A Situational Analysis of Communication Effect: a New Product Purchase By Mail Order

ABSTRACT - The impact of message contents and message sources on consumers' post-purchase satisfaction was examined in the present study. Message contents were controlled in terms of the degree of consumer expectation. Message sources were of two kinds, publicity and paid advertisement. Results revealed interesting contrasts of the impact of different message contents and sources.

Citation:

C. Whan Park and William J. Bahr (1980) ,"A Situational Analysis of Communication Effect: a New Product Purchase By Mail Order", in NA - Advances in Consumer Research Volume 07, eds. Jerry C. Olson, Ann Abor, MI : Association for Consumer Research, Pages: 650-654.

Advances in Consumer Research Volume 7, 1980     Pages 650-654

A SITUATIONAL ANALYSIS OF COMMUNICATION EFFECT: A NEW PRODUCT PURCHASE BY MAIL ORDER

C. Whan Park

William J. Bahr

[This research was supported by grants from the University of Kansas, School of Business Research Fund provided by the Fourth National Bank and Trust Company, Wichita, Kansas.]

[C. When Park is Associate Professor at the School of Business, University of Pittsburgh. William J. Bahr graduated from the University of Kansas, School of Business.]

ABSTRACT -

The impact of message contents and message sources on consumers' post-purchase satisfaction was examined in the present study. Message contents were controlled in terms of the degree of consumer expectation. Message sources were of two kinds, publicity and paid advertisement. Results revealed interesting contrasts of the impact of different message contents and sources.

INTRODUCTION

As pointed out by Olson and Dover (1976), much has been written about the importance of consumer expectation, particularly its effect upon consumer satisfaction (Anderson & Hair 1972; Howard & Sheth 1969). However, only a few studies have empirically investigated this topic (Anderson 1973; Cohen & Goldberg 1970; Olshavsky & Miller 1972; Olson & Dover 1976). In reviewing such previous empirical work, the authors of this paper identified two major problems which they felt had tended to inhibit understanding in the consumer expectation area but which they hoped to overcome in their own research discussed below.

The first problem to be resolved was control of subject expectation. Previous manipulations usually involved a written communication about specific characteristics of a product, the implicit assumption being that exposure to information about product attributes would create an expectancy regarding the product. However, expectancy formation is a function of several factors, only one of which is message content. Perhaps the most important factor is that which involves a motivational dimension such as individual need for or interest in the product under investigation, a factor which has been emphasized as critical in influencing cognitive change (Brock et al. 1970; Bruner 1957; Katz 1968; Krech & Crutchfield 1972). The authors believed controls for this factor to be critical in a proper manipulation of expectancy.

The second problem to be resolved was related to the range of the expectation level experimentally manipulated. Previous investigation of the relationship between the level of expectation and the post-purchase product evaluation had been conducted at two different levels of expectation, high and low (Olshavsky & Miller 1972). However, in real life communication situations, consumers, in order to avoid negative consequences, would probably not engage in purchasing behavior were their expectations low. If this is the case, a truly high variance in expectation level, such as high and Low expectation, might not exist among those who engage in product purchasing behavior.

Instead, the initial level of expectation of the purchasers most probably would have a distribution skewed toward a positive level of expectation. This implies that just a slight difference (such as mild and high) in expectation level, controlled only with extreme difficulty in an experimental setting, could be very critical in influencing consumers' post-purchase product evaluation.

To avoid the problems discussed above, the present study was conducted in a natural setting which involved an innovative product and the consumers who purchased and used it. The product was entirely new and had not been introduced to the market prior to conducting the present study. Thus, the consumers' beliefs or expectations about overall product performance were possible only through messages controlled for both content and source. Furthermore, since the respondents participating in the present study were actual purchasers of the new product, it could be assumed that they all had high motivation and that any motivational differences among them were eliminated by excluding moderate-to-low motivated persons.

Theoretical Perspective of the Present Study

The objective of the present study is to develop and test a situational model of communication effect based upon assimilation and contrast theory (Berkowitz 1960; Bruner 1957; Sherif & Hovland 1961). In determining the degree of assimilation or contrast in any communications situation, message source, the authors believe, plays a dominant role (Hovland et al. 1953; Hewgill & Miller 1965; Powell & Miller 1967; Wilding & Bauer 1968).

