What Do You Learn Standing in a Supermarket Aisle?

K. W. Kendall, Simon Fraser University
Ian Fenwick, Dalhousie University
ABSTRACT - During a five day period, over 2,300 shoppers were observed in supermarket aisles for 44 hours to determine the prevalence of label reading, the characteristics of the shoppers who read the labels, and the information format trade offs they would be willing to make. For the attention foods, 43% viewed the products for more than 8 seconds and had a mean viewing time of over 38 seconds. Discriminant analysis correctly classified between 75 and 90% of the defined shopper classifications. Grabbers and lookers showed marked differences in utility for both amount and type of information.
[ to cite ]:
K. W. Kendall and Ian Fenwick (1979) ,"What Do You Learn Standing in a Supermarket Aisle?", in NA - Advances in Consumer Research Volume 06, eds. William L. Wilkie, Ann Abor, MI : Association for Consumer Research, Pages: 153-160.

Advances in Consumer Research Volume 6, 1979      Pages 153-160

WHAT DO YOU LEARN STANDING IN A SUPERMARKET AISLE?

K. W. Kendall, Simon Fraser University

Ian Fenwick, Dalhousie University

ABSTRACT -

During a five day period, over 2,300 shoppers were observed in supermarket aisles for 44 hours to determine the prevalence of label reading, the characteristics of the shoppers who read the labels, and the information format trade offs they would be willing to make. For the attention foods, 43% viewed the products for more than 8 seconds and had a mean viewing time of over 38 seconds. Discriminant analysis correctly classified between 75 and 90% of the defined shopper classifications. Grabbers and lookers showed marked differences in utility for both amount and type of information.

The Provision of Consumer Information

Since long before the Fair Packaging and Labeling Act of 1966 in the U.S. and the revised Packaging and Labeling Act in Canada, consumer groups had demanded further information be available on food products so consumers could make better food choices. Better information has come in the form of unit pricing, ingredient information lists, open dating and the presentation of nutrient content, etc. The provision of this information was heralded as the solution to the information gap for consumers.

However, some early work on information load in a laboratory setting (Jacoby, Speller and Kohn, 1974a, 1974b), suggested that more information produced "dysfunctional consequences" for consumers. A key issue involved the definition of information, e.g. more brands as more information or more information on each product. Several researchers have pointed to this problem (Russo, 1974; Summers, 1974; Wilkie, 1974). In fact, Russo (1974) Suggests that more information is better since evoked set size is usually found to be quite small (Howard, 1977).

Friedman (1972), reviewing the empirical literature on consumer use of information aids in supermarkets and on food labels, found most studies to have identified relatively few regular users of the three aids most extensively researched: unit pricing, open dating and nutrient labeling. Friedman's conclusions are somewhat misleading. At the time of the review a lot of the aids were relatively new. Furthermore, the dependent variables were structured in a normative sense of "best choice". Friedman defines regular use by whether or not consumers could correctly identify or reason through specific names of the new aids. More recently, the same author (Friedman, 1977) has reported little reason to alter those largely negative findings.

Reviewing similar work, Ross (1974) has suggested that a hierarchical relationship exists. Many people are aware of the information aids but few report using them. Day (1976) and Jacoby, et al (1977) echo this hypothesis of a hierarchy of effects. Day points to the generally large number of consumers who appear aware of the information in supermarkets and on the labels of food but the small number of people who claim to use this information. Interestingly, the hierarchy has not been tested in the stores. Instead, empirical studies rely on survey self-reports or laboratory situations with highly motivated subjects. There is little actual verification in the field.

The low usage of label and supermarket information aids is frequently blamed on the consumers' lack of comprehension of such information as noted by Daly (1976) and Jacoby, et al (1977). This has been true of open dating methods (Taylor, 1976), ingredient names (Warland and Herrmann, 1971), unit pricing measures (King and Gideon, 1971) and nutrient information (FDA, 1973, 1975; Kendall 1977).

