Message Characteristics and Audience Characteristics: Predictors of Advertising Response

ABSTRACT - The use of advertising response--the immediate, initial reaction to an advertisement--as a measure of advertising effectiveness is receiving increasing attention from researchers. Interest has been intensified by theories which specify a positive relationship between advertising response and brand attitude. a is paper examines the efficacy of two potential predictors of ad response--namely, the cognitive complexity test which provides a measure of audience characteristics and the cloze procedure which provides a measure of advertising structure. By monitoring reactions to print ads, it is found that both predictors are positively associated with advertising response. Successful predictions are made to a validation sample.

Citation:

George M. Zinkhan and Claude R. Martin. Jr. (1983) ,"Message Characteristics and Audience Characteristics: Predictors of Advertising Response", in NA - Advances in Consumer Research Volume 10, eds. Richard P. Bagozzi and Alice M. Tybout, Ann Abor, MI : Association for Consumer Research, Pages: 27-31.

Advances in Consumer Research Volume 10, 1983      Pages 27-31

MESSAGE CHARACTERISTICS AND AUDIENCE CHARACTERISTICS: PREDICTORS OF ADVERTISING RESPONSE

George M. Zinkhan, University of Houston

Claude R. Martin. Jr., The University of Michigan

ABSTRACT -

The use of advertising response--the immediate, initial reaction to an advertisement--as a measure of advertising effectiveness is receiving increasing attention from researchers. Interest has been intensified by theories which specify a positive relationship between advertising response and brand attitude. a is paper examines the efficacy of two potential predictors of ad response--namely, the cognitive complexity test which provides a measure of audience characteristics and the cloze procedure which provides a measure of advertising structure. By monitoring reactions to print ads, it is found that both predictors are positively associated with advertising response. Successful predictions are made to a validation sample.

INTRODUCTION

Recently, researchers have become interested in assessing the effectiveness of an advertisement by measuring advertising response (Shimp 1981; Bartos 1981). This paper examines two hypothetical predictors of at response that are coincidentally improvements in copy testing techniques. Essentially, message characteristics are assessed through use of the cloze procedure; and audience characteristics are assessed through use of the cognitive complexity test.

ADVERTISING RESPONSE

Advertising response can be broadly defined as the immediate, initial reaction that a receiver experiences upon being exposed to an advertisement. Another term for this phenomenon is attitude toward the ad (Mitchell and Olson 1981). In other words, does the reader or viewer respond favorably to an advertisement at the time of exposure?

Bartos (1981) explains that recent interest in the concept of advertising response is based upon a link between liking an advertisement and forming a good opinion toward the advertised product or service. The suggestion is that this link between liking and good opinion is so strong that liking, or advertising response, can be a major strategic tool in building a brand's personality. The implication is that any advertisement that offends a potential customer could create a serious negative bias toward the advertised product.

Shimp (1981) has clearly delineated between this attitude toward the advertisement (ATTA) and the attitude toward the product or brand (ATTB). The essence of using the attitude toward the advertisement approach is to have the consumer left with a favorable positive attitude after having processed the content. Underlying this is the tacit assumption that consumers are hedonistically motivated by the desire to feel good (Shimp 1981). Also there is Bartos' (1981) conceptualization that ATTA can influence ATTB, especially in terms of an attitude transfer mechanism. Indeed, Morrison and Dainoff (1972) have reached the conclusion that ATTA is strongly related to the effect that advertisers are usually interest in--sales.

The relationship of ad response to the two hypothetical predictors can be illustrated through an examination of two components of the communication process: message and receiver. In order to understand the process of perception, both the message characteristics and the audience characteristics must be taken into account. The latter dimension, receiver complexity, can be tapped by the cognitive complexity test. The complexity of the advertisement, the stimulus object, can be examined by the cloze procedure.

COMPLEXITY OF THE ADVERTISEMENT

Concept

Individuals approach a creative work, including advertisements, with certain expectations. Once individuals begin to experience the stimulating object, these expectations are altered and furthered by clues they detect regarding the structure of the object. Holbrook (1978) has demonstrated the link between expectation and perception or identification. Repeated exposure to this process (expectation-identification, expectation-identification) can lead to enjoyment. As shown by Berlyne (1960), conflict is generated in an individual due to unfulfilled expectation. The decrease in conflict which follows reductions in incongruity is held to be reinforcing. This is basically the application of a theory of consumer aesthetics known as requiredness.