If a message source is editorial publicity, individuals are thought to attribute the content of the product endorsement to the product (i.e., the product must be accurately described). In other words, due to its high credibility, a publicity message is expected to have great weight in product performance evaluation. Therefore, everything else being equal, the latitude of acceptance (due to a tendency to confirm an initial belief) is expected to be greater than the latitude of rejection. As a result, it is hypothesized that a publicity message creating a high level of expectation would lead to a more favorable product evaluation, due to an individual's confirmation tendency, than would a publicity message creating a low level of expectation.

If a message source is an advertisement, however, individuals are thought to more often attribute the message content to the advertiser's personal disposition (i.e., to sell the product). Thus, due to its low credibility, an advertisement message is expected to have little weight in product performance evaluation. With an advertisement, then, the latitude of rejection (due to a tendency to disconfirm an initial belief) is expected to be greater than the latitude of acceptance. If such is the case, an advertising message with a mild level of expectation may be easier to confirm than one with a high level of expectation. Thus, a high-expectation advertising message may lead to a higher level of disappointment after product usage than would a low-expectation advertising message. A similar suggestion was proposed by Howard and Sheth (1969). The authors of this paper, therefore, hypothesized that of two advertising messages, the one conveying the milder level of expectation would lead to the higher level of post-purchase production evaluation.

METHODOLOGY

The product used in this study was a recently patented haircutter, and there was no variation of the actual quality level of the product. The respondents were those individuals who saw an advertisement or publicity item for this product and purchased the product by mail order. It should be noted that, since the product had not previously been marketed, the respondents had access to information concerning the product only through these publicity and advertising messages.

An address coding scheme allowed the authors to identify the publication in which the customers had seen the ordering information. A file of names, addresses, and publication affinity was thus obtained. Approximately 290 of these purchasers were contacted by the authors within a period of two to four months after the purchase and asked to participate in the study.

The cover letter sent to the prospective respondents by the first author indicated that, as a faculty member, he had been studying the process through which information about new ideas and products was disseminated throughout the population. The respondents were further told that, in appreciation for their cooperation, they would be sent checks for three dollars immediately upon receipt of their completed questionnaires.

A total of 105 questionnaires were returned. Of these, 6 were incomplete and unusable for research purposes. [Two out of these six respondents indicated that they were exposed to both magazine publicity items. Therefore, they were excluded from data analysis.] Information from the 99 usable respondent questionnaires revealed that 35 subjects had seen the publicity message in ISL (Inflation Survival Letter), 34 had seen the publicity message in FE (Free Enterprise), and 30 had seen the advertisement in the Star.

Message Selection

Two publicity items (see Table 1) had been selected from a pool of four different publicity items (from four different magazines) for several major reasons. First, according to the authors' judgment, the two items could be viewed as creating significantly different product expectation levels. Second, the authors judged the two magazines (Free Enterprise and Inflation Survival Letter) to be very similar in terms of editorial subject matter dealing mainly With business, investment, and consumer purchasing aspects in general.

Since the authors could not control the publicity items, the advertisement message was controlled so that it would be as close as possible to one of the publicity messages (Inflation Survival Letter). It must be noted that, since the manufacturer had to deliver his message directly to the public, the exact wording of the advertisement message could not be identical to the publicity message. The advertisement was placed in the Star national weekly tabloid. This advertisement was run only once.

QUESTIONNAIRE DESIGN

The questionnaire sent to the mail order customers con-rained several series of questions. First of all, the respondents were asked to indicate the following: their prior usage experience with other haircutters on a binary scale of yes or no, whether or not they had used the new haircutter at least once before responding to the questionnaire, and the source from which they had obtained information about the haircutter. The publicity respondents were then asked to indicate their confidence in the credibility of the publicity item in their magazine. A 5-point confidence scale was used for this question ("extremely confident" [5] to "not confident at all" [1]). This question was viewed as important in determining any significant difference between the two groups of respondents (FE and ISL groups) with respect to their attitude toward publicity in general.

