The Relationship of Product Type, Preferred Evaluative Criteria, and the Order of Consumption to the Evaluation of Consumer Products

John N. Hallaq, University of Idaho
Kathy Pettit, Washington State University
ABSTRACT - Consumers tasted and evaluated two unidentified grocery products and the order of consumption was orthogonally manipulated. Various hypotheses were tested and the results indicate that preferred evaluative criteria are significantly dependent on some demographic characteristics of respondents. MANOCOVA, in which certain demographic characteristics were employed as covariates, showed that product ratings were significantly different when the ratings for a product were made in the company of particular other products (showing complementarity) as well as when the order of consumption was altered. However, the ratings did not differ significantly when the product was consumed with another brand of the same product class as opposed to being with a product of a different product class. Promotional implications are discussed.
[ to cite ]:
John N. Hallaq and Kathy Pettit (1983) ,"The Relationship of Product Type, Preferred Evaluative Criteria, and the Order of Consumption to the Evaluation of Consumer Products", in NA - Advances in Consumer Research Volume 10, eds. Richard P. Bagozzi and Alice M. Tybout, Ann Abor, MI : Association for Consumer Research, Pages: 600-604.

Advances in Consumer Research Volume 10, 1983      Pages 600-604

THE RELATIONSHIP OF PRODUCT TYPE, PREFERRED EVALUATIVE CRITERIA, AND THE ORDER OF CONSUMPTION TO THE EVALUATION OF CONSUMER PRODUCTS

John N. Hallaq, University of Idaho

Kathy Pettit, Washington State University

ABSTRACT -

Consumers tasted and evaluated two unidentified grocery products and the order of consumption was orthogonally manipulated. Various hypotheses were tested and the results indicate that preferred evaluative criteria are significantly dependent on some demographic characteristics of respondents. MANOCOVA, in which certain demographic characteristics were employed as covariates, showed that product ratings were significantly different when the ratings for a product were made in the company of particular other products (showing complementarity) as well as when the order of consumption was altered. However, the ratings did not differ significantly when the product was consumed with another brand of the same product class as opposed to being with a product of a different product class. Promotional implications are discussed.

INTRODUCTION

The consumer's evaluation of the product after consumption is an important determinant of the success of that product in the marketplace. Accumulated evidence indicates that while the consumer can evaluate one brand more favorably than another when they are identified, he/she frequently cannot tell the difference between those two brands if they are presented without identifying marks (Allison and Uhl, 1964; Bowles and Pronko, 1948; Pronko and Bowles, 1948, 1979; Pronko and Herman, 1950; Prothro, 1953). However, other studies indicate that the consumer is capable of correctly identifying two brands of soft drink beverage when told in advance which brands will be included (Lane, Zychowski and Lelii, 1975; Thumin, 1962, 1972). The above evidence indicates that frequently physical attributes like sweetness, bitterness, and aroma are not as crucial in determining evaluations as brand name and image associated with the products.

It is also known that in paired-comparison product tests the order of presentation of the products may affect their evaluations and that the order of presentation may interact with the type of the product (Daniels and Lawford, 1974). Consequently, the order of presentation and the type of the produce must be controlled in any experimental research.

The purpose of the present study is to vary the order of presentation and the type of product orthogonally and examine whether the consumer changes his/her ratings of the product as a result of this manipulation. The study accomplished this through examining the various hypotheses presented below.

.&n important variable is subject age; some consumer products are heavily used and preferred by certain age groups. When two different products are consumed at the same time, it is reasonable to expect that their evaluation may depend upon the age group to which the consumer belongs and a number of other consumer characteristics. Such demographic characteristics as the levels of education and income may affect preferences for different types of products. For instance, blue collar workers may prefer beer to cocktails.

In addition, different products and brands may be preferred because of differentially preferred evaluative criteria. Thus, different brands of the same product class may be rated favorably due to this variation in preferred evaluative criteria among consumers. The above reasoning leads to the first hypothesis, stated in the null form:

Hypothesis 1: Preferred product evaluative criterion (criteria) is (are) independent of the demographic characteristics of the respondent.

The type of products one consumes together, temporally, may affect the evaluation of the products. That is, certain products may be mutually more compatible than others, and thus, they tend to complement each other in consumption. Consequently, the product evaluated, as well as the accompanying product tasted, are expected to interactively affect product evaluations. This can be summarized in the following hypothesis, stated in the null form:

Hypothesis 2: The type of product evaluated, the type of accompanying product with which it is tasted, and the two interactively, will not affect ratings of the primary product tested.

