Search For Nutrition Information: Synthesis and Empirical Test

Lawrence F. Feick, University of Pittsburgh
Robert O. Herrmann, The Pennsylvania State University
Rex H. Warland, The Pennsylvania State University
ABSTRACT - A model of search for nutrition information is proposed and tested. The model represents an extension of the work of Stigler (1961). In the model we hypothesize that costs of search are the opportunity cost of time while benefits are derived from the extent to which information has a monetary value and-a preventive health value, and the extent to which consumers regulate current diet. Other variables affecting costs and benefits are also included in the model. A regression estimation testing the model yielded a good fit but did not support the inclusion of the cost variable as a significant determinant of search in the multivariate setting.
[ to cite ]:
Lawrence F. Feick, Robert O. Herrmann, and Rex H. Warland (1983) ,"Search For Nutrition Information: Synthesis and Empirical Test", in NA - Advances in Consumer Research Volume 10, eds. Richard P. Bagozzi and Alice M. Tybout, Ann Abor, MI : Association for Consumer Research, Pages: 624-629.

Advances in Consumer Research Volume 10, 1983      Pages 624-629

SEARCH FOR NUTRITION INFORMATION: SYNTHESIS AND EMPIRICAL TEST

Lawrence F. Feick, University of Pittsburgh

Robert O. Herrmann, The Pennsylvania State University

Rex H. Warland, The Pennsylvania State University

[Funding for this research was provided by Competitive Grant 5901-0410-9-0299-0 of the U.S. Department of Agriculture.]

ABSTRACT -

A model of search for nutrition information is proposed and tested. The model represents an extension of the work of Stigler (1961). In the model we hypothesize that costs of search are the opportunity cost of time while benefits are derived from the extent to which information has a monetary value and-a preventive health value, and the extent to which consumers regulate current diet. Other variables affecting costs and benefits are also included in the model. A regression estimation testing the model yielded a good fit but did not support the inclusion of the cost variable as a significant determinant of search in the multivariate setting.

INTRODUCTION

The literature exploring the determinants of information search dates at least to Stigler (1961). In its baldest form, Stigler's model stated the extent of search undertaken by a consumer is a function of returns to and costs of search, with the optimal amount of search given by equating costs and returns at the margin.

Stigler viewed returns to search as the quantity purchased times the price reduction expected from an additional unit of search. Thus, returns would be greater, the larger the dispersion of prices in the market and the greater the quantity of the item purchased.

Stigler saw cost of search as principally a function of the opportunity cost of time expended in the search effort. Search cost, he felt, could be taken as being proportional to the number of sellers canvassed.

The basic prediction of the Stigler model that search is inversely related to costs and directly related to returns has been confirmed empirically. Lanzetta and Kanareff (1963) and Bucklin (1966) have found search decreases as cost of information increases. Rudell (1979) and Moore and Lehmann (1980) have found increased opportunity cost of time as inferred from subjective reports of time pressure resulted in decreased search. Research also indicates search increases with increased payoff to search. Katona and Mueller (1955), Bucklin (1966), and Newman and Staelin (1972) all found increased search for higher priced products.

However, other research has lead to contradictory results. Farley (1964), tested the hypothesis that since nigher income individuals have a greater opportunity cost of time and thus should search less, they should be more loyal. Goldman and Johansson (1978) applied the Stigler .Model to predicting consumer search for gasoline prices. They found estimated coefficients on costs and return variables to have the correct signs (i.e., negative and positive, respectively), but to not be very important in explaining search behavior. They concluded variables not typically included in the Stigler model, such as past experience, habit persistence, etc., were more important determinants of search than their measures of costs and returns.

ELABORATION OF THE STIGLER MODEL

Results suggesting the importance of other variables have lead researchers to hypothesize more elaborate models to explain search behavior. Engel, Blackwell, and Kollat (1978) have summarized some of the research into the determinants of search. While maintaining Stigler's distinction between perceived value and costs of information they have broadened both terms to include more noneconomic variables. They conclude perceived value of search is affected by: 1) perceived quality and quantity of existing information, 2) ability to recall the information, 3) perceived risk, and 4) confidence in ability to make decisions. They state the components of search costs as: 1) decision delay, 2) expenditure of time and money, 3) information overload, and 4) psychological costs. In addition, they note individual differences such as age, income, personality, etc., also appear to affect search.