Requiredness applies to aesthetic objects which involve a sequential string of individual components (e.g., music, literature, and dance). These are objects which cannot be perceived all at once in a gestalt-type manner, but instead reveal their segments only one at a tine. Print advertisements can be viewed as a sequential string of words and can also be viewed as aesthetic objects. Aesthetic objects which possess this sequential string ordering may have their form partially predetermined by the style in which the creator has chosen to work (Wallendorf, et al. 1980). After a sequence of words had occurred, the next word is, to a certain degree, predetermined or "required" by the pattern or structure of preceding words. One way that a writer can create structure in a work is by conforming to a predefined or traditional pattern (such as a sonnet or a testimonial or an advertising jingle). Or the writer can structure the work in a new way, all the while leaving cues to guide the audience. To create a successful prose passage, then, the writer must set up a certain structure. The audience, upon discovering this underlying structure, derives enjoyment. a is component of requiredness may be what distinguishes successful or good pieces of writing from writing of lesser qualitY.

As implied above, this principle of requiredness operates with respect to advertising. Indeed, as an individual is able to predict or anticipate the structure of an advertisement (its complexity), the enjoyment or aesthetic appreciation for the advertisement may result. This leads to the first hypothesized predictor of advertising response.

H1: The better a receiver can predict the structure of a particular advertisement, the more the receiver will enjoy the advertisement.

We point out that Morrison and Dainoff (1972) found that arousal properties of an ad include: novelty, complexity, irregularity, incongruity, surprisingness and uncertainty. They stress also that when advertising complexity and advertising enjoyment are plotted on a two-dimensional scale, the result is an inverted U-shaped curve (see Figure 1). In other words, for optimal enjoyment levels, an ad must reach and surpass a certain level of complexity.

FIGURE 1

JUDGMENTS OF PLEASANTNESS AND AD COMPLEXITY

Measurement

The cloze test measures an individual's ability to provide deleted words in a passage of writing. As such, the cloze procedure can be viewed as a measure of verbal uncertainty (Holbrook 1978). The cloze procedure is similar to other measures of readability (such as the Flesch test, Gunning's Fog Index, or the Dale-Chall formula); but it is superior in that cloze scores are the result of a read r interacting with a written passage, rather than representing a score based on the length of sentences or the length of words within sentences.

The cloze procedure measures the likeness between the patterns a writer has used and the patterns the reader is anticipating while reading. The cloze method does not deal directly with specific meaning. Instead, it repeatedly samples the extent of likeness between the language patterns used by the writer to express what he meant and those possibly different patterns which represent readers' guesses at what they think the writer meant (Taylor 1953). The cloze procedure seems particularly well suited to operationalize the theory of requiredness and the readers' "comprehension of advertising structure," as outlined above.

The cloze procedure can be administered as follows. After a random start, every sixth word in the passage is deleted. The mutilated passages are then reproduced with all missing words replaced by standard-size blanks. Subjects are asked to close up the gaps" in the passages by guessing the identities of the missing words and writing their guesses in the corresponding blanks. Each time a subject correctly guesses a missing word he scores one point; an individual's -cloze score for any particular passage is simply the total number of words guessed correctly (Taylor 1956). MacGinitie (1960) and Alderson (1979) have found that the amount of context between cloze gaps does not have any significant effect on the predictability of the deleted word, providing that at least five words of context are available. In other words, it is best to delete every sixth word.

COMPLEXITY OF THE RECEIVER

Concept

Cognitive complexity refers to the number and sophistication of cognitive structures that a given individual may possess. A distinction is made here between cognitive content and cognitive structure. me content of cognition consists of concepts of objects and their attributes (Scott 1962). The structure of cognition refers to the relationships among these objects. Cognitive complexity, therefore, is based not on the content of cognition but rather on the presence of structures for organizing this content.

Berlyne (1960) has shown that for a stimulus object to be enjoyable that object must surpass a certain minimal level of complexity. It makes sense that different individuals would have different complexity thresholds. What is complex for one individual in a particular situation may not necessarily be experienced as a complex stimulus object by a second individual (Morrison and Dainoff 1972). This is where cognitive complexity comes into play as a measure of individual difference. It seems logical that those with complex cognitive structures would require more "complexity" in a stimulus than would those of simpler cognitive structure. Figure 2 represents this notion graphically, where pleasantness ratings are plotted against the complexity of the stimulus object (ad).

