Advances in Consumer Research Volume 8, 1981 Pages 477-483
FOOD LABEL INFORMATION: WHAT CONSUMERS SAY THEY WANT AND WHAT THEY NEED
Edwin C. Hackleman, Food and Drug Administration
This paper addresses the problem of isolating the important nutrient information desired by consumers on the food label. Since consumers have been criticized for their lack of nutrition knowledge, subject areas for additional learning are studied, and personal profiles based on learning interests are developed. Finally, comparisons between nutritional expert groups and household food buyers are mode on the basis of desired nutrient information.
Considerable attention in recent years has been devoted to the nature and types of information desired by consumers on food product labels. Actually, the concern lies within the broader realm of interest in the design of what Bettman (1975) dubs "information environments" in which consumers are being exposed to larger arrays of data describing all sorts of products. Wilkie (1975) and more recently McNeil and Wilkie (1979) and Westbrook and Fornell (1979) in their discussions related to formats for durable and energy product labels have shown how such information can affect appliance purchase decisions. One can conclude that most of the investigations surrounding the issue have addressed alternative formats for supplying consumers with information about products and brands to improve their ability to choose what to buy at the marketplace.
The literature abounds with studies concerned specifically with food product label information, including units of measurement concerns (Stokes and Haddock 1971), and the presentation of nutrition information for supermarket products (e.g., Bucklin and Asam 19731 Lenahen, et. al. 19731 Ruses 1975, 19791 Goodwin and Etgar 1980). Bettman (1975) has proposed that a concentrated list of nutrient information, conveniently displayed in a prominent place adjacent to all brands of a product, would overcome a number of limitations to information processing imposed by separate labels of nutrient information. Such a scheme utilizing a matrix display in an effort to reduce memory demands and mental search costs was studied experimentally by Goodwin and Etgar (1980) yielding inconclusive results, apparently because of a lack of understanding of nutritional information by consumers to begin with.
Still, regardless of many consumers' lack of nutrition knowledge, there is considerable evidence that more detailed nutrition data will be referred to by some segment of shoppers and can result in purchases of more nutritious products (Stokes 1972). Research contracted for by the Food and Drug Administration has shown that nutrition-labeled products are noticed and used by enough supermarket customers to judge the labeling worthwhile (Adamsy 1972). Hence, as Barksdale and French (1974) and Day (1976) predict the future will see the label increasingly used for the dissemination of nutrition infuriation, a strategy several large processors have already used to promote a competitive differential advantage.
Closely related to the problems of what to include on the food label is the literature concerned with information load and its effect on consumer brand choice (e.g., Russo 1974; Wilkie 1974; Summers 1974; Jacoby, et. al. 1974; Scammon 1977). These studies have addressed the quantity of product information display and its effect on a variety of selected response variables such as aided and unaided recall, "correctness" of choice, confusion of information, satisfaction with formats, and so on. Although results have been mixed, the tendency has been to suggest that too little information is worse than too much, thus paralleling the public policy of greater disclosure. Most of these studies, however, have operated under controlled experimental conditions and have not examined subject differences which would necessarily influence both the amount and type of information desired on the food label.
A major purpose of this paper, therefore, will be to further investigate food label information components desired by consumers, what types of label information to which they tend to pay most secession, and what food, health, and nutrition-related issues they desire to learn more about. An attempt will also be made here to utilize segmentation techniques to uncover relationships between desired label information, learning interests, and personal characteristics of chief food buyers and their households.
Data Base and Background
The Food and Drug Administration in cooperation with the U.S. Department of Agriculture and the Federal Trade Commission launched a major effort in 1978 to evaluate consumer usage of and satisfaction with information on food labels. Three agencies conducted a series of five hearings across the nation, designed to solicit consumer requests for revisions to the label. Several media releases were issued covering the food hearings and stressing the need for consumer feedback to encourage comments by mail. Over 9000 letters were obtained, and 450 individuals testified at the hearings, thus indicating a high degree of interest in the subject of food inhaling.
During the same year the FDA's' Consumer Studies Division, Bureau of Foods, began a comprehensive program of consumer research to examine both the quantity and content of information on the food label to allow recommendations for making the label more understandable and useful to consumers. The first phase of this program generated the data base for this study. A national survey, conducted by Response Analysis Corporation, yielding a probability sample of 1374 respondents. Personal interviews averaging just over 60 minutes each took place at home utilizing a questionnaire developed by the FDA and refined through a series of 18 developmental pretests, followed by a 50-case pilot study from two geographic locations. Upon completion of the interviews with the chief food buyers another questionnaire was dropped off for completion and mail-in at the respondents' convenience. A total of 864 questionnaires were recovered.
