Evaluation of Food Labeling Policies Through Measurement of Consumer Utility

ABSTRACT - Food labeling continues to be a controversial topic among consumers, nutritionists, manufacturers, educators, and agencies of the government. The difficulties that have been experienced in past in terms of food labeling awareness, usage, and understanding as well as the consumers right to know more about the composition of food products will effect future food labeling policy decisions. The method of conjoint analysis presented in this paper offers a way of evaluating consumer preferences for existing and alternative labeling policies and method of inferring the degree of satisfaction that would be derived under alternative policies by any given segment of consumers as well as by the market in general.


Roger Best and Jim McCullough (1978) ,"Evaluation of Food Labeling Policies Through Measurement of Consumer Utility", in NA - Advances in Consumer Research Volume 05, eds. Kent Hunt, Ann Abor, MI : Association for Consumer Research, Pages: 213-219.

Advances in Consumer Research Volume 5, 1978      Pages 213-219


Roger Best, University of Arizona

Jim McCullough, University of Arizona


Food labeling continues to be a controversial topic among consumers, nutritionists, manufacturers, educators, and agencies of the government. The difficulties that have been experienced in past in terms of food labeling awareness, usage, and understanding as well as the consumers right to know more about the composition of food products will effect future food labeling policy decisions. The method of conjoint analysis presented in this paper offers a way of evaluating consumer preferences for existing and alternative labeling policies and method of inferring the degree of satisfaction that would be derived under alternative policies by any given segment of consumers as well as by the market in general.


Underlying public policy decisions concerning food product labeling is the assumption that greater quantities of more detailed information will permit consumers to make improved brand choice decisions (French and Barks-dale, 1974). Motivation for better food labeling was outlined at the 1969 White House Conference on Food, Nutrition, and Health, and the United States Food and Drug Administration (FDA) has subsequently required additional label information. Current policy requires one of the following formats for food products.

1. Ingredient and Nutrition Labeling: For products which have been fortified with additional nutrients and/ or food products making a nutritional claim.

2. Ingredient Labeling: For all other food products. (Nutrition labeling of these products is voluntary but if disclosed must follow the format prescribed by the FDA).

Under this policy labeling formats have had limited success. Consumer understanding and usage of label information has been disappointing. Nevertheless, there are many who favor even more detailed disclosure of food product composition and nutrition. Existing as well as proposed labeling alternatives are complex requirements which are likely to present difficulties for food manufacturers, technologists, and consumers. The objectives of this study were to: 1) discuss current and proposed food labeling with respect to consumer usage, understanding, and ability to process food labeling information, 2) empirically measure consumer utilities for existing and proposed changes in food labeling, and 3) infer the degree of consumer and overall market satisfaction with alternative food labeling policies.


Ingredient labeling found on most food products provides little more than content information. Listings of ingredients in descending order based on quantity is not sufficient for purposes of making an accurate brand choice decision in many instances. For example, individuals on restricted salt diets need to know the amount of salt in a bread product, not simply that there is less salt in the bread than flour, water, and sugar. In response to this type of need, the FDA is considering a percentage ingredient labeling requirement for most packaged foods and has already proposed percentage ingredient labeling for baby food.

The requirement for percentage ingredient labeling is not a simple change in policy and will require specification of a standard format. If the policy dictates percentage ingredient labeling, the FDA must decide whether the policy should require listing the percentage of only the "characterizing" ingredient or ingredients or the percentage of all ingredients? Or, should the policy require a format listing percentage ingredients down to some cutoff value such as 1% or 2%, after which the other ingredients would be listed in descending order? Whatever the exact format, this requirement must also take into account problems created by this type of labeling in terms of production and quality control as well as from the standpoint of consumer information processing. Percentage ingredient labeling will increase manufacturing costs and these costs will almost certainly be passed on to consumers, adding to already increasing food prices. Imposition of these costs must be reconciled with the increased benefits provided by a percentage ingredient label and the net welfare or satisfaction of all consumers directly affected by such a requirement.

Nutrition labeling has been required for the past three years for those food products to which nutrients have been added or which make a nutritional claim. However, despite the fact that most agree that nutrition labeling was an important step toward presenting consumers with better information for making brand choice decisions (Beloian, 1973; USFDA, 1974b), nutrition labeling has been a controversial policy that has been disappointing in many ways (Daly, 1976).

