Cue Utilization in Product Perception

Robert E. Burnkrant, University of California, Berkeley
ABSTRACT - Cue utilization in product perception is viewed as complex information processing. It involves a process of making inferences about products from the configuration of cues available. It is maintained that this type of information processing is more likely to be engaged in with complex rather than simple products. It is also suggested that the consistency in meaning of cues associated with a product is directly related to the intensity of the perception of that product. Preliminary results of a study to test these hypotheses are briefly reviewed. The relevance of attribution theory to product perception is discussed.
[ to cite ]:
Robert E. Burnkrant (1978) ,"Cue Utilization in Product Perception", in NA - Advances in Consumer Research Volume 05, eds. Kent Hunt, Ann Abor, MI : Association for Consumer Research, Pages: 724-729.

Advances in Consumer Research Volume 5, 1978      Pages 724-729

CUE UTILIZATION IN PRODUCT PERCEPTION

Robert E. Burnkrant, University of California, Berkeley

[The author would like to acknowledge the support of the Center for Research in Management Science, the Committee on Research and the Institute of Business and Economic Research of the University of California, Berkeley. He would also like to thank Mr. Scott Vitell for research assistance provided this project.]

ABSTRACT -

Cue utilization in product perception is viewed as complex information processing. It involves a process of making inferences about products from the configuration of cues available. It is maintained that this type of information processing is more likely to be engaged in with complex rather than simple products. It is also suggested that the consistency in meaning of cues associated with a product is directly related to the intensity of the perception of that product. Preliminary results of a study to test these hypotheses are briefly reviewed. The relevance of attribution theory to product perception is discussed.

Much of the information people acquire about products comes from the processing of product related cues. These cues provide a basis for making inferences about the product and give it meaning which frequently goes beyond and departs from the specific arguments made in advertising appeals. As such, they become important determinants of how people behave toward products.

An individual in a purchasing context is frequently faced with the problem of evaluating the worth of a product which he cannot possibly objectively evaluate from physical manipulation and observation of its characteristics alone. Among the strategies employed by the consumer is processing information contained in messages (i.e., advertisements) about the product. This has been given considerable attention in the literature. Much less attention has been devoted to another strategy. The consumer may turn to cues given by the product, the context in which it is found and the people who use it. These cues can provide a basis for inferring the characteristics of the object under consideration. This information processing involves making inferences about the product from the selective processing of product related cues.

It is important for marketers to come to grips with some of the factors which will determine the extent to which inferences will be made in particular situations and to identify cues people utilize when they make inferences. This will permit marketers to employ cues more consciously and appropriately with products and in situations in which these cues are likely to have an affect on consumers' product evaluations. This requires that we determine which types of products and situations will lead to cue processing. It requires a concerted effort to determine which product and other related factors are used by people as cues. It is also necessary to consider how various combinations of these cues will affect the inferences people make. The study to be described here addresses two of these issues. It attempts to identify a product dimension which will be directly related to the use of cues in product evaluation, and it attempts to assess the affects of variations in cue consistency on product evaluation.

The process of inferring the characteristics of an object from indirect information is clearly more complex (i.e., involves more cognitive effort) than the direct processing of observable product attributes. As a re-suit, it would be expected to be engaged in only when available direct information is insufficient to permit satisfactory performance on the impending task. The assessment of product differences in cue utilization naturally turns, then, to a consideration of dimensions which will distinguish between products requiring the individual to infer from indirect information and those which do not require this type of information processing.

Product familiarity has been found to be related to cue utilization in product evaluation. Monroe (1976) found that price played a greater role when subjects were unfamiliar with the brand than when they were familiar with it. Similar findings were also obtained by Smith and Broome (1966). In Jacoby, Olson and Haddock (1971) the influence of price ceased when price was paired with an already known brand.

It is suggested that if we control for familiarity, the tendency to make inferences from such cues as price should be directly related to product complexity. What is meant here by product complexity is the typical or mean number of dimensions by which the product is represented in the minds of its potential customers. It is expected that some products are largely represented in only very few, perhaps one or two, dimensions. These products are considered to be relatively simple. On the other hand, some products are likely to be represented by most people in several dimensions, and these products are considered to be complex.

