Consumer Factors Influencing the Use of Nutrition Information Sources

Citation:

Guijing Wang and Stanley M. Fletcher, University of Georgia Dale H. Carley (1995) ,"Consumer Factors Influencing the Use of Nutrition Information Sources", in NA - Advances in Consumer Research Volume 22, eds. Frank R. Kardes and Mita Sujan, Provo, UT : Association for Consumer Research, Pages: 573-581.

Advances in Consumer Research Volume 22, 1995      Pages 573-581

CONSUMER FACTORS INFLUENCING THE USE OF NUTRITION INFORMATION SOURCES

Guijing Wang, University of Georgia

Stanley M. Fletcher, University of Georgia

Dale H. Carley, University of Georgia

[The authors wish to thank two anonymous referees for comments and suggestions that led to improvements of an earlier version of this manuscript.]

ABSTRACT -

Consumer usage of nutrition information sources is investigated in a cost-benefit framework using the 1987-88 Nationwide Food Consumption Survey data. Results show that food expenditure, family size, age, education and enduring importance of nutrition are major predictors in positive direction in the usage for all sources. Income has a significant effect on the usage in varied directions across the sources. The sources consumers used and preferred may be different. Hence, consumers may be willing to pay for but unwilling to use obtained nutrition information. These findings have important implications for the design of nutrition information programs.

In the past several decades, much research has been conducted on consumer usage of nutrition information. As early as 1970's, Jacoby, Chestnut, and Silberman (1977) pointed out that majority of consumers claimed they wanted and were willing to pay for nutrition information. However, they found that most consumers neither acquired nutrition information when making purchase decisions nor comprehended most information they received. Up to 1990's, Moorman (1990) reported that 20 years research indicated that adequate nutrition information and its use were far from satisfactory. In order to enhance the efficiency of nutrition information, it is necessary to understand how consumers perceive and use such information (Lenahan, et al. 1973).

Feick, Herrmann, and Warland (1986) conducted a probit analysis of consumer usage of different nutrition information sources in a cost-benefit framework. Based on their data obtained from telephone interviews with 1,265 women between the ages of 20 and 59, they found some variables such as concerns on future health and experience on reading food advertisement had consistent effect on the usage across six information sources. Other variables such as present health, age, and education had varied effects on the usage across information sources.

Moorman (1990) investigated the effects of consumer characteristics such as nutrient familiarity, education, age, and stimulus characteristics such as information format and content on the utilization of nutrition information. She pointed out that effectively designed nutrition disclosures facilitate information utilization. Bass (1991) conducted a study on consumer use and satisfaction of food labels. She found that food labels, the most used source of nutrition information (Navder 1993), was indeed used for various reasons with various satisfaction levels. Burton and Biswas (1993) examined the effects of changes in labels required by the nutrition labeling and education act of 1990 (NLEA) on nonstudents aged 18 years or older. Their survey results showed consumer attitude and perception of nutrition likelihood were strongly related to nutrition information recommended by the NLEA.

Cole and Balasubramanian (1993) provided some insights about nutritional information usage and age differences in consumer search for and use of the information. They concluded that elderly adults tended to search less intensely and less accurately than young adults. This finding appears in conflict with that of Moorman (1990) who found that age is positively related to information processing and comprehension accuracy. Moorman and Matulich (1993) examined the effects of various consumer characteristics on health information acquisition behaviors. Their results showed that consumer health motivation and ability characteristics are not always critical precursors of the information acquisition behaviors. Jensen and Kesavan (1993) analyzed the relationship between the sources of information, consumer attitudes on nutrition, and consumption of dairy products. They found that nutrition massages related to calcium affected demand for dairy products.

The above mentioned studies are among the most significant attempts in exploring consumer usage of nutrition information. However, some of them only focused on the usage of nutrition information sources of a specific population group (Feick, Herrmann and Warland 1986, Burton and Biswas 1993, Cole and Balasubramanian 1993). Others typically investigated a specific information source such as food labeling or the relationship between nutrition information and food consumption (Bass 1991, and Jensen and Kesavan 1993).