A number of studies, including some by the FDA have indicated that consumers want nutrition labeling, claim they would use it, but seldom actually use it or understand it correctly in brand choice decisions (Lenahan et. al., 1973; USFDA, 1974a; Daly, 1976; Day, 1976). In an early study Lenahan and his associates found that only 25% of consumers they studied were aware of the nutrition label, 15% understood it, and 10% used it in a purchase decision. An FDA study found that 33% of the consumers they interviewed looked for ingredient labeling and only 5% looked for nutrition labeling. In another earlier study, Asam and Bucklin (1973) reported that the greatest impact of detailed nutrition labeling was in a more favorable perception of product quality. In a more recent review article, Day (1976) has noted that the only apparent effect of nutrition labeling has been to increase consumer confidence in nutritionally labeled food products. The value of nutritional labeling is questioned even further by Peterson's (1976) study of consumer perceptions of nutritional value as a function of nutrition labeling, price, and color of the product in which he found that the color of a bread product was more important than price and nutritional labeling combined in determining consumer perceptions of nutritional value.

Regardless of these findings, the view held by some (Bymers, 1972) is that "the use the consumer makes of nutritional information is peripheral to the main issue of the consumers right to know." Proponents of this view advocate policies leading to increased volume and detail of food labeling information. This raises a more pragmatic question: Do consumers have the capacity and ability to process this type of information load?

Lambert (1976) evaluated the question of capacity by taking 57 branded food products that could be considered in making up a single days menu. Using linear programming he processed the information such that the selection of brands would meet minimum daily nutritional requirements while minimizing the cost of the purchase. Facing some 54,000 pieces of information it took an IBM Model 370/ 165 Computer system 7 1/2 minutes of computer time to determine an optimal selection of brands. Since shoppers are faced with even greater purchasing tasks, it seems incomprehensible that the average consumer has capacity to process such large volumes of complex information accurately. With respect to choice accuracy Jacoby and his associates (1974a, 1974b) suggest that providing the consumer with increasing amounts of information may lead to a problem they have termed "information overload." When the consumer becomes overloaded with information his brand selection becomes dysfunctional and the accuracy of his choice is diminished. Considering all of these factors it seems quite possible that policies aimed at increased disclosure of the composition and nutritional value of food products may actually increase the cost of food products and lessen the ability of consumers to make improved brand choice decisions. While the question of choice accuracy under varying types and amounts of food label information has not been resolved, the purpose of this study was to measure and evaluate consumer utility for alternative labeling formats and amounts of information.


In this study variations in label format and information load were used to create 15 different bread labels. Bread was selected as a stimulus since it represents a familiar product and one for which the existing labeling format is well standardized. As shown in Table 1 the format of the bread labels was varied in five ways to present labels as ingredients only, ingredient plus nutrition, nutrition only, percentage ingredient plus nutrition, and percentage ingredient labeling. The information load was varied in three ways by constructing labels which contained either 4, 8, or 12 items of information. In each case two of the items of information were price and net weight. Other items were added to the label in the order they would normally appear on an actual label.



Presenting these labeling stimuli to a consumer and obtaining their rank order preference for these different labels provides the information needed for monotonic analysis of variance, or conjoint analysis of preference data. Given the assumption that consumer utility for alternative formats and amounts of information is an additive function, this analysis of preference data provides part-worth utility functions for each component of an additive utility function. Thus, from a consumer's ranked preference for alternative labels created by variations in format and amount of information, the following utility function can be derived separately for any given consumer:

U = UF + UL


U = a consumer's overall utility for a given label.

UF = a consumer's part-worth utility for a given food label format.

UL = a consumer's part-worth utility for a given amount of labeling information (i.e., information load).

These utilities can be determined individually and used to identify different consumer segments in the market (Assael, 1976) on the basis of similar utility functions. Recognizing the idiosyncratic nature of utility functions representative of different consumer segments can aid a policy makers ability to understand the labeling preferences of any given segment, recognize the diversity of preferences for food labeling represented in the market, and infer the satisfaction or dissatisfaction of any given consumer segment as well as the overall market with alternative food labeling programs.

One-hundred and forty female household shoppers were selected at random from the telephone directory of a city with a population of around 500,000. Although this sampling frame slightly reduces the number of high and low income consumers included in the sample, no attempt was made to compensate for this bias.

Each shopper was interviewed in a 20-minute long in-home interview. During the interview the shopper was presented seven food labels that varied in terms of format and amount of information. The food label shown in Figure 1 is typical of the food labels presented to shoppers. An overlapping Latin-Square design was used to reduce the number of stimuli that needed to be evaluated by the consumer yet retain enough information about alternative formats and amounts of information such that preferences for these stimuli could be processed and part-worth utilities derived. Without this procedure it is quite likely that the consumer would be overburdened with this particular type of evaluative task.



The shoppers were asked to evaluate each of the seven bread labels they were presented and then rank the labels in the order they would prefer them as labels they would use in evaluating a brand of bread. After ranking the labels, the participants were instructed to use a Likert-type scale to evaluate each label on the basis of several attitudinal statements which are shown in Table 2. Following this they were asked to evaluate several more general attitudes that were thought to be common among bread purchasers (these are shown in Table 3). Finally, the participants were asked several demographic questions relating to the occupation of the major wage earner, family structure, and questions concerning their consumption of bread products.

Ninety-nine usable questionnaires were obtained. There was no evidence that respondents whose results could not be analyzed were different than the 99 consumers examined. Individual preference rankings were analyzed using MONANOVA (Kruskal and Carmone, 1968) and part-worth utility function derived for each of the 99 consumers. Consumers with like or similar part-worth utility functions for label format and amount of information were aggregated using a cluster analysis program (Diehr, 1974). Discernable clusters were then evaluated in terms of differences in attitude, demographics, or consumptive characteristics that might be unique to consumers with like preferences for bread labeling.


The results of this consumer evaluation task and mathematical derivation of part-worth utilities for alternative food label formats and amounts of information provided three very distinct consumer groups. This is best seen in Figure 2 by graphs of the part-worth utility functions for each of the three consumer groups.

Group I (N = 35) was termed "label avoiders." These 35 consumers provided preferences which were analyzed and converted to part-worth utilities that showed a decreasing utility for more information and a great deal of utility for existing labeling whether it be ingredient or ingredient plus nutrition labeling. The fact that these consumers prefer a bread label with essentially two items of labeling information (shown in Figure 3)led us to infer from their part-worth utility for information load that they may in fact prefer no labeling, just price and net weight. These consumers were heavy users of white sandwich bread and predominantly blue collar families. With respect to other consumers, however, these consumers held attitudes that were more optimistic about how other consumers used food labels as shown in Table 3.

Group II (N = 36) was termed "information seekers." These 36 consumers exhibited a very strong preference for increasing amounts of information. However, they would be most satisfied with the existing format of labeling information as their utility for more detailed information diminished very acutely as shown in Figure 2. Thus, given this overwhelming preference for more information with existing format the food label in Figure 3 for the group termed "information seekers" depicts the label this group preferred the most. This group of consumers were predominantly white collar workers and consumed very little white sandwich bread, but considerably more whole wheat bread, and were more pessimistic about other consumers use of food labels.









Group III (N = 28) was termed "label users." This consumer segment preferred greater amounts of information in greater detail. This would be a group least satisfied with existing labeling requirements and the one most likely to be satisfied with an FDA move to percentage ingredient plus nutrition labeling. Based upon their part-worth utilities for more and detailed labeling information these consumers were termed "label users" and their preferred food label based on what was presented in this study is shown in Figure 3. Members of this group were also predominantly white collar workers, moderate to heavy users of white sandwich bread, and also pessimistic about other shoppers use of food product labeling as shown in Table 3.

The analysis of preference data and its value in identifying discernable differences in consumer utility for alternative food labels could be very useful in evaluating the satisfaction of specific consumer groups as well as overall market satisfaction under alternative labeling policies. In dealing with the food labeling question, for example, examination of the part-worth utilities for various labeling alternatives among different consumer groups such as those shown in Figure 2 can provide a clear indication of the consumer satisfaction in each of these groups for policy decisions that alter the amount and/or type of labeling information provided to these consumers.

Since it is possible to segment markets on the basis of differing utilities for labeling information, it is then possible to infer the approximate satisfaction or utility each segment of the market would obtain under different food labeling policies. To make this type of inference the following measure of consumer satisfaction can be used to determine labeling satisfaction for any given consumer or consumer segment in the market.


Market satisfaction can then be inferred by aggregating the satisfaction of each consumer segment with respect to the market share associated with the relative size of each consumer segment as shown below.


