The Effect of Nutrition Information Format on Cognitive Response, Product Evaluation and Choice

Fredrica Rudell, Iona College
ABSTRACT - A controlled experiment was designed to test the hypothesis that dimension coding (processing information by attribute) promotes greater cognitive response and nutritional evaluation change by consumers than object coding (processing by brand). Although no significant differences in evaluation or choice were found, dimension coders had significantly more expressed reactions to the information, while object coders relied more on prior knowledge and experience.
[ to cite ]:
Fredrica Rudell (1984) ,"The Effect of Nutrition Information Format on Cognitive Response, Product Evaluation and Choice", in NA - Advances in Consumer Research Volume 11, eds. Thomas C. Kinnear, Provo, UT : Association for Consumer Research, Pages: 374-378.

Advances in Consumer Research Volume 11, 1984      Pages 374-378

THE EFFECT OF NUTRITION INFORMATION FORMAT ON COGNITIVE RESPONSE, PRODUCT EVALUATION AND CHOICE

Fredrica Rudell, Iona College

[This research was supported by a grant from The City University of New York PSC-CUNY Research Award Program. The author also wishes to thank three anonymous reviewers for helpful comments.]

[Fredrica Rudell is Assistant Professor of Marketing, Hagan School of Business, Iona College, New Rochelle, New York 10801.]

ABSTRACT -

A controlled experiment was designed to test the hypothesis that dimension coding (processing information by attribute) promotes greater cognitive response and nutritional evaluation change by consumers than object coding (processing by brand). Although no significant differences in evaluation or choice were found, dimension coders had significantly more expressed reactions to the information, while object coders relied more on prior knowledge and experience.

INTRODUCTION

A major concern in the field of consumer behavior is the consumer decision-making process, and the role of information in that process. The growing body of consumer information processing research focuses on consumer behavior related to the search for, initial processing of, and attitude change resulting from use of information. "Detailed analyses of consumer response to information" has been identified as a priority area in research on consumer choice (Bettman 1978).

This paper reports on a study designed to investigate the effects of information format on cognitive response and product evaluation. Nutrition information, important for its public policy implications and itself a subject of much recent research, was chosen as the stimulus. For examples of work on acquisition and effects of nutrition information, see Asam and Bucklin (1973), Brucks, Mitchell and Staelin (1981), French and Barksdale (1974), Jacoby, Chestnut and Silberman (1977), Lenahan et al. (1973), Rudell (1979), Scammon (1977), Stanley (1977), Stokes and Haddock (1972).

INFORMATION FORMAT

Consumer researchers have investigated determinants of information acquisition strategy used by consumers as well as the effects of format on choice. The major strategies are object coding (processing by brand) and dimension coding (processing by attribute). As Bettman (1975) pointed out, the former represents the actual presentation of individual products, such as on a package label, but the latter has been found to facilitate choice among alternatives (e.g., Russo, Krieser and Miyashita 1975). Chestnut and Jacoby (1977) applied the terms "spectator" and "participant" behavior to the alternative acquisition strategies, and tried to relate the consumer's choice of strategy to past purchasing experience (Jacoby, Chestnut and Fisher 1978).

Bettman and Kakkar (1977) rejected the hypothesis that consumers have a preferred strategy for acquisition of information, and showed that task environment constrains the consumer's processing method. Rudell (1979) found that about one-third of subjects who acquired information during three choice tasks seemed to have a "preferred" strategy. Discriminant analysis revealed that consistent dimension coders were better educated, tended to shop more frequently, and considered more sizes of products than subjects who practiced object coding. Capon and Burke (1980) also related strategy to individual differences, finding attribute processing to be more prevalent for mid/hi SES than for low SES subjects. Biehal and Chakravarti (1982) found that information processing activities and memory organization were not strictly bound by information format, but were also influenced by learning goals (e.g., to make a choice vs. to remember facts for later questioning).

Regarding the effect of information format or acquisition strategy on product evaluation, Russo, Krieser and Miyashita (1975) found that alteration of unit pricing displays to facilitate dimension coding resulted in changes in purchase patterns. Using nutritional information, Goodwin and Etgar (1980) found virtually no difference in choice or attitude effects across subjects exposed to three format conditionsCindividual brand, matrix, and joint matrix and brand. However, Rudell (1979) measured product evaluations before and after exposure to nutrition information, and found dimension coding to have a significant influence on nutritional evaluation change.