This study adds to the literature in that determinants of consumer usage of various nutrition information sources and whether consumers obtain nutrition information from their preferred sources are investigated. In the next section, the current usage levels of nutrition information sources and factors which may affect consumer benefits and costs of information search are discussed. Then a model is specified for analyzing whether consumers use these information sources and whether the source used is their preferred. The model is applied into the 1987-88 Nationwide Food Consumption Survey (NFCS), the most recent and complete household food practice data source available. The estimation results and their possible implications are discussed. Finally, a conclusion of this research is provided.

INFORMATION SOURCES AND FACTORS AFFECTING CONSUMER USAGE

In the 1987-88 Nationwide Food Consumption Survey (NFCS), consumers were asked whether they obtained nutrition information from any of eight sources during the last year and which source is their most preferred. The eight sources and their current usage levels are highlighted in table 1.

The table 1 shows that newspapers, magazines or books, and food packages or labels are the most often used sources. More than 45 percent of consumers used them. This is followed by the sources of a doctor, nurse or health professional, and radio or television which are used by more than one third consumers. The sources of food company publications, government or health organization publications, a nutritionist, dietitian, home economist, or extension agent, and relatives or friends are rarely used as nutrition information sources by consumers (less than a quarter consumers).

Another set of question in the NFCS is to ask what is consumers' most preferred information source on nutrition. The survey results are also reported in table 1. These results show that a doctor, nurse or other health professional is most likely to be a consumer's most preferred information sources if he/she uses them. The proportions of consumers who prefer food company publications, government or health organization publications, relatives or friends, radio or television, and food packages or labels are very low. These facts may indicate the credibility of these sources.

TABLE 1

NUTRITION INFORMATION SOURCES AND THEIR CURRENT USAGE LEVELS

In order to comprehend the consumer usage of these information sources, it is necessary to investigate the factors affecting consumer information search. Russo and Leclerc (1991) conducted a research in the success of product information programs using a cost-benefit analysis. They claimed that from the consumer's perspective the cost-benefit framework is a useful guide to program design. According to Stigler's cost-benefit approach, consumers obtain the amount of information up to the point at which the anticipated marginal benefits equal marginal costs of information search. Feick, Herrmann, and Warland (1986) provided a detailed discussion of possible benefits and costs of information search. Following their procedure, the perceived benefits of search for nutrition information are affected by the enduring importance of nutrition to the individual and his enduring interest in nutrition. They used the self-reported individual's present health such as poor and excellent as the measurement of the importance of and interest in nutrition. In this study, the consumption level of fat and cholesterol are used to capture the effect of enduring importance of nutrition.

In the last several decades, consumers have become increasingly aware of health and food nutrients. Fat and cholesterol are probably the most well known nutrients which affect consumer food purchase decisions in the United States. Consumers who consume more fat and cholesterol can be assumed to be less concerns about the importance of nutrition. Hence, they may also be less likely to search and use more nutrition information. One of the hypothesis is that the amount of consumer search for nutrition information depends on the enduring importance of nutrition.

Feick, Herrmann, and Warland also discussed the appropriateness of age, married status of individuals, and involvement of children as major factors influencing the benefits of nutrition information. Besides the age, household head status and household size are used as the factors affecting consumer information search instead of the married status and involvement of children in this study. The household head status means whether or not a household is headed by both male and female.

Costs of information search are mainly determined by opportunity cost of time and efficiency of search. Feick, Herrmann, and Warland explained the appropriateness of marginal wage (income) and household size as measures of the opportunity cost. The efficiency of search are determined by education and experience (age).

In addition to the above factors, food cost at-home consumption, region and urbanization of a consumer resides, and race are used as explanatory variables in this study. These variables have not been addressed properly in the literature, partially because of data limitations. Food expenditure is hypothesized to affect consumer information search in a positive direction. Urbanization and region a consumer resides may be associated with the availability and accessibility of various information sources. Moreover, consumers in different areas may use nutrition information differently based on beliefs or traditional food consumption behaviors. Thus the benefits and costs of information search are expected to be different among residence areas. Race is postulated to affect consumer information search because the value, attitude, and perception of nutrition information may differ among racial groups.

