Linear Effects in Cognitive Response to Advertising

Larry Percy, Creamer Inc.
Martin R. Lautman, Associates for Research in Behavior, Inc.
ABSTRACT - This paper presents the results of an exploration of linear effects in cognitive response on cognitive structure. Specifically, elicited cognitive response to advertising stimuli were related to beliefs and purchase intent associated with the advertised product. A strong linear relationship was found between supportive arguing responses and related cognitive structure measures.
[ to cite ]:
Larry Percy and Martin R. Lautman (1981) ,"Linear Effects in Cognitive Response to Advertising", in NA - Advances in Consumer Research Volume 08, eds. Kent B. Monroe, Ann Abor, MI : Association for Consumer Research, Pages: 434-436.

Advances in Consumer Research Volume 8, 1981      Pages 434-436


Larry Percy, Creamer Inc.

Martin R. Lautman, Associates for Research in Behavior, Inc.

[The authors would like to thank Ms. Carol Farquhar, Northwestern University, for her assistance in designing and analyzing this study.]


This paper presents the results of an exploration of linear effects in cognitive response on cognitive structure. Specifically, elicited cognitive response to advertising stimuli were related to beliefs and purchase intent associated with the advertised product. A strong linear relationship was found between supportive arguing responses and related cognitive structure measures.


Interest in the joint cognitive response-cognitive structure model for studying the impact of persuasive communication has grown since its early presentation by Lutz and Swasy (1977). A number of studies have been reported which attempt to link the number of cognitive responses to a communication stimulus (usually in the form of advertising) with a corresponding effect upon elements of cognitive structure. Olson, Toy, and Dover (1978), for example, studying, specifically, five levels of price information, demonstrated a strong relationship between the number of counterarguments and support arguments as related to the amount of information contained in several advertisements. Wilson and Muderrisoglu (1979, 1980) have utilized much the same procedure, applying the model to an analysis of comparative advertising. In each case, strong support for the notion of the number of cognitive responses elicited from an advertising communication stimuli mediating cognitive structure was shown.

All of this research, however, has been concerned with the total or average number of cognitive responses elicited by the communication stimuli. Accepting the basic hypothesis that there is a relationship on the macro level between the number of cognitive responses and the effect upon cognitive structure measures, it is the purpose of this paper to explore beyond the total effect and consider the functional relationship between different magnitudes of cognitive response upon cognitive structure. More specifically, what will be explored is whether respondents exhibiting greater cognitive activity, operationally defined as offering a greater number of cognitive responses, also exhibit a corresponding change in cognitive structure.


Selection of Experimental Stimuli and Respondents

The methodology was designed to test the theory of individual intensity in cognitive response by utilizing both specific communication test stimuli as well as the general experience of respondents. The specific experimental stimuli used consisted of two television commercials for a well-known brand of frozen food, with a high positive image. Each commercial featured a variety of products: one with several people featured; one only a couple. In addition, one group of respondents was not exposed to a commercial and was asked only to think shout advertising for the product with which they were familiar.

To ensure a realistic response, subjects were carefully screened to be certain only those people who belong to the product's target market, and who had experience with the product category, were included. A total of 300 respondents were interviewed (female household heads, 18+, with family incomes over $15,000), 100 in each of three widely distributed geographical markets. One-third of the respondents in each market received each of the treatment conditions.

Respondents in each city were randomly assigned to treatment cells. Those assigned to the first two cells were exposed to one of two television commercials. Each respondent was exposed independently, and asked to watch the commercial very carefully, leaning everything from it that they felt the advertiser wanted them to know. Next, following the general procedure advanced by Greenwald (1968) and discussed by others (cf. Calder, Insko, and Yandell 1974; Wright 1973 and 1974), respondents were asked to relate all of their thoughts while viewing the commercial. Respondents in the third cell were not shown any specific advertising; rather, they were asked for all of the things that come to mind when thinking of looking at or reading advertisements for the product. In all treatment cells thoughts were elicited and recorded one at a time.