Furthermore, the order of tasting may also affect the ratings of the product. The following null hypothesis will be tested to examine this possibility:

Hypothesis 3: The type of product and the order in which it is tasted relative to an accompanying product will not interactively affect the rating of the product.

While in many paired-comparison experiments researchers typically present two brands of the same product class (e.g., soup, cola drink, or beer), the consumer in real-life usually tastes one brand of one product class (e.g., beer) and one brand of another product class (e.g., potato chips). In order to simulate this realistic consumption situation and to increase the external validity of this study, the present experiment had some subjects consume and evaluate two different products at the same time. Other subjects sampled and rated two different brands of the same product. A final hypothesis, stated in the null form, will be investigated:

Hypothesis 4: The ratings of a product when tasted with another product do not differ from those of the first product when tasted with another brand of the some product.

METHODOLOGY

Two brands of three different grocery products, orange Juice, hot dogs, and potato chips, were selected for the tasting experiment. The choice was made to maximize similarity to previous research and because brands of these products are difficult to distinguish from each other without prior identification. They are also frequently purchased so that most consumers have a high level of familiarity with them and with the evaluative criteria on which to judge them. Lastly, it was assumed, a priori, that certain of these products were mutually more compatible and complementary than others.

Arrangements were mate with one of the national supermarket chains to permit interviewers (students) to conduct the tasting experiments inside the supermarket. Sessions were scheduled every day of the week for one week. Each pair of experimenter-interviewers spent about two hours and had between 20 to 30 shoppers taste the food. Although some sessions were held during the morning hours, most of them were scheduled during the afternoon hours when shopping traffic is heavier in order to gather data from a larger number of shoppers and increase external validity of the study. The experimenter-interviewers administered the taste experiment using two products simultaneously. They either used two brands of the same product (e.g., two brands of orange juice), or two different products (e.g., a brand of orange juice and a brand of hot dogs). They rotated the order of tasting so that each product had approximately an equal number of consumers who tasted it first or second. The brands of orange juice were mixed at home to the concentration recommended on the container (equivalent for both brands), carried in unidentified large containers, and served in small paper cups. The hot dogs were cooked at home, cut into four pieces each, and kept warm at the supermarket. The potato chips were brought to the supermarket in unidentified containers.

Before a consumer was allowed to taste a product, he/she was asked to fill out the questionnaire which requested information concerning his/her brand preferences, preferred evaluative criteria on which the preferences were based, and demographic characteristics. Over four-hundred consumers completed tastings.

Since a convenience sampling procedure was followed, care must be exercised in generalizing the findings of this study to other consumers. Alternatively, one could have randomly selected about 400 households, visited each, and conducted the tasting experiment. However, the cost of this procedure was considered too prohibitive. Furthermore, the controls over the tasting procedure in the supermarket would have been difficult to reproduce in separate households, so the internal validity of this study would have suffered if the latter procedure had been employed.

RESULTS

The experiment involved tasting two brands of the same product or two different products. Following tasting, each product was evaluated on a seven-point scale with a number of criteria. A score of one indicated the least favorable evaluation on a.criterion and a score of seven the most favorable. The following symbols were used to identify the different products:

Orange Juice: OJA and OJB,

Hot Dogs: HDA and HDB, and

Potato Chips: PCA and PCB.

Below is a brief summary of information concerning characteristics of the consumers who participated in the study. The experiment was conducted in a "college town" where most of the students resided in apartments outside the university. As a rule, the participants did the majority of the shopping for their households, were almost equally divided by sex, and a little less than one-half of them were non-students. The respondents had an average of two years of college, about one-half were married, and the average family size was about two and oneChalf individuals. usually with at least one employed.

The average age was 34 years and the average 1976 income before taxes was $8,773.

Although not formalized as a hypothesis, our results support earlier findings by other researchers which indicate that consumers were unable to differentiate between different brands of the same product class. Both brands of each product were preferred by some consumers.

Table 1 provides the data necessary to test the first hypothesis, showing information gathered on preferred criteria for purchase of the three types of products used in the experiment. Although respondents expressed a variety of reasons for brand selection (including multiple reasons), the dominant factors were price, taste, and brand availability. Reasons for brand selection were cross-classified with age, sex, and income of the consumers, using the Chi-square statistic.