TABLE 1

PUBLICITY IN TWO MAGAZINES

After the respondents completed the questions mentioned earlier, they were then asked three major questions concerning the message content and the performance of the haircutter. First, the respondents were queried about their expectation levels upon deciding to order the haircutter. A 7-point scale ranging from "very high expectation" (7) to "average" (4) to "very low expectation'' (1) was used. Second, the respondents were asked to indicate how satisfactory a job they thought the haircutter performed. A 7-point scale ranging from "very satisfactory" (7) to "neutral" (4) to "very dissatisfactory" (1) was used. Third, the respondents were asked to indicate how disappointed they were after having used the haircutter. A 7-point scale ranging from "very disappointed" (1) to "average" (4) to "not disappointed at all" (7) was used. The authors had expected that the responses for the level of satisfaction and level of disappointment would be highly correlated, since they seemed to measure a similar dimension. Both were used, however, in order to check for the validity of the two measures.

Validation of Control on Message Content and Source

One limitation of the present study was that the measures of the respondents' expectations before purchase and their product performance evaluations after purchase were obtained at the same time. Ideally, in this mail order case, product expectation measures should have been obtained after message exposure but before purchase. For obvious reasons, however, the authors could not identify nor thereby contact those individuals purchasing the product before their actual purchases were made and their orders were received in the mail. It was at first thought that a post-purchase -- pre-trial expectation measure could be substituted. Due to the small sample size, however, to obtain two different measures at two different points in time from the same respondents through mail questionnaires was deemed to be an extremely difficult, if not impossible, task. [It should also be noted that a two-period survey could have caused respondent sensitization (due to testing effect).] Therefore, as the only viable alternative, the authors decided to use consumer recall of pre-purchase expectations. In addition, student samples were used for the purpose of validating the authors' judgment about message contents and sources.

To validate the authors' judgment about the resemblance of the three message contents in terms of expectation level, three groups of students were individually exposed to one of the three messages in Table 1. After being exposed both to message content and message source, each student was asked to indicate the degree of his expectation level for the product introduced in the particular message, based upon a 7-point scale ranging from "very high expectation" (7), to "average" (4), to "very low expectation" (1). In addition, each student was asked to indicate the degree of general believability of publicity items and advertisements. A 5-point scale ("highly believable" [5] and "not believable at all" [1]) was used for this measure.

Examination of the data obtained revealed that the authors' judgment was consistent with that of the students'. The students indicated that the two publicity messages were different in terms of creating expectation (for FE, mean = 4.18, SE = 0.20; for ISL, mean 3.61, SE = 0.22, t = 1.88, d.f. = 73, P < 0.06). Furthermore, the students viewed the Star advertisement message as significantly different from the FE publicity message (for Star, mean = 3.44, SE = 0.21, t = 2.50, d.f. = 71, P < 0.01), while they did not view the Star advertisement message as significantly different from the ISL publicity message (t = 0.55, d.f. = 68, P < 0.58). [It should be noted that both the FE publicity message and the Star paid advertisement carried pictures of the product while the ISL publicity message did not. To identify its possible impact, a fourth group of students was tested. This group received a message similar to the one received by the ISL student group. The only difference between the two student groups was that, in the case of the fourth group, the picture of the product was included in the message. Data analysis revealed no significant differences between the responses of these two groups (for the group with a picture, mean = 3.61, SE = 0.22, N = 32; for the group without a picture, mean = 3.31, SE = 0.25, N = 36; t = -.0189, d.f. = 66, P < 0.38).] Additional analysis was then performed to ascertain the students' general attitude concerning differences in believability for the two different types of message sources. The data revealed that the students indicated a significantly higher believability for publicity than for advertisement (for publicity, mean 3.08, SE = 0.06; for advertisement, mean = 2.67, SE = 0.05, t = 4.74, d.f. = 138, P < 0.000). This again confirmed the importance of message source influence in the communication process.

RESULTS

Data analysis of the present study involved comparing the three groups of consumer respondents with respect to relevant criterion variables. The respondents in each group were exposed to only one type of message and it was not possible for them to interact among themselves.

The first analysis examined the two groups which were exposed to the two different publicity messages (FE and ISL). The two groups were, first, compared with respect to initial expectation toward product performance.

As revealed in Table 2, the levels of initial expectation for the two groups were different (P < 0.06). This difference was consistent with the authors' judgment and the students' perception about the messages in terms of establishing expectation levels.

Given the difference in the initial level of expectation, the two groups were then compared with respect to degree of satisfaction (as an attitudinal measure) with the product and the level of disappointment with the product. As revealed in Tables 3 and 4, the results showed an interesting difference between the two groups for the two criterion variables. In the case of the degree of satisfaction, the respondents in the FE group demonstrated a significantly higher satisfaction than those in the ISL group (P < 0.001). Likewise, the FE respondents felt less disappointed than the ISL group (P < 0.07).

