Defining the Problem: the Scope of Consumer Concern With Food Labeling

ABSTRACT - Some of the major data-gathering approaches used by the Federal Government in its examination of consumer perceptions of food labeling are described, as well as some of the overall findings. These primary findings are discussed in terms of their implications for redesigning food labels to make then more appropriate to consumer needs and are efficient as communication tools.

Citation:

James T. Heimbach (1981) ,"Defining the Problem: the Scope of Consumer Concern With Food Labeling", in NA - Advances in Consumer Research Volume 08, eds. Kent B. Monroe, Ann Abor, MI : Association for Consumer Research, Pages: 474-476.

Advances in Consumer Research Volume 8, 1981      Pages 474-476

DEFINING THE PROBLEM: THE SCOPE OF CONSUMER CONCERN WITH FOOD LABELING

James T. Heimbach, Food & Drug Administration

ABSTRACT -

Some of the major data-gathering approaches used by the Federal Government in its examination of consumer perceptions of food labeling are described, as well as some of the overall findings. These primary findings are discussed in terms of their implications for redesigning food labels to make then more appropriate to consumer needs and are efficient as communication tools.

.   .    .   . 

Food is a uniquely important commodity. It is an absolutely necessary one: no one can avoid consuming it and very few people in this country can avoid buying at least some of their food through the manufacturing/distribution network.

At the same time, food is highly complex and includes a vast variety beyond that which we can directly grow, pick or kill. Foods are composed of a plethora of substances. Some of these we need; some we should avoid; others we should limit. Different people have different needs for these substances, and different people have different tolerances for them.

Buying food is a complex activity, because the appearance, the texture, even the taste will not usually tell us much about its important characteristics. How can we help people to know the nature of the foods they buy?

The primary medium, of course, is the food label. Food labels carry a truly incredible amount of information about the contents of the container: what it is, how such is there, how much it costs, who made it, what ingredients are in it, what nutrition it provides, when it was packed or when it should be consumed, etc.

But if this panoply of information cannot be understood, if it cannot be used by people who want to use it to make the decisions they want to make, then it is worthless.

A problem, however, is that we often do not know what information consumers really want or why they want it. There has been extensive speculation, for example, that people are becoming increasingly interested in the nutritional values of their foods as one aspect of a more general involvement with physical fitness. Possibly related to this, there appears to be a sort of "back to nature" movement, involving both an outright fear of technology and a desire for more "natural" products. Assuming these trends do exist, to what extent do they translate to need for and use of label information and of what kinds of information?

Another factor is the consumer's "right to know." Virtually all consumer advocates, and possibly many consumers, feel that a person buying a product has a right to know all the characteristics of that product, whether or not he or she has any specific use for the information or will make any decisions based upon it. To what extent does this philosophy affect labeling concerns?

What about the effects of medical successes which have drastically increased life expectancy? Has this trend increased awareness and concern about low--level long-term threats to health -- the kind of thing that does no immediate harm, but may cause problems after sixty or seventy years' exposure? Are there implications for food labeling from this phenomenon ?

These factors, and others, mean that we cannot rely on the old food label system to provide in any adequate form the information consumers want and need in ways they want and need. What is needed is to look at the food label as a holistic entity, existing and functioning in a dynamic way against the backdrop of the concerns and ambitions of our consumer society.

In 1978, the Food and Drug Administration (FDA), along with the U.S. Department of Agriculture (USDA) and the Federal Trade Commission's Bureau of Consumer Protection (FTC) embarked upon a comprehensive program to evaluate the effectiveness of current food labeling and to revise food labels as necessary. The subjects of primary interest are the listing of ingredients and the nutrition label, although the entire label - as a whole - comes under scrutiny.

This paper addresses the first phase of the program: the evaluation of the nature and scope of consumer concern with food labeling and identification of problems. Later papers will discuss the remaining phases of the research and development program.

Data have been incorporated in this paper from a number of sources. The primary source is the FDA 1978 Consumer Food Labeling Survey (Heimbach and Stokes 1979), which included personal interviews with 1974 "primary food shoppers" (the person in each household who does moat of the food shopping). The design was a three-stage national probability cluster sample, and the field-work was performed by Response Analysis Corporation of Princeton, NJ, in the fall of 1978.

The second major source of data is a series of public hearings held in five cities across the country in the fall of 1978, along with associated written comments from consumer groups, food and nutrition professionals, and food companies and trade associations. Over 450 individuals testified at the FDA/USDA/FTC-sponsored hearings, and the agencies' widely-publicized request for consents attracted an additional 9,400 written comments. Several hundred comments were subjected to content analysis, from which a coding system was developed and applied to the total of 9,864 oral and written comments by a team from the three agencies (U.S., Department of Health, Education and Welfare 1979).

Two additional sources of data were also used. In the fall of 1978, the then-Commissioner of FDA appeared on QUBE, the interactive cable TV system in Columbus, Ohio. In the course of his discussion of food labeling, he asked a number of multiple-choice questions about viewers' opinions. Over three thousand viewers responded, "touching in" on their home consoles, and QUBE's computer produced instantaneous tabulations (Hecht 1979). Finally, another FDA-sponsored national consumer survey completed in February, 1980, by Market Pacts, Inc., also included a number of questions dealing with food labeling (Heimbach 1980).

A few words would be in order about the data-collection techniques. There is nothing new, unusual, or controversial about using survey-research data generated from national probability samples. The QUBE data and those from the hearings and written comments, on the other hand, are known to be biased: there is a large degree of respondent self-selection in these data sources. Those who chose to watch the QUBE program on food labeling and to respond to questions, and even more those who took the time to testify at public hearings or to write letters to the Government, are different: they are more interested in food labeling than the "average" consumer, they are more concerned and dissatisfied with labeling, and they tend to be more literate and more highly educated. What then is the value of these forms of data collection?

There are two answers to this question. The first is political: these people are the "squeaky wheels" who make the system work, and to whom the Government must be responsive. It is no news to anyone that governments are more influenced by a minority who care devoutly about an issue than by a majority who do not care at all or who care very little.

The second point is that these are often the consumers with real needs. Most of us can - and do - walk into a grocery store and pull packages off the shelf with little regard for anything but the price. Many people, however, are not so fortunate: there are diabetics and prediabetics who must control their intake of sugars, hypertensives who must limit - sometimes drastically - their sodium consumption, people with cardiovascular disease who must watch cholesterol and saturated fats, pregnant and lactating women who must increase their intake of iron, elderly people who suffer calcium deficiency, people with allergies or intolerances for almost any substance imaginable, and so on. There are also numerous people who believe that it is healthier for them to eat more "natural" foods and to avoid artificial flavors and colors and preservatives; who are we to tell them they cannot do this?

Some of these needs can be extremely rare statistically and yet still be important for national policy. There are over 200,000,000 people in this country: a need which exists for only 0.01% of the public still represents over 20,000 people. Yet what is the probability that even one person with a given such low-frequency need will appear in a survey of 1,500 people? It is easy to calculate: the answer is 14% - about one chance in seven. (Even if it is assumed that each respondent can report needs of anyone in the household, the probability is still under 50%.)

Consequently, survey research using any reasonable sample sizes simply will not help us to identify many food problems or their associated labeling needs. At the same time, there is an obvious problem with relying upon needs expressed by a self-selected sample: these writers not only possess needs, they also possess literacy and the ability to process information, Many of the things they propose would not be comprehensible to other consumers who may share the problem but not the education or intelligence.

A good example is percentages. A large proportion of those who wrote in with suggestions advocated percentage ingredient labeling. But many people cannot cope with percentages. For example, in our 1979 survey we showed respondents a square divided into four quadrants, one black and three white, and asked what percentage of the square was white. Only 46% said 75% (and only 23% of those who have not completed high school). Similarly, only about two consumers in three know that if a serving provides 25% of your needs for vitamin A it would require four servings to get 100% (Heimbach 1980).

Finally, a brief statement of the obvious before we get into actual data. "Concern" with food labeling, like most other things, is not an all-or-none phenomenon, but rather a continuum. We have found that the Likert-scale approach ("extremely concerned", "somewhat concerned", etc.) does not satisfactorily express this continuum, probably because it is too tied to social desirability. Rather, we have taken the approach of moving from general free-response questions to specific direct questions, with the assumption that relatively uncued expressions of concern or of label-information needs represent stronger feelings than do those elicited only by direct questioning.

The first question to examine is that of the extent and depth of general dissatisfaction with current food labeling. The 1978 survey began by asking respondents:

"Aside from prices, please tell me about any particular problems, difficulties or concerns which you have with food these days."

It is assumed that mentions of dissatisfaction with labeling expressed in response to this exceedingly general question -which does not even mention food labels - reveal rather strong concern with existing labeling. Overall, about 8% of respondents had something to say about labeling, most often about ingredient listing. This climbs to 10% of those who have a food-related health problem in the household, and to about 20% of the highly-educated respondents who (as revealed in later questioning) pay most attention to labels).

The most frequently voiced complaints dealt with desired information not always being available; especially calorie content, listing of ingredients, and open-dating. There were also scattered mentions of displeasure with the complexity of label information and the use of chemical and technical terminology.

A little later in the interview the following question was asked:

"Are you satisfied with the kind or amount of information that's printed on food packages and cans, or can it be improved on in some way?"

This question, being considerably more direct than the previous one, naturally elicited responses from consumers with lesser degrees of concern or dissatisfaction with food labeling. One third (33%) of all respondents indicated less than complete satisfaction with current labeling. A follow-up question revealed that again the ingredient list (or occasionally the absence of an ingredient list) accounted for the majority of all complaints, although the desire to have information about calorie content on all products was also one of the most frequently offered single responses.

The largest source of dissatisfaction with the ingredient list is its failure to include all ingredients, i.e., the fact that it sometimes uses generic terms like "spices" or "artificial flavors." This problem was mentioned by nearly 20% of those who expressed dissatisfaction with any aspect of food labeling or about 5% of all consumers.

The next most frequent request with regard to ingredient listing is to provide quantitative information to tell how much of each ingredient - especially sugar - is in the product. This was mentioned by about 16% of those who were dissatisfied with current food labeling.

Finally, the third most frequent problem with the ingredient list - named by about 13% of those who complained about food labeling - is that the ingredient list is too complicated and uses too much technical or chemical terminology.

As an aside, it might be noted that this type of problem recurs throughout the research on labeling. Some people want more information, including information of an unavoidably technical nature, while others - or often the same people -want it all simplified. The two desires are obviously somewhat incompatible. The major goal of the label development program is to somehow provide necessary technical information in sufficient detail for potential users, and at the same time avoid overloading those who will be so repelled by so much information that they will abandon attempting to look at labels at all.

The only frequently mentioned desire other than those related to the ingredient list is to have information about calorie content available on all foods, expressed by 15% of those who were dissatisfied with current labeling.

These are not strikingly large numbers, but it must be noted that they represent only consumers with rather high levels of dissatisfaction. When those questions are asked more directly the proportion advocating changes rises substantially. Demand for quantitative ingredient listing, for example, climbs to 74% under direct questioning,

The reasons for these desires for more detailed information about ingredients are not far to seek. One of the most disturbing single findings from the research program is that over half (54%) of all consumers - and 70% of those who ever look at ingredient lists - report using the ingredients information for the specific purpose of avoiding consumption of a particular ingredient. In some cases the person only desires to limit consumption, while in others the avoidance is total; in either case, fear of adverse health consequences from consuming the ingredient is the dominant concern, although weight-loss diets, taste, nutritional concerns, etc., are also mentioned occasionally.

By a large margin the ingredient of most concern - and the most avoided - is sugar, which over 25% of all consumers report trying to avoid totally or to some degree. [This, by the way, again points to the differences in levels of concern expressed in response to open vs. direct questioning: when consumers were asked directly if they were attempting to modify their consumption of any of a list of two dozen substances, 80% claimed that they are making an effort to reduce their intake of sugar.] This is followed by salt, preservatives, artificial colors and flavors, and fats, and there is at least one person attempting to avoid or reduce consumption of almost any substance which ever occurs in food.

At this point let us put the probability sample aside and turn to the 10,000 comments from consumers who chose to write to FDA or testify at the hearings. One mild surprise, considering the wholly unrepresentative nature of this sample, was the high correspondence which actually existed between the two approaches.

Naturally, all of the people who testified or commented are concerned to some extent about food labeling -- usually to a pretty great extent -- and consequently exhibit far higher frequencies of specific complaints, but their overall pattern of responses was surprisingly similar to that obtained in the consumer survey.

TABLE 1

PERCENT OF CONSUMERS REPORTS VARIOUS NEEDS REGARDING LABELING

Again, the ingredient list was the major source of problems -- it was discussed by 85% of commenters -- with the nutrition label second, being addressed by 52%. Indeed, the rank-order correlation between the top specific concerns from the two approaches was .49, with only two specific complaints accounting for most of the difference.

The first of these is the demand for ingredient listing to be extended to all products, which was the 2nd most frequent desire among commenters but the 6th among all consumers. The other is the need to simplify the ingredient list, which was virtually never mentioned by commenters but was the 4th most frequent concern of those in the consumer survey. In point of fact, this was the single complaint which was mentioned more frequently by "average consumers" in absolute as well as relative terms.

In many cases, it is not clear whether differences in desires for label changes between survey respondents and commenters are due primarily to the higher level of sophistication of the latter or to the fact that people with diet-related health problems are over-represented in this group. (Unfortunately, many letters did not explain the state of health of the writer nor his/her reason for writing; however, as noted in the report on the analyses of comments (DHEW 1979), "Those who wanted more ingredient information on food labels often cited a specific health problem as the basis for their request.") For example, commenters were both absolutely and relatively more likely to demand that the sodium or salt content of a food be stated on the package, and that the specific type of fat or oil be given (as opposed to the current "contains one or more of the following"). These data are useful primarily to those suffering from hypertension or cardiovascular diseases, but the survey did shows distinct correlation between education and income and desire for all type of additional label information.