TABLE 1

PREFERRED EVALUATIVE CRITERIA FOR THE MOST FREQUENTLY PURCHASED BRAND OF EACH PRODUCT ACCORDING TO DEMOGRAPHICS (CROSSTABULATION)

The results in Table 1 indicate that preference criteria for purchase of brands of the three products used in the experiment differed among different consumer groups. There were two exceptions: (1) No significant differences were observed between the high and low income groups on any of the three products. This may be due to the fact that these three products are all convenience goods and constitute a small portion of income. (2) No significant differences were observed among respondents of different ages, levels of education, or income concerning the purchase of potato chips. One possible explanation for this lack of difference is the fact that this product may not be purchased as frequently as the other two and may be less important to the sampled consumers. Table 1 shows that no criterion seems to dominate strongly in the purchase of orange juice among both the younger and older participants. However, in the purchase of hot dogs and potato chips, the taste criterion tends to be dominant. If all three products are combined, taste still dominates. The dominance of taste as the primary evaluative criterion is similar for both sexes. Thus, one can conclude with high confidence that taste is a very critical consideration in the case of hoc dogs and potato chips, a definite advantage for promotional purposes. Conversely, a variety of reasons seem nearly equally important to consumers of orange juice, which can make the task of the marketer more difficult for reaching target markets with optimal appeals.

The analysis in Table 1 was based on the use of the "most frequently purchased" brands within the three product categories. Our evidence indicates that they can be used as "proxies" for "preferred brands" with high reliability for drawing conclusions. This statement is based on the fact that when the "most frequently purchased" brands were cross-classified with the "preferred brand," an overlap of 86%, 92% and 93% was obtained for orange juice, hot dogs, and potato chips, respectively. In other words, for 86% of buyers of orange juice, the Preferred brand and the most frequently purchased brand were the same.

In order to test Hypotheses 2 and 3, a MANOCOVA was performed in-which age, sex, and income were employed as covariates. Multivariate results for the series of dependent measures were obtained, as well as univariate results for ratings of product color, taste, and overall quality.

The results of the analysis can be found in Table 2. Specifically, a 3 x 3 x 2 factorial design was used in which the effects of type of product (product), type of accompanying product (other), and order of tasting the product (order) were employed as independent measures. Equivalent numbers of observations of the two different brands of each product were randomly selected and assigned to treatment conditions; the effective sample size was reduced from 405 to 396.

Hypothesis 2 can be addressed by examining the main effects of "product" and "other" (accompanying product), as well as the product by o her interaction. For the series of evaLuations, the main effect of product was highly statistically significant (F = 4.74, p < .001). Additionally, the main effect of product was significant concerning the ratings of color (F = 10.31, p < .001), taste (F = 9.80, p < .10) and overall quality (F = 5.23, p < .01). Thus, mean ratings were significantly different among the products.

The main effect of other (accompanying product) was also significant in the multivariate case (F = 2.62, p < .05). It was statistically significant in univariate tests of ratings of color (F = 4.57, p < .05) and overall quality (F = 6.02, p < .01), as well. Such results imply that the type of product consumed in conjunction with another product can affect ratings of that product.

Most interesting, however, was the emergence of a highly significant "product-other" interaction. Such a finding obtained concerning the series of product judgments (F = 2.47, p < .01), as well as for each of the individual dependent measures. Specifically, the product-other interaction significantly affected ratings of color (F = 6.00, p < .001), taste (F = 3.38, p < .01), and overall quality (F = 4.65, p .001). The product-other interaction regarding ratings of overall quality is presented in Figure 1. Somewhat surprisingly, the lowest obtained ratings were for hot dogs when tasted with potato chips--a rather common combination. Thus, the data support rejection of Hypothesis 9 and indicate that certain products are complemented by others and are rated more positively when tasted together, while other products together tend to detract from the favorableness of the evaluation.

FIGURE 1

INTERACTIVE EFFECT OF TYPE OF PRODUCT AND ACCOMPANYING PRODUCT TASTED (OTHER) ON EVALUATIONS OF OVERALL PRODUCT QUALITY

Hypothesis 3 stated that the type of product and the order in which it was tasted relative to an accompanying product will not interactively affect product ratings. As can be seen from Table 2, the multivariate result of the product-order interaction concerning the series of product ratings was statistically significant (F = 2.40, p < .05). Furthermore, ratings of taste (F = 2.33, p < .10) and overall product quality (F = 5.67, p < .01) were significantly affected by product and order interactively. Figure 2 presents the product-order interaction concerning ratings of overall product quality. Order of tasting made little difference for both orange juice and hot dogs. However, potato chips were judged as higher in overall quality when tasted first, rather than second. Thus, one must reject this hypothesis and conclude that product type and order of casting interactively affect product evaluations.