According to the preceding results, it appeared that message content significantly affected respondent post-purchase product performance evaluation. Furthermore, the results suggested that post-purchase product performance evaluation varied in the same direction as the level of initial expectation created by message content (confirmation based upon the assimilation principle). This result was consistent with a study in which Olshavsky and Miller found that the higher the initial expectation, the more favorable the product performance evaluation.

Some readers may argue that the differences of the two publicity groups with respect to key criterion variables might not be attributable solely to differences in the levels of created expectation but might also be attributable to the unique difference between the two groups with respect to their general attitude toward publicity. Data analysis, however, revealed that the two groups indicated almost identical confidence in the truthfulness of the publicity news. There was no statistical difference (t = 0.32, d.f. = 64, P < 0.75) between the FE group (mean = 3.50, SE = 0.19) and the ISL group (mean = 3.58, SE = 0.19).

The second analysis was concerned with the impact of message source upon the respondents' post-purchase product evaluation, given the similar level of expectation. For this analysis, the ISL group (publicity) and the Star advertisement group were compared. As revealed in Table 2, the two groups indicated very similar expectation levels after being exposed to their respective messages (P < 0.75). Again, this was consistent with the authors' and students' perceptions about the two messages with respect to expectation level.

TABLE 2

T-TESTS WITH RESPECT TO THE LEVEL OF INITIAL EXPECTATION

TABLE 3

T-TESTS WITH RESPECT TO THE DEGREE OF SATISFACTION

The two respondent groups were then compared with each other with regard to degree of satisfaction and level of disappointment. An examination of Tables 3 and 4 showed interesting results. First, with respect to the degree of the respondents' satisfaction, the ISL group was significantly lower than the advertisement group (P < 0.000). Second, in terms of the level of disappointment, the ISL group felt more disappointed than the paid advertisement group (P < 0.000).

The above result suggested that, given the two similar message contents, the message source greatly influenced the respondents' post-purchase product evaluation. More specifically, when the advertisement was moderate (mild) in creating expectation, the respondents felt a significantly higher satisfaction than in the case with a publicity item similar in content. A possible explanation for this result is that in the case of the advertisement, respondents might have disconfirmed their initially hypothesized beliefs (i.e., expectations). Since the product actually performed better than their initial beliefs, the respondents were favorably disposed toward the product's performance. In the case of publicity in ISL group, since they tended to believe the content of the publicity item, the respondents might have confirmed their initial beliefs (mild expectations).

TABLE 4

T-TESTS WITH RESPECT TO THE LEVEL OF DISAPPOINTMENT

The third analysis examined the differences between the FE publicity group and the Star advertisement group. As revealed in Table 2, there was, first, a difference in the initial expectation between the two groups. The group which was exposed to the Free Enterprise message indicated a higher expectation than the group which was exposed to the advertisement message, although the difference did not reach a conventional significance level (P < 0.11). Second, with respect to the degree of satisfaction and disappointment, the two groups showed interesting but unexpected differences. The group which was exposed to the Star advertisement message revealed a significantly higher degree of satisfaction (P < 0.01) than the group which was exposed to the Free Enterprise publicity.

It should be noted that group differences for the key criterion variables might be attributed to unique differences among the three groups of the respondents rather than to any communication effects. However, all respondents who participated in the present study were innovators, according to most definitions of the term and, therefore, might share certain psychological characteristics common to innovators. Furthermore, the three groups of the respondents had similar product usage experiences with other haircutters, which suggested that prior product usage experience (or familiarity with product use) did not play a causal role. Finally, it might be argued that differences in socioeconomic characteristics of the respondents might have caused the group differences in post-purchase product evaluation. However, when partitioned with respect to such key variables as the level of education, income, age and sex, the three groups still revealed the same general pattern of results reported earlier in the study. In summary, all preceding results suggest that the significant differences among the three groups for several criterion variables can be attributed to the message content and message source differences.

DISCUSSION

The results of the present study suggest that when a message source is publicity, there is a strong assimilation effect operating between message content and product performance evaluation, which leads to confirmation of the inferred belief. The more favorable the message content is, the more satisfied the respondent seems to feel about product performance. Furthermore, the results of the present study suggest that, if the message source is advertising and if the message is not creating a high level of expectation, a strong contrast effect takes place, which leads to disconfirmation of the initially hypothesized belief (i.e., in the sense that the product is viewed to perform better than expected).