This treatment of product complexity is adapted from work focusing on cognitive complexity. Researchers concerned with this topic have treated cognitive complexity as an individual difference variable and have made predictions about the effects of complexity on information processing (e.g., Kelly, 1957; Schroder, Driver and Streufert, 1967). It has been found that "most persons are relatively 'complex' in some areas of cognition and relatively 'simple' in others" (Gardner and Schoen, (1962). It follows that most people may be "complex" with regard to some products and "simple" with regard to other products. If this is true this distinction should have implications for information processing, including the tendency to base a products' perception on inferences.

Schroder, Driver and Streufert (1967) point out that "the more judgmental space contains dimensions of information that are not objectively or directly given by the situation (for example, as instructions or as direct outputs from the environment), the more we can assume the presence of higher and more flexible levels of integration'' (p 27). The reverse should also be true. That is, the more complex the perception the more it should contain dimensions of information not objectively or directly given by the situation. This indirect information must come from an inference process. As a re-suit, we would expect that an inference process plays a greater role in the formation of perceptions about complex products than simple products. If people have employed an inference process more readily in the evaluation of complex products than simple products, we would expect, based on learning, that people would more readily make inferences from indirect information about complex than simple products.

It is necessary, in order to test this proposition, to distinguish between products in terms of their complexity. Cognitive complexity has been measured in a number of different ways (c.f., Vannoy, 1965). One of the most frequently employed techniques has been Kelly's Role Construct Repertory Test (Rep Test). People are asked to provide the ways in which known stimuli (i.e., parent, teacher, child) are similar to and different from one another. People who provide many differences in response to this task are considered more cognitively complex regarding this stimulus domain than those who provide only few differences. Bieri, et. al. (1966) modified the Rep Test by providing bi-polar adjective scales rather than having the subject employ his own construct dimensions. Their results with the modified test have been consistently positively correlated with the results obtained from Kelly's original instrument. This modified Rep Test will be employed here to measure product complexity.

PRODUCT RELATED CUES

If we are interested in determining the effects of cues on product perception we must employ cues which do not, by their very presence, bring to mind previously formed, well-defined product perceptions. In studying cue utilization we are interested in determining the effects of cues on the formation of product perceptions. Use of these types of cues will permit us to explicate the nature of the process by which cues are combined to give meaning to the goods and services in the consumer's environment. Two cues which seem to meet these criteria are price and advertising. While the former has traditionally been employed in studies of product perception the latter cue is new.

Price

It has generally been found that, when price is the only cue, the subject perceives the higher priced product as being higher in quality (i.e., Tull, Boring and Gonsior, 1964; McConnell, 1968). When price is combined with ether cues an interaction is frequently obtained (i.e., Andrews and Valenzi 1971; Jacoby, Olson and Haddock, 19711 Valenzi and Andrews, 1971).

Advertising

Nelson (1974) suggests that one of the most important pieces of information the consumer takes from advertising is merely the understanding that the brand is advertised. He contends, "the consumer believes that the mere a brand advertises the more likely it is to be a better buy" (p 732). That is, he suggests consumers make inferences about a brand from knowledge of the magnitude of advertising support afforded that brand. The magnitude of advertising support given a product may be viewed, therefore, as a cue which permits inferences about that product without directly conveying product beliefs.

In the natural environment this belief that one brand is more heavily advertised than another is the result of exposure to many more advertisements for one brand than another. Nelson's contentions about the effects of this varied exposure could be subjected to experimental investigation by providing subjects in one condition with many advertisements for a brand while providing those of another condition with only a few or no advertisements. If actual advertisements were provided, however, the cue properties of advertising magnitude would be confounded with specific beliefs presented in the advertisement. One way to avoid this problem would be to give people information about the magnitude of advertising support without actually showing them the advertisement. This is the procedure which will be followed here.

THEORETICAL ORIENTATION

It has been maintained that product perceptions are formed by cue processing in accordance with a general inference process. Therefore, theories which account for the conditions under which and the extent to which people make inferences should be relevant to the prediction of product perception. Attribution theory is appropriate to this spectrum. While it has been used primarily to account for inferences made to other individuals, it is equally applicable to inanimate objects. That is, a person should employ the same process when making inferences about objects that he uses when attributing characteristics to other individuals. One of the most basic principles emerging from work in attribution theory is that attributions are directly related to the consistency of the meaning of cues associated with that object (c.f., Kelley, 1967). A stronger inference will be made when the attributional implications of the cues are consistent than when they are inconsistent. If high price is combined with high advertising support we would expect a stronger inference than if it is combined with low advertising or if only one cue is provided. We would expect then that a stronger, more extreme attribution will be made when cues are combined consistently than when they are combined inconsistently or when only one cue is provided.