FIGURE 2

JUDGMENTS OF PLEASANTNESS AND AD COMPLEXITY

The slope of the cognitively complex person's curve is similar to that of the cognitively simple; but the curve of the former is shifted more to the right. This leads to the second of our hypothetical predictors of ad response:

H2: The higher an individual's level of cognitive complexity, the more that individual wilL enjoy complex ads.

Measurement

Cognitive complexity is operationalized in a manner outlined by Scott (1962) using two stages. First, in free format, respondents list all of the important attributes that they can think of associated with a product. Second, respondents put their attributes into groups that are similar. A respondent can make as many or as few groups as seem appropriate. A measure of cognitive complexity can be obtained according to a formula derived from information theory (Attneave 1959):

EQUATION    (1)  and   (2)

where n is the total number of attributes, n_i is the number that appears in a particular combination of groups, and p_i = n_i/n.

H may be treated as an approximate measure of the dimensional complexity of the cognitive domain referring to a particular class of attributes. It is a purely structural property, because it does not depend on the content of the attributes, but on the relations (similarity or dissimilarity) among them (Scott 1963). H, then, provides a measure of cognitive differentiation.

An additional measure R, or the index of relative entropy, may be used to correct for varying numbers of attributes initially listed by different subjects:

R = H/log₂(n),    (3)

where n is the number of attributes listed by the subject. While H represents the absolute complexity of the subject's category system, R may be interpreted as the complexity relative to the number of attributes to be comprehended. R thus tends to correct downwards the cognitive differentiation scores of subjects who name a large number of attributes, without fully distinguishing among them Scott (1962). Consequently, a subject's R score is used as a measure of cognitive complexity in this investigation.

THE ADVERTISEMENTS

In our study, the stimulus object consisted of two 150 word long print-advertisements for fictitious calculator brands--the Computron R-55 and the El-tronic 38. [The ads were written by professional copywriters at Hanish Associates in Florida. he layout work was done by the advertising department at Jacobson's in Jackson, Michigan; the production work was done by Winkleman's advertising department in Detroit, Michigan.] Since a fictitious brand was used, this constitutes a new product situation. m e actual product information contained in the ads was determined on the basis of two pretests; the two advertised products were designed to be similar to one another in terms of attributes. The two calculator ads were also similar in terms of sentence length and word difficulty. The two ads were designed to differ from one another in terms of structure--one ad being complex in structure and the second ad being more simple in structure.

In order to insure that the two ads are comparable in quality, a panel of four advertising experts rated the ads on multiple-item scales to assess product positioning and copy execution. The two ads do not significantly differ along these dimensions.

SAMPLE

Two-hundred-thirty-eight undergraduate students were selected to take part in the study. All of the subjects were paid $10 for their participation in the survey. Students seem to be appropriate for the purposes of the present study since they are especially interested in calculators as a product class.

The use of student subjects as a sample required the researchers to be especially careful in their research design (Goodwin and Etgar 1980). A problem occurs if students are used in studies of advertised products primarily used and/or purchased by consumers in advanced life-cycle stages. Yet the student sample may be entirely acceptable and consistent with improved external validity if the product classes advertised are salient to them. It is important to employ subjects who are representative of the target for the chosen product Burnkrant (1978). It is our judgment that, with respect to calculators, students do represent such a target audience.

PROCEDURE

Questions were administered in three different sessions. Cognitive complexity was measured in the first session; the cloze procedure was administered during the second session. In the third session, respondents were exposed to for advertisements--two of which were for calculators. After exposure to each of the four ads, the respondents answered questions about the ads designed to assess ad response.

Since there was a chance that the data gathering instruments would affect one another, several checks were performed. To test for ordering effects, half of the respondents were exposed to the Computron ad first and half were exposed to the El-tronic ad first. (one-way analysis of variance--performed on ad response scores-indicates no significant (p < .05) ordering effect for either the Computron response scores (F = 1.029; df: 1, 138) or the El-tronic response scores (F = 1.0238; df: 1, 138).

In order to determine if exposure to the cloze procedure could affect ad response scores, a second group of 85 subjects was exposed to the ads and tested for ad response without taking part in a cloze procedure. Analysis of variance results show no significant (p < .05) difference between these two groups in terms of Computron response scores (F = .963; df: 1, 223) and in terms of El-tronic response scores (F = 1.130; df: 1, 223).