As a result, the most recent emphasis has been on methods of presenting information and information processing formats (Kendall, 1977; Scammon, 1977; Bettman and Kakkar, 1977; Russo, Krieser, and Miyashita, 1975). It should be emphasized that many public policy decisions are presumably based on survey self-report evidence (FDA, 1973, 1975) and assume the hierarchy outlined by Ross (1974) and Day (1976). However, Day himself does not pretend to present an extensive literature review. And it is not at all clear that the evidence does support the hierarchy hypothesis.

A more extensive literature review on consumers' importance ratings for information aids shows that price information tends to be most important followed by brand name and then nutrition values in the aggregate data (FDA, 1975, 1973; CRI/FDA, 1972; Lenahan, et. al., 1972; Darden and French, 1971). These studies also show a trend more recently of evaluating nutrition cues more heavily than price users.

Even more emphatic are the reports on use, or probability of use, of information items on the labels of food products. Two points are of extreme interest. First, the more recent the study, the higher the reported use of different information aids (Jacoby, et. al., 1976; FDA, 1975; Babcock and Murphy, 1973; FDA, 1973; Lichtenstein, 1972; Lenahan, et al. 1972; CRI/FDA, 1972). For example, the CRI/FDA (1972) study found only 24% of the respondents reported checking the ingredients list. By 1975, this figure had more than doubled (FDA, 1975). [For further details on these figures, see Kendall and Fenwick, Simon Fraser Univ.-Working Paper, 1978.] Second, claimed usage rates are not as much below the importance ratings as hierarchy models would Suggest.

Furthermore, the Redbook Nutrition Study (1974) found almost half their sample claiming regularly to read food content information, and a further 45% claiming "occasionally" to read such information. For new products claimed readership was rather higher, 60% regularly reading, 36% occasionally reading. More recently, an even higher figure is reported, 66% regular reading (Redbook Nutrition Study, 1976).

However, there appears to be little evidence from the market place to determine how many consumers look at the information on products, the kind of people who actually look at labels or their preferences for label information formats in the store. Best and McCullough (1977) have looked at the consumer utility of label information with conjoint measurement, but did not tie their preference measures to any strong behavior measures. Furthermore, Day (1976) suggests that there is a pressing need for field research rather than laboratory studies in this area.

The studies reported here were designed to accomplish three tasks. First, the actual number of people who look at labels and the amount of time they spend looking at food products in a regular supermarket setting was determined. The emphasis on the time construct has recently been noted in the literature (Jacoby, Szybillo and Berning, 1976). Time estimates from actual shopping observations should also be helpful in verifying laboratory studies on information processing of information formats.

The second purpose of the studies was to determine the characteristics and situations of those consumers who look at food products and those that do not look at food products. Finally, the studies were developed to measure the utility of different information formats for a new convenience food product for the different shopper groups.

Methodology

The data analyzed here were collected in two separate studies. Both were run in Halifax, N.S. where the only required label information is a list of ingredients in descending order of proportion (Health Protection and Food Laws, 1970). The studies were conducted in two of the larger supermarkets in the metropolitan area, one from each of the major chains. Although basic procedures were identical in both studies, the products examined differed and there were minor differences in the questionnaires used. In what follows the particular products studied will be referred to as the attention foods or products.

In both studies two researchers, dressed as supermarket personnel, were stationed in an aisle and recorded the time spent looking at the attention foods by every shopper entering that aisle. A sample of the shoppers who had passed through the aisle was then interviewed by other researchers in another part of the store. Stratified random sampling was used, strata being defined by the time spent looking at the attention foods.

The first study, referred to here as AISLE I, concentrated on rice and pasta products. These foods were chosen because previous studies had reported that particular attention was given to their labels (see for example FDA, 1975). Observations were taken throughout shopping hours on a Monday and a Thursday. A total of 1,328 shoppers was timed and 141 were interviewed during the two day period. For the second study, AISLE 2, the attention products were canned meat/fish and powdered (dehydrated) soup. The canned meat/fish products were chosen because observations during the earlier study was suggested that shoppers paid particular attention to those foods. Powdered soup was selected as representing a new convenience food, heavily prompted at the time by mailed samples, and marketed at least partly on a health basis. Observations were taken throughout shopping hours on a Wednesday, Thursday, and Friday. For the second study, a total of 1,053 shoppers was timed and 147 were interviewed over the three day period.

The personal interview questionnaire on the AISLE 2 study was shorter than that for AISLE I to allow respondents time to perform a label ranking task. For this respondents were shown nine labels for a new food product, each presenting the same product information but in different formats. The subjects were asked to place the labels "in order of their helpfulness to the subject for making a decision to buy a new food product." The nine labels formed an orthogonal design embodying three levels for each of four types of information brand name, ingredient information, nutrition information and open dating information. The levels used for each type of information are detailed in Table I. This fractional factorial design affords extreme economy in treatments ( full factorial design would involve 81 product labels) allowing the ranking task to be kept within individual capabilities. Clearly, the cost of such parsimony is the assumption of no interaction effects; however, previous studies suggest such interactions are rare (see for example Green and wind, 1975; Green, 1974). An artist's sketch of one of the labels used appears in Figure I.

The Time Construct and Shopper Classifications

The major dependent variable in the studies reported here is the time each shopper spent examining the attention products. Jacoby, Szybillo and Berning (1976) have reviewed the literature on time and consumer behaviour. As they point out very convincingly, no major conceptual treatment nor systematic empirical effort has yet been focused on this subject. Our studies are an attempt to look at the time construct in relation to consumer behaviour unobtrusively. There appears to be no such studies reported in the literature; if there is one it is the exception.

TABLE I

DESIGN FOR CONJOINT ANALYSIS

FIGURE I

ILLUSTRATIVE LABEL

The time construct is emphasized here both because little is known about how consumers allocate their time in the supermarket, and because the time spent viewing products is one operational measure of the use consumers make of labels; extremely short viewing times suggest virtually no use of label information whereas extended product viewing implies considerable processing of label information and product comparisons. This is further supported by Gatewood and Perloff (1973) who have suggested that speed in processing supermarket and label aids would seem to be a desirable end product of consumer information systems.

Figure 2 shows the frequency distribution of time spent examining the attention products for all shoppers entering the relevant aisle throughout both studies. The distribution is heavily skewed towards short product viewing times, but there is a sizeable tail with sixteen individuals actually examining the products for more than two minutes. Given that the interest is in product viewing time as a measure of the use and interest in label information and that the distribution in Figure 2 is so heavily skewed, it is preferable to treat it as a classificatory rather than a continuous variable.

The obvious dichotomy was between "grabbers" who simply snatched the product from the shelf with virtually no discernible product viewing time, and "lookers" who observed the products for at least one second. In view of the wide range of viewing times by "lookers" (from one second to 5 1/2 minutes) this is a very crude classification. Consequently, for parts of the analysis this group was subdivided into "scanners" who observed the products under study for eight seconds or less and "label readers" who considered the products for more than eight seconds. The eight second cut-off aimed to distinguish between shoppers employing a routinized response and those in a limited or an extensive problem solving stage of decision making (see Howard, 1977).

FIGURE 2

FREQUENCY DISTRIBUTION OF PRODUCT VIEWING TIMES

As discussed previously there is little prior research on the role of time in consumer behaviour and the 8 second cut-off time demands some explanation. In the design of this study estimates of reading times were obtained using a group of undergraduate students. (For details of this methodology see Kendall and Fenwick, 1978.) The results suggested an average scanning time of about eight seconds.