DEPENDENT VARIABLES

If we view price and advertising as cues that people use to attach meaning to the products in their environment, it seems that dependent variables employed in the assessment of these products should give the consumer an opportunity to provide this meaning.

Osgood and his associates have found across a large group of studies, that the meaning of an object can be accounted for by the first three factors derived from a factor analysis of responses to a set of bi-polar adjective scales (i.e., Osgood, Suci and Tannenbaum, 1957). The object's position in a space comprised of these three dimensions is considered to be the object's meaning. The intensity of a person's position on a dimension is indexed by the distance of his position on that dimension from the midpoint. It is suggested, therefore, that the individual's position will be more extreme when cues are combined consistently than when they are combined inconsistently or when only one cue is provided.

METHOD

A pilot study was performed in order to select products differing significantly in complexity. This was followed by a laboratory experiment in which the implications of the previously suggested hypotheses were tested. In both studies the subjects were females recruited from civic organizations in the San Francisco Bay area. They volunteered in return for a donation made in their name to the civic organization or group from which they were recruited. The donations were $1.50 each for those participating in the pilot study and $3.00 per person for those taking part in the experiment.

Initial Study

Questionnaires were administered in a group setting to 61 subjects. Each subject received an instrument designed to measure cognitive complexity. Brands of black and white television, headache remedy, toothpaste and soft drink were used as stimuli. The brands of each product class were placed on rows in a separate rating matrix The columns of the matrices were occupied by bi-polar adjective pairs. The matrices provided the vehicle for measuring product complexity. They were developed in accordance with the procedure set out by Bieri, et. al. (1966). The order in which matrices were presented was randomized.

Experiment

Subjects arrived in the laboratory at the time of their pre-designated appointments in groups of up to eight people. After a general introduction by the experimenter, they proceeded to their cubicles. All treatments and questionnaires were administered individually on laboratory terminals while subjects were in these cubicles.

Subjects. One-hundred forty-four female volunteers participated in the experiment. Of this total, 118 subjects completed the task. The remainder failed to complete the experiment due to computer breakdowns and other malfunctions.

Independent Variables. Three levels of price (low, high and no-information control) were combined with three levels of advertising support (low, high and no-information control) and four products (black and white television, toothpaste, soft drink and headache remedy). Each subject evaluated each product under one of the price-by-advertising conditions. The order in which products were presented was counterbalanced.

Subjects were told that a new brand (in each of the given product classes) was about to be introduced to the local market. Variations in price and advertising support associated with this brand constituted the manipulations of the independent variables.

Price was manipulated by providing subjects with information about the suggested retail price of the new brand. Subjects in the high condition were presented with a price which was found to be higher than most brands in the given product class. Subjects in the low price condition were presented with a price which was found to be lower than that charged for most brands in the given product class. The control condition was given no price information.

In the low advertising condition it was stated that the brand will be "only lightly advertised." It was pointed out that advertisements will appear occasionally in local newspapers. In the high advertising condition it was pointed out that the brand will be "heavily advertised" and that it will appear frequently on "national network television" and in "national news and women's magazines". The control condition was given no information about advertising.

Product was manipulated by providing subjects with four different product classes. The products employed were chosen to represent both high and low complexity.

Dependent variables. After receiving a product description subjects were asked to evaluate the product on a set of 7-point bi-polar adjective scales. This was followed by a question which asked how likely they would be to buy the product if they were in the market for it. Finally, they were asked to provide their feelings about how high or low the product's price was and how lightly or heavily it was advertised.

This procedure was repeated three times (once for each remaining product). Subjects were then asked to respond to a series of post-experimental questions. These included a number of probes for demand characteristics.

After leaving their cubicles the subjects gathered in a second waiting area. Their reactions were elicited. The purpose of the experiment was explained to them and any misperceptions were corrected. They were thanked for their participation and asked not to divulge anything about what went on during the experiment.

