1000

Asia Pacific J Clin Nutr (1996) 5: 70-78

Which nutrition information do shoppers want on food labels?

Anthony Worsley, BSc(Hons) PhD

Department of Community Medicine, The University of Adelaide, Adelaide

---

Two surveys examined supermarket shoppers’ views of food label nutrition information terms. Approximately four out of five of the respondents were women. The first study, conducted among 941 shoppers in Sydney, Melbourne and Adelaide showed that information about cholesterol, fats, additives and health claims was perceived to be most important among 15 items of food label information. Extensive differences between the perceptions of members of different demographic groups were observed.

Principal components analysis of the ratings data derived three components which were named Positive Nutrition, Additives and Cholesterol, Calories and Claims. Women had significantly higher scores on all three components.

The second study of 631 shoppers in Sydney examined their ratings of the usefulness of 25 food and nutrition terms. The results confirmed the findings from the first study; information about negative as well as positive food constituents was perceived to be most useful and important. The study showed:

1. Divergence between consumers’ and experts’ views of the usefulness of label information; consumers were less interested in energy content, dietary fibre and more interested in a variety of other constituents such as cholesterol and flavourings.
2. Differences in desired label information between groups of consumers according to their gender, educational background, and other demographic characteristics.
3. Principal components analysis of the ratings broadly confirmed the findings of the first study: attitudes toward food label information were distributed along five components named Positive Nutrients, Additives, Fats, Salt and Sugars, and, Unfamiliar Concepts. Several statistically significant but small demographic differences were found.

The studies’ findings suggest that there is a need for food product labels which more fully reflect consumers’ perceptions of foods, especially information about "additives" and "negative nutrients. Negotiation is required between the different perspectives of consumer groups, regulators, nutrition educators and industry personnel about label design and content.

---

Introduction

There is a great deal of interest in food labels on the part of consumers, regulatory bodies and 1000 nutritionists. In the USA, the Nutrition Labeling and Education Act has mandated a revised form of nutrition information panel whilst the EC has issued a directive about the need for adequate nutrient labelling^l,2. Among researchers, many workers have assessed consumer reactions to a variety of nutrient information formats^3-14. Most of these studies have been limited to the presentation of the narrow range of nutrient information prescribed by Codex Alimentarius, ie energy, fat, carbohydrate, protein, salt content, and occasionally, vitamin and mineral content.

Some investigators, however, have examined consumers’ responses to more varied types of nutrition information on food labels^l5-21. These workers used quite different questions to assess, essentially, the perceived importance of the various nutrients. It is clear that fat, cholesterol, sugars and calories (with dietary fibre, salt/sodium) have been among the chief interests of many consumers in several countries.

Food opinion surveys, over the past two decades, have shown that consumers are interested in many aspects of food and nutrition. Several of them lie outside the orthodox nutrition agenda For example, many consumers are highly concerned about food additives and contaminants such as pesticide residues as well as some environmental effects of food production^l5,l6,22-38. These concerns appear to co-exist in consumers’ minds with more orthodox views about the dangers of high fat and salt diets for example³⁹.

More recently, several clusters of consumers’ concerns about food and health have been identified, including concerns about: safety and quality; additives; disease; general food system problems and regulatory concerns, as well as concern for helpless people and animals⁴⁰. Groups of consumers are likely to interpret food labels and nutrient meanings in different ways according to these and related standpoints. For example, Crawford and Baghurst have shown that the concept of fat is closely related by men to heart disease but for women it is linked more to personal appearance⁴¹. Similarly, ‘calories’ have been associated with weight control and ‘energy’ with health and vigour^39,42.

Since most food labels are inspected during shopping^2l,43 it is important to assess shoppers’ views of label information. To date, research has concentrated on consumers’ views of nutrients. Few opportunities have been provided to gauge shoppers’ perceptions of the relative importance of nutrients, other food constituents and health messages on food product labels. In addition, their responses to novel types of health and nutrition information which might be put on food labels have not been examined. This information is required before major revisions of food label regulations are made. Therefore two studies, were conducted in order to assess consumers’ views of label information items and the interrelationships between them. The first study assessed their views of ideal label information. The second study was a broad replication which assessed shoppers’ perceptions of the usefulness of information items.

Study 1

Aim: To examine shoppers’ desires for particular kinds of nutrition and health information on the ‘ideal’ food label.

Method

A short questionnaire was designed after discussions with small groups of consumers and after inspection of the food labelling literature. Respondents were asked whether they wanted each of 15 possible types of nutrition and food information on food labels (Figure 1, Table 2 for full details of the items; responses were Yes, No, Not sure) and to select 1000 the two most important items from this list.