This crude estimation does not appear to be too inconsistent with the literature. On an average package of food, a reader can identify about seven chunks of information: price, brand name, contents, ingredients, directions, nutritive composition (or where to get it), weight, manufacturer's name and address, and a few aesthetic phrases about the glories of the product. Peterson and Peterson (1959) and Murdock (1961) suggest that percentage of correct recall tends to level off at about eight seconds for three consonant trigrams. This would be equivalent to three chunks of information which the average consumer might choose to look at on the label of a food package. The above time referents are based on a retention task. Most food shoppers are not trying to retain the information; they are trying to make a decision with the information which would entail some retrieval time.

Recent work with chronometric analysis in consumer behaviour (Gardner, Mitchell, and Russo, 1977) for low/ high involvement products showed that response times for brand and attention evaluation statements averaged just under four seconds. Furthermore, Johnson and Russo (1977) have shown that mean recall times for a brand probe and an attribute probe for cooking oil was just over 12 seconds. However, the latter again refers to recall time and most shoppers are not concerned with this task. One could argue that a mean time might be 8 seconds merely by combining the above two studies. But it would be more realistic to take the first study and double the four seconds since the shopping is not a forced choice situation and shoppers do not feel that they are being tested. Consequently, the cut-off between scanners and readers of 8 seconds does receive some justification, both in the literature and from pilot empirical work. Clearly this is an important area for further field research.

RESULTS AND DISCUSSION

Prevalence of Product Reading

During the five days covered by the studies, 2,381 shoppers passed through the attention aisles. Summary statistics for the various days, products and shopper categories appear in Table 2. Three points are of particular interest.

First, comparing results across product groups, rice and pasta buyers are four times more likely to simply grab the product than are canned meat/fish and soup buyers. This finding conflicts completely the FDA (1975) study which reported grain produces (bread, rice and pasta) as the most "label read" food products. The FDA study, relied, of course, on claimed label usage in a personal interview situation. It would appear that such claims may not be a reliable indication of behaviour and should have little input to decision making.

Second, over both studies (i.e., including the rice and pasta products), only 25% of buyers simply snatched the product from the shelf. The other 75% showed at least some interest in gathering product information. This contrasts with the self-report surveys which (with the exception of the Redbook Studies 1974, 1976) generally report around 30-50% claimed product information usage. Furthermore, of those showing some interest, over half spent more than 8 seconds with the products, giving a mean product viewing time of 38 seconds. Although no specific information was gathered on the particular product/package/label attributes that warranted this attention time, the extent of exposure suggests much more information processing, or product comparisons, than is normally assumed (for example see Friedman, 1977, reviewed earlier).

TABLE 2

SUMMARY STATISTICS OF PRODUCT VIEWING TIME

Third, looking in more detail at the difference in reading time between products, canned meat/fish recorded the longest mean time (42 seconds) with the newer product, dehydrated soup, having the shortest (34 seconds) -- a difference Significant at the .025 level. The label reading literature would suggest the reverse effect. Viewing times are hypothesized as being shorter for the "older" more established products (see for example Howard, 1977; Buck and Jacoby, 1974; Ross, 1974). Interestingly all these studies have concentrated either on pure theory (e.g., Howard, 1977) or else have been confined to forced-choice laboratory situations (e.g., Buck and Jacoby, 1974). Clearly there are more complex factors at work than has been popularly supposed.

TABLE 3

STANDARDIZED DISCRIMINANT COEFFICIENTS

Who Looks At Labels?

In answering this question the primary distinction made was between "grabbers, who spent negligible time with the product, and the rest (i.e., scanners and label readers). Non-buyers were excluded completely from the analysis. Two-way discriminant analyses were performed separately for the AISLE I and AISLE 2 data.

First, restricting the analysis as far as possible to variables common to both data sets, a stepwise analysis was performed for AISLE I. This found six variables (chosen from a total of 15) which could correctly classify 75% of the sample. [The dangers of evaluating a discriminant function by correct classification rate are well known (for example, see Morrison, 1969). In this case validations using the jack-knife method (Tukey, 1958; Fenwick, 1978) are currently being performed. Furthermore, in no group was the correct classification rate more than 3% below the average reported.] These six variables were: days of the week, time of the day, age of the shopper, marital status, shopping party size and use of a shopping list.