RESULTS

Manipulation Checks

Complexity. The data from the pilot study were scored for cognitive complexity in accordance with the procedure developed by Bieri, et. al. (1966). Cognitive complexity is measured by comparing each rating in a row with the rating adjacent to it (i.e., for the same brand) in the other rows of the matrix. A score of one is given for every exact agreement of ratings on any one brand. This matching procedure is carried out in the same way for all possible pairwise comparisons in the matrix. The scores for each comparison are totaled to yield a final complexity score on the product class for each individual. Given an 8x13 matrix, the maximum, or simplest score obtainable would be (13(13-1)/2)8 or 624.

A cognitive complexity score was calculated in this manner for each subject and each product class. Orthogonal comparisons were made across products using multiple t-ratios. TV (x=140.3), and soft drink (x= 159.5) were found to be significantly (p< 0.001) more complex than headache remedy (x=242.3) and toothpaste (x=208.8).

Advertising. After evaluating a given brand, subjects were asked how heavily advertised they believed the brand would be relative to other brands in the given product class. Responses were recorded on a 9-point equal appearing interval scale which varied from "Very Heavily Advertised" at the high end to "Very Lightly Advertised" at the low end. An analysis of variance was performed for each product class on the responses to this question. The manipulation of advertising was found to be significant (p< 0.01) in the expected direction for toothpaste, soft drink and headache remedy. It was found to approach significance (p<0.10) for television.

Price. Subjects were asked how high in price they believed the test brand would be relative to other brands in the given product class. Responses were recorded on a 9-point equal appearing interval scale with the end-points labeled "Very High" and "Very Low" respectively. Analysis of variance was performed for each product class on the results of this question. The manipulation was found to be significant (p< 0.01) in the expected direction for each of the four product classes.

Experiment

A principal components analysis was performed on a with-in-cell correlation matrix of bi-polar adjective scale scores. This adjusts the correlations to remove treatment effects. The components analysis was performed separately for each product class with varimax rotation of the first three components. Items loading more heavily than 0.500 on a given component were taken as representing that perceptual dimension. Items representing each of the first three factors for each product are shown with their loadings in Table 1. They represent the meaning or perception a la Osgood, Suci and Tannenbaum of each given product.

Individuals' scores on each factor were derived by summing their scores on items representing each factor. The items were scored +3 to -3, but, due to the fact that the earlier hypotheses concerned intensity or strength of perception, absolute values were summed to provide the dependent scores.

A separate 3x3 (advertising by price factorial multivariate analysis of variance was performed for each product class. The dependent variables were individuals' scores on each of the three factors. A significant (p<0.01) price main effect was obtained for television and for soft drink (see Table 2). No other significant main or interaction effects were obtained. Univariate analyses show that in both cases the price effect was due to factor 3. Multiple comparison tests (i.e., Tukey's Test, Kirk, 1968) indicated that the effects were due to the difference between the control and high price conditions. This can also be seen by referring to the table of means (Table 3).

TABLE 1

ITEMS REPRESENTING FIRST 3 FACTORS FOR EACH PRODUCT

TABLE 2

SUMMARY TABLE: MULTIVARIATE ANALYSES OF VARIANCE

TABLE 3

MARGINAL MEANS FOR SIGNIFICANT EFFECTS

DISCUSSION

These preliminary results provide some support for the contention that cue utilization in product evaluation is directly related to product complexity. Price was related to product perception only for the two products found to be relatively complex. Further work is necessary, however, to substantiate this finding.

The results failed to support Nelson's contention that the amount of advertising provided a product is a cue to the product's value. The advertising manipulation check was found to be significant for three of the four products. Even though subjects differed significantly in their belief about how heavily the product was to be advertised, this difference did not affect their perception of the product.

The failure to find a significant cue effect for advertising prevented the testing of the hypothesis suggested earlier that the strength of the inference is directly related to the consistency in meaning of provided cues.

It would be desirable to employ a multi-method approach to the measurement of product complexity. The results provided by the Bieri, et. al. method could be compared with those obtained when other methods are employed. This could provide evidence to support the convergent validity of an instrument to measure product complexity. Vannoy (1965) found some evidence to support the convergent validity of instruments used to measure cognitive complexity. Work is in progress comparing results of the Bieri, et. al. approach with those obtained through factor analysis.