In summary, the above factors which affect the benefits and costs of information search should be included in studying the determinants of consumer usage of nutrition information. Feick, Herrmann, and Warland also explained different effects of these factors on the extent of information search from various sources are present. For example, reading books is time consuming. The opportunity cost of time should be more important in analyzing the information search from this source. Listening radio or watching television may be a kind of entertainment. The effect of opportunity cost of time on obtaining nutrition information from this source may be negligible. In spite of this fact, the same set of explanatory variable are used for all the information sources for the comparability across these sources. In addition, utilizing different sets of explanatory variables for each information source requires a number of assumptions about the source and the information available from them (Feick, Herrmann and Warland 1986).

THE MODEL

Consumers maximize their utility derived from various nutrition information sources. The amount of information searched is optimized according to the rule of Stigler's marginal benefits and costs of information search (1961). Consumers, the utility maximizers, make decisions on whether they search nutrition information based on the utility derived from various alternatives. The utility level derived from an alternative depends on consumers' socio-economic and demographic characteristics such as their education, race, age, family size, and income which associate with consumer benefits and costs of information search.

Assume that the utilities derived from the two choices of an individual who uses and does not use a nutrition information source are U_i0 and U_i1, respectively. They can be expressed as:

EQUATION (1)

where U_i0 and U_i1 are average utilities from the two choices, respectively, e_i0 and e_i1 are random errors. The average utilities are used since the individual's socio-economic characteristics do not vary across alternatives.

The utilities, U_i0 and U_i1, are random variables. The consumer i will use this source only if U_i0>U_i1. Define a latent random variable as Y^*=U_i0-U_i1>0, a linear statistical model can be expressed as:

EQUATION (2)

where Y^* is consumer usage of this source, x is a vector of explanatory variables, ¯ is a vector of coefficients to be estimated, and e^* is the error term. The dependent variable, Y^*, is unobservable. The observable variable, Y, is a discrete (binary) choice variable and is defined as:

EQUATION

The values of Y_i are determined by the following relationship:

EQUATION

Due to the difficulties of the standard regression technique for the discrete choice model, several alternative procedures such as probit and logit models are proposed (Maddala 1983). The logit model is utilized in this study.

In the logit model, the interest is in the probability of Y_i (a random variable) taking a value of 1 or 0 and what factors affect this probability. The probability for Y_i=1 is given by:

EQUATION (3)

The probability distribution of e_i* is assumed to be determined by a logistic distribution and the P_i can be estimated by:

EQUATION (4)

where x is a vector of independent variables with ¯ as the corresponding regression coefficients to be estimated. Its log likelihood function is expressed as:

EQUATION (5)

where T is the number of households.

The utility derived from an individual's preferred nutrition information source should be different from the utility derived from a source which is not his/her preferred source. For an information source used by the individual, two possibilities exist. They are 1) he/she uses the source which is his/her preferred source, and 2) he/she uses the source but it is not the preferred source. Whether the information source used by the consumer is his/her preferred source is a measure of success of consumer information search.

The success can be modeled and estimated in the exact same way as the logit procedure described above by the following assumptions. Assume that the utilities derived from the two possibilities of an information source used by a consumer i is his/her preferred source and not his/her preferred source are U_i0 and U_i1, respectively. The observable variable, Y, is a discrete (binary) choice variable and is defined as:

EQUATION

DATA SOURCE

The data source used in this study is the 1987-88 Nationwide Food Consumption Survey (NFCS) which was conducted between April 1987 and August 1988 sponsored by the United States Department of Agriculture. The survey is a self-weighing, multistage, stratified, area probability sample representative of the households in the 48 conterminous States of the United States. There are 4,250 housekeeping households which are used for this study. The housekeeping households are defined as at least one member had ten or more meals from household food supply during the survey week.

Consumer usage and preference of nutrition information from eight popular sources has been discussed previously. The survey contains many important economic and socio-demographic variables. These variables are important determinants of consumer benefits and costs of information search. They are used as explanatory variables in the logit analysis. The variable definitions and selected sample statistics are presented in table 2.

RESULTS

The logit model was used to estimate the determinants of consumer search for four nutrition information sources [the other four sources were not analyzed by the logit procedure because there were not sufficient consumers using them (Table 1).]. The estimation results are presented in table 3. The majority of the estimated parameters are statistically significant at the 0.05 significance level for all the four information sources. The likelihood ratio chi-square statistics suggest the overall significance of the models. Within the sample, more than 62 percent of the observations are correctly predicted for all the information sources.

The estimated coefficient of food expenditure at-home is statistically significant and positive in all equations. This is expected because the more food a household consumes at-home, the more it searches for nutrition information in order to make better food choice. Because no previous studies incorporated food expenditure as an explanatory variable, no comparison can be made for this finding. The coefficient of income, a major component of opportunity cost of time, presents an interesting pattern. When incomes increase, the probability of obtaining nutrition information from a doctor, nurse or other health professional decreases. The opposite is held for the equations of books and labels, respectively. This pattern suggests that high income consumers may visit doctors and health professionals less often than low income consumers. In addition, they may be more accessible to the sources of newspapers, magazines, and books, and understanding the nutrition information provided in food packages and labels better than do the others. Inconsistent results were found by Feick, Herrmann, and Warland (1986). They reported that income was not a major determinant for the use of all information sources.

Family size is not a significant predictor of information search from doctor and television while it is a significant predictor in positive direction in the equations of books and labels. These findings are consistent with those of Feick, Herrmann, and Warland. Their results suggested that number of children is a significant and positive determinants of usage of labels. The coefficient of age is statistically significant and positive in equations of doctors and books. For elderly households, they may visit doctors more often than young households. Therefore, they are more likely to obtain nutrition information from doctors than the latter. Additionally, elderly households may own high human capital and more readily access to newspapers, magazines, and books.

The age effect on information search is consistent with Feick, Herrmann, and Warland's find. They found that age was a significant and positive predictor of information search from magazines and newspapers, and books and pamphlets. Moorman (1990) also found that age is positively related to ability to process and comprehension accuracy of nutrition information. Different finding about the age effect was found by Cole and Balasubramanian (1993). They reported older consumers had difficulties in using nutrition information. They interpreted this conflict as the results of greatly different age subjects used in different studies. In addition, their small sample (79 consumers), interview methods, and grouping of consumers (younger consumer aged between 20 and 59 years and older consumer aged between 60 to 89 years) may be responsible to this contradiction. The controversial conclusion was also obtained by Moorman and Matulich (1993). They pointed out that age has a negative main effect on information acquisition.

Education increases the amount of nutrition information search from every sources. This is consistent with the finding of Feick, Herrmann, and Warland. They interpreted this finding as nutrition information was too technical or confusing to many consumers. Because previous research suggested that effectively designed nutrition disclosures facilitate information utilization (Moorman 1986), the information format and disclosures need to be improved. Consumer education program in improving comprehension of nutrition information may be necessitated.

Race is not a significant determinants for information search from doctors, but it is from the sources of television, books, and labels. The probabilities of obtaining nutrition information from television, books, and labels are higher for white households than for other racial groups. Both male and female headed households search more nutrition information from all sources than single headed households. This is a different finding from Feick, Herrmann, and Warland. They reported that marital status was not a significant predictor of information search from any source.

Consumer resident areas are also influencing consumer search for nutrition information from some sources. Compared with suburban residents, central city and non-metro consumers search more information from the sources of radio or television, and newspapers, magazines, or books. Non-metro households obtain nutrition information from food packages and labels more often than others. Compared with west region, households in northeast search information less often from the sources of radio or television, newspapers, magazines, or books, and food packages or labels. Midwest residents are more likely to obtain information from a doctor, nurse, or other health professional, and food packages or labels, but less likely to use the sources of radio or television, and newspapers, magazines, or books. Because most previous studies used regional survey data, no comparison can be made on these findings. The regional differences in information usage may indicate that consumer's accessibility or believes on information vary across regions.