People who spent some time looking at the products tended to be older, married, without a shopping list, in smaller shopping parties, and to be shopping later in the week and later in the day than grabbers (see coefficients in Table 3). Interestingly, self-reported in-store label reading behaviour did not discriminate well. If anything, grabbers tended to claim more label-reading than did lookers! Thus, self-reported label reading was found to be an unreliable indicator of true behaviour, at least for the rice and pasta products studied in AISLE I.

Widening the predictor set to include three variables not measured in the AISLE 2 study - satisfaction with current label information, extent of support for mandatory federal label regulation, and the degree of menu planning - improved the correct classification rate to 90%. People who looked at labels tended to plan menus further ahead, to support mandatory federal regulation of label information, and to be dissatisfied with the current label information offered (see Table 3). Although the latter two effects are intuitively appealing, it is less obvious why menu planners tend to examine labels when those with shopping lists do not. One possible explanation is that menu planners spend the time with labels to determine if the product fits into the menu scheme. This would indicate the consumer uses either direction or recipe information, should it be available. The shopping list person on the other hand simply wants the specific product, thus precluding the need for information in the store.

Having successfully distinguished grabbers from the product lookers, a three-way discrimination was attempted. Those who looked at products were divided into scanners (looking for 1-8 seconds) and readers (looking for more than 8 seconds). The same variables described previously produced a 74% correct classification rate using two discriminant functions (see Table 3, column 3). Interestingly, the first discriminant function was almost identical to that estimated in the two-way analysis, and label-readers were simply like the lookers discussed above, but more so! On the second discriminant function however, grabbers and readers tended to have similar scores, with the scanners well apart from both. It seems then that in some respects label readers have more in common with grabbers than scanners. One possible explanation for this would be that grabbers are currently processing little information, but if alerted, say by a health scare, search for product information, i.e., become readers. Scanners on the other hand are already gathering and processing the available information in under 8 seconds.

Applying the original six variables to the AISLE 2 data set produced an overall average 70% correct classification rate. Although using the same variables, unfortunately the AISLE 2 discriminant function showed some sign changes (see Table 5). For the AISLE 2 products (canned meat/fish and powdered soup) people who look at the products, although still older and married, tend to have a shopping list and to shop earlier in the week, earlier in the day and to be in larger shopping parties. Part of the explanation for this is due to the differences in products and the different store chain used in the second study.

As always when using search procedures, invalidated results must be treated with caution. It is impossible to say at this stage whether these sign changes show unstable discriminant functions or identify true differences in the characteristics of label readers for different products and different stores. Research is proceeding on this point using the jack-knife method (Tukey, 1958) to test the stability and significance of the discriminant results.

Substitution of years of full time education for age improved the correct classification to 75%. Those looking at labels were better educated than the grabbers as the literature would suggest (Daly, 1976). A three-way discrimination between grabbers, scanners and readers could not be successfully performed for AISLE 2, correct classification being well below 50%.

In summary, the question of who looks at labels is not easily answered. Although the same sets of variables identify product lookers in both studies, the direction of influence of these variables differs.

Shoppers' Trade-Offs For Label Formats

In the final part of the AISLE 2 study shoppers were asked to rank order 9 labels for a new food product (see Figure 1) according to their helpfulness in making a buying decision. These 9 labels formed a fractional factorial representation of a 34 design as noted in Table 1. Interest centered on a new food product since regulatory authorities are particularly interested in assisting consumers with new product choices (U.S., 1974). Preference rankings were analyzed for a group of lookers using the MONANOVA program. [The preference rankings of individuals within these groups were treated as replications, a single set of trade-offs being estimated for the whole group. The alternative procedure of estimating trade-offs for each individual and then clustering the results (e.g., Best & McCullough, 1977) makes interpersonal comparisons of individuals' trade-offs which are strictly inadmissible and produces less well constrained solutions.]