Further assessment of differences in cue utilization across product complexity levels would also be desirable. This could be accomplished by assessing the effects of cues on subjects' evaluations across a larger sample of products found to vary in terms of product complexity. If supportive results were obtained in these other studies it would make an alternative hypothesis that the results are due to product differences other than the typical complexity of its representation in cognitive structure less likely than it is from an assessment of only four products.

The failure to obtain stronger results in the study may be due, in Dart, to use of non-student women as subjects. These women, many of whom were older did not seem as comfortable and familiar as are students with the environment and tasks to which they were subjected. Most studies of price effects have employed students as subjects (Olson, 1976). It is likely, however, that different groups of people differ in their use of cues. It is also likely that complexity of a product's representation in cognitive structure differs by group. It is, therefore, important in this type of study to employ subjects representative of the target market for the chosen products. The women employed in this study seem to fit this criterion.

Olson (1976) has maintained that a theoretical orientation to future product perception research is necessary if we are to gain an understanding of cue utilization in product perception. It is believed here that attribution theory can contribute to this goal. Our consideration of the implications of cue consistency represents an initial basic hypothesis related to attribution theory. Many other hypotheses are also derivable from this work. In future research the theory could be adapted in a much more complete fashion. This could contribute greatly to our understanding of how, when and with what effect people employ cues in product evaluation.

Kelley (1967, 1973) employs four criteria in his treatment of attributional validity:

1. Distinctiveness: the impression is attributed to the thing if it uniquely occurs when the thing is present and does not occur in its absence.

2. Consistency over time: each time the thing is present, the individual's reaction must be the same or nearly so.

3. Consistency over modality: his reaction must be consistent even though his mode of interaction with the thing varies.

4. Consensus: attributes of external origin are experienced the same way by all observers.  (Kelley, 1967, p 197).

Subjects in attribution studies are typically given a behavior sequence such as: John is the only person who laughs at the clown (i.e., McArthur, 1972). Subjects are then asked to attribute John's behavior to something about John (a person attribution), something about the clown (an entity attribution) or something about the situation (circumstance). When consensus (hardly anyone else laughs at the comedian) and distinctiveness (John also laughs at most other comedians) are low and consistency is high (John has almost always laughed at the comedian) an attribution will be made to the person. When consensus (almost everyone laughs at the comedian), distinctiveness (John does not laugh at almost any other comedian) and consistency are high an attribution will be made to the entity (McArthur, 1972).

We may view an entity as a product (or brand). A product attribution will then be most likely when consensus, consistency and distinctiveness are high. Similarly we may expect the strongest product attribution when consensus, consistency and distinctiveness are high. If everybody buys brand B (high consensus); if John, for instance, does not buy any other brand (high distinctiveness); and if John always buys brand B (high consistency) then a strong attribution should be made to brand B. Presumably, the attributor would conclude that brand B is the best brand to buy.

We may adapt the theory even further by viewing consistency (in meaning) of cues related to the product as consistency over modality. That is, if every cue associated with the product provides a consistent implication then we have an instance of high consistency over modality. We may also wish to consider product class heterogeneity to represent distinctiveness. If high consistency and high distinctiveness are combined with high consensus we would expect a strong attribution to the product.

Further research would, of course, be necessary to substantiate these propositions. If people make inferences about products the way they do about other people then attribution theory should be relevant to the prediction of these inferences. Research directed at providing support for the propositions made here should contribute to a greater understanding of how and when people use cues in their evaluation of products. It should permit the marketer to do a better job of consciously and appropriately employing cues.

REFERENCES

I. R. Andrews and E. R. Valenzi, "The Relationship Between Price and Blind-Rated Quality for Margarines and Butter," Journal of Marketing Research, 7(August, 1970), 393-395.

M. Bieri, A. L. Atkins, S. Briar, R. L. Leaman, H. Miller and T. Tripodi, Clinical and Social Judgment: the Discrimination of Behavioral Information (New York: Wiley, 1966).

R. W. Gardner and R. A. Schoen, "Differentiation and Abstraction in Concept Formation," Psychological Monