Consumer Research on Food Labels
Below are research abstracts of consumer research studies conducted or supported by the Center for Food Safety and Applied Nutrition.
Performance Characteristics of Seven Nutrition Label Formats.
1996. Alan S. Levy, Sara B. Fein, and Raymond E. Schucker. Journal of Public Policy and Marketing 15(1):1-15.
The authors evaluate seven nutrition label formats to determine consumer comprehension and acceptance of displayed information. They test comprehension of five tasks: comparing two products, judging healthfulness, verifying claims, estimating servings needed to meet the daily requirements for a nutrient, and balancing nutrients in a daily diet. Performance scores were higher on some tasks--particularly dietary management ones--for formats that displayed nutrient amounts in percentages than for those that displayed nutrient amounts in metric units, even when interpretational aids were included on the metric formats. The two most preferred formats were metric formats with an interpretational aid. The findings have an important impact on decisionsabout the final nutrition label format required by the Food and Drug Administration.
More Effective Nutrition Label Formats Are Not Necessarily Preferred.
1992. Alan S. Levy, Sara B. Fein, Raymond E. Schucker. Journal of American Dietetic Association 92(10):1230-1234.
An experimental design was used to compare performance and preference for five nutrition label formats. Four performance measures--accuracy and false-positives in identifying nutrient differences, time required, and correctness in judging which product was more nutritious-- were derived from a product-comparison task. A sample of 1,460 food shoppers over 18 years old was recruited by a shopping mall-intercept method. Results of the study demonnstrated that preferences and performance do not necessarily agree. The Control format, which had no nutrition profile information, performed the best but was liked the least. The Adjectival format, which provided nutrition profile information in the form of descriptive adjectives, was the most preferred. Results also showed that listing Daily Reference Values or nutrition profile aids increased preference but either did not affect performance or decreased it, depending on the specific aid and performance measure. Formats that some subjects liked for having adequate information others disliked for being hard to use. Formats that some subjects liked for being easy to use others disliked for having inadequate information. Age, education, and race were related to all of the performance measures except judgment of relative nutrition. Only gender was related to preference. Results of the study are useful as guidance for the development of consumer education materials.
Nutrition Labeling and Public Health: Survey of American Institute of Nutrition Members, Food Industry, and Consumers.
1982. James T. Heimbach and Raymond C. Stokes.American Journal of Clinical Nutrition 36:700-708.
Five hundred thirty-one members of the American Institute of Nutrition, 177 persons from the food industry, and 107 consumers from a Food and Drug Administration mailing list responded to a survey dealing with nutrition labeling of foods. They identified obesity and heart disease as the major diet-related national health problems and chose information about calories, sodium, fat, protein, iron, calcium, and carbohydrates as most useful to the public.
Focus Group Sessions on Formats on Nutrition Labels.
1992. Christine J. Lewis and Elizabeth A. Yetley. Journal of the American Dietetic Association 92(1):62-66
Four consumer focus group sessions, with a total of 40 participants, were conducted to gather information on the utility and appropriateness of selected components of nutrition label formats.. The formats reviewed were bar graphs, pie charts, numeric listings, and adjectival descriptors such as high and low. Participants were asked to compare food labels using various format types and to discuss the utility and interpretability of the formats. The outcomes suggested that these consumers did not find pie charts useful. They considered bar graphs confusing or unnecessary when numeric values were provided. Participants expressed concern that adjectival descriptors could be misleading. The numeric listing format they considered the most useful consisted of two columns of numbers: one listing the amounts of food components present in a serving of the food, and a second listing either the percentage of the label reference value (e.g., the US Recommended Daily Allowance) or the quantity established as the label reference value. Participants repeatedly stressed their interest in a simple label. The results form one component of the Food and Drug Administration's efforts to evaluate nutrition label formats and will be used in conjunction with ongoing experimental and quantitative research studies.
The Impact of a Nutrition Information Program on Food Purchases.