When the data from the survey are compared with that from the analysis of comments in terms of the overall patterns, however, the impact becomes quite clear. Commenters show both absolutely and relatively more desire for information about disease-linked food characteristics: sodium and potassium content, cholesterol, saturated fat, and specific sources of fats. Survey respondents, on the other hand, exhibit greater concern with the overall information load: they want such things as calorie information and a list of ingredients to appear on more products, but they also stress that the label is already too complicated and too technical and should not be made more so.

There is much mots detailed information about consumer desires available from the survey, the hearings, and the written comments, but this is not the place for discussion of all of the findings, and some of them will be addressed in following papers. Rather, let us look at the overall conclusions and their implications.

First, most consumers do not appear to be strongly motivated by a philosophy of "right to know." They are essentially pragmatic, and the information they want on food labels is that for which they have use. There is some recognition that too much information can be as bad as too little.

Secondly, they are not often motivated to use current label infuriation or to request additional information by nutritional considerations. Indeed, the desire to obtain good nutrition from food appears to take a back seat to the fear of consuming substances which are perceived to pose a threat to one's health. Probably the key finding of this phase on the research program was that, where food labels are concerned, consumers are oriented almost wholly toward avoidance of risk rather than seeking of benefits. Use of the ingredient list is primarily directed toward avoiding possibly hazardous ingredients rather than toward obtaining desired ones: sugar, salt and preservatives dominate the reports of what ingredient-list information is most important. Of the twelve types of information on the nutrition label rated as "very useful" by at least 50% of survey respondents, only three - protein, vitamin C, and iron -are sought while the other nine are avoided. (This is particularly significant when it is noted that thirty-eight nutrients were rated, and these nine were the only "things to be avoided" in the group -- the other 29 are all substances seen as "good for you.") Given a choice of having information about things people often get too little of vs. information about things people often get too much of, survey respondents chose the latter by a 2-to-1 margin. [This discussion is not to say that consumers do not seek foods with good nutritional characteristics, only that they do not often use food labels to do it. My hypothesis is that consumers decide tentatively to purchase a food on other grounds, such as advertising, package appearance, taste, nutrition, friends' recommendations, etc., and use the label as a final hurdle; i.e., they check the label to make sure there is no good reason not to buy the product.]

These two points together point to the conclusion that the determination of the information to include on food labels should be based primarily upon consideration of the relationship of the information to health problems and concerns. That is, top priority must be placed upon information about substances which are known to impact upon diet-related diseases and diabetes. The next priority should be those substances to which people are most commonly or most severely allergic or intolerant, and to those - such as preservatives - which do not appear to be related to health, but which many people are concerned about regardless. Other information, while its inclusion might be laudable, must be considered dispensable in comparison.

Third, information about these substances must be presented in the correct level of detail. For example, fats must be broken into saturated vs. polyunsaturated because of the known health effects, while sugars are pretty much sugars: despite much publicity about fructose and other varieties of sugars, they are all about the same to diabetics - and to teeth.

Similarly, the amount of sugar, or sodium, or saturated fat is of concern and should be available, while that of preservatives or artificial colors is not and should not. Beside the fact that avoiders of these are not working under a set tolerance level in the way, say, hypertensives are with sodium, the amounts are tiny in any one package and thus require complicated and confusing quantitative statements.

Finally, it is evident that many consumers find currant labeling incomprehensible. Only 10% have any idea of what "U.S. FDA" means (accepting any statement to the effect that it refers to the amount people should consume as correct), a rather basic necessity for understanding of much of the nutrition label; only about 75% know what "net weight" means; words like "riboflavin," "hydrogenated," and "polyunsaturated" are confusing to nearly everyone (Heimbach 1980). And this refers to the total population: the plight of the socially and educationally deprived is obviously far worse. Consumers themselves are aware of this when they demand that labeling be simplified, although they rarely realize quite how wide the gap is, probably because they have simply tuned out much of the terminology. Ignoring something is the sincerest form of rejection.

It is clear that severe constraints must be placed upon the amount of information placed upon labels, that this information must be directed toward real diet-related problem, that it must not be presented at a level of detail beyond what is really needed, and that every effort must be made to find alternative ways of expressing the information in terms that will be comprehended by at least a majority of the population.

The next five papers will discuss in more detail the Government's efforts to fulfill these goals.

REFERENCES

Health, Education and Welfare, Dept. of (1979), "Food Labeling: Report on the Analysis of Comments." Washington, D.C.: DHEW.

Hecht, A. (1979), "Consumers Speak on Cue, On 'Qube'." FDA Consumer, March.

Heimbach, J. T. (1980),"Report on the 1979 FDA Consumer Survey." (In preparation. )

Heimbach, J. T. and R. C. Stokes (1979), "FDA 1978 Consumer Food Labeling Survey." Washington, D.C.: Department of Health, Education and Welfare.

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