Figure 3 shows the part-worth utility functions for each of these groups. Almost all the functions show marked differences between grabbers and lookers. Figure 3a shows the part worths for different forms of nutrition information. Results are consistent with the literature although the Food Equivalent Method performed less well than would be expected from other results (Babcock and Murphy, 1973). Lookers tended to value the more extensive FDA/FTC Method, while the grabbers preferred the singular representation of the Jacobsen (1973) Scoreboard Method. Similar results were shown for these methods in media presentations (Kendall, 1977; Kendall and Krane, 1978).

Figure 3b notes the trade-offs for different ingredient information. It is not surprising that both groups have little liking for the present method of presenting ingredient information since it is not easily understood (see G. A. O., 1975). The grabbers again prefer the succinct warning/comment statement since presumably it takes less time and is more convenient. The lookers have the highest utility for the explanation of the ingredients and appear to think that a warning/cogent does not add to their information handling capacities.

Figure 3c shows the part worths for open dating formats. The "sell before date" appears to offer the highest utility for both shopper groups. However, the difference in utilities over all three formats is fairly small indicating the low importance of this piece of information to both groups.

Finally Figure 3d shows shoppers' reactions to brand name information. For the typical grabber private brand name has highest utility and the no brand name label introduced by some major retailers (e.g., Carrefour, Jewell and Dominion, see Ebert, 1977) is least preferred. In contrast product lookers actually prefer no brand products. While it is understandable that product lookers do not need the additional information offered by brand name - they can acquire the necessary information from other items on the product label - it is difficult to see why they should actually prefer the no brand name as an information aid. Presumably other characteristics of no brand products (e.g., low price) led to its higher preference ranking.

FIGURE 3

PART WORTH UTILITY FUNCTIONS

Examining the range of utility covered by each information aid gives an indication of the importance of that kind of information to shoppers (Green and Wind, 1975). Table 4 summarizes the results. For the average grabber, brand name and ingredient information constitute over 60% of the utility range, whereas for lookers nutrition and ingredient information account for almost 70% of the total range in utility.

Of particular interest is the difference in the order of importance of information items between the two groups. For the average grabber brand name is the most important information source, followed by ingredient information. The average looker, however, is most concerned with nutrition information, although ingredients are a close second. The importance of nutrition information when buying a new food product is most interesting as Canada, unlike the U.S.A., has no mandatory legislation on nutrition information. Ottawa, and Consumer and Corporate Affairs, has been reticent about such legislation and little empirical research has been reported (see Liefield and Bond, 1974; Liefield, et al., 1975).

The only other reported work on label information trade-offs appears to be that by Best and McCullough (1977). That study collected its data by in-home interviews and could not, therefore, relate results to strong behavioral measures. Furthermore, most of the information formats tested in that paper are neither in use nor even proposed. Finally, the Best-McCullough study sums utilities across individuals, an operation which is not meaningful with MONANOVA output.

CONCLUSIONS

First, this study found self-reports of label reading to be misleading. Food products for which other studies had found high self-reported label reading turned out not to be read frequently. Individuals who reported label reading in this study spent, if anything, less time with the product than did those who claimed little reading. For public policy purposes any studies based on self-reported label reading should be treated with caution.

TABLE 4

SUMMARY OF UTILITY RANGES FOR INFORMATION AIDS

Second, a large majority of shoppers spent enough time with the products as to imply considerable label reading. Only 25% simply snatched the product from the shelf. The rest had time enough to process at least some label information and some had time to process large amounts of information. However, product class strongly influenced consumers' propensity to view the product.

Third, although using a limited number of descriptor variables, this study suggests that it is not as much the type of person but the situational variables which determine label reading. Day of the week and time of day play a large role in determining readership. Degree of menu planning and possession of a shopping list also determined readership. The literature takes little account of these variables and further research should look into such explanations.