Other writers have not maintained the distinction between costs and returns but have instead grouped similar variables affecting search into categories (Newman 1977; Bettman 1979; Moore and Lehmann 1980).

The Model of Search for Nutrition Information

A number of researchers have explored the determinants of search for nutrition information (Lenahan et al. 1971; Jacoby et al. 1977; Rudell 1979; Klopp and MacDonald 1981). However, no model has emerged to begin to unify the constructs found important in empirical work. This paper sketches our attempt to develop a model of search for nutrition information. The model of nutrition information search we will propose is based on an elaboration of Stigler's work. Most of the balance of this paper presents our attempt to explicate relevant constructs for the model. In addition, we briefly describe a test of the model.

Costs of Search for Nutrition Information

As noted, the economics literature views the cost of information search primarily as the opportunity cost of time. However, we should also consider explicit costs of search such as travel expenses and payment for information if they are relevant. Nutrition information, though, is generally available without any explicit cost. Most nutrition information is provided jointly with another product--on food labels and in magazine articles, for example--or is available gratis in government and extension pamphlets, etc. The important cost involved in obtaining this information is the time expended in search. Thus, in our model of nutrition information search, we will follow the economics literature in defining search costs as the opportunity cost of time. Stigler's theory of search contends the extent of search is negatively related to its cost. Thus, we expect that as the opportunity cost of time increases, search for nutrition information decreases.

Returns to Search for Nutrition Information

The Stigler model assumes returns to search are reductions in the minimum price the buyer of a good must pay. The consumer ascertains price offers by canvassing sellers in the market then buys the good from the lowest priced seller. Even in Stigler's model, the extent of search is a function of individual's perception of price dispersion in the market, not the actual dispersion. Buyer's extent of search is determined by equating perceived marginal returns to search and marginal costs. Thus, the extent of search is not necessarily a function or the amount of payoff but of the perception of payoff.

In the model of search for nutrition information, individuals are hypothesized to search more, the greater their perceived returns to search for nutrition information. These returns are a function of the perceived value of nutrition information to the individual. This value, we hypothesize, comes principally from three areas.

First, there is a monetary return to search. The monetary return to nutrition information search is defined as the extent to which the consumer feels able to obtain more nutrition per dollar expenditure by searching. The more useful information is in obtaining the best food value, the greater should be search.

However, it seems clear that individuals obtain other benefits from search in addition to simply getting the best nutritional value for their expenditure. They might search more, for example, if they were concerned about the effect of nutrition on their future well-being. In addition, consumers could search more irrespective of their future concern because of special diet needs imposed by weight or health Problems.

Thus the second benefit of information search could be described as preventive health. Venkatesan (1978) has argued consumers' nutrition-oriented behavior should be seen as examples of preventive health behavior: actions undertaken by healthy individuals to prevent undesirable future health consequences. Nutrition information search can be viewed as a preventive health action. This point of view implies the more the individual feels there is a link between current diet and future well-being, the greater will be search.

Finally, the value of nutrition information, and thus the extent of information search, should be affected by the extent to which individuals restrict their diets. That is, consumers who regulate the amounts of certain kinds of foods or nutrients in their diet will be more likely to search simply because diet restrictions force an individual to pay more attention to the content of foods they eat. These restrictions would not necessarily be forced on the consumer by, for example, health problems, but could be undertaken voluntarily by someone who simply watches what they eat. In either case, it is hypothesized consumers who constrain or restrict their diets will be more likely to search since they would presumably need more information on the content of their diet.

VARIABLES AFFECTING COSTS AND RETURNS

Perceived Characteristics of Nutrition Information

Bettman (1979) has emphasized information must be available and accessible for the individual to search. If no relevant information exists, search is fruitless. Even if it is available, however, information must also be in a form in which it can be processed for it to have value to the consumer. If information is perceived as incomprehensible or difficult to obtain, consumers would be less likely to search than if the information were seen as being available and understandable.

The hypothesized relationship between these characteristics of nutrition information and search is positive. That is, the more consumers feel nutrition information is available and comprehensible, the greater would be their search.