Late in 1979 another survey instrument was prepared to collect expert information from members of the American Institute for Nutrition (AIN), the Milk Industry Foundation (MIF), the Grocery Manufacturers of America (GMA), the American Meat Institute (AMI), the Food Marketing Institute (FMI), the American Bakers' Association (ABA), the National Food Processors Association (AFPA), and a segment of consumers on the FDA mailing list who, in the past, have exhibited heavy interest in such issues as nutrition labeling. A total of 820 questionnaires were recovered, including 536 AIN members, 177 industry representatives, and 107 from the consumer segment.
The personal interviews and mail questionnaires addressed a number of concerns, including the following:
1. Consumer awareness and usage of current label information;
2. Problems encountered with usage of label Information;
3. Experiences with food buying and usage;
4. Needs and desires for additional or revised information;
5. Knowledge and beliefs about food and nutrition;
6. Areas of interest for learning ears about food-related issues;
7. Evaluations of media for learning more about food and nutrition;
8, Sources and types of untruthful information received about food;
9. Diet-related health problems in the U.S.
Just over 64% of the total sample claimed they paid attention to nutrient-label information. This group was subsequently asked to rate the information about 38 specific nutrients as being either "very useful," 'of some use," or 'of little or no use." A rating scale for each nutrient was derived by weighting these responses by 100, 50, and 0. In addition, the usefulness raw scores were subjected to principal components factor analysis in an effort to reveal underlying dimensions along which respondents may have perceived the nutrients. The average rating scores of the nutrients clustering in each relevant factor were then compared to personal characteristics of chief food buyers and their preferences for other aspects of the food label.
Just under 63% of the total sample returned the nail questionnaire dropped off after the completion of the personal interview. One of the major sections of the questionnaire concerned a list of 17 food-related issues to which respondents indicated their degree of interest in learning more about.
Initially the items were ranked for a rough measure of overall interest and attention using a rating scale derived by weighting each response, similar to the procedures followed for nutrition information usefulness. Next, the items were subjected to principal-axis factor analysis to uncover any patterns in the data and to allow both sets of items to be rearranged to a smaller set of factors accounting for the inter-relationships. Factor scores derived from the analysis were regrouped into categories, and cross-tabulated with other personal characteristics of the respondents.
Table 1 lists the simple frequency tabulations for each category of usefulness applied to the nutrients by the respondents. The nutrients are listed in the rank ordering derived from the rating scheme which closely parallels the fraction selecting the nutrient "very useful.' Hence, a top-box analysis would produce virtually the same ordering of the nutrition information. The range of interest is quite wide for those items currently on the label. Calories were rated "very useful" by over 80% of the respondents, whereas riboflavin obtained 19%. Sugar and starch content, both not on the label, received substantial support and were above the median of those items not on the label. Respondents also desired, by over a 2-to-1 margin from a separate question, to have carbohydrates broken down into sugar, starch, and fiber.
Table 2 shows the varimax-rotated factor pattern derived from the first three factors extracted from the usefulness ratings given to the 38 nutrients. These three factors accounted for nearly 52% of the total variance in the system. All but three of the communalities exceed .4, indicating that nearly all the nutrients are closely related to at least one of the three principal components. Vitamin K and Calories possess the two lowest communalities, the latter perhaps because it is not really a nutrient. One would expect both of these variables to load higher on an additional factor.
USEFULNESS OF NUTRIENTS BY THOSE PAYING ATTENTION TO THE NUTRITION LABEL
A THREE-FACTOR, VARIMAX ROTATED SOLUTION TO FOOD LABEL NUTRIENTS RATINGS
Factor III appears to be a "who cares?" factor, since those items loading highly on it are many with which consumers are not familiar and tend to judge unuseful. The other two factors are more ambiguous, primarily because of the overlap between macronutrients often judged undesirable by consumers (e.g., saturated fat, sugar, calories, carbohydrates) and between micronutrients often judged desirable (e.g., iron, vitamin A, vitamin C, calcium). Two clear exceptions are fiber, a desired macronutrient, and sodium, an undesired micronutrient. Although protein is considered desirable, it is not clear from the double loading on factors I and II whether it is perceived as a macro- or a micronutrient.