After all thoughts were recorded, a series of questions were asked relating to selected cognitive structure measures. Each respondent was asked how likely she .would be to buy the product advertised, measured over a five-point Likert type scale; then how strongly she felt the product reflected specific beliefs associated with the advertising, using a seven point balanced attitude scale.

Analytic Plan

Initially, each of the cognitive responses elicited from a respondent were classified in the manner suggested by Wright (1973). Supportive arguments were classified as those thoughts, beyond a simple copy-point playback, which evidenced a favorable reaction to the product, its use, or the belief messages communicated. Counterarguments were those thoughts that were generally directed against the product or its use, and those that refuted message claims. In addition to these critical responses, three other categories were recorded: positive and negative ad-related thoughts which addressed only the executional elements, and neutral statements. Then, rather than averaging the total number of responses elicited for each classification, those respondents offering only one thought within a classification, two thoughts, etc. were grouped. In this way it is possible to preserve the frequency of individual cognitive responses by type of response.

Each of these distributions was compared over the three treatment cells. It was hypothesized that there should be no significant difference in the general distribution of cognitive response between cells, at least for the important supportive and counterarguments, since the two commercial stimuli used had been pretested and both communicated the same general message beliefs; and, importantly, these were the same message beliefs generally communicated through the advertising for this brand over the last several years (a campaign which included one of the two commercial tested here). As a result, those respondents not directly stimulated by a specific commercial should nonetheless have exhibited the same, or at least similar general cognitive structures, to those exposed to the commercials.

Incidence of responses were then compared with cognitive structure measures of belief and intention to buy in order to establish whether a linear relationship might mediate cognitive structure.


Table 1 shows the mean number of cognitive responses offered by classification for each treatment cell studied. The high number of supportive arguments vs. other types of cognitive responses is, no doubt, a function of the existing highly positive image of the product.



It is generally accepted (cf. Slash, Toy, and Dover 1978) that positive and negative ad-related responses, as well as neutral responses, do not mediate cognitive structure. For this reason, along with the fact that very few respondents provided positive or negative ad-related responses, only supportive arguments and counterarguments will be analyzed. Furthermore, because of the very low total number of counterarguments elicited, and the fact that 34 of the 41 respondents offering a counterargument gave only one, the statistical analysis will further be restricted to only the supportive arguments.

The distribution of supportive arguments between treatment cells is shown in Table 2. Little difference in either the overall distribution or average number of arguments elicited appears to be in evidence. A chi-square analysis of the distribution indicated no significant difference between the distributions (X2 = 4.90, d.f. = 8, p > .05).



These results would suggest that there is no significant difference in the general difference in the general distribution of cognitive responses (or at least, in this case, supportive arguments) among treatment cells. Therefore, the subsequent analysis permits us to look at total response over cells where appropriate.

Belief measures for six constructs reflected in the advertising for this product were measured via a seven point attitude differential scale. Mean beliefs averaged over the six constructs for respondents offering one through five or more supportive argents are shown in Table 3. The one inversion, between three and four supportive arguments was primarily the result of one belief construct--the only one showing low overall belief.



A 3 x 5 ANOVA of this attitudinal data (see Table 4) indicated only Support Arguments to be statistically significant--this the result of a significantly increasing linear trend as shown in Table 3. As the number of support arguments offered by a respondent increased so did his positive attitudinal evaluation of the product.



Finally, the intention to purchase also appears to be strongly mediated by the number of support arguments offered by a subject. Intention did increase linearly, as shown in Table 5.



An ANOVA of this data indicated that there was a significant difference of this particular measure between treatment cells, this was occasioned by a mean purchase intent of 3.84 in Cell 1 (those who saw one of the commercials) vs. 4.04 and 4.06 (among those seeing the second commercial and those seeing no commercial, respectively). One commercial was less effective than either the other commercial or the control in generating purchase interest. However, there was no interaction between this difference and the number of support arguments elicited; and, there was a significant linear trend in support arguments, as indicated in Table 6. As the number of support arguments offered by an individual respondent increased so did her interest in buying the product.