MEASURING ADVERTISING RESPONSE

Aesthetic response to advertising focuses on the emotional component of communications effects. Unlike measurements of learning and comprehension, it does not directly concern itself with the retention of claims, slogans, or other factual material. Aesthetic response has to do with how people feel after seeing an ad rather than what they know. Some ads are intended to be enjoyed; but all ads engender some type of emotional reaction from the audience. Illustrating this measurement procedure is the technique employed in our study.

In the third session, immediately after reading an advertisement, respondents indicated how enjoyable the ads were to read. Four items were used to tap this notion of advertising response: enjoyment, likability, persuasiveness, and interestingness. An eight point Likert-type scale was used, and these items were summed to form a composite index. Thus, one enjoyment score was obtained for each of the two calculator ads investigated. Estimates for coefficient alpha associated with these indices--.91 and .90--were quite substantial.

RESULTS

As discussed earlier, concerning the cloze procedure, it is best to delete every sixth word. With such deletion, there are, in total, 43 blanks which a respondent attempts to "close up' in each of the two calculator ads used here. The Kuder-Richardson Formula 20 (KR-20) can be used to assess the reliability of the cloze procedure. Coefficient alpha for the more simply written ad is .76, while coefficient alpha for the more complex ad is .71. This alpha of .71 is the lowest attained in this study, but it is still well above the lower limit of .50 recommended by Nunnally (1978).

While advertisers are interested in the reading difficulty of their ads, they are also interested in discriminating between an effective message and a less effective message. The objective of a discriminant analysis is to classify people (or objects), by a set of independent variables, into one or more mutually exclusive and exhaustive categories. In this case the dependent variable, advertising response, is dichotomized as follows:

0,   Subject prefers simple ad.

1,   Subject prefers complex ad.

The candidate dependent variables include the subjects' cognitive complexity scores and cloze procedure scores. In this case, discriminant analysis is used to discriminate between those who prefer the Computron ad (simpler in structure) and those who prefer El-tronic ad (more complex in structure). Or, in other words, a discriminant function is formed to differentiate between those who enjoy the Computron ad more than the El-tronic ad from those who enjoy the El-tronic ad more than the Computron ad. A sample of 140 is used to estimate the model; and the remaining 98 respondents serve as a validation sample.

The candidate independent variables include: cognitive complexity scores, cloze procedure scores (complex ad), and cloze procedure scores (simple ad). The results of this discriminant analysis are presented in Table 1. An F test indicates that the assumption of equal covariance matrices is not significantly violated. The Mahalanobis D2 is sizable (.660) and significant beyond the .001 level.

The ad response variable is coded so that a success (1) is equivalent to superior affect for the El-tronic 38 advertisement; a failure (O) is equivalent to superior affect for the Computron R-55 advertisement. The weights and signs of the discriminant function are consistent with the two hypotheses previously developed.

Afifi and Azen (1972) and Rao (1952) outline a way to test the significance of discriminant variables by using a stepwise procedure. At each successive step, the conditional distribution of each variable not entered given the variables entered is considered. m e next variable to enter is the one for which the mean values of the conditional distribution in the two populations are "most different." This difference can be measured by a one-way analysis of variance F statistic. The F statistics for the three independent variables-cognitive complexity, cloze (simple ad), and cloze (complex ad) are displayed in Table 1. Since all three univariate F statistics are significant beyond the .10 level, support is found for both hypotheses.

The directionality of the discriminant weights indicates the following. The more cognitively complex a person is with respect to calculators, the more that person is likely to prefer the complex ad over the more simply written ad. Similarly, the better that a person is able to predict the structure of an advertisement, the more that person tends to prefer that particular ad. In sum, evidence is found which supports the two hypotheses related to advertising response.

TABLE 1

DISCRIMINANT FUNCTION: AD RESPONSE AS THE DEPENDNT VARIABLE

As mentioned previously, a holdout sample of 98 is used to assess predictive validity. Table 1 presents the classification results. 59.2 percent are correctly classified in terms of their relative preference for the Computron vs. the El-tronic advertisements. This is an improvement over the expected proportional chance criteria of 50 percent (Morrison 1969).

DISCUSSION

The results indicate support for the two predictors of t advertising response: the structure of the advertisement itself and the complexity of the receiver. This adds to our understanding of advertising response-theorized recently as an alternative to other measures of advertising effectiveness. With the increasing interest in creating a positive attitude toward the advertisement itself the research here indicates how that response or liking can be enhanced.