Factor I contains "top of the label" nutrients' total fat, saturated and polyunsaturated fat, total carbohydrates, sugar, starch and fiber, calories and cholesterol. A conspicuous omission is protein which loads slightly higher on factor II. All the nutrients on factor I are frequently rated "bad for you," except for fiber and polyunsaturated fat, the letter of which receives mixed reviews. The presence of these two nutrients weighs against the argument of naming this factor, "undesirable nutrients." On the other hand, the presence of sodium and the absence of protein weighs somewhat against the label, "macronutrients."
Factor II envelopes all the lettered vitamins (A, B-6, B-12, C, D, E, and K), thiamin, protein, and minerals such as iron, calcium, phosphorus, and possibly potassium. An alternate name may also apply here, "desirable nutrient." The only argument negating this label is that sodium also loads somewhat highly here, despite a much higher loading on factor I.
Factor III is dominated by items many respondents have hardly heard of: inositol, biotin, selenium, molybdenum, folacin, pantothenic acid, riboflavin, and so on. Also, several of these nutrients comprise elements or substances people are aware of but are not sure how useful they may be. Nutrients such as potassium, phosphorus, niacin, iodine and zinc would be examples of these.
Average usefulness scores were computed using the nutrients which loaded highest on a given factor. These scores, in turn, were broken down by the personal characteristics of households as shown in Figure 1. The slopes of the three lines permit comparisons between the demographically-defined segments. Highly educated food buyers tend to find all three types of nutrients more useful than the lower educated. High school graduates have equivalent interest in macronutrients and important micronutrients as those with some college or more. Those with college experience, however, tend to have a keener interest in unfamiliar micro-nutrients.
Perceived usefulness of macronutrients rises with age and does not decline until the food buyer reaches the senior citizen category. However, there is a steady decline in perceived usefulness of important micronutrients as age increases. This may, in part, be explained by an increase in interest in macronutrients and important micronutrients by food buyers with children. In addition, food buyers with household members taking vitamin or mineral supplements display greater perceived usefulness of all nutrients. Female food buyers find information on macronutrients and important micronutrients more useful than do male food buyers - indicating that obtaining nutritionally balanced diets for the family is regarded primarily as the woman's responsibility.
PERCEIVED USEFULNESS OF NUTRIENTS BY DEMOGRAPHIC SEGMENTS
Table 3 shows the 17 consumer learning interest items, rank ordered by their overall rating derived from a five-point Likert scale. There appears to be substantial interest in general for all 17 items, as the modal response for all of them was "very interested." The heaviest interest, however, enter on relationships between food intake and health or diseases such as cancer, heart disease, blood pressure, and the "Delaney Clause" -- a law which prohibits the sale of any food product containing an additive causing cancer in animals or man.
A principal-axis factor analysis of the original data matrix uncovered a two-factor solution accounting for over 66% of the variance. Table 4 shows the varimax-rotated factor matrix and the communalities associated with the loadings on the first two factors. All but one of the communalities exceed .5, indicating that nearly all the variables are closely related to one of the two principal components. "Benefits and risks of reducing diets" may possess a higher loading on a factor not extracted here. Factor I is a food additive information factor, collecting together interest in the safety, regulations, types, rationale, and benefits/ risks of food additives. Factor II is a health-disease concern factor, highlighting a strong consumer interest in those issues connecting food rich heart disease, cancer, high blood pressure, weight reduction and balanced nutrition.
A RANK ORDERING OF INTEREST SCORES ON VARIOUS FOOD-RELATED ITEMS
It is suggested by these findings that two major groups can be isolated based on their learning interest orientations in the directions described by the two factors. To further describe them, factor scroes were derived from the principal components to measure the propensity of each respondent to have an interest in these directions. The sample was split into thirds along each dimension and cross-tabulated against a host of descriptive variables to obtain a profile.
A cross-tabulation analyses yielding a number of significant differences is summarized in Table 5. Concentrating first on factor I, individuals who are more likely to be interested in obtaining information on food additives would also more likely be overweight, 46 to 59 years old, reside in a larger household, and have earned a college education. These individuals either are or live with a special dieter(s) and vitamin taker(s). They are less satisfied with the freshness of their food, tend to have greater knowledge of nutrition, are extensive users of the ingredients list, report greater incidence of stale or spoiled food, and are heavy users of open dates on perishable food labels. On the other hand, individuals who are disinterested in obtaining information on food additives would also more likely be overweight, 46 to 59 years old, reside in smaller households, and have at most a high school education. [The implication here is that overweight, middle-aged respondents were sharply divided into their interests, one group having much interest and the other having little or none.] These individuals are not or do not live with a special dieter(s) and vitamin taker(s). Although they claim to have little knowledge of nutrition and do not pay much attention to the list of ingredients, they are more satisfied with the freshness of their food, report fewer stale or spoiled food experiences, despite little reported usage of dates on food labels.