Several differences in methodology may account for the different results of these two studies. Goodwin and Etgar offered subjects five hypothetical brands of breakfast cereal, while Rudell used three pairs of products (whole and skim milk, white and whole wheat bread, and real and imitation bacon). Goodwin and Etgar assigned subjects to information format conditions, while Rudell recorded subjects' actual strategies. Dependent variables also differed. For instance, Goodwin and Etgar measured accuracy of memory and choice, attitude toward mode of information presentation, and satisfaction with choice, while Rudell measured evaluation and choice change after exposure to information, and number and source of thoughts during decisions. With respect to subjects, Goodwin and Etgar recruited unpaid housewives using stratified random sampling, while Rudell used PTA groups who earned money for participation, though both samples were demographically representative of the population.

Any of these differences might have contributed to the differences in findings. However, what was common to Russo, Krieser and Miyashita and Rudell was the voluntary nature of the format manipulation. In essence, subjects read the information in the manner they found most helpful (though presumably the centralized unit pricing display encouraged dimension coding). Goodwin and Etgar assigned subjects to specific format conditions, regardless or personal preference. In fact, Goodwin and Etgar found higher averages on all response variables and a significantly stronger sense of commitment to their choice among subjects in the third treatment condition who received both types of information displays. This may indicate the importance of allowing consumers to encode and process information in their own way.

Cognitive Response

Much attention has been paid in recent years to the mediating role of cognitive response to persuasive communication (Wright 1980). Greenwald (1968) contended that persisting effects of communication might be explained in large part by the rehearsal and learning of cognitive response to persuasion, rather than learning of the content of the persuasive communication itself. Collection of thought verbalization has been extended to nonpersuasion situations as well, and analyzed for insight into information processing (e.g., Bettman and Park 1980). Even if there are no measurable effects or information format on product evaluation and choice, cognitive response data might reveal differences in thought processes, especially the salience of information and how it is integrated into the decision.

Applying Greenwald's thought categories (externally originated, recipient modified and recipient generated) to thought lists generated during decision tasks, Rudell (1979) found that sources or thoughts seemed to be influenced by information acquisition strategy. Specifically, dimension coding had a larger beta coefficient than object coding in all regression equations describing both the total number of externally originated and the total number of recipient modified thoughts generated during tasks. When reanalyzed at the individual task level, dimension coders had more externally originated and recipient modified thoughts, on average, than object coders (t test significant at .05). It appeared that dimension coding facilitated memorization of and response to information during the decision process. This would be consistent with findings of Russo, Krieser and Miyashita (1975) and Rudell regarding the effects of such a format on evaluation and choice.

HYPOTHESES

There is some evidence that dimension coding of information promotes evaluation change at the individual and aggregate level, and greater use of information as reflected in thoughts. However, a dimension coding strategy is also associated with individual differences such as education, SES and shopping habits which may predispose consumers to more thoughtful choice. While Goodwin and Etgar (1980) did force subjects into specific coding strategies (except for Treatment III, as noted above), no cognitive response data were collected. Therefore, an experiment was designed to test for the effects of information format on decision-related thoughts and product evaluations, while controlling for demographic and other characteristics (through randomization). The hypotheses to be tested were:

1. Subjects receiving nutrition information presented by attribute will have more information-related (externally originated and recipient modified) thoughts than subjects exposed to the same information presented by brand.

2. Nutritional evaluation and choice change will be greater for subjects receiving information in the attribute format than for subjects exposed to the brand format.

METHODOLOGY

Experimental Design

A "before-after with control group" design was used to test the hypotheses. Subjects were randomly assigned to three treatment groups. Experimental Group I received nutritional information in brand (object coded) format, Experimental Group II received the same information in attribute (dimension coded) format, and a control group received no nutrition information. This third condition provided a useful "baseline" of cognitive response and evaluation change.

Experimental Procedure

The questionnaire was administered to groups of 10 to 20 respondents at a time. Each subject was randomly assigned to a treatment (or control) group and was given a loose-leaf binder containing the appropriate questionnaire. Respondents were instructed to read and complete the questionnaire in the order in which it appeared, and cautioned not to turn back to any previous page. After a few questions about cold cereal usage, the subject was asked to imagine that he/she was out of cereal and was given two actual brands to choose from at the store. The two actual brands were chosen for the experiment because of their parity in terms of market share (approximately 42) and nutritional quality.