1985. Alan S. Levy, Odonna Mathews, Marilyn Stephenson, Janet E. Tenney and Raymond E. Schucker. Journal of Public Policy and Marketing 4(1)1-13.
A quasi-experimental repeated measures design using a matched set of 20 test and comparison supermarkets in the Washington, D.C. and Baltimore, Md., metropolitan areas was used to evaluate a nutrition information program called "Special Diet Alert" (SDA) introduced by Giant Food, Inc. into Washington, D.C. stores in March 1981. The objective of the SDA program was to help supermarket shoppers find products for special diet needs by providing brand-specific (i.e., individual product level) shelf markers that identified products considered low or reduced in sodium, calories, cholesterol, and fat, supplemented by take-away information booklets available from a rack in the store which listed SDA brand names and specific nutrient values. Market shares of these products were tracked over the two-year evaluation period in Washington, D.C. and Baltimore stores. The pattern of differential sales trends across 16 individual food categories was complicated, but sales of shelf-marked products increased on the average 4 to 8 percent more over the two-year evaluation period in Washington, D.C. than in Baltimore, Md. stores. The average magnitude of effect attributable to SDA was modest in comparison with other factors influencing consumer purchases, highlighting the need for powerful evaluation designs to assess the effectiveness of information programs that operate in the context of many other more powerful influences. Further research is needed to determine which aspects of the SDA program were critical to its success, but one obvious difference between SDA and other in-store nutrition information programs that have been reported in the literature was the use of individual brand-specific shelf markers to deliver nutrition information to shoppers rather than prominently displayed sectional posters and detailed educational pamphlets.
Evaluating the Life Cycle of a Product Warning: Saccharin and Diet Soft Drinks.
1984. Robert G. Orwin, Raymond E. Schucker, and Raymond C. Stokes. Journal of Evaluation Review 8(6):801-822.
The effect of the saccharin warning label on sales of diet soft drinks was modeled with an autoregressive integrated moving average (ARIMA) process. Retail price trends and attendant publicity were modeled concurrently to separate these effects from those due to the warning. Results indicated that the label produced a small yet statistically significant reduction in sales, with an abrupt onset and, thus far, permanent duration. Reasons for the absence of decay effects, limitations of interpretability, and ideas for improving future evaluations of warning labels are discussed.
Does Nutrition Labeling Lead to Healthier Eating?
1997. Mario F. Teisl and Alan S. Levy. Journal of Food Distribution Research. 3(28):19-26
Nutrient labeling is found to significantly affect consumer purchase behavior; some evidence that consumers may act as if they hold nutrient (or health risk) budgets is found. Providing nutrient information may allow consumers to more easily switch consumption away from 'unhealthy' products in those food categories where differences in other quality characteristics (e.g., taste) are relatively small between the more and less 'healthy' products, toward 'unhealthy' products in categories where differences may be relatively large (i.e., a 'substitution effect'). If this substitution effect is large, nutrient labeling may not change the overall consumption of 'unhealthy' nutrients and thus may not lead to significant changes in health risk.
Consumers' Ability to Perform Tasks Using Nutrition Labels.
1998. Alan S. Levy and Sara B Fein. Jounal of Nutrition Education 30(4):210-217
Consumers' ability to perform common nutritional label use task components, which revealed the effect of prior knowledge, were analyzed for relationships with demographic characteristics, label reading, and health status. Data were from a mall intercept experimental study. Most consumers (78%) accurately compared two products, 58% accurately evaluated nutrient level claims, 45% comprehensively balanced nutrients over a daily diet, and 20% accurately calculated the contribution of a single food to a daily diet, a task that required complex math. The subjects who performed significantly poorer were over 55 years of age, nonwhite, and less educated than those who performed best. Not reading food labels and having a diet-related health condition were also related to poorer performance. Task component analysis showed that all types of subjects shared the same response tendencies when making nutrition judgments. The findings suggest that dietary guidance for consumers will be more effective if it does not require quantitative tasks but relies instead on tasks that are easier for consumers.
Contact Linda.Verrill@fda.hhs.gov for more information about the above studies.