Figure 1. Study 1: Shoppers’ choices of the most desired health information for food labels.

In addition questions were asked about the respondents’ views and use of current nutrition information on food labels and their concerns about food and health issues (reported elsewhere) as well as their demographic characteristics.

The study was conducted during March and April 1991, among clustered samples of supermarket shoppers in Sydney, Adelaide and Melbourne. Care was taken to select numbers of supermarkets according to the approximate market shares of the main retail chains. Sixty shoppers were randomly selected from each supermarket during peak shopping times according to a predetermined protocol (available from the author and similar to that reported elsewhere⁴⁴). The shoppers’ voluntary cooperation was elicited and the general purpose of the survey was explained. They were invited to complete the questionnaire at home and to return it via a free-post envelope. Their names and addresses were recorded so that up to two reminders could be sent to non-respondents at two weekly intervals thereafter. In all, 600 shoppers were selected from ten supermarkets in Sydney, 220 from four in Adelaide and 480 from eight in Melbourne.

The shoppers’ ratings of the desirability of the items on food labels were subjected to principal components analysis^45,46 in order to examine the interrelationships between the shoppers’ perceptions of the items. An alternative technique, non-ordinal multidimensional scaling^47-49 was also used to examine these relationships. The two techniques yielded broadly similar findings but for the sake of simplicity only the results of the principal components analyses will be reported here. Details of the multidimensional scaling findings are available from the author. The respondents’ scores on each of the principal components were calculated⁴⁶ and their dependence on several demographic variables Sex, Age, Presence/Absence of children under 18 years, Educational level (see Table 1) and Employed/Unemployed status were examined through a series of multiple regression analyses⁴⁶.

Results

Response rate and demographic characteristics.

Nine hundred and forty one shoppers returned completed questionnaires, a response rate of approximately 75%. The demographic characteristics of the pooled sample are shown in Table 1.

The most important and desired food label information.

Additives, health claims, cholesterol, irradiation and the amount of fat were seen to be the most important items of the listed food label information, and, the amounts of starch, complex carbohydrates, energy and protein were perceived as the least important items (Figure 1).

Overall, the percentages of respondents who rated the items as desirable closely paralleled the choice data. Bivariate analyses of the desirability ratings revealed several statistically significant differences between the demographic groups.

More women than men wanted the total amount of sugar, and the amounts of added sugar, dietary fibre, salt and starch, as well as health claims, and details of irradiation status on food labels (Table 2).

The youngest age group was least interested in the amount of starch or whether the food had 1000 been irradiated. (Amount of starch: 55% of 18-32 year olds, 64% of 33-44 year olds, 71% of 44 years and older, p< 0.001; irradiation status: 73% of 18-32 year olds, 84% of 33-44 year olds, 87% of 44 years and older, p< 0.0004.).

The tertiary educated group reported least interest in health claims, and, the amounts of starch, complex carbohydrate, dietary fibre and protein. However, the least and most educated groups expressed the greatest interest in irradiation status (Table 2).

Table 1. Demographic characteristics of the samples. Study 1 Sydney, Adelaide, Study 2, Sydney Melbourne (n=941, response rate 75%) (n=631, resp. rate 74%) Sex n % n % Women 729 77.8 550 87.2 Men 208 22.2 70 11.1 Not stated     11 1.2 Age groups         18-32 310 34.0 208 33.8 33-44 310 34.0 195 31.4 over 44 293 32.1 214 34.8 Marital Status         Single/divorced/sep 307 32.9 142 22.5 Married/de facto 626 67.0 484 76.7 Not stated 8 0.8 5 0.8 Educational groupings Year 8-10 360 38.3 268 42.4 Year 11/12/Tech qual 300 31.9 200 31.7 Tertiary 266 28.3 152 24.1 Not stated 15 1.6 11 1.8 Dependents under 18 years No 497 52.8 272 43.1 Yes 418 44.4 335 53.1 Not stated 26 2.8 24 3.9 Shopping Status         Main shopper 587 62.4 484 76.7 Joint shopper 263 27.9 101 16.0 Not main shopper 85 9.0 42 6.7 Not stated 6 0.6 4 0.9 Employment categories In paid work 634 67.4 346 54.8 Unpaid work in home 286 30.4 260 41.2 Not stated 21 2.2 25 4.4 Respondents were asked: ‘Getting down to details, what health information would you like to see on food products? Circle one answer next to each item. Circle ? if you are Not Sure.’ Then followed the list of items above; the responses Yes, No and ? were headed "Do you want it?’ The figures in bold at the head of each column are the numbers of respondents 1000 in each group. The figures in the columns are the percentages of each group endorsing the items. *p<0.05; **p<0.01; ***p<0.001; ****p<0.0001.