After rating both cereals on several five-point scales (including nutritional value) and indicating a preliminary choice, subjects in EG I and II turned the page to find nutrition information bound into their questionnaires. (See Table 1 for the actual information provided.) Those in the object coding condition had complete information about one brand, followed by a page of information about the other (similar to the nutrition information panels on cereal packages). Subjects in the dimension coding condition had eight pages--information about both brands on one dimension per page (e.g., protein content for both cereals in grams per serving).

TABLE 1

NUTRITION INFORMATION

The control group had no information, but proceeded directly to the next page of the questionnaire, which asked all respondents to list, "as quickly as possible, any thoughts which come to mind about this decision." No time limit was specified, but subjects seemed to spend just a few minutes on this task, with an average of three thoughts per subject. The author and another coder independently classified the thoughts into three source categories. Agreement was reached on 516 of the 559 thoughts (922), and the remainder were jointly categorized after some discussion. (See Table 2 for examples.)

TABLE 2

EXAMPLES OF THOUGHT CODING

Finally, subjects were asked to re-evaluate the two cereals and indicate a final choice. The rest of the questionnaire consisted of demographic questions, a nutrition quiz. and manipulation checks.

Subjects

165 subjects (55 per group) were recruited from PTA's and clerical staff and adult students of the college. Subjects were offered five dollars to participate in "research on consumer decision-making." 79.7,0 of the sample were students (all living in their own households, not with parents, to ensure some familiarity with shopping tasks).

While approximately one-third of the subjects were male, the modal respondent was a married female between 95 and 99 years of age with some college education who worked 35 hours per week and still did all of the grocery shopping(!). Subjects were randomly assigned to treatment groups, and data analysis revealed no significant differences in demographics, shopping and consumption habits, or prior nutritional knowledge.

Operationalization of Variables

Dependent variables were:

1. Source of thoughts - Externally originated (from nutrition information), recipient modified (illustrations, qualifications, reactions to information), and recipient generated (not referring to nutrition information). Total number of thoughts was also counted.

2. Nutritional evaluation change - Each cereal was rated twice on five-point scale of nutritional value. Absolute value of difference between two ratings was calculated for each cereal, and combined for each subject's nutritional evaluation change score.

3. Choice change - Whether or not brand switching occurred after exposure to information and/or thought listing and re-evaluation.

RESULTS

The first hypothesis concerned the effect of information format on information-related thoughts. Table 3 shows the mean number of each type of thought across treatment groups, and the t-test results for the brand and attribute format conditions.

TABLE 3

MEAN NUMBER OF THOUGHTS

Not surprisingly, the control group receiving no information differed significantly from both treatment groups on almost all thought variables, having no externally originated or recipient modified thoughts, more recipient generated thoughts, and fewer total thoughts. Having no information to recall or respond to, this group brought their own knowledge and experience to bear on the decision, but had less to think about than the groups with information

Comparing the brand and attribute treatment groups, there was no significant difference in externally originated or total thoughts, but the number of recipient modified and recipient generated thoughts were significantly different. Processing by attribute (dimension coding) resulting in only slightly more recall of the information itself, but significantly more response to nutrition information and significantly less reliance on previous knowledge and experience.

The second hypothesis concerned the effect of information format on nutritional evaluation and choice change. Results for the three groups are-shown in Table 4.

TABLE 4

NUTRITIONAL EVALUATION AND CHOICE CHANGE

As expected, the control group receiving no information had significantly less evaluation change (t significant at .0001) and choice change than either experimental group. However, no significant differences were found between object and dimension coders with respect to total nutritional evaluation change (t=.3433, p = .7320) or choice (10 switchers in each group). (The differences in nutritional evaluation change for individual brands were not significant either.) This may be attributable to the brands used in the experiment, which were fairly equivalent in terms of nutritional quality.