The final hypothesis tested concerned whether ratings of products would differ when two brands of the same product were tasted together, as compared to casting two different types of products. To do this, three difference-of-two means tests were individually performed on the scores for the three dependent measures: color, caste, and overall quality. Each test included testing the mean scores on the individual dependent measures for cases where products were the same and where products were not the same. The t-values for color, taste, and overall quality were 0.78, 0.50, and 0.84, respectively. Thus, the data do not support rejection of the null hypothesis.

TABLE 2

THE EFFECTS OF TYPE OF PRODUCT, TYPE OF ACCOMPANYING PRODUCT, AND CONSUMPTION ORDER ON PRODUCT EVALUATIONS: MULTIVARIATE AND UNIVARIATE RESULTS

FIGURE 2

INTERACTIVE EFFECT OF TYPE OF PRODUCT AND ORDER OF TASTING ON EVALUATIONS OF OVERALL PRODUCT QUALITY

CONCLUSIONS AND IMPLICATIONS

Results of the current study showed that preferred product evaluative criteria varied among products, as well as among consumers with differing demographic characteristics. In testing the effects of type of product, type of accompanying product, and order of tasting on product evaluation, demographic characteristics were used as covariates.

The current findings supported the notion that the type of product tasted and the type of accompanying product do, interactively, affect ratings of that product. Most surprisingly, the lowest product evaluations recorded, in terms of ratings of color appeal, taste, and overall quality, were those of hot dogs when accompanied by potato chips. Although it was believed, a priori, that complementarity of products might affect evaluations, the combination in question was not expected to be rated unfavorably.

Thus, it may be important to the marketing practitioner to perform research as to which products tend to complement his/her own most favorably in an actual consumption situation. Future research should seek to determine expected desirability of particular combinations, as well as actual ratings following tasting. That is, in order to design advertising which is effective in stimulating product trial, it is the expected appeal or complementarity of the products which is important. However, to facilitate repurchasing, the complementarity or the products in actual consumption may be paramount.

Although the results of the current study can not be generalized to other products and consumers, they nonetheless suggest that some promotional advantages may be gained by the manufacturer who offers cooperatively sponsored premiums with another manufacturer whose products complement his/her own in consumption.

Further research is needed to suggest a more general theoretical model of complementarity among grocery products. The currently employed products varied as to color, sweetness, saltiness, and consistency. Future studies should perhaps vary these characteristics more systematically in order to study the effects of each. Additionally, as suggested by the current study, order of consumption should probably be systematically varied along with the previously mentioned attributes.

REFERENCES

Allison, R.I. and Uhe, K.P., "Influence of Beer Brand Identification on Taste Perception," Journal of Marketing Research, August 1964, Vol. I, No. 3, 36-39.

Bowles, J.W., Jr. and Pronko, N.H., "Identification of Cola Beverages: II A Further Study," Journal of Applied Psychology, 1948, 32, 559-564.

Daniels, P. and Lawford, J., "The Effect of Order in the Presentation of Samples in Paired Comparison Product Tests," Journal of Market Research Society, April 1974. Vol. 16. No. 2. 127-133.

Lane, S.H.1 Zychowski, J. and Lelie, K., "Cola and Diet Cola Identification and Level of Cola Consumption," Journal of Applied Psychology, 1975, Vol. 60, No. 2, 278-279.

Pronko, N.H. and Bowles, J.W., Jr., "Identification of Cola Beverages: I First Study," Journal of Applied Psychology, 1948, 32, 304-312.

Pronko, N.H. and Bowles, J.W., Jr., "Identification of Cola Beverages: III A Final Study," Journal of Applied Psychology, 1949, 33, 605-608.

Pronko, N.H. and Herman, D.T., "Identification of Cola Beverages: IV Postscript," Journal of Applied Psychology, 1950, 34, 68-69.

Pronko, N.H., "Identification of Cola Beverages Overseas," Journal of Applied Psychology, 1953, 37, 494-495.

Thumin, F.J., "Identification of Cola Beverages," Journal of Applied Psychology, 1962, 46, 358-360.

Thumin, F.J., "Application of the Paired-Comparison Method to New Product Development," The Proceedings of the Annual Convention of the American Psychological Association, 1972, Vol. 7, Pt. 2, 735-736.

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