The Perceptual Task in Acquisition of Package Information

Stephen Calcich, University of South Carolina
Edward Blair, University of Houston
ABSTRACT - This paper concerns the perceptual task involved in consumers' acquisition of product information from packages. It reports a study that relates the time needed to acquire package information with a perceptual skill called disembedding, The results show that disembedding skill does correlate with acquisition time and that there are substantial differences among consumers in the length of time needed to acquire package information. The results also show that acquisition time varies across types of information in a manner consistent with their perceptual accessibility, and that acquisition time does not vary across product categories with equally complex information displays. These results have some interesting implications for the study of consumers' information processing.
[ to cite ]:
Stephen Calcich and Edward Blair (1983) ,"The Perceptual Task in Acquisition of Package Information", in NA - Advances in Consumer Research Volume 10, eds. Richard P. Bagozzi and Alice M. Tybout, Ann Abor, MI : Association for Consumer Research, Pages: 221-225.

Advances in Consumer Research Volume 10, 1983      Pages 221-225

THE PERCEPTUAL TASK IN ACQUISITION OF PACKAGE INFORMATION

Stephen Calcich, University of South Carolina

Edward Blair, University of Houston

[Data for this study were gathered with funds provided by a grant from the Office of Research Administration, University of Houston.]

ABSTRACT -

This paper concerns the perceptual task involved in consumers' acquisition of product information from packages. It reports a study that relates the time needed to acquire package information with a perceptual skill called disembedding, The results show that disembedding skill does correlate with acquisition time and that there are substantial differences among consumers in the length of time needed to acquire package information. The results also show that acquisition time varies across types of information in a manner consistent with their perceptual accessibility, and that acquisition time does not vary across product categories with equally complex information displays. These results have some interesting implications for the study of consumers' information processing.

INTRODUCTION

Product packages present complex stimuli to any consumer wishing to use the information contained on them. Consider, for example, that a box of Kellogg's Corn Flakes has more than 45 different sizes and styles of print in addition to pictures and illustrations. It seems likely that a fairly substantial perceptual effort is a precondition to acquiring and using information under these circumstances. If so, understanding the nature of the perceptual task, the extent to which individuals differ in their abilities to perform it, and the correlates of these individual differences may be helpful in designing product packages and in designing and interpreting research into consumers' information processing. This paper reports some data relevant to the first two issues.

A key perceptual ability involved in acquiring information from a product package is that of disembedding. A rood package can be thought of as an organized visual field which a consumer must break up in order to deal with its separate parts, such as product information. Disembedding ability is measured by a psychological assessment procedure called the Embedded Figures Test (EFT) (Witkin 1950; current manual Witkin et al. 1971). Validation studies involving this instrument show that it does tap a perceptual ability (see, for example, Widiger et al. 1980), and relates to performance in a wide array of perceptual situations which share the requirement of perceptual disembedding --requiring a subject to deal with part of an organized field independently of the field. It does not relate to performance in perceptual situations which do not have this requirement (see, for example, Karp 1963).

Previous research using the EFT has shown substantial variance among people in disembedding ability. In addition, lower disembedding ability has been demonstrated for older people (Comalli 1965; Schwartz and Karp 1967) and for members of cultural groups that discourage autonomous behavior (Dawson 1967: Berry 1966). Differences in disembedding ability have been associated with general cognitive style (Witkin et al. 1971), and lower disembedding ability has been linked with slower decisions and less impulsivity (Loo 1977). If acquisition of package information does involve a perceptual disembedding task, then these findings suggest that the perceptual costs of acquiring information in any given purchasing situation will vary across consumers and that this variance may be associated with differences in information use and behavior.

Previous research on disembedding also shows that the amount and organization of information in a complex stimulus affects the difficulty of disembedding specific elements (Witkin 1950). This finding suggests that different product categories will present different levels of perceptual difficulty to the extent that their packages contain different amounts or organizations of information. It also suggests that the difficulty of disembedding various types of information will v. rv according to their relative positions and emphasis on packages; for example, brand name, package weight, and price generally are more prominently displayed than individual pieces of nutritional information on a food package.

THE STUDY

Hypotheses

The study reported here was designed to explore the perceptual task in package information acquisition. Two issues were of principal interest; whether this task does indeed seem to be a disembedding one, and whether its difficulty varies across different types of information. The associated hypotheses were:

H1: A positive relationship exists between an individual's score on the Embedded Figures Test and the time taken to acquire specific pieces of information from a package.

H2: Different types of information will take different amounts of time to acquire.