A TWO-FACTOR, VARIMAX ROTATED SOLUTION TO LEARNING INTERESTS IN FOOD-RELATED ITEMS
RELATIONSHIPS BETWEEN LEARNING INTEREST FACTOR GROUPS AND THEIR PERSONAL CHARACTERISTICS
Interest in factor II, concerns for information related to health or disease, generates a somewhat different profile. A higher interest in these issues is shared by those who tend to be about the right weight or slightly under, white-collar workers, those living in smaller households, earning relatively high annual incomes ($15K to over $25K) and in possession of a college education. In addition, individuals in this group tend to be less satisfied with their food's freshness, and although they tend to use the ingredients list and dates on food labels, they claim to know little about nutrition. In contrast, consumers reporting a disinterest in health-disease information are more likely to be overweight, blue-collar workers, from larger households, earning relative low annual incomes ($5K to $15K) and possessing at most a high school education. These people are more satisfied with their food's freshness, although they tend not to use dates on food labels, pay little attention to the list of ingredients, and claim to have little knowledge of nutrition.
The overall ratings of the nutrient information supplied by the chief food buyers (see Table 1) were used to rank the items from 1 to 38 a8 shown in Table 6. Similar racing scores were obtained from the mail questionnaires which tapped industry experts, AIN members, and consumers heavily involved with the subject. Ranks from each of these groups are also shown in Table 6. Priority agreement between the judgmental groups seems high (Kendall's W = .95) overall, and pairwise agreement between the chief food buyers and each expert segment is also high. The lowest agreement (Spearman's R = .89) is obtained in a comparison with AIN members. Despite the apparent high overall agreement certain nutrients stand out from others as being judged quite differently. Vitamin C, sugar, and starch are given much higher priority by chief food buyers than by the expert groups. On the other hand, the experts tend to rate sodium (salt), calcium, iron, potassium, and riboflavin substantially more useful than household food buyers Judge these nutrients.
A COMPARISON OF RANKS GIVEN TO ALL NUTRIENT INFORMATION ITEMS BY CONSUMERS AND EXPERT GROUPS
Tables 7, 8 and 9 contain the rank orderings of the usefulness rating scores bestowed by each group upon the macronutrients, important micronutrients and unfamiliar micronutrients derived from the factor analysis. Priority agreement within these three dimensions is again relatively high, but there appears to be substantially less agreement between the groups in the case of macronutrients (Kendall's W = .83) as shown in Table 7. In particular, AIN members show the greatest disagreement with chief food buyers (Spearman's R = .50). The rank for sugar shows the greatest deviation between chief food buyers and the experts, as households tend to give this nutrient higher priority. Turning to important micronutrients shown in Table 8, there is a tendency for minerals (iron, calcium, and potassium) to be ranked higher in priority by the experts than by chief food buyers, whereas the latter tend to rank vitamins (C, E, B-12) higher than the experts. One would have expected the households to agree the least with the experts on the unfamiliar micronutrients (Table 9), but agreement between the three expert groups and households was surprisingly high. The only nutrient which seems out of place in this list is choline/lecithin, which received a higher ranking by chief food buyers.