Finally, ingredient and nutrient data are valued by consumers. Even those simply grabbing the product from the shelf admit the usefulness of ingredient information. For the majority of shoppers nutrition and ingredient information dominate their label preferences for new products. However, a public policy decision has to be made on the target market for such information. It is not enough to say that there are different criteria for different groups. If information is to appeal to grabbers, who could be in need of more guidance, ingredient information should contain simple warnings, and nutrition data should be in the scoreboard form. On the other hand, for lookers, who are more likely to process the information, more detailed formats are appropriate. There is no real evidence that this issue has yet been confronted by public authorities.

REFERENCES

S. Addelman. Symmetrical and asymmetrical fractional factorial plans. Technometrics, 1962, 4, 21-58.

Babcock, M. J. & Murphy, M. M. Two nutritional labeling systems. Journal of the American Dietetic Association, 1973, 62, 155-161.

R. Best, & J. McCullough. Evaluation of food labeling policies through measurement of consumer utility. In Hunt, K. (Ed.). Advances in Consumer Research, 1977, 5, 213-219.

M. Buck & J. Jacoby. Time spent acquiring product information as a function of type of innovation and information load. Purdue Papers in Consumer Psychology, No. 136, Purdue University, 1974.

R. W. Chestnut. The Expenditure of Time in the Acquisition of Package Information. Unpublished Masters Thesis Purdue University, 1975.

CRI/FDA. R. C. Stokes & R. Haddock. Interim report of the first two phases of the CRI/FDA nutritional labeling research program. Washington, D.C. Consumer Research Institute, 1972.

P. A. Daly. The response of consumers to nutrition labeling. Journal of Consumer Affairs, 1976, 10, 170-178.

R. W. Darden & W. A. French. Product label information: Dimensions of meaning and evaluation of importance. Combined Proceedings of the Spring and Fall Conferences of the AMA, 1971, 647-652,

G. S. Day. Assessing the effects of information disclosure requirements. Journal of Marketing, 1976, 40, 42-52.

L. K. Ebert. No-Brand lines bow at Jewel Stores. Advertising Age, 31 October 1977, 2.

FDA, U.S. DHEW. Food and Nutrition: Knowledge and Beliefs, Report I, 1973-4 (RAC #3697). Washington, D.C. 1973.

I. Fenwick. A user's guides to conjoint analysis in marketing. European Journal of Marketing, 1978, 12, forthcoming.

M. Friedman. Consumer use of Informational aids in supermarkets. The Journal of Consumer Affairs, 1977, 11, 78-89.

M. Friedman. Consumer responses to unit pricing, open dating, and nutrient labeling. Proceedings of the 3rd Annual Conference of the Association for Consumer Research, 1972, 3, 361-369.

GAO, Comptroller of the U.S. Food Labeling: Goals, Shortcomings and Proposed Changes, Washington, D.C. 1975.

M. P. Gardner, A. A. Mitchell & J. E. Russo. Chronometric analysis: An introduction and application to low involvement perception of advertisements. In Hunt, K. (Ed.). Advances in Consumer Research, 1977, 5, 581-589.

R. D. Gatewood & R. B. Perloff. An experimental investigation of three methods of providing weight and price information to consumers. Journal of Applied Psychology, 1973, 57, 81-85.

P. E. Green. On the design of choice experiments involving multifactor alternatives. Journal of Consumer Research, 1974, 1, 61-68.

P. E. Green & Y. Wind. New way to measure consumers' judgments. Harvard Business Review, 1975, 53, 107-118.

J. A. Howard. Consumer Behavior: Application of Theory. New York: McGraw Hill, 1977.

M. Jacobson. Nutrition Scoreboard: Your Guide to Better Eating. Center for Science in the Public Interest, Washington, D.C. 1973.