For instance, a widespread apothegm regarding advertising copy is "keep it simple!" The assumed premise is that many receivers have such limited vocabularies or low thresholds of boredom that the copywriter must use readily understood words (Anderson and Jolson 1980). But as this research shows, it is not only the words which must be kept simple. The structure of the sentences must also be kept simple. It makes sense that a highly repetitive or predictable ad would be more readily assimilated by the entire target audience than would a less repetitive or predictable ad. This is where the cloze test fits in. The cloze test provides a measure of the complexity, repetitiveness, or predictability of the advertisement. Cognitive complexity, combined with the cloze test, conveys to the manager some idea as to who would be bored by an ad that is too simple or repetitive. In other words, the individual abilities of an ad reader to like advertising copy or structure can be taken into account. Once this process is better understood, it may be possible to improve a person's attitude toward the advertisement itself and thus trigger a positive attitude toward the brand (Shimp 1981) and ultimately influence purchase (Smith and Swinyard 1982).

REFERENCES

Afifi, A. A. and S. P. Azen (1972), Statistical Analysis: A Computer Oriented Approach, New York: Academic Press.

Alderson, J. Charles (1979), "The Effect on the Cloze Test of Changes in Deletion Frequency," Journal of Research in Reading, 2, (No. 2): 108-119.

Anderson, Rolph E. and Marvin A. Jolson (1980), 'Technical Wording in Advertising: Implications for Market Segmentation,' Journal of Marketing, 44, (Winter): 57-66.

Attneave, Fred (1959), Applications of Information Theory to Psychology, New York: Holt-Dryden.

Bartos, Rena (19231), "Ads that Irritate May Erode Trust in Advertised Brands," Harvard Business Review, May-June: 133-140.

Berlyne, D. E. (1960), Conflict Arousal and Curiosity, New York: McGraw-Hill.

Burnkrant, Robert L. (1978), 'Cue Utilization in Product Perception," in H. Keith Hunt, ed., Advances in Consumer Research: 724-728.

Goodwin, Stephen and Michael Etgar (1980), "An Experimental Investigation of Comparative Advertising: Impact of Message Appeal, Information Load, and Utility or Product Class,-- Journal of Marketing Research, 17, (May): 187-202.

MacGinitie, W. 21. (1960), "Contextual Constraint in English Prose," unpublished Ph.D. thesis: Columbia University.

Mitchell, Andrew A. and Jerry C. Olson (1981), "Are Product Attribute Beliefs the Only Mediator of Advertising Effects on Brand Attitude?' Journal of Marketing Research, 18, (August): 318-332.

Morrison, Donald G. (1969), 'On the Interpretation of Discriminant Analysis," Journal of Marketing Research, 6.

Morrison, Bruce John and Marvin J. Dainoff (1972), "Advertisement Complexity and Looking Time," Journal of Marketing Research, 9, (November): 396-400.

Nunnally, Jun C. (1978), Psychometric Theory, New York: McGraw-Hill.

Rao, C. R. (1952), Advanced Statistical Methods in Biometric Research, New York: Wiley.

Scott, William A. (1962), 'Cognitive Complexity and Cognitive Flexibility," Sociometry, 25: 405-414.

Scott, William A. (1963), "Conceptualizing and Measuring Structural Properties of Cognition,' in O. J. Harvey , ed., Motivation and Social Interaction Logritive Determinants, New York: Ronald Press.

Shimp, Terrence (1981), "Attitude Toward the Advertisement as a Mediator of Consumer Brand Choice," Journal of Advertising, 10, (No. 2): 9-15.

Smith, Robert E. and William R. Swinyard (1982), "Information Response Models: An Integrated Approach," Journal or Marketing, 46, (Winter): 81-93.

Taylor, Wilson L. (1953), "Cloze Procedure: A New Tool for Measuring Readability," Journalism (Quarterly, 30, (Fall): 415-433.

Taylor, Wilson L. (1956), "Recent Developments in the Use of 'Cloze Procedure'," Journalism Quarterly, 33, (Winter): 42-48, 99.

Wallendorf, Melanie, George Zinkhan, and Lydia S. Zinkhan (1980), "Cognitive Complexity and Aesthetic Preference," in Morris B. Holbrook and Elizabeth Hirschman, eds., Consumer Esthetics and Symbolic Consumption, Ann Arbor: Association for Consumer Research: 52-59.

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