The first hypothesis is a direct reflection of the notion that higher disembedding ability, as indicated by lower (faster) scores on the EFT, will lead to faster information acquisition. The second hypothesis reflects the fact that various types of information are positioned and emphasized differently on packages, and the notion that these differences affect perceptual difficulty. Both hypotheses operationalize perceptual difficulty through the time needed to acquire information.

A third issue of some interest concerned whether the difficulty of disembedding varies across different product categories. No formal hypothesis was stated on this issue because the limited study planned did not allow adequate testing of product differences. However, it was planned to use multiple product categories with similar amounts and presentations of information on their packages, and check whether information acquisition times were similar across these homogeneous (with respect to information display) packages despite other product differences that might exist.

Method

A correlational and experimental study was conducted with 40 female food shoppers in a Houston shopping center to test these hypotheses. The shopping center used is in a predominantly working class and ethnically mixed area of the city. Participants received a modest cash inducement for completing an Embedded Figures Test, an information acquisition task, and a brief background questionnaire.

The EFT consisted of twelve trials plus a practice trial. In each trial, the participant was shown a simple figure or shape for ten seconds. She then was shown a complex figure and asked to pick out the simple figure. She was instructed to tell the tester when she found the simple figure within the complex figure and to trace around it with a stylus. Participants were told that they were to be timed on each trial and to proceed as quickly as possible. The first simple figure could be referred to as many times as desired (simple and complex figures were not shown together) without penalty, as the watch was stopped when such a reference was requested. When a participant indicated that she had found a simple figure, the time elapsed was noted but the watch allowed to run. If her tracing was correct, the noted time was recorded and the next trial was begun. If her tracing was incomplete or inaccurate, she was told: "No, that's not it," as the watch continued to run. If the simple figure was not found within three minutes, the solution time was recorded as 180 seconds and the next trial was begun. A total EFT score was calculated by summing solution times for the twelve trials.

The information acquisition task consisted of four trials. In each trial, the-participant- was given two cards, one listing a piece of product information and one listing five brands of a packaged food product. They were instructed to find the piece of information, in the brand order presented, on five packages from the same product category placed together on a table in front of the participant. Each trial presented packages from a different product category and required finding a different piece of information. When the information had been located, the participant was to read it aloud. If she was incorrect, she was told: "No that's not correct," and had to continue until correct identification. As with EFT, subjects were told to progress as quickly as possible because they were being timed. Elapsed time was recorded in each product category when the correct information had been found on the fifth package, so that each participant provided four scores, one from each trial, that could be associated with the different product categories or with the different types of information sought.

Product categories used in the acquisition task were breakfast cereal, instant potatoes, instant rice, and macaroni and cheese These categories were chosen on the following criteria: 1) that at least five different brands be available; 2) that packages contain similar amounts and presentations of information across categories; and 3) that packages contain enough common information across categories to allow use of the same type of information across categories. Four national brands and one store brand were used in each category. Every participant in the study was given these products in the same order (cereal, potatoes, rice, macaroni) because the limited sample size did not allow manipulation of product order and because any order effects would have worked against the hypothesis of no product differences.

The types of information sought were price, servings per package, grams of protein per serving, and percentage of the U.S. recommended daily allowance of iron per serving. These are listed on all food packages and are considered especially relevant for broad groups of consumers, but were chosen primarily because they differ in display position. Price is prominently by itself on the top or bottom of the package, protein and iron are given with other nutritional information in a clearly marked area on the side of the package (with protein listed first and iron buried further down the list), and servings per package stands by itself in a relatively obscure position.

To prevent practice or order effects from influencing the test of differences among acquisition times for types of information, participants were assigned randomly to four information order treatment groups. These groups sought product information across the four product categories in the following sequence: price for the first product, servings for the second, iron, then protein: servings, protein, price, then iron: protein, iron, servings, then price; and iron, price, protein, then servings. The four sequences comprise a 4 X 4 X 4 Latin Square repeated measures design of groups by products by information types.

RESULTS AND DISCUSSION

The first hypothesis, that a positive relationship exists between EFT scores and the time taken to acquire package information, was tested by correlating EFT scores with acquisition task times across the 40 study participants. Five correlations were computed; a correlation between EFT scores and total package task time (summed over the four products and information types) and four correlations between EFT scores and acquisition times for each information type. These correlations are shown in Table 1.

Across all 40 participants, EFT scores correlate .46 with price acquisition time, .39 with servings per package acquisition time, .55 with iron acquisition time, .39 with protein acquisition time, and .51 with total acquisition time. All five correlations are positive, as expected, and significant beyond the .01 level. These relationships are especially impressive when one considers that ten of the 40 participants had extreme difficulty with the Embedded Figures Test and went close to the 36 minute time ceiling. The ceiling effect for these ten participants attenuated the correlations between EFT scores and acquisition times; when the ten are removed from the data, the five correlations reach .50 to .70. These adjusted correlations also are shown in Table 1.

