The Effect of Framing and Advertising Sequencing on Attitude Consistency and Behavioral Intentions

Michael A. Kamins, University of Southern California
Lawrence J. Marks, Kent State University
ABSTRACT - This study investigates the effect of framing and advertising sequencing on attitude consistency and behavioral intent. The results revealed that the sequence of exposure differentially impacted subjects' derivation of attitude and purchase intention consistency when using the unframed ad as the basis for product evaluation. As compared to viewing the unframed ad first, prior exposure to the framed ad had a stronger impact on the effect of the unframed ad on the formation of attitude consistency and behavioral intention. This -is discussed in the context of the availability-valence hypothesis.
[ to cite ]:
Michael A. Kamins and Lawrence J. Marks (1987) ,"The Effect of Framing and Advertising Sequencing on Attitude Consistency and Behavioral Intentions", in NA - Advances in Consumer Research Volume 14, eds. Melanie Wallendorf and Paul Anderson, Provo, UT : Association for Consumer Research, Pages: 168-172.

Advances in Consumer Research Volume 14, 1987      Pages 168-172

THE EFFECT OF FRAMING AND ADVERTISING SEQUENCING ON ATTITUDE CONSISTENCY AND BEHAVIORAL INTENTIONS

Michael A. Kamins, University of Southern California

Lawrence J. Marks, Kent State University

ABSTRACT -

This study investigates the effect of framing and advertising sequencing on attitude consistency and behavioral intent. The results revealed that the sequence of exposure differentially impacted subjects' derivation of attitude and purchase intention consistency when using the unframed ad as the basis for product evaluation. As compared to viewing the unframed ad first, prior exposure to the framed ad had a stronger impact on the effect of the unframed ad on the formation of attitude consistency and behavioral intention. This -is discussed in the context of the availability-valence hypothesis.

INTRODUCTION

Recent research clearly indicates that knowledge which is activated and available during evaluative processing strongly influences attitudinal judgments and attitude-behavior consistency. From both a theoretical and practical perspective it is valuable to understand what factors influence knowledge activation (i.e., information availability) and to further explore the effects of activated knowledge on attitude structures. This paper briefly reviews two sets of factors which have typically been suggested as influencing information availability-prior product-related experiences and stimulus factors--and explores the effects of sequence of advertisement processing.

Fazio and Zanna (1981) have presented a compelling argument for the idea that the relation between people's attitudes and their behavior will be stronger when they have formed their attitudes from direct behavioral experience with the attitude object than when they have only indirect experience. One reason for this is that direct experience seems to make more information about the object available to the person. That is, direct experience with an object appears to strengthen the associations that are made in memory about an object and so this attitudinal information will be more accessible and utilizable when behavior is contemplated (Kisielius and Roedder 1983). In a marketing application of Fazio and Zanna's ideas, Smith and Swinyard (1983, 1985) found that attitudes are more extreme and confidently held when subjects are exposed to product trial as opposed to advertising. Moreover, these authors observed that attitude behavior consistency was significantly higher for subjects exposed to the trial as opposed to the advertising condition.

Advertisers seldom have the opportunity to manipulate the type of experience which the readers have. However, advertisers can and do manipulate stimulus factors which can influence the amount and type of information which their readers activate while processing an advertisement. Kisielius and Sternthal (1986) suggest vividness of message information may be one such factor. They use the availability-valence hypothesis to explain the effects of message vividness on attitudinal judgments. According to this hypothesis, judgments regarding an attitude object are a function of the favorableness of the information which is most available during attitudinal judgments. Their work shows that vividness may affect the degree of message-relevant information which is retrieved from memory. When this information is favorable the attitude judgment will be positive. On the other hand, information which does little to create cognitive elaboration (i.e., pallid stimuli) results in idiosyncratic associations with the brand. Such idiosyncratic associations are likely to be less favorable than the associations which reflect message-related content.

Another potential stimulus factor which can influence the availability of information is message framing. Edell and Staelin (1983) define an unframed message as one in which the pictorial content of the message is not related to verbal copy. A framed message is one in which the pictorial information is restated in verbal form. It seems likely that knowledge structures activated by processing framed messages would contain more product-relevant (and less idiosyncratic) associative pathways than those activated by the processing of unframed messages. This would result in a greater degree of accessibility to message-related information previously stored in memory. If the knowledge made available is favorable to the advocacy message (or if it leads to favorable cognitive responses), the resultant judgment would be positive; however, if the knowledge made available is negative (or encourages negative cognitive responses) the evaluation would be unfavorable.

A third factor which can influence the availability of information is the order in which messages are seen. Research by Webb (1979) and Aaker, Stayman, and Hagerty (1986) and Brooker (1981) suggests that the order of advertising exposure impacts message evaluation. Along these lines, the availability-valence hypothesis might be extended to help explain why advertising exposure sequencing differentially impacts judgments. If the knowledge activated by the first ad in a sequence remains available in the consumer's memory (i.e., establishes a "mint set"), then this knowledge may influence judgments about other ads in ensuing exposures. On the other hand, the current exposure in the sequence may have the most effect on consumer evaluation judgments due to its recency in memory. That is, the knowledge available to form judgments is strongly influenced by the information in the current advertisement. This notion is consistent with research on memory schemata which indicates that an activated schema (a mind set) can strongly influence how information is processed (e.g., Bransford and Johnson 1972, Dooling and Lachman 1971).

As has been noted, activated knowledge influences not only judgments, but attitude-behavior consistency (e.g., Kisielius and Roedder 1983). This suggests that understanding attitude-behavior consistency may require investigation of the factors which affect knowledge availability. The purpose of the current paper is to explore some of the factors which influence knowledge availability and their impact on attitude structure and attitude-behavior consistency. This should help to clarify when attitude-behavior consistency may be expected to occur.

METHODOLOGY

The data which are reported here for investigatory purposes were collected as part of a study investigating a related issue the effects of level of expertise and sequence of exposure on evaluative ratings of framed and unframed pictorial print advertisements (Harks, Kamins, and Murphy, 1986). The focus of the current study, however, specifically considers the effects of message framing and sequencing on structure and attitude-behavioral intention consistency.

Subjects

Subjects recruited from a large West Coast city participated in this study. This sample includes a cross section of subjects who represent a full spectrum of experience with the product presented in the ads. In order to obtain a broad range of experience, 116 questionnaires were distributed with self-addressed stamped envelopes. The questionnaires were left with the subjects to be filled out and mailed back later. Of these, 86 were returned and 60 were usable. This group contained 38 males and 22 females.

Stimulus Materials

The advertisements were developed to represent framed and unframed stimuli for computer software. The framed advertisement contained the headline "If you were able to do your work faster... what would you to with your new, free time???" Moreover, the ad contained twelve specific product claims and a small picture of two diskettes labeled with the name of the software (T-1000). The picture presented in the at serves to reinforce the copy claims by linking the text to the visual image. That is, even without the advertising copy the picture indicates that the product is computer software and is called T-1000. The advertisement contained the tag line "T-1000 Software by MYKRO-COMP makes work go quicker. Call 1-800-SOF-WARE for more information."

In contrast the unframed advertisement was designed so that the picture was a source of distraction and not directly linked to any claims made in the ad. Specifically, the ad contained the same headline and tag line as the framed ad (with R-1000 in place of T-1000) but presented a large color picture of a male model engaged in town-hill skiing. Without exposure to the at copy the picture alone would convey no information about the advertised product. None of the twelve specific copy claims made in the framed ad were made in the unframed ad.

Procedures

Data was collected through the use of a cover letter and questionnaire. Each questionnaire contained both the framed and unframed at. However, in half of the questionnaires, the framed ad appeared first whereas the reverse was the case for the remaining questionnaires. The cover letter explicitly instructed subjects to read the first advertisement, and answer questions about it and then read the second at and respond to questions relating to it. This approach toes not directly control the sequencing of exposure to the treatments; however, it seems safe to assume that very few subjects (if any) completed the questionnaire from back to front. Thus, we assume that the subjects viewed the ads in the proper sequence.

In this experiment each type of at can be viewed either with or without prior exposure to the other ad, depending on its position in the sequence. For example, when the unframed at is viewed and rated first (i.e., in the unframed-framed sequence) the measures are independent of exposure to the framed ad. However, when the unframed ad is evaluated in the second position (i.e., in the framed-unframed sequence) evaluations are influenced by the knowledge activated during prior exposure to the framed ad. Therefore, when an ad is viewed first there is no sequence effect the ratings are a control for when it is viewed second. A final section appearing after exposure to both ads involved the collection of data measuring attribute importance judgments and demographic information.

Dependent Variable Measures

Measures of four specific dependent variables were taken (see Appendix). The first variable measured overall global attitude (Ao) on a seven point scale for the product advertised in each of the ads (excellent/poor). The next dependent variable measured the subjects' evaluation of how satisfactorily the two advertised products performed on each of eight product-related attributes (Bi). The third dependent measure was the importance judgment of each of these attributes (Wi). These two measures were later multiplied and then summed in accordance with the adequacy-importance model of attitude (Engel, Blackwell, and Miniard 1986). The final dependent measure related to purchase intention (Pl) which was measured on a seven-point scale (definitely would purchase/definitely would not).

Analysis Technique

A general path analysis model (Duncan 1960, Wright 1960) assuming a weak causal order and closure among the variables was used to model the linkages between the component measures of attitude (SWi x Bi) and the measure of global attitude (SWi x Bi--->Ao) and attitude-behavioral intention consistency (Ao-->PI). Path analysis was applied to the dependent measures under two distinct conditions. The two conditions analyzed related to the sequencing of exposure to the framed and unframed ad. That is, whether the framed or unframed ad was viewed first in the two exposure sequence.

HYPOTHESIS

This study proposes that type of ad (framed vs. unframed) and order of presentation interact effecting both the subjects' attitude consistency toward the advertised products and the strength of the attitude-behavioral intent relationship. When a subject is exposed to the framed-unframed sequence and analysis is done on the product ratings based on the unframed ad, it is hypothesized there will be strong associations between weighted attribute attitude and global attitude as well as a strong causal and direct relationship between each of these measures and purchase intention. The previously discussed extension of the availability-valence hypothesis suggests that this would occur because the processing of the framed at first will activate existing knowledge structures. Once activated, this knowledge would effect the subjects' processing of the information in the unframed ad. The result of this cognitive activity will be reflected in a consistency between attitude measures and ultimately purchase intention.

Subjects exposed to the unframed-framed sequencing and who rate the product after exposure to the unframed ad are hypothesized to have weaker associations between both attitude measures and between the attitude measures and purchase intent. Again, using the extension of the availability-valence hypothesis, exposure to the unframed ad should lead to the activation of relatively few message-relevant associations in memory.

Formally stated, we expect:

H1: When the unframed ad is the basis for product evaluation in the framed-unframed exposure sequence, strong attitude consistency and relatively strong, direct, causal linkage between attitude measures and purchase intention will result.

H2: When the unframed ad is the basis for product evaluation in the unframed-framed exposure sequence, weaker direct attitude consistency and weaker direct causal linkage between attitude measures and purchase intention will be observed relative to that found in H1.

When subjects utilize the framed ad as the basis for product evaluation in either sequence it is hypothesized that activation of knowledge associations will result in strong causal linkages within attitude measures and between each of the attitude measures and purchase intention. When the framed ad appears first, it seems obvious that many message-related pathways will be activated in memory. When the unframed ad is viewed first followed by the framed ad, evaluations of the framed ad are still influenced by the activation of message-related associations. However, prior processing of the unframed ad results in activation of a schema containing relatively few message-related associations and so the linkages within attitude components and between attitude and purchase intention should be weaker than when the framed ad is viewed first. Therefore the following hypotheses are proposed:

H3: When the framed ad is the basis for product evaluation in the framed-unframed sequence relatively strong attitude consistency and relatively strong causal linkage between attitude measures and purchase intention will be observed. All linkages should be positive.

H4: When the framed ad is the basis for product evaluation in the unframed-framed sequence relatively weaker attitude consistency and causal linkage between attitude measures and purchase intention should be found relative to those observed in H3. All linkages should be positive.

RESULTS

The general path analysis models testing the four hypotheses are presented in Figure 1 (a)-(t). The statistics presented represent standardized regression coefficients between the variables in the model.

FIGURE 1

PATH MODELS DESCRIBING RELATIONSHIPS BETWEEN ATTITUDE MEASURES AND PURCHASE INTENT

The data in Figures 1a and 1b show direct support for H1 and H2. When the unframed ad -is the basis for product evaluation, the linkage both within attitudinal measures (SBi x Wi-->Ao) and between each of these measures and purchase intention (SBi x Wi--->PI and Ao--->PI) is stronger for the framed-unframed sequence. That is, for this framed-unframed sequence linkages are significant at an alpha less than or equal to .01 while for the unframed-framed sequence the level of significance drops to .05 for the linkage within attitudinal measures and between Ao and PI. Additionally, the linkage between SBi x Wi and PI is non-significant.

A Chow test (Chow 1960), was applied to determine if the relationship between identical pair-wise constructs (e.g., SBi x Wi--->Ao, for (a) and (b) in Figure 1) differed as a function of sequencing for those subjects who utilized the unframed ad as a rating basis. A single significant result was observed, indicating that the regression model of the relationship between EBi x Wi and PI differed as a function of sequencing (F-2.186, d.f.=25, 33, p<.05). This significant result was determined to be a function of both the difference between beta coefficients (t=1.758, d.f.=56, p<.05) and intercepts (t=10.192, d.f.-57, p<.001). First, this means that attitudinal components were more strongly related to purchase intent when the framed ad was viewed first in a sequence using the unframed ad as a rating basis. Second, it indicates that initial attitudes were higher for those subjects who viewed the unframed ad first and used the unframed ad as a rating basis.

For the unframed-framed sequence approximately 22.6% of the variation in the SBi x Wi-->Ao relationship is explained, while for the framed-unframed sequence 61.12 of the variation in this relationship is accounted for. Likewise, SBi x Wi explains 11.O of the variation in purchase intention (PI) in the unframed-framed sequence while explaining about 652 of the variation in the framed-unframed sequence. Finally, attitude toward the product (Ao) explains 20.7 of the variation in purchase intention (PI) in the unframed-framed sequence while accounting for 52.1 of the variation in the framed-unframed sequence.

Hypotheses 3 is supported by the data in Figure 1c, as when the framed ad is the basis for product evaluation linkages both within attitude measures and between the attitude measures and purchase intention are significant. Hypothesis 4, on the other hand, is not supported. This is because the linkages between the attitude measures and purchase intention are equally significant (alpha less than .01) for each of the sequences represented in Figures 1c and 1d. In addition, the magnitude of the correlation coefficients shows little visible difference between these sequence conditions.

A Chow test applied to determine if the relationship between identical pair-wise constructs differed as a function of sequencing for those subjects who used the framed ad as a rating basis found no significant differences. This provides further evidence that sequencing did not differentially effect the relationship between constructs.

DISCUSSION

The results of the study suggest that the order of sequencing in which the subjects viewed the framed and unframed advertisements did influence the consistency among attitude components and attitude-behavioral intention consistency. However, sequencing and ad format had a greater differential impact on attitude consistency and attitude linkage to behavioral intent when the unframed ad was used as the basis for product evaluation.

In fact, the strongest linkages between attitude measures and between the attitude measures and purchase intent were observed when the unframed ad was both utilized as the basis for product evaluation and appeared second in the exposure sequence. This finding provides strong support for the discussed extension of the availability-valence hypothesis. The framed ad served as a knowledge activation device impacting on the degree of influence the unframed ad had in the formation of product attitude and behavioral intention. When the unframed ad appeared first in the sequence, and was used as a basis for evaluation, attitude toward the product (SBi x Wi) was not significantly linked to purchase intention (PI) because the unframed ad did not cause substantial knowledge activation.

Finally, sequencing had a relatively small effect on the measures when the framed ad was used as a basis for product evaluation. This suggests that viewing the framed ad at either point in the exposure sequence may serve to activate the knowledge necessary to create a high degree of consistency among attitude components and a strong attitude-behavioral intent consistency.