Table 2. Study 1: Health information wanted by shoppers on food labels. Sex and Educational Group Differences Information % Women 721 % Men 203 P Number of calories or kilojoules 86 76 * Health claims (eg. reduced fat) 94 87 *** Amount of fat 93 89 NS Amount of added sugar 89 80 ** Total amount of sugar 87 76 **** Details of additives 95 91 NS Details of vitamins and minerals 84 80 NS Amount of starch 66 54 < 1000 p align="center">** Amount of complex carbohydrate 73 65 NS Amount of dietary fibre 86 71 **** Amount of protein 84 76 ** Whether food has been irradiated 83 77 * The salt content of the product 91 83 *** Amount of cholesterol 90 88 NS Amount of energy in the product 69 67 NS 1000 Information Education Groups   <16yr <18yr Tert P   353 300 [266]   Number of calories/ kilojoules 86 84 80 NS Health claims (eg reduced fat) 97 93 88 *** Amount of fat 93 92 92 NS Amount of added sugar 88 86 85 1000 NS Total amount of sugar 87 83 82 NS Details of additives 93 92 97 NS Details of vitamins and minerals 84 80 85 NS Amount of starch 74 58 54 **** Amount of complex carbohydrate 77 67 68 * Amount of dietary fibre 87 77 82< 1000 /p> ** Amount of protein 87 78 81 ** Whether food has been irradiated 86 75 83 **** Salt content of the product 88 88 91 NS Amount of cholesterol 90 89 89 NS Amount of energy in product 72 66 66 NS

Interrelationships between shoppers’ perceptions of food label information.

The principal components analysis yielded three components which accounted for 49.9% of the intercorrelation matrix variance. Items to do with "Positive Nutrition" loaded on the first component, "Negative Nutrition" items on the second and "Cholesterol, Claims and Calories" items on the third (Table 3). Calories/kilojoules loaded on both the Positive Nutrition and the Cholesterol, Claims and Calories factors to moderate extents. Inspection of the multi-dimensional scaling findings confirmed this and suggested that calories/kilojoules were seen by women as having links to positive nutrients (as "energy") as well as links to Fat (as "fattening").

The regression analyses showed that the demographic factors explained minimal amounts of the variance in the component scores. However, women had higher scores than men on all three components (Table 6).

Discussion

The high demand for information about fats, calories and cholesterol confirm findings from previous surveys of consumers’ responses to label information^l5-2l. In addition the widespread demand for information about additives, irradiation status and health messages is consistent with the results of a number of surveys of consumers’ health concerns^{20,22-24,26-36,40}. The strong preference for health claims on labels was supported by the finding from Study 2 in which respondents indicated that examples of seven health (and nutrient) claims (eg ‘low cholesterol’) were ‘helpful’⁴³.

The results show that there is interest in ‘orthodox’ nutrition issues such as fat and salt intake but these are accompanied by interest in other issues (eg additives, irradiation, perhaps calories/kilojoules) which are not part of the ‘orthodox’ nutritional agenda. In addition important nutritional concepts such as complex carbohydrate, protein, starch and energy appeared to be of little relevance to the shoppers.

This interest in negative ingredients is consistent with the work of Payne et al who found consumers check labels (on household chemicals) first for things that may harm them⁵⁰. Regulatory authorities may need to consider whether they should include such information into new food label designs.

The support for health claims which are currently not allowed in many countries (eg Australia, New Zealand, the European Community), again supports the information processing viewpoint espoused by several workers^4,50,51. Health claims quickly draw attention to key product attributes or benefits. The prominence of health claims on the third ‘Cholesterol, Claims and Calories’ component suggests they are associated in consumers’ minds with warnings about ‘negative’ nutrients less familiar. The lesser interest of the tertiary educated in health claims and the greater interest of early school leavers in health logos demonstrated in Study 2⁴³ suggests they may have greater utility for 1000 people with less knowledge of, or confidence about, food ingredients.

Two items on the current Codex label were not popular: energy and protein, nor were two synonyms of carbohydrate - complex carbohydrate and starch. This suggests either that these are redundant terms which should be removed, or, that regulatory authorities should investigate ways of raising the public’s awareness of their importance.