Individual Differences in Abilities and Experience

Individual differences in experience with information and abilities to obtain, process, and understand information have been shown empirically to affect the extent of consumer search. Characteristics and abilities which make the individual a better, more efficient information processor can be seen as reducing the cost of information (by reducing the time necessary to gather or process information) or as enhancing the value of information (making it more understandable, hence more useful). The characteristics of individuals which are hypothesized to affect the search for nutrition information are:

- specific experience with and interest in food and food-related information, and

- knowledge and self-assessed knowledge of nutrition.

Bucklin (1969) has found greater familiarity with, interest in, and use of media was related to greater information search on the part of consumers. Rudell (1979) found a significant, positive relationship between frequency of shopping and the amount of information search. Consumers can become better information users with more experience and involvement with information. Specifically, nutrition information search should be greater the more experience the consumer has with the media used to convey food information.

Individuals who are more involved in meal planning and food preparation also seem to be more likely to be nutrition information seekers. Rudell, for example, has characterized the information users in her experimental research as being ". . . very involved in the food consumption process" (1979, p. 43). Moore and Lehmann (1980) have found people interested in gourmet cooking used more information prior to making a food choice in an experimental setting. Klopp and MacDonald (1981) indicate food label readers are more likely to plan meals further in advance.

At least two possible reasons exist for the link between interest in food, cooking, and meal planning and nutrition information search. First, an interest in food per se may simply prompt an interest in nutrition. Individuals who are involved in the purchasing and preparation process may also be interested in food's nutrition content. Second, an interest in food would bring an individual into contact with food information (recipes, etc.), and perhaps with nutrition information as well, since these types of information are often presented together. Because familiarity with nutrition information should improve processing ability, the individual would be more likely to search because of the improved processing ability regardless of actual interest in nutrition. In our model of nutrition information search we expect individuals should search more, the more involved they are with food and food preparation.

Finally, last in the category of individual differences which affect search is the consumer's knowledge of nutrition. The marketing literature is clear on the predicted relationship between knowledge and search-increased knowledge leads to decreased search (see, for example, Engel, Blackwell, and Kollat 1978, p. 939).

The individual searches less because more of the information needed is stored in memory. In Bettman's (1979) terms, internal search replaces external search. In the marketing literature, where information seeking has typically been measured prior to a purchase, knowledge has often been operationalized as experience with a product. Katona and Mueller (1954) and Newman and Staelin (1972) found increased experience with a durable to be negatively related to information search. Moore and Lehmann (1980) have reported the same result for a nondurable (bread).

However, search for nutrition information does not occur exclusively or even most importantly prior to a purchase. Defining search as the information seeking before a purchase is more appropriate for major or infrequent purchases. Nutrition information is gathered from many sources over time and influences food choice decisions and-habitual food behavior over a lifetime. There are two reasons for expecting the relationship between knowledge and search to be positive rather than negative.

First, because search is not necessarily undertaken to gain information prior to a purchase, knowledge need not be seen as a substitute for search. Knowledge can instead be seen as a reflection of interest in nutrition and of a past history of search. Thorelli and Engledow have noted this relationship in a more general context and have stated, "There seems, in general to be a snowballing effect to education, experience, and information usage. We found, contrary cc earlier hypotheses, the greater the experience . . . the greater the information search" (Thorelli and Engledow 1980, p. 18).

The second reason for expecting a positive relationship between knowledge and search involves the link between processing ability and the extent of search. Many consumers find nutrition information too technical (Woman's Day 1978) or unclear (Redbook 1976). Because consumers rind nutrition information difficult to understand, prior experience with information should make them more effective processors and thus make information more valuable to them and reduce the cost of acquiring useful information

Some empirical work supports this position. For example, Klopp and MacDonald (1981) found the self-assessment of nutrition knowledge to be the most significant predictor of rood label reading. In the model of search for nutrition information, the anticipated relationship between search and knowledge is positive. Consumers who know more and who think they know more should search more.

DATA COLLECTION

Data used to test the model were obtained using a survey administered through telephone interviews conducted with 1,382 adult women selected by random digit dialing. Interviews were conducted in October; November, and December, 1980, with respondents residing in the 48 coterminous states. The respondent set for the sample was limited to women between the ages of 20 and 59 who were not pregnant or nursing at the time of the interview. These restrictions were imposed because it was felt the potentially special diet needs of the excluded categories of women might affect their nutrition concern and their search for nutrition information. However, the use of this restricted respondent set was not seen as a serious limit to the external validity of this study since the women included in the study are important, if not the most important, consumers of nutrition information

MEASUREMENT

The variables used to test the model came from a set of questions included on the survey. These variables, the survey questions from which they came, their means and variances, and the range of values which they could assume are included as Table 1. In addition, Cronbach's alpha, a weighted average inter-item correlation, is reported for variables which were indexes.

Search

Search, the dependent variable in the study, was measured using a set of six Likert-type questions that asked respondents how often they: 1) read pamphlets and books, 2) talked to doctors and health professionals, 3) read newspapers or magazine articles, 4) had discussions with family and friends, 5) read ingredient and nutrition labels, and 6) watched TV programs about nutrition. Previous research (Jalso et al. 1965; Woman's Day 1978) has indicated these actions are major sources of nutrition information for consumers. Respondents were asked to characterize their participation in these behavior as frequently, sometimes, or never. Responses for each behavior were then coded 3, 2, or 1, respectively, and the composite index, SRCH, was formed by summing the scores for each of the six items.

The use of ad index s a measure of information search has historical precedent. Constructed in the same way as SRCH in this study, an index used by Katona and Mueller (1954) weighted search by both the number of types of information sources used and by the intensity of use of each. Newman and Staelin (1972) have also used indices as measures or search. Their indices were constructed by awarding points to consumers based on specificity of information obtained and number of sources used.

Cost

Because the theoretically appropriate measure of the opportunity cost of time--the marginal wage rate--is difficult to measure, proxies have typically been used in empirical work. These proxies have included the average wage rate, labor force participation, and measures asking respondents the extent to which they feel time-pressed. These latter measures emphasize the perception of time pressure, an approach suggested by Schary (1971). Rudell (1979) and Moore and Lehmann (1980) have employed perceptual time pressure measures and estimated significant, negative coefficients for them in multiple regression equations using extent of search in an experimental setting as a dependent variable. A Likert-type perceptual measure of the opportunity cost of time, X1, was used in this study.

Returns

The monetary returns to search variables was measured as an item listed as X2. The preventive health returns to search variable, X3, was an index summing the scores on three questions. The model also hypothesizes individuals derive benefits from nutrition information search if they restrict their diets. Restrictions were measured both as an index of the kinds of restrictions consumers could make, X4, and reasons for restrictions, X5 and X6.

Other Variables

Variables X7 and X8 were Likert-type measures of the extent to which individuals felt nutrition information was available (X7) and understandable (X8). Variable X9 was an index or the consumer's experience with food-related (but non-nutrition) information and X10 was an index of consumer enjoyment of food preparation and meal planning. Consumer's nutrition knowledge was measured by X11, an item which measured their assessment of their own knowledge.

TABLE 1

VARIABLES USED IN THE STUDY

RESULTS

Regression results testing the proposed model of nutrition information search are presented in Table 2. An F test for the model was significant, p < .01, and the model explained 44 percent of the variance in SRCH, the index of information seeking.

In order to facilitate comparisons among the exogenous variables, we have listed standardized regression coefficients for the model. In examining these coefficients, the variables measuring the self-assessment of nutrition knowledge, X11, and the limiting of certain foods in the diet, X4, appear to be the most important determinants of information search included in the model. Three other variables which proved to be strongly significant determinants of search were the measures of the monetary value of nutrition information seeking, X2, the use of food shopping information, X9, and the enjoyment of cooking and meal planning, X10. Finally, the measure of preventive health benefits of nutrition information, X , and the measure of whether the respondent was on a health-constrained diet, X6, were also significant determinants of search. The remaining variables--the opportunity cost of time, a weight-constrained diet, and the availability and understandability of information--were not significantly related to search in the multivariate model.

TABLE 2

ESTIMATED STANDARDIZED REGRESSION COEFFICIENTS OF THE MODEL OR SEARCH FOR NUTRITION INFORMATION

DISCUSSION

The regression results suggest a number of conclusions about the model. The significant variables in the regression can be broadly grouped into three categories: variables reflecting the benefits of search (X2, X3, X4, X6), a variable reflecting a past history of search and perhaps salience of or involvement with nutrition (X11), and variables reflecting interest in or involvement with the process of food purchasing and preparation (X9 and X10). The variables reflecting how much trouble consumers nave obtaining and understanding nutrition information and the opportunity cost of time measure did not appear to significantLy affect the extent of search. Perhaps the original costs and returns model we proposed may be better described as a model of perceived benefits se nutrition information. In this model, salience of nutrition to the consumer seems to greatly influence the extent of search. In addition, involvement and interest in food and food shopping, apart from interest in nutrition. appear to affect the extent of search for nutrition information

Several implications emerge from the findings. It is not surprising that individuals search more if they see a greater health value in nutrition and nutrition information. It seems less obvious that search would be so strongly related to the perceived monetary value of the information. The results suggest that policy intended to increase nutrition information search could profitably emphasize the monetary value of the information as an alternative or supplement to emphasizing its heaLth value.