The correlational findings observed in this study must be interpreted with caution. Stronger direct linkages between model components do not necessarily mean that framed ads should precede other advertising appeals in order to obtain more positive product attitudes or higher purchase intention. As reported in prior analysis of this data, initial exposure to a framed ad in an advertising exposure sequence results in lower attitude and product evaluations than initial exposure to an unframed ad (Marks, Kamins, and Murphy 1986). This was evident in the findings of the Chow test for the intercept terms across sequences for subjects who used the unframed ad as a rating basis. Thus, the high correlations reported in the current study could reflect both a less positive attitude and purchase intention. This means that when advertisers use framed appeals, they must take great pains to ensure that their product is presented and perceived in a positive light, since such appeals may well encourage processing of ad content in an active (and perhaps critical) fashion due to the establishment of an active mind set (i.e., the activation of many message-related pathways).

LIMITATIONS AND FUTURE RESEARCH

As was noted, the data presented in this paper was part of a study done to investigate a related area. It would have been desirable to have had multiple measures of overall attitude toward the object and behavioral intention. In addition, given the assumptions made about the activation of knowledge, the measurement of cognitions would have been appropriate (although such measures could well have biased the results if there is little initial knowledge activation encouraged by the ad). Also, the study would have been more realistic if actual behavior was measured. In spite of these problems, the results are interesting and suggest directions for future research.

The results of this study provide a foundation for further investigation of the effects of stimuli and sequencing of advertisement exposure on the activation of knowledge and the effect of knowledge activation on attitude formation and attitude-behavioral intent consistency. Future research will have to consider whether the observed effects hold true for actual behavior, for longer sequences of ads, for media other than print, and for other types of products.

APPENDIX

DEPENDENT MEASURES

REFERENCES

Aaker, Davit A., Douglas M. Stayman, and Michael R. Hagerty (1986), "Warmth in Advertising: Measurement, Impact and Sequence Effects," Journal of Consumer Research, 12 (March), 365-381.

Bransford, John D. and Marcia R. Johnson (1972), "Contextual Prerequisites for Understanding: Some Investigations of Comprehension and Recall," Journal of Verbal Learning and Verbal Behavior, 11(December), 717-726.

Brooker, George (1981), "A Comparison of the Persuasive Effects of Mild Humor and Mild Fear Appeals," Journal of Advertising, 10(4), 29-40.

Chow, Gregory C. (1960), "Tests of Equality Between Sets of Coefficients in Two Linear Regressions," Econometrica 28 (July), 591-605.

Dooling, D. J. and R. Lachman (1971), "Effects of Comprehension on Retention of Prose," Journal of Experimental Psychology, Volume 88, 216-222.

Duncan, Otis D. (1960), "Path Analysis: Sociological Examples," American Journal of Sociology, 72, 1-16.

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

Engel, James F., Roger D. Blackwell, and Paul W. Miniard (1986), Consumer Behavior, Fifth Edition, Chicago: The Dryden Press.

Fazio, Russell H. and Mark P. Zanna (1981), "Direct Experience and Attitude-behavior Consistency," Advances in Experimental Psychology, Volume 14, 161-202.

Kisielius, Jolita and D. L. Roedder (1983), "The Effects of Imagery on Attitude-behavior Consistency," in Advances in Consumer Research, Volume 10, R. Bagozzi and A.M. Tybout (eds.), Ann Arbor, MI: Association for Consumer Research. 72-74.

Kisielius, Jolita and Brian Sternthal (1984), "Detecting and Explaining Vividness Effects in Attitudinal Judgments," Journal of Marketing Research, 21 (February), 54-64.

Marks, Lawrence J., Michael A. Kamins, and Donna Murphy (1986), "The Effects of Level of Expertise on the Processing of Framed and Unframed Pictorial Print Advertisements," in 1986 AMA Educators' Proceedings, Terence A. Shimp, Subhash Sharma, et al. (eds.), in press.

Smith, Robert E. and William R. Swinyard (1983), "Attitude-behavior Consistency: The Impact of Produce Trial Versus Advertising," Journal of Marketing Research, 20(August), 257-267.

Smith, Robert E. and William R. Swinyard (1985), Cognitive Response to Advertising and Trial, Working Paper, Indiana University, Bloomington, Indiana.

Webb, Peter H. (1979), "Consumer Initial Processing in a Difficult Media Environment," Journal of Consumer Research, 6 (December), 225-236.

Wright, S. (1960) "Path Coefficients and Path Regressions: Alternative of Complimentary Concepts?" Biometrics, 16, 189- 202.

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