Advertising Implications and Limitations of the Present study

The main advertising implication of the study's findings is that an advertisement creating a moderate level of consumer expectation would enhance consumers' post-purchase product evaluation. However, an advertisement generating a moderate degree of expectation should not be interpreted as one with negative information but as one with just enough positive information to generate sufficient demand for the products or services.

It must be realized that the present study is concerned only with the respondents' product performance evaluation, given the demand for a product. A certain degree of conflict always exists between the generation of demand based upon message content and the enhancement of post-purchase product satisfaction. This implies that an advertiser walks a tightrope between the increase of demand by generating a high degree of expectation and the increase of product-performance dissatisfaction. It must be realized, too, that the results of the present study by no means suggest that message content and source solely determine the consumer's evaluation of product performance. However, the study suggests that, while product quality determines the general level of product evaluation, message content and source are expected to determine its specific level.

One of the limitations of the study is that it did not directly compare an advertisement creating a moderate degree of expectation with one creating a high degree of expectation. To have created an additional group for exposure to a high expectation advertising message in a national magazine would have been very costly and ethically uncertain. Perhaps this limitation can be overcome in future studies.

Finally, it should be noted that the present study is a quasi-experimental study which always contains alternative explanations of results. Specifically, the non-random and self-selecting assignment of subjects to conditions, unequal response rate over the three groups, and, consequently, possible group dissimilarity in important yet unknown psychological characteristics may have confounded the results of the present study. This also should be examined in future studies.

REFERENCES

Anderson, Ralph E. and Hair, Joseph F. (1972), "Consumerism, Consumer Expectations, and Perceived Product Performance," Proceedings, Third Annual Conference of the Association for Consumer Research, 812-815.

Anderson, R. E. (1973), "Consumer Dissatisfaction: The Effect of Disconfirmed Expectancy on Perceived Product Performance," Journal of Marketing Research, 10, 38-44.

Berkowitz, Leonard (1960), "The Judgmental Process in Personality Functioning," Psychological Review, 16, 130-142.

Brock, T. C., Albert, S. T. and Becker, L. A. (1970), "Familiarity, utility, and Supportiveness as Determinants of Information Receptivity," Journal of Personality and Social Psychology, 14, 292-301.

Bruner, J. S. (1957), "On Perceptual Readiness," Psychological Review, 64, 123-152.

Cohen, J. B. and Goldberg, Marvin E. (1970), "The Dissonance Model in Post-Decision Product Evaluation," Journal of Marketing Research, 7, 315-321.

Hewgill, M. and Miller, G. (1965), "Source Credibility and Responses to Fear-Arousing Communications," Speech Monographs, 32, 95-101.

Hovland, Carl I., Janis, I. L. and Kelly, H. H. (1953), Communication and Persuasion, New Haven: Yale University Press.

Howard, J. A. and Sheth, J. A. (1959), The Theory of Buyer Behavior, New York: John Wiley and Sons.

Katz, Daniel (1968), "Consistency for What? The Functional Approach," in Theories of Cognitive Consistency: A Sourcebook, Abelson, Robert P., et al. (eds.), Chicago: Rand McNally & Co., 179-191.

Kretch, David and Crutchfield, R. S. (1972), "Perceiving the World," in Behavioral Science Foundations of Consumer Behavior, Cohen, Joel B. (ed.), New York: The Free Press.

Olshavsky, Richard W. and Miller, John A. (1972), "Consumer Expectations, Product Performance, and Perceived Quality," Journal of Marketing Research, 9, 19-21.

Olson, Gerry C. and Dover, Philip (1976), "Effects of Expectation Creation and Disconfirmation on Belief Elements of Cognitive Structure," Proceedings, Sixth Annual Conference of the Association for Consumer Research, 168-175.

Powell, F. and Miller, G. (1967), "Social Approval and Disapproval Cues in Anxiety-Arousing Communications," Speech Monographs, 34, 152-159.

Sherif, Muzafer and Hovland, Carl I. (1961), Social Judgment: Assimilation and Contrast Effects on Communication and Attitude Change, New Haven: Yale University Press.

Wilding, John and Bauer, Raymond A. (1968), "Consumer Goals and Reactions to a Communication Source," Journal of Marketing Research, 73-77.

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