DISCUSSION

The findings indicate that dimension coding of nutrition information does not result in significantly greater product evaluation and choice change than object coding. However, as stated above, this may be due to the fact that the brands of cereal offered to subjects were nutritionally comparable. When Russo, Krieser and Miyashita (1975) changed unit pricing format, consumers were able to perceive obvious differences in price and acted accordingly. Even then, consumers exhibited loyalty to certain subcategories of brands (e.g., nationally advertised brands). With cereals, extreme differences in nutritional quality are less likely to exist, individual consumers may differ in their calculations of nutritional benefits (e.g., desiring iron vs. avoiding sodium), and subcategory loyalty may be at work (e.g., "children's" vs. "adult" cereals). It was clearly demonstrated, however, that provision of nutrition information in either format leads to significantly more attitude and behavioral change than the absence of such data.

The current study also confirms Goodwin and Etgar's (1980) findings of no format effects on attitude and choice. Although the two studies differed somewhat in methodology and stimuli (e.g., one time choice vs. measuring change in choice; 5 hypothetical vs. 2 real brands of cereal), both controlled for demographics and other respondent characteristics. The Rudell (1979) study which found format effects on nutritional evaluations used hypothetical choices between pairs of products and, more importantly, allowed subjects to choose their own information acquisition strategies. It would appear that the differences in product evaluations between dimension and object coders owed more to individual differences than to format effects.

Regarding cognitive response to nutrition information, the findings indicate that information format might affect the decision-related thoughts of consumers independently of education, prior knowledge and other individual differences. Specifically, attribute format appeared to prompt more thoughts and greater recall of nutritional facts (as reflected in externally originated thoughts), perhaps by highlighting the differences between brands on dimensions of interest. Significant differences in recipient modified thoughts implied that dimension coding was more likely to prompt subjects' reactions to the information, perhaps because this format emphasized the product comparison task, or perhaps because of its novelty.

The significant difference in recipient generated thoughts implied that object coding may have forced subjects to rely more heavily on past knowledge and experience. Perhaps comparison of nutritional data was more difficult in this format, so subjects "gave up" and chose on other bases, such as previously experienced sogginess. This is consistent with Biehal and Chakravarti's (1982) observation that memory retrieval of product information learned during choice showed higher levels of attribute-based processing. Subjects were in a sense prevented from "doing what comes naturally."

The results of this experiment might also be related to Bettman's (1979) observations about the effects of consumer experience on the form of processing. He points out that consumer research using experienced shoppers and real brand names seemed to encourage processing by brand, while inexperienced subjects choosing from hypothetical products were more likely to employ attribute processing. In the current experiment, equally experienced subjects were forced to process information about real brands by attribute or by brand. It appears that subjects in the brand processing treatment were more likely to fall back on their knowledge and experience than were subjects in the attribute processing mode.

Based on these results, marketers and public policy makers who wish to change consumer decision processes and break old habits might try to encourage attribute processing, which appears to facilitate integration of nutrition information in the decision-related thoughts. However, evaluation and choice change are not assured.

At least three factors will limit the generalizability of the current study. First, although drawn from many age, education and income levels, and both sexes, the subjects were self-selected, paid volunteers, and therefore not fairly representative of the consumer population at large. Secondly, the use of a before-after design raises the possibility of demand bias. Although it is important to measure the marginal effects of information presented in different format, future experiments should control for this by including an after-only treatment group.

Finally, since this was a laboratory experiment, one hesitates to generalize about real-world shopping conditions. Information format may become more important in the distracting environment of a supermarket (as Russo, Krieser and Miyashita found with unit pricing), or it may become totally irrelevant. Future research might test the effects of nutrition information format in the field, as Russo is doing (personal communication).

Additional future research in this area should extend to more types of products, with varying levels of importance. Although nutritional claims are common in cereal and bread ads, these products may not be important enough to consumers to rate serious nutritional comparisons. On the other hand, a new food product (such as textured vegetable protein) may merit investigation, especially since there is less information stored in the consumer's memory. An attempt was made in Rudell (1979) to distinguish three levels of choice, with mixed success. The "EPS" (extensive problem solving) case of real vs. imitation bacon turned out to have the lowest perceived importance, while the "RRB" (routinized response behavior) example of whole vs. skim milk generated most information search and greatest cognitive response. Future studies which overcome the limitations of that research may have different findings.

REFERENCES

Asam, Edward H. and Bucklin, Louis P. (1973), "Nutrition Labeling for Canned Goods: A Study of Consumer Response," Journal of Marketing, 37 (April), 32-37.