The results here indicate that there is a strong reason to suspect a linear effect mediating the relationship between cognitive response and cognitive structure. Overall belief increased with the number of support arguments offered, as did intention to purchase. Because of the high positive image nature of the product involved, one is tempted to believe these results are even more significant. Since both belief and interest are already high, to stimulate even greater interest as a function of advertising based cognitive response is surprising. Unfortunately, owing to that same high positive image, linear effects among counterarguments could not be tested because of sample size limitations. However, the trends, at least for purchase intent, suggested the predicted linear functional relationship. That is, more counterarguments resulted in lover purchase interest.

It is reasonable to predict that a similar linear relationship might be observed for products with less positive images. Here, however, one is equally likely to suspect that the quality of support arguments, not only their existence and frequency, might be of significance.

The elicitation measure used in this study differs considerably from that commonly employed in pretesting and on-air testing of advertising. Most typically, tabulations are obtained of the percentage of respondents recalling various messages. Occasionally, responses are tabulated by type of response (ad-related, sales point related, etc.). Rarely, if ever, is a coding scheme employing cognitive response categories used. Yet, the data presented here strongly suggest that the mean number of responses within cognitive response categories (a "recall-type" measure) may be a significant correlate of a motivational construct -- purchase interest. Whether these responses must be precursors of shifts in purchase interest, however, remains a critical question.

The findings of this study suggest that segmentation of groups by number of responses might reveal if there is a critical number of support arguments correlated with purchase interest; and/or, if any specific arguments or combinations of support arguments are integral to promoting purchase interest, given that the critical number of responses have been elicited. Additional commercial tests (with larger samples) would be necessary to explore these questions.


Calder, Bobby J., Insko, Chester A., and Yandell, Ben. (1975), "The Relation of Cognitive and Memorial Processes to Persuasion in a Simulated Jury Trial," Journal of Applied Social Psychology, 4, 62-93.

Greenwald, A. G. (1968), "Cognitive Learning, Cognitive Responses to Persuasion, and Attitude Change," in A. G. Greenwald, T. C. Brock, and T. M. Ostrome (Eds.), Psychological Foundations of Attitude, New York: Academic Press.

Lutz, Richard J. and Swasy, John L. (1977), "Integrating Cognitive Structure and Cognitive Response Approaches to Monitoring Communications Effects," in W. Perreault, Jr. (Ed.), Advances in Consumer Research Vol. IV, 363-371.

Olson, Jerry C., Toy, Daniel R., and Dover, Philip A.(1978), "Mediating Effects of Cognitive Responses to Advertising on Cognitive Structure," in H. Keith Hunt (Ed.), Advances in Consumer Research Vol. V, Ann Arbor, Michigan: Association for Consumer Research.

Toy, Daniel R., Dover, Philip A. and Olson, Jerry C.(1977), "Communication Effects on Cognitive Responses, Cognitive Structure, and Attitude," Paper presented at the American Psychological Association 85th Annual Convention, San Francisco.

Wilson, R. Dale, and Muderrisoglu, Aydin (1979), "An Analysis of Cognitive Responses to Comparative Advertising," Pennsylvania State University Working Series in Marketing Research, Number 84.

Wilson, R. Dale and Muderrisoglu, Aydin (1980), "An Analysis of Cognitive Responses to Comparative Advertising," in Jerry C. Olson (Ed.), Advances in Consumer Research Vol. VII, Ann Arbor, Michigan: Association for Consumer Research.

Wright, Peter L. (1973), "The Cognitive Process Mediating Acceptance of Advertising," Journal of Marketing Research, 10, 53-62.

Wright, Peter L. (1974), "On the Direct Monitoring of Cognitive Response to Advertising," in G. C. Hughes and M. L. Ray (Eds.), Buyer/Consumer Information Processing, Chapel Hill: University of North Carolina Press, 200-248.