A COMPARISON OF RANKS GIVEN TO MACRONUTRIENTS BY CONSUMERS AND EXPERT GROUPS
A COMPARISON OF RANKS GIVEN TO IMPORTANT MICRONUTRIENTS BY CONSUMERS AND EXPERT GROUPS
A COMPARISON OF RANKS GIVEN TO UNFAMILIAR MICRONUTRIENTS BY CONSUMERS AND EXPERT GROUPS
It would appear that the nutrition information on the food label can be revised to include a list of nutrients the consumer would find more useful in evaluating the contents of the package. There are several nutrients not included in most nutrition information lists which consumers would like to see (e.g., sugar, starch and fiber content, separate from total carbohydrates, and vitamin D). This may be one of the reasons why nearly 26% of the sample desired to see more label information, whereas slightly over 2% desired to see less. On the other hand, some of the "standard" information appearing on nutrition labels is judged to be of little or no use by the majority of consumers (e.g., riboflavin, niacin, thiamin). In addition, many producers include a number of micronutrients on the label which the consumer cares very little about (e.g., pantothenic acid, zinc, magnesium, phosphorus, chromium, copper, biotin, selenium, inositol, etc.). These findings may help explain why nearly 27% of the chief food buyers found the label information to be confusing and why over 34% expressed dissatisfaction with the current content of the food label information. What is desired, therefore, is a finer tuned, more relevant list of nutrients and not merely a more exhaustive list contributing mass to the information load. A typical criticism has been that consumers as a whole know very little about food and nutrition and are, therefore, incapable of properly evaluating nutritional information on food labels. Since over 73% of the respondents indicated they were interested in learning more about nutrition and over 77% were interested in learning more about food additives, their lack of knowledge appears overshadowed by the desire to learn more. The learning interests focus heavily on the effect of food intake on health and disease problems--the things which generate news headlines. However, a surprisingly strong support in learning more about food additives was shown here, although consumption side affects may be noticeable only in the very long run, and many are completely safe for use. Results indicate a powerful similarity between what nutrients consumers find most useful (and useless) on the label and what the experts judge they need to see (or not see). If the chief food buyers in households are to be criticized for their lack of nutritional knowledge, one cannot in the same breath criticize their ability to decipher which nutrients they should examine most carefully when considering a purchase.
Adamy, C.F. (1972), "The Retailer's Experience," Food and Cosmetic Law Journal, 27, 263-270.
Bettman, James R. (1975), "Issues in Designing Consumer Information Environments," Journal of Consumer Research, 2, 167-77.
Bucklin, L. and Asam, E. (1973), "Nutrition Labeling for Canned Goods: A Study of Consumer Response," Journal of Marketing, 37, 32-37.
Day, G. S. (1976), "Assessing the Effects of Information Disclosure Requirements," Journal of Marketing, 40, 42-52.
French, W. A. and Barksdale, H. C. (197A), "Food Labeling Regulations: Efforts Toward Full Disclosure," Journal of Marketing, 38, 14-20.
Goodwin, S. and Etgar, M. (1980), "Alternative Organizational Formats for Nutritional Information: Processing and Policy Normative Perspectives," Marketing in the 80's: Changes and Challenges, 46, ed. R. P. Bagozzi, et. al., Chicago: American Marketing Association, 412-15.
Heimbach, J. T. and Stokes, R. C. (1979), "FDA 1978 Consumer Food Labeling Survey," Washington, D.C.: Food and Drug Administration, Bureau of Foods.
Jacoby, J., Speller, D. E., and Kohn, Carol A. (197A), "Brand Choice Behavior as a Function of Information Load," Journal of Marketing Research,11, 63-9.
Jacoby, J. "Brand Choice is a Function of Information Load: Replication and Extension," The Journal of Consumer Research, 1, 33-42.
Lenahen, R. J., et. al., (1973), "Consumer Reaction to Nutritional Labels on Food Products," Journal of Consumer Affairs, 7, 1-14.
McNeil, D. L., and Wilkie, W. L. (1979), "Public Policy and Consumer Information: Impact of New Energy Labels," Journal of Consumer Research, 6, 1-11.
Russo, J. E. (1974), "More Information is Better: A Reevaluation of Jacoby, Speller, and Kohn," Journal of Consumer Research, 1, 68-72.
Scullion, D. (1977), "Information Load and Consumers," Journal of Consumer Research, 4, 148-55.
Stokes, R. C. (1972), "The Consumer Research Institute's Nutrient Labeling Research Program," Food and Drug Cosmetic Law Journal, 27, 249-262.
Stokes, R. C. and Haddock, R. (1972), Interim Resort for the First Two Phases of the CRI/FDA Nutritional Labeling Research. Washington, D.C.: Consumer Research Institute.
Summers, J. O. (1974), "Less Information is Better?" Journal of Marketing Research, 11, 467-68.
Westbrook, R. A. and Fornell, C. (1979), "Patterns of Information Source Usage Among Durable Goods Buyers," Journal of Marketing Research, 16, 303-12.
Wilkie, W. L. (1973), "Analysis of Effects of Information Load," Journal of Marketing Research, 11, 462-6.
Wilkie, W. L. (1975), How Consumers Use Product Information: Assessment of Research in Relation to Public Policy Needs. Washington, D.C.: U.S. Government Printing Office.