Second, the significance of X9 and X10, suggests one way to get consumers tuned into nutrition information may be to encourage their interest in food, cooking, meal planning, shopping, and their price awareness. The results imply cultivating these non-nutrition interests could lead to increased search for nutrition information.

Finally, the strongly significant positive relationship between the self-assessment of nutrition knowledge and nutrition information search lends support to Thorelli and Engledow's notion of a "snowballing effect to education, experience, and information usage." One or the implications of this result is that nutrition education programs, which may have as their goal diet improvement. can affect not only current food choice, but also the search for more information.

REFERENCES

Bettman, James R. (1979), An Information Processing Theory of Consumer Behavior, Reading, MA: Addison-Wesley.

Bucklin, Louis P. (1966), "Testing Propensities to Shop," Journal of Marketing, 30 (January). 22-28.

Bucklin, Louis P. (1969), "Consumer Search, Role Enactment, and Market Efficiency," Journal or Business, 42 (October), 416-438.

Engel, James F., Roger D. Blackwell, and David T. Kollat (1978), Consumer Behavior, 3rd ed., Hinsdale, IL: The Dryden Press.

Farley, John D. (1964), "Brand Loyalty and the Economics of Information," Journal of Business, 37 (October), 370-379.

Goldman, Azieh and J. K. Johansson (1978), "Determinants of Search for Lower Prices: An Empirical Assessment of the Economics of Information Theory," Journal or Consumer Research, 5 (December), 176-186.

Jalso, Shirley B., Marjorie M. Burns, and Jerry M. Rivers (1965), "Nutrition Beliefs and Practices," Journal of the American Dietic Association, 47 (October). 263-268.

Katona, George and Eva Mueller (1954), "A Study of Purchase Decisions in Consumer Behavior," in Consumer Behavior, Lincoln Clark, ed., New York: New York University Press.

Klopp, Pamela and Maurice MacDonald (1981), "Nutrition Labels: An Exploratory Study of Consumer Reasons for Nonuse," Journal of Consumer Affairs, 15 (Winter), 301-316.

Lanzetta, J. T. and V. T. Kanareff (1962), "Information Cost, Amount of Payoff and Level of Aspiration as Determinants of Information Seeking in Decision 'laking," Behavioral Science, 7 (October), 459-473.

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

Moore, William L. and Donald R. Lehmann (1980), "Individual Differences in Search Behavior for a Nondurable," Journal of Consumer Research 7 (December), 296-307.

Newman, Joseph W. and Richard Staelin (1979), "Prepurchase Information Seeking for New Cars and Major Household Appliances," Journal of Marketing Research, 9 (August), 249-257.

Newman, Joseph W. (1977), "Consumer External Search: Amount and Determinants," in Consumer and Industrial Buying Behavior, Arch G. Woodside et al.. eds., Amsterdam. Holland: North Holland.

Redbook Magazine (1976), "The Redbook Nutrition Report," presented at the Symposium on Nutrition in Marketing, San Francisco.

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

Schary, P. B. (1971), "Consumption and the Problem of Time," Journal of Marketing, 35 (April), 50-55.

Stigler, George J. (1961), "The Economics of Information," The Journal of Political Economy, 69 (June), 213-225.

Thorelli, Hans B. and Jack L. Engledow (1980), "Information Seekers and Information Systems: A Policy Perspective," Journal of Marketing, 44 (Spring), 9-27.

Venkatesan, M. (1978), "Consumer Behavior and Nutrition: Preventive Health Perspectives," in Advances in Consumer Research, 5, H. Keith Hunt, ed., Ann Arbor: Association for Consumer Research, 518-520.

Woman's Day Magazine (1978), Nutrition: A Study of Consumers' Attitudes and Behavior Towards Eating at Home and Out of Home.

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