TABLE 2

VARIABLE DEFINITION AND SELECTED SAMPLE STATISTICS (N=4250)

TABLE 3

ESTIMATED RESULTS OF CONSUMER INFORMATION USAGE (N=4192a)

Another plausible finding is that enduring importance of nutrition is a significant predictor in positive direction for all information sources. [Only fat consumption is included in logit analysis since cholesterol consumption is highly correlated with the fat consumption.] Consumers with low level of enduring importance of nutrition search less information from all sources than do the other consumers. This is consistent with previous studies. Feick, Herrmann, and Warland (1986) found that the importance of food consumption on future health was a significant predictor on nutrition information search from every sources in positive direction. Moorman and Matulich (1993) concluded that health motivation increases the amount of health information usage. Burton and Biswas (1993) also reported that the information about cholesterol and fat are major determinants of consumer perception of nutrition information. This finding is useful in consumer education program. For example, efforts in educating consumers about the effects of food consumption on future health will improve consumer usage of nutrition information.

TABLE 4

ESTIMATED RESULTS OF THE SUCCESS OF INFORMATION USAGE (N=4192a)

Turning to the question of whether consumers achieve their preferred information sources, the logit model was estimated again to analyze the determinants of success of information usage. The estimation results are presented in table 4. Generally, the models are not fitted as well as those of determinants of consumer usage. Out of a sample of 1385 households who search nutrition information from the source of radio and television, there are only 162, below 12%, households used their preferred information source. These facts suggest that nutrition information appeared in food advertisement in radio and television programs may not help food promotions for the food industry. Including nutrition information in television and radio programs may not be an effective way in educating consumers to make right food choices because they are more likely to ignore the information from this source.

The relatively poorer estimation results of the success equations suggest that majority consumer socio-economic and demographic characteristics such as income, family size, age, and household head status are not significant determinants in the success of consumer search for nutrition information. Although consumers do obtain nutrition information from various sources, it is far from the satisfactory of using them. This finding may imply that consumers have a difficulty time in finding their preferred information sources. This may be a major reason why consumers want and are willing to pay for nutrition information but neither acquire such information when making purchase decisions or comprehend most nutrition information once they receive it (Jacoby, Chestnut, and Silberma 1977). If consumers did not obtain nutrition information from their preferred sources, the usefulness of nutrition information they obtained and the benefits of searching nutrition information may be severely limited. For example, they may simply do not believe or understand the information they obtained so that they do not use the information in making food purchase decisions.

CONCLUSIONS

The current levels of consumer usage of nutrition information and factors related to the usage were discussed. Then the determinants of consumer usage of nutrition information were investigated by a qualitative response model. Data from a national survey, the 1987-88 Nationwide Food Consumption Survey (NFCS), was analyzed. Majority results are consistent with previous studies such as Feick, Herrmann, and Warland (1986), and Moorman (1990). Compared with previous studies, this study has several unique features. First, a more comprehensive data, the most recent and complete national survey, serves as the database while most previous studies used rather limited regional survey data. This makes the inclusion of food expenditure and regions of residence as explanatory variables possible.

Second, the consumption of fat and cholesterol is used to capture the effect of enduring importance of nutrition on information usage. In the literature, self-reporting health conditions and information index were often used and biased research results might be obtained. For example, self-reporting health condition is greatly affected by consumer characteristics. Some consumers may like to overstate while others may like to understate their health conditions. Finally, factors influencing consumer behavior in information search are thoroughly investigated. After estimating the determinants of consumer usage of nutrition information, this study further verifies whether consumers use their preferred sources.

The factors influencing consumer usage of nutrition information are identified successfully by the logit model. Food expenditure at-home, family size, age, education, household head status, and importance of nutrition are major predictors in positive direction in consumer information usage for all information sources. Income has a significant effects on information search but that varies across the sources. For example, high income consumers are more likely to use the source of newspapers, magazines, or books, and food packages or labels while they are less likely to use the source of a doctor, nurse, home economist or extension agent than do the low income consumers. The results also indicate that consumer information usage varies across urbanization areas and regions. These findings are useful for consumer education programs such as identifying the targets of consumer education programs and the design of information disclosures.

The results suggest that factors influencing consumer decision on information usage are not necessary to be significant determinants of the success of consumer information search. Consumers may not be able to find their preferred information sources. This finding reinforces the results of previous studies such as consumers' willingness to pay for nutrition information but unwilling to use it when making purchase decisions (Jacoby, Chestnut, and Silberman 1977, and Feick, Herrmann, and Warland 1986). Thus there is much need in improving nutrition information disclosures to facilitate consumer information utilization.

REFERENCES

Amemiya, T. (1981), "Qualitative Response Models: A Survey," Journal of Economic Literature, 19:1483-1536.

Arabsheibani, G.R. (1993), "The Determinants of Success in Educational Choice," Economics Letter, 42:411-417.

Bass, V.F. (1991), "Food Labeling and Consumer Satisfaction," Journal of Home Economics, 83(Spring):2-5.

Burton, S. and A. Biswas (1993), "Preliminary Assessment of Changes in Labels Required by the Nutrition Labeling and Education Act of 1990," The Journal of Consumer Affairs, 27(Summer):127-145.

Cole, C.A. and S.K. Balasubramanian (1993), "Age Differences in Consumers' Search for Information: Public Policy Implications," Journal of Consumer Research, 20(June):157-169.

Cole, C.A. and G.J. Gaeth (1990), "Cognitive and Age-Related Differences in the Ability to Use Nutritional Information in a Complex Environment," Journal of Marketing Research, XXVII(May):175-84.

Cramer, J.S. (1991), The Logit Model: An Introduction for Economists. Edward Arnold, A Division of Hodder and Stoughton Limited.

Daly, P. (1976), "The Response of Consumers to Nutrition Labeling," The Journal of Consumer Affairs, 10(Winter):170-178.

Feick, L.F., R.O. Herrmann and R.H. Warland (1986), "Search for Nutrition Information: A Probit Analysis of the Use of Different Information Sources," The Journal of Consumer Affairs,20(Winter):173-192.

Hauser, J.R., G.L. Urban and B.D. Weinberg (1993), "How Consumers Allocate Their Time When Searching for Information," Journal of Marketing Research, XXX(November):452-66.

Jacoby, J., R.W. Chestnut and W. Silberman (1977),"Consumer Use and Comprehension of Nutrition Information," Journal of Consumer Research, 4:119-128.

Jensen, H.H. and T. Kesavan (1993), "Sources of Information, Consumer Attitudes on Nutrition, and Consumption of Dairy Products," The Journal of Consumer Affairs, 27(Winter):357-76.

Lambert-Lagace, L. (1983), "Media, Nutrition Information, and Consumer Reactions," Journal of Nutrition Education, 15:6-7.

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

Levy, A.S., O. Matthews, M. Stephenson, J. E. Tenney and R.E. Shuker (1985), "The Impact of a Nutrition Information Program on Food Purchases," Journal of Public Policy and Marketing, 4:1-13.

Maddala, G.S. (1983), Limited-dependent and qualitative variables in econometrics, Cambridge University Press, New York, NY.

Moorman, C. (1990), "The Effects of Stimulus and Consumer Characteristics on the Utilization of Nutrition Information," Journal of Consumer Research, 17(December):362-74.

Moorman, C. and E. Matulich (1993), "A Model of Consumers' Preventive Health Behaviors: The Role of Health Motivation and Health Ability," Journal of Consumer Research, 20(September):208-228.

Navader, K.P. (1993), "Food and Nutrition Labeling: Past, Present and Future," Journal of Home Economics, 85(Summer):43-50.

Russo, J.E., R. Staelin, C.A. Nolan, G.J. Rusell and B.L. Metcalf (1984), "Nutrition Information in the Supermarket," Journal of Consumer Research, 13(June):48-70.

Russo, J.E., and F. Leclerc (1991), "Characteristics of Successful Product Information Programs," Journal of Social Issues, 47(1):73-92.

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

Urbany, J.E. (1986), "An Experimental Examination of the Economics of Information," Journal of Consumer Research, 13(September):257-271.

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