving different types of content and consideration of overt behavioral measures may provide useful insights. Differences among individuals' personal relevance links to brands or products and the possible impact of this variable on the Aad-behavioral intention relationship should be studied.

Research has shown that in some cases, the consumer's feelings toward the advertisement mediates the consumer's feelings toward the object. However, while the exact nature of the relationship between feelings toward the advertisement (Aad) and toward the object (Ao) is still under investigation (see for example Mackenzie et al. 1986), it has been shown that feelings toward the advertisement (Aad) can make a significant contribution toward feelings toward the advertised product or issue and behavioral intention. Thus, the incorporation of feelings regarding the advertised issue should extend our understanding of the role of feelings evoked by an advertisement in the persuasion process.

This study is not without its limitations. Although the results are interesting, care should be exercised in the interpretation and generalization of the findings. The study utilized a single ad design with a single exposure. Thus, it is possible that the results might have been influenced by impression management biases. This may be true in the behavioral intention (BI) results in that when subjects are asked sensitive questions like the "likelihood of drunk driving" they might give socially desirable answers. An alternative operationalization of the BI measure which does not entail the potentially sensitive "driving under the influence" wording is, for example, to measure the likelihood of driving after a specific number of drinks/body weight are consumed in a specific time period (Lastovicka, Murry, Joachimstaller, Bhalla, and Scheurich 1987). These authors found greater agreement with this objective measure than with the measure of driving drunk. Future research should therefore investigate this measurement issue together with the possible "subject effects" and the influence of personal relevance on "subject effects" or demand artifacts.

APPENDIX 1

CONTENT OF PUBLIC SERVICE ANNOUNCEMENT

In 19--, (STATE) adopted new tougher drunk driving laws, and we told you that if we pulled you over you'd better be sober. Some of you listened, but a lot of you did not. Those of you who still drink and drive, despite all the warnings, are at a terrible risk, a risk of killing yourself, or worse killing someone else, and the risk you're taking is unacceptable. We are going to do everything in our power to stop you. As a consequence of the death and destruction caused by drunk drivers, the combined police forces of (STATE) are coming out in force. There will be more officers, more cars, more sobriety checkpoints. If you drive drunk we want you to remember this, that you are on a collision course with every law enforcement officer in (STATE). The heat is on this summer. Don't drink and drive! This is a message from the Governor's task force to reduce drunk driving.

REFERENCES

Anderson, Carol H. (1986), "Hierarchical Moderated Regression Analysis: A Useful Tool For Retail Management Decisions," Journal of Retailing, 62, 186-203.

Borgida, E. and B. Howard-Pitney (1983), "Personal Involvement and the Robustness of Perceptual Salience Effects," Journal of Personality and Social Psychology, 45, 560-570.

Burton, Scot and Donald R. Lichtenstein (1988), 'The Effect of Ad Claims and Ad Context on Attitude Toward the Advertisement," Journal of Advertising, 17, 3-11.

Chaiken, Shelly (1980), "Heuristic Versus Systematic Information Processing and the Use of Source Versus Message Cues in Persuasion," Journal of Personality and Social Psychology, 39, 752-766.

Cohen, Jacob and Patricia Cohen (1975), Applied Multiple Regression/Correlation Analysis for the Behavioral Sciences, Hillsdale, N.J.: Lawrence Erlbaum Associates, Publishers.

Cushing, Peter and Melody Douglas-Tate (1985), "The Effect of People/Product Relationships on Advertising Processing," in Psychological Processes and Advertising Effects, Linda A. Alwitt and Andrew A. Mitchell, eds. Hillsdale, NJ: Lawrence Erlbaum Associates, Publishers, 241-259.

EdelL Julie A. and Richard Staelin (1983), "The Information Processing of Pictures in Print Advertisements," Journal of Consumer Research, 10, 45-61.

Gardner, Meryl P. (1983), "Advertising Effects on Attributes Recalled and Criteria Used for Brand Evaluations," Journal of Consumer Research, 10, 310-318.

Gorn, Gerald J. (1982), "The Effects of Music in Advertising on Choice Behavior: A Classical Conditioning Approach," Journal of Marketing, 46, 94-101.

Lastovicka, John L., John P. Murry, Jr., Erich A. Joachimstaller, Gaurav Bhalla and Jim Scheurich (1987), "A Lifestyle Typology to Model Young Male Drinking and Driving," Journal of Consumer Research, 14, 257-263.

Leippe, Michael R. and Roger A. Elkin (1987), "When Motives Clash: Issue Involvement and Response Involvement as Determinants of Persuasion," Journal of Personality and Social Psychology, 52, 269-278.

Lutz, Richard J. (1985), "Affective and Cognitive Antecedents of Attitude Toward the Ad: A Conceptual Framework," in Psychological Processes and Advertising Effects, Linda A. Alwitt and Andrew A. Mitchell, eds. Hillsdale, N.J.: Lawrence Erlbaum Associates, Publishers, 45-63.

Mackenzie, Scott B., Richard J. Lutz and G. Belch (1986), "The Role of Attitude Toward the Ad As A Mediator of Advertising Effectiveness: A Test of Competing Explanations," Journal of Marketing Research, 23, 130-143.

Pedhazur, Elazar J. (1982), Multiple Regression in Behavioral Research, New York: Holt, Rinehart and Winston, Inc.

Peter, Paul J. and Jerry C. Olson (1987), Consumer Behavior: Marketing Strategy Perspectives, Homewood, IL: Richard Irwin, Inc.

Petty, Richard E., John T. Cacioppo and David Schumann (1983), "Central and Peripheral Routes to Advertising Effectiveness: The Moderating Role of Involvement," Journal of Consumer Research, 10, 135-146.

Petty, Richard E. and John T. Cacioppo (1981), "Issue Involvement As A Moderator of the Effects on Attitude of Advertising Content and Context,"in Advances in Consumer Research, Vol. 8, Ken Monroe, ed. Ann Arbor, MI: Association for Consumer Research, 20-24.

Petty, Richard E., John T. Cacioppo and Rachel Goldman (1981), "Personal Involvement As A Determinant of Argument-Based Persuasion," Journal of Personality and Social Psychology, 41, 847-855.

Petty, Richard E. and John T. Cacioppo (1979), "Issue Involvement Can Increase or Decrease Persuasion by Enhancing Message-Relevant Cognitive Responses," Journal of Personality and Social Psychology, 37, 1915-1926.

Richins, Marsha L. and Peter H. Bloch (1986), "After the News Wears Off: The Temporal Context of Product Involvement," Journal of Consumer Research, 13, 280-285.

Saunders, David (1956), "Moderator Variables in Prediction," Educational and Psychological Measurement, 16, 209-222.

Shimp, Terence A. (1981), "Attitude Toward the Ad As a Mediator of Consumer Brand Choice," Journal of Advertising, 10, 9-15.

Sivacek, John and William D. Crano (1982), "Vested Interest As A Moderator of Attitude-Behavior Consistency," Journal of Personality and Social Psychology, 43, 210-221.

Smith, Robert E. and William R. Swinyard (1988), "Cognitive Responses to Advertising and Trial: Belief Strength, Belief Confidence and Product Curiosity," Journal of Advertising, 17, 3-14.

Wells, William D., Clark Leavitt and Maureen Mcconville (1971), "A Reaction Profile for TV Commercials," Journal of Advertising Research, 11, 11-17.

Zedeck, Sheldon (1971), "Problems With the Use of Moderator Variables," Psychological Bulletin, 76, 295-310.

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