The derivation of three principal components suggests that label designers should take into account consumers’ wishes for information about food constituents which yield benefits ("Positive Nutrition") as well as those which are perceived to pose some risk to them ("Negative Nutrition", "Cholesterol, Claims and Calories"). The similar loadings of several items on the same component suggests that consumers do not discriminate finely between them. Consumer knowledge may be a rough and ready affair.

Study 2

Aims: Study 1 examined shoppers’ wishes for a small range of nutritional terms on food labels. The main aim of the second study was to examine shoppers’ perceptions of the usefulness and importance of a wider range of information about nutrients and other food constituents. This included proposed label information as well as information which is often present on food labels.

Method

A short questionnaire included questions about the usefulness of 25 nutritional terms (Figure 2), as well as questions about the respondents’ use and understanding of health messages and ingredients lists (reported elsewhere⁴⁹) and demographic information. After rating the usefulness of the nutritional terms the respondents chose up to three most important items from the list.

Table 3. Study 1: Summary of the principal components analysis of shoppers’ ratings of desired label content. Positive Nutrition   Amount of complex carbohydrate 74 Amount of protein 74 Details of vitamins and minerals 69 Amount of starch 67 Amount of dietary fibre 65 Amount of energy in the product 62 Number of calories or kilojoules 43 Eigenvalue (% variance) 5.13 (34.2) Cronbach’s alpha 0.83 Negative Nutrition   Amount of added sugar 70 Total amount of sugar 69 Salt content of the product 62 Amount of fat 59 Details of additives 53 Whether the food has been irradiated* 41 Amount of cholesterol 38 Eigenvalue (% variance) 1.33 (8.9) Cronbach’s alpha 0.72 Cholesterol, Claims, Calories   Amount of cholesterol 60 Health claims (eg reduced fat, natural, no additives)** 59 Amount of fat 52 Number of calories or kilojoules 50 Eigenvalue (% variance) 1.01 (6.8) Cronbach’s alpha 0.59 * Cronbach’s alpha =0.75 if this item deleted. **Cronbach’s alpha =0.61 if this item deleted.

Figure 2. Study 2: Shoppers’ first choices of nutrient label information.

The questionnaires were administered to 900 shoppers at 15 supermarkets in Sydney during May and June 1991, using similar methods to those employed in Study 1. As in the first study the resulting data were examined via contingency table analyses, principal components and multidimensional scaling analyses of the inter-item correlation matrices, followed by regression analyses of the respondents’ scores on the principal components.

Results

Response rate and demographic characteristics.

Six hundred and thirty one shoppers returned completed questionnaires, a response rate of approximately 74 percent The respondents were demographically similar to those of Study 1 (Table 1).

The shoppers’ views of the usefulness and importance of nutrition information.

Again, the shoppers indicated that they found some items of food information more useful than others. Cholesterol, pesticides, preservatives, additives, total sugars and total fats headed the list of ‘very useful’ information whilst details of emulsifiers, complex carbohydrates, energy used in production and drapule were least popular. (Drapule is a fictitious term included to assess acquiescence set - the tendency to agree with items irrespective of their meaning.) These aggregate ratings were similar to 1000 the rankings which were derived from shoppers’ choices of the most important items out of the list of 25 items (Figure 2). The rank order of these was similar to that found in Study 1.

Generally, women indicated that food and nutrition information was more useful than men did (Table 4); a similar gender difference to that found in Study 1. For example, more women rated over half of the items as "very useful". Similar sex differences were observed in the choices of important nutrient information of men and women. More men chose cholesterol, total fats, total sugars and price per hundred grams as items of most importance.

Table 4. Study 2: Shoppers’ views of the usefulness of nutrition information on food labels (% ‘Very Useful’): Sex, Age and Education Group differences. Food constituents Sex   Age Groups     F 545 M 70 p 18-34 207 34-44 195 >44 215 p Calcium 39 23 ** 37 33 42 < 1000 p align="center">NS Carbohydrate 32 28 NS 31 27 36 NS Complex carbohydrate 27 23 NS 27 21 32 ** Cholesterol 66 63 NS 56 63 77 ***** Colourings 53 29 *** 53(18) 58(14) 43(24) ** Dietary fibre 50 30 ** 43 48 51 NS Emulsifiers 28 16 * 26(31) 26(15) 29(25) *** Energy content (calories/kJ) 43 32 NS 45 37 43 NS "E" (food additives) numbers 48 28 **** 42(29) 51(15) 45(19) *** Energy used in production 9 6 NS 9 8 10 * Total Fats 57 53 NS 51 55 65 ** Flavourings 47 29 ** 48 48 n NS Iron 36