Utilizing these measures of satisfaction, consumer and market satisfaction with alternative bread labeling programs can be inferred for the consumer groups identified in this study. For any given consumer segment there exists a combination of format and information load that will provide consumers in that segment with maximum satisfaction. When a labeling policy corresponds to a consumer's preferred format and information load, consumer satisfaction will be maximum and indexed as 1.00 using the formula presented above. Labeling programs that are less desirable will result in less consumer satisfaction and a consumer index of satisfaction less than one. In some cases when a labeling program offers a format and information load that are essentially the inverse of the consumer's utility for labeling, the index for consumer satisfaction will take on a negative value. Weighting the consumer satisfaction index of each consumer segment by the relative size of each segment leads to an inference of overall market satisfaction. It also is 1.00 when maximized and can take on negative values under labeling programs that a majority of consumers would find very undesirable.

Utilizing these formulas and the part-worth utility functions shown in Figure 2, the following analysis is provided to illustrate the satisfaction or dissatisfaction of each consumer segment and overall market satisfaction under three different food labeling policies.

Program 1: Ingredient Labeling

A policy of ingredient labeling would correspond to a food labeling program for bread that would provide ingredient labeling on six ingredients. With the addition of price and net weight this corresponds to an information load of eight items. Consumer satisfaction for each consumer segment and overall market satisfaction under this type of program could be inferred as:


Program 2: Ingredient plus Nutrition Labeling

In this study this type of bread labeling policy would correspond to a format of ingredient and nutrition labeling with an information load of 12 items. Consumer and overall market satisfaction under this policy could be inferred as:


Program 3: Percentage Ingredient plus Nutrition Labeling

This type of bread labeling policy would require (on the basis of this study) 12 items of information that were in a format of half percentage ingredients and half nutritional information. Under this program consumer and overall market satisfaction could be inferred as:


Based on these inferences a food labeling policy of ingredient plus nutrition labeling would provide the greatest market satisfaction of the three programs evaluated. For consumers in groups I and II their satisfaction was greatest with this type of labeling policy; group II in fact would have maximized their satisfaction for bread labeling under this type of program.

Ingredient labeling produced the least overall market satisfaction, yet it is a permissible food labeling alternative for food products not containing nutrients or making a nutritional claim. Except for consumers in group I all other consumers would be considerably dissatisfied with this type of bread labeling policy.

Program 3, a policy of percentage ingredient plus nutrition labeling would achieve a moderate overall market satisfaction while maximizing the utility for consumers in group III. However, overall market satisfaction with this type of bread labeling policy is hindered by consumers in group I who prefer less information in a simpler format. If the consumers in group I actually prefer no labeling other than price and weight, or are simply indifferent toward labeling format, then the consumer satisfaction inferred from group I is misleading and should be discounted on the basis of no interest in bread labels. If this were the case, then program 3 would provide the greatest satisfaction for bread labeling among consumers represented in this study. On the other hand, if consumers in group I prefer less information in a simpler format because they do not use it but realize that a greater quantity of more detailed information will lead to higher bread prices, then their utilities for simpler bread labels, and subsequent satisfaction under alternative programs can not be ignored in evaluating overall market satisfaction under these programs.


This study examined consumer responses to variations in bread labeling format and information load using consumer preferences for alternative bread labels and analysis of these preferences using the method of conjoint analysis. Application of partial preference analysis to identify differences in consumer responses offers a useful method for assessing the impact of policy decisions on the consumer and market in general. Analyzing the cost of policy decisions can be made with relative ease in monetary terms, but assessment of its impact on the consumer satisfaction can be considerably more difficult. The methodology presented in this paper offers some benefit to those needing to analyze the latter effect of alternative policy decisions.

This type of analysis is particularly useful as a segmentation tool that differentiates the market on the basis of utility. With respect to food labeling this procedure allows researchers and policy makers to evaluate existing and alternative labeling programs with regard to the idiosyncratic preferences of different consumer segments, recognize the diversity of these preferences, and infer the likely satisfaction of dissatisfaction of any given consumer segment and overall market under any specific food labeling policy. In some instances it may be possible to specify different labeling requirements based upon differences in product usage in different consumer segments of the market. Though this type of marketing program would contribute to greater consumer and market satisfaction, it is probably not feasible in most situations.