H1 clearly is supported. Acquisition of package information varies in difficulty with consumers' perceptual disembedding abilities. As one might expect, the relationship between disembedding ability and acquisition speed may vary according to how deeply the information is embedded: the adjusted percentages of variance explained by the EFT in acquisition times for types of information ranges from 25% for the most prominently displayed item, price, to 442, for the least prominently displayed items, servings and iron. These percentages leave plenty of room for other determinants of acquisition time, but are larger than commonly encountered in behavioral research and deserve attention (especially the one-half of total acquisition time variance explained).

An issue of interest in the analysis of acquisition times was the extent to which these items varied across consumers regardless of what caused this variance. Substantial variance was observed. The fastest study participant completed the total acquisition task in 139 seconds (for 20 pieces of information, each on a separate package) while the slowest participant took 572 seconds; a ratio of more than four to one. Mean time across all participants was 264 seconds with an interquartile range of 156 seconds to 388 seconds.

TABLE 1

PEARSON CORRELATIONS BETWEEN EFT AND PACKAGE SCORES

TABLE 2

MEAN TIMES (IN SECONDS) FOR PACKAGING EXPERIMENT

The second hypothesis, that acquisition time will vary by types or information, was tested with a repeated measures analysis of variance among the four types of information used in this study. The rows or Table 2 show mean acquisition times by produce category for the four types of information, along with an overall mean time needed to acquire each type of information for the "average" product. (Remember that each figure represents total time taken for five separate packages.) Not only are the overall differences among the types of information significant at the .01 level (F = 22.109, 3 d.f.), but pairwise comparisons show every pair significantly different. Price was acquired easiest at an overall mean time of 41.2 seconds, protein next fastest at 61.1 seconds, iron third at 75.0 seconds, and servings per package slowest at 86.3 seconds. The ordering or these acquisition times is consistent with the apparent perceptual accessibility of each type of information (as described in Methods).

H2 clearly is supported. Acquisition times do vary among types of information in a manner consistent with the position and prominence of their displays.

Table 2 also shows, in its columns, mean times needed to acquire the different types of information for the four product categories, along with an overall mean time needed to acquire the "average" piece of information for each product. There are some differences among the overall means, with a range of 59.6 seconds to 72.4 seconds, but the F value associated with these differences is only 2.144 with three degrees of freedom and is not significant at the .10 level. A test of the products by types of information is no: shown because no such interaction was expected and the design confounds this interaction with group differences; however, the F value associated with the interaction is only .381.

It is interesting to note that acquisition time does not vary significantly across these products, despite differences that may exist regarding factors other than their information displays. However, this finding provides only weak evidence for the general premise that acquisition time relates to a package's informational complexity. The packages used in this study represent only one level of complexity, and one would like to see data that incorporate more and/or less complex packages to confirm that acquisition times vary across but not within complexity levels. Also, further analysis of the data shows a significant difference in comparing cereals and potatoes with rice and macaroni; a difference explainable by practice (cereal and potatoes came first), difficulty of package manipulation (cereal and potatoes are larger), and product differences as well as chance.

CONCLUSIONS

The results of this study provide evidence that consumers vary substantially in the time needed to acquire package information, and that acquisition time re aces to a consumer's perceptual disembedding ability and the position and prominence of a piece of information.

Several limitations on these results must be acknowledged. Regarding the specific completion times shown for the package task, one must consider that participants had to handle a different package for each piece of information sought; essentially, attribute-based acquisition with real packages. Regarding the relationship between completion times for the package task and the EFT, one must consider that EFT scores correlate as high as .6; with other measures of cognitive spatial abilities (Widiger et al. 1980). Thus, while the data are consistent with an interpretation that package information acquisition requires disembedding, discriminant validity has not been established for this point. Regarding the specific acquisition times for different types of information, one must consider that a variety of factors such as attribute importance to consumers, consumers' familiarity with the package, and others are likely to have some impact on these times.

Despite these limitations, the general points remain that acquisition of package information imposes a perceptual task on consumers, that consumers vary substantially in their abilities to do this task, and that the difficulty of the task varies across pieces of information on actual packages. These points have some interesting implications.

One implication for future research on information acquisition and integration is that it may prove useful to measure participants' perceptual skills. The Embedded Figures Test probably is too time consuming for many researchers, but a short form of the test or some other assessment procedure offers hope for improving explanation of differences in acquisition or integration among consumers with similar utility of information structures.