Bettman, James R. (1975), "Issues in Designing Consumer Information Environments," Journal of Consumer Research, 9 (December). 169-177.

Bettman, James R. (1978), "Issues in Research on Consumer Choice," in Advances in Consumer Research, Vol. VI, ed. William L. Wilkie, Ann Arbor, MI: Association for Consumer Research, 214-217.

Bettman, James R. (1979), An Information Processing Theory of Consumer Choice, Reading, Mass.: Addison-Wesley Publishing.

Bettman, James R. and Kakkar, Pradeep (1977), "Effects of Information Presentation Format on Consumer Information Acquisition Strategies," Journal of Consumer Research, 3 (March), 233-240.

Bettman, James R. and Park, C. Whan (1980), "Effects of Prior Knowledge and Experience and Phase of the Choice Process on Consumer Decision Processes: A Protocol Analysis," Journal of Consumer Research, 7 (December), 234-248.

Biehal, Gabriel and Chakravarti, Dipankar (1982), "Information-Presentation Format and Learning Goals as Determinants of Consumers' Memory Retrieval and Choice Processes," Journal of Consumer Research, 8 (March), 431-441.

Brucks, Merrie, Mitchell, Andrew A. and Staelin, Richard (1981), "The Effect of Nutritional Information Disclosure in Advertising: An Information Processing Approach," unpublished working paper, Graduate School of Industrial Administration, Carnegie-Mellon University, Pittsburgh,

Capon, Noel and Burke, Marian (1980), "Individual, Product Class, and Task-Related Factors in Consumer Information Processing," Journal of Consumer Research, 7 (December), 314-326.

Chestnut, Robert w. and Jacoby, Jacob (1977), "Consumer information Processing: Emerging Theory and Findings, in Consumer and Industrial Buying Behavior, ed. A.G. Woodside, J.N. Sheth, and P.D. Bennett, New York: North-Holland, 119-133.

French, Warren A. and Barksdale, Hiram C. (1974), "Food Labeling Regulations: Efforts Toward Full Disclosure, Journal of Marketing 38 (July), 14-19.

Goodwin, Stephen A. and Etgar, Michael (1980), "Alternative Organizational Formats for Nutritional Information: Processing and Policy Normative Perspectives," in Marketing in the 80's: Changes and Challenges (1580 Educators Conference Proceedings), ed. Bagozzi, Richard P., et al., Chicago: American Marketing Association, 412-415.

Greenwald, Anthony G. (1968), "Cognitive Learning, Cognitive Response to Persuasion, and Attitude Change," in Psychological Foundations of Attitudes, ed. A.G. Greenwald, T.C. Brock, and T.M. Ostrom, New York: Academic Press, 147-170.

Jacoby, Jacob, Chestnut, Robert W. and Fisher, William A. (1978), "A Behavioral Process Approach to Information Acquisition in Nondurable Purchasing," Journal of Marketing Research 15, 532-544.

Jacoby, Jacob, Chestnut, Robert W. and Silberman, William (1977), "Consumer Use and Comprehension of Nutrition Information," Journal of Consumer Research 4 (September), 119-128.

Lenahan, R.J., Thomas, J.A., Taylor, D.A., Call, D.L. and Padberg, D.I. (1973), "Consumer Reaction to Nutritional Labels on Food Products," Journal of Consumer Affairs 7 (Summer). 1-12.

Rudell, Fredrica (1979), Consumer Food Selection and Nutrition Information, New York: Praeger.

Russo, J. Edward, Krieser, Gene and Miyashita, Sally (1975), "An Effective Display of Unit Price Information," Journal of Marketing, 39 (April), 11-19.

Scammon, Debra L. (1977), "'Information Load' and Consumers," Journal of Consumer Research 4 (December), 148-155.

Stanley, Thomas J. (1977), "Nutritional Information and Preferences for Breakfast Cereals," Journal of Consumer Affairs, 11 (Winter), 121-126.

Stokes, Raymond C. and Haddock, Rafael (1972), "Interim Report of the First Two Phases of the CRI/FDA Nutritional Labeling Research Program" working paper, Washington, D.C.: Consumer Research Institute.

Wright, Peter (1980), "Message-Evoked Thoughts: Persuasion Research Using Thought Verbalizations," Journal of Consumer Research, 7 (September), 151-175.

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