In this study we have demonstrated how preference data could be analyzed using the method of conjoint analysis and used as a tool for evaluating alternative food labeling programs. Though the results in many ways concur with previous studies, additional research is planned to validate the results reported here. In the next two stages of our research we plan to sequentially: 1) evaluate in a choice situation the choice consistency between a consumers labeling preference and the brand of bread selected when only the labeling format is varied, and 2) evaluate the degree to which a consumer with differing bread labeling preferences will pay for additional costs of information. Beyond this additional research needs to focus on the relationship between consumption and preference as well as the understanding and use of labeling on more complex products such as over-the-counter drugs, ointments, and remedies.

Food labeling will continue to be a controversial topic among consumers, nutritionists, manufacturers, educators, and agencies of the government. The difficulties that have been experienced in past in terms of food labeling awareness, usage, and understanding as well as the consumers right to know more about food products will effect future food labeling policy decisions. The method of conjoint analysis presented in this paper offers a way of evaluating consumer preferences for labeling under existing and alternative policies and a method for inferring the degree of satisfaction that would be derived under policy alternatives by any given segment of consumers as well as by the overall market in general.


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

H. Assael, "Segmenting Markets by Response Elasticity," Journal of Advertising Research (1976).

A. Berloian, "Nutrition Labels: A Great Leap Forward," FDA Consumer (September 1973), 1.

G. Bymers, "Seller-Buyer Communication: Point of View of a Family Economist," Journal of Home Economics, 64 (February 1972), 59.

P. Daly, "The Response of Consumers to Nutrition Labeling,'' Journal of Consumer Affairs, 10 (Winter 1976), 170-78.

G. S. Day, "Assessing the Effects of Information Disclosure Requirements," Journal of Marketing, 40 (April 1976), 42-52.

G. Diehr, "Cluster Analysis Program (Revised," College of Business Administration, University of Washington, (1974), mimeographed.

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

J. Jacoby, D. Speller, and C. Kohn, "Brand Choice as a Function of Information Load: Replication and Extension,'' Journal of Consumer Research (December 1974), 68-72.

J. B. Kruskal and F. J. Carmome, Jr., "MONANOVA, a FORTRAN IV Program for Monotone Analysis of Variance," Bell Telephone Laboratories (1968), mimeograph.

A. L. Lambert, "Nutrition Information: A Look at Some Processing and Decision-Making Difficulties," in W. D. Peneault, Jr., ed., Advances in Consumer Research, 4 (Fall 1976), 126-32.

R. J. Lenahan, J. A. Thomas, O. A. Taylor, D. L. Call, and D. I. Padberg, "Consumer Reaction to Nutritional Labels on Food Products," Journal of Consumer Affairs, 7 (Summer 1973), 1-14.

C. S. Martinsen, J. G. Ostrander, and J. M. McCullough, "Consumer Attitudes Toward Preservatives in Bread and Other Foods," Bakers' Digest (1975).

R. A. Peterson, "Consumer Perceptions as a Function of Product Color, Price and Nutrition Labeling," in W. D. Peneault, Jr., ed., Advances in Consumer Research, 4 (Fall 1976), 61-63.

U.S. Food and Drug Administration, "Consumers Talk about Labeling," FDA Consumer (February 1974a)

U.S. Food and Drug Administration, "The Food Labeling Revolution," FDA Consumer (April 1974b)



Roger Best, University of Arizona
Jim McCullough, University of Arizona


NA - Advances in Consumer Research Volume 05 | 1978

Share Proceeding

Featured papers

See More


Penny for Your Preferences: Leveraging Self-Expression to Increase Prosocial Giving

Jacqueline R. Rifkin, Duke University, USA
Katherine Crain, Duke University, USA
Jonah Berger, University of Pennsylvania, USA

Read More


That's Just Plain Creepy: Understanding Consumer Responses to Personalized Food Products That Resemble People

Freeman Wu, Vanderbilt University, USA
Adriana Samper, Arizona State University, USA
Andrea Morales, Arizona State University, USA
Gavan Fitzsimons, Duke University, USA

Read More


Love is Blind: How Sensory Liking Impacts Perceptions of Unbranded Products

Jennifer L Stoner, University of North Dakota
Maria A Rodas, University of Minnesota, USA

Read More

Engage with Us

Becoming an Association for Consumer Research member is simple. Membership in ACR is relatively inexpensive, but brings significant benefits to its members.