A second implication is that information display boards, whatever the specific format and procedure, may circumvent important perceptual demands made by actual packages and may affect the use of information by altering the perceptual costs of acquiring different types of information (or information from packages of differing complexity). Prior research on acquisition time has intimated that the impact on consumers' decision processes and choices may be considerable (Chestnut 1975). While information display boards offer an undeniably attractive opportunity to observe information acquisition directly, it seems necessary that research using display boards be bolstered by more research that uses actual packages for stimulus display.

The lengths of time needed by consumers to perform the acquisition tasks used in this study have at least two implications. First, the substantial variance among consumers brings into question the use of student subjects, who are likely to have higher perceptual skills and, therefore, may use information differently than other consumers. Second, the lengths of time needed b! even the fastest study participants bring into question the likelihood that tasks of this nature are performed in daily shopping.

One study that timed subjects on information processing tasks using information like that used in the present study found a lowest mean processing time among experimental groups of 97 seconds, given five alternatives and six attributes (Bettman and Zins 1979). This time included information acquisition from simplified presentations. The mean time needed in the present study to identify four attributes across five brands from actual packages was at least 86 seconds (the average time needed to get the most difficult single attribute). It seems reasonable to sum these two mean scores for a time of acquisition and processing, bearing in mind that the processing times were gathered from students. The resulting time of more than three minutes, not including a choice deliberation period, would seem unreasonably long for the average shopper to spend comparing brands in a real food shopping situation. Even though shoppers will spend less time on any one decision as the number of choices to be made increases and can rely on previously learned choices to reduce shopping time, one is forced to suggest that these data, for food products, lend credence to Kassarjian's (1978) contention that in most cases no prepurchase choice process exists, only that of a routinized or habitual nature.

The implication of this research for public policy makers is clear. The present study supports those who have called for simplified information presentation in grocery stores and/or on packages (Jacoby and Bettman 1976; Russo 1975). Even if appropriate information is provided on food packages and consumers are educated and motivated to use it, there still may be those -who lack the perceptual abilities needed to acquire information in this context within a time period that-is feasible for a normal grocery shopping trip.

This study is seen as a small step toward a better understanding of the perceptual task involved in consumers' use of product package information. Certainly, additional research on perceptual abilities. as well as individual differences on other cognitive abilities and cognitive styles, will be a valuable addition to the marketing literature.

REFERENCES

Berry; J.S. (1966), "Temne and Eskimo Perceptual Skills." International Journal of Psychology, 1, 207-229.

Bettman, James R., and Zins, Michael A. (1979), "Information Format and Choice Task Effects in Decision Making," Journal of Consumer Research, 1, 141-153.

Chestnut, Robert W. (1975), "Expenditure or Time in the Acquisition of Package Information," unpublished master's thesis, Purdue University.

Comalli, P.E. (1965), "Life Span Developmental Studies in Perception: Theoretical and Methodological Issues," presented at Eastern Psychological Association Meetings, Atlantic City, New .Jersey.

Dawson, J.L.M. (1967), "Cultural and Physiological Influences Upon Spacial-Perceptual Processes in West Africa," Part 1 and Part 2, International Journal of Psychology, 2, 115-128, 171-185.

Jacoby, Jacob, and Bettman, James R. (1976), "Patterns of Processing in Consumer Information Acquisition," in Advances in Consumer Research, Vol. 3, ed ., Beverlee B. Anderson, Cincinnati: Association for Consumer Research, 315-320.

Karp, S.A. (1963), "Field Dependence and Overcoming Embeddedness," Journal of Consulting Psychology, 97, 294-302.

Kassarjian, Harold H. (1978), "Presidential Address, 1977: Anthropomorphism and Parsimony," in Advances in Consumer Research, Vol. 5, ed., H. Keith Hunt, Ann Arbor, MI: Association for Consumer Research, pp. xii-xiv.

Loo, Robert (1977), "Note on the Achromatic Item in the Embedded Figures Test," Perceptual and Motor Skills, 44, 1258-1260.

Schwartz, D.W. and Karp, S.A. (1967), "Field Dependence in a Geriatric Population," Perceptual and Motor Skills, 24, 495-504.

Widiger, Thomas A., Knudson, Roger M., and Rorer, Leonard J. (1980), "Convergent and Discriminant Validity Or Measures of Cognitive Style and Abilities." .Journal of Personality and Social Psychology, 39, 116-129.

Witkin, Herman A. (1950), "Individual Differences in Ease of Perception of Embedded Figures," Journal of Personality, 19. 1-15.

Witkin, Herman A., Oltman, Philip K., Raskin, Evelyn, and Karp, Stephen A. (1971), A Manual for the Embedded Figures Test, Consulting Psychologists Press, Inc., Palo Alto CA

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