J. Jacoby, R. W. Chestnut & W. Silberman. Consumer use and comprehension of nutrition information. Journal of Consumer Research, 1977, 4, 119-28.

J. Jacoby, D. E. Speller & C.A. Kohn. Brand choice behavior as a function of information load. Journal of Marketing Research, 1974a, 11, 63-69.

J. Jacoby, D. E. Speller & C.A. Kohn. Brand choice as a function of information load: replication and extension. Journal of Consumer Research, 1974b, 1, 33-42.

J. Jacoby, G. S. Szybillo & C. K. Berning. Time and consumer behavior: An interdisciplinary overview. Journal of Consumer Research, 1976, 2, 320-339.

E. J. Johnson & J. E. Russo. The organization of product information in memory identified by recall times. In Hunt, K. (Ed). Advances in Consumer Research, 1977, 5, 79-86.

K. W. Kendall. An experimental study to investigate methods of presenting nutrient values in the media. Unpublished Doctoral Dissertation, The University of Iowa, 1977.

K. W. Kendall & R. Krane. Television, nutrition and product choice. Simon Fraser University Working Paper, 1978.

K. W. Kendall & I. Fenwick. What do you learn standing in a supermarket aisle?, Simon Fraser University Working Paper, 1978.

J. King & J. Gideon. Unit pricing for new grocery system Report prepared for Delchamps, Inc., Mobile, Alabama,

22 December 1971.

R. J. Lenahan, J. A. Thomas, D. L. Call, & D. I. Padberg. Consumer reaction to nutritional information on food product labels. Search Agriculture, 1972, 2, 37-53.

J.P. Liefield & E. Bond. Consumer information schemes -Do we need them, do we want them, will we use them. The Canadian Consumer, June 1974, 16-18.

J.P. Liefield, E. A. Gullett, M. Sanderson, T. A. Watts & J. H. Sabry. Attitudes of a sample of consumers towards nutritional labeling. Canadian Institute Food Science Technology Journal, 1975, 8, 46-50.

D. G. Morrison. On the interpretation of discriminant analysis. Journal of Marketing Research, 1969, 6, 156-163.

B. B. Murdock. The retention of individual items. Journal of Experimental Psychology, 1961, 62, 816-625.

L. R. Peterson & M. J. Peterson. Short term retention of individual terms. Journal of Experimental Psychology, 1959, 58, 193-198.

Redbook Nutrition Study. Conducted for Redbook Magazine by the Gallup Organization Inc., 53 Bank Street, Princeton, NJ May 1976.

Redbook Readership Survey. How do you feel about food, what do you know about food? Conducted by the Editorial and Research Departments of Redbook Magazine, 1974.

I. Ross. Applications of consumer information to public policy decisions. In Sheth, J. N. & Wright, P. E. (Ed). Marketing Analysis for Social Problems. Illinois: University of Illinois Press, 1974, 42-76.

J. E. Russo. More information is better. Journal of Consumer Research, 1974, 1, 68-72.

J. E. Russo, G. Krieser & S. Miyashita. An effective display of unit price information. Journal of Marketing, 1975, 39, 11-19

D. L. Scammon. Information load and consumers. Journal of Consumer Research, 1977, 4, 148-155.

J. O. Summers. Less information is better. Journal of Marketing Research, 1974, 11, 467-468.

E. F. Taylor. Food processors and open data labeling. National Food Situation, 1976, 155, 27-31.

J. W. Tukey. Bias and confidence in not quite large samples. Annals of Mathematical Statistics, 1958, 29, 614.

United States, FTC. Food advertising: Proposed trade regulation rule and staff statement. Federal Register, 39, No. 218, 29842-39862, 11 November, 1974.

R. H. Warland & R. O. Hermann. The new wave of imitation foods: Problems ahead. Journal of Consumer Affairs, 1971 5, 56-59.

W. L. Wilkie. Analysis of effects of information load. Journal of Marketing Research, 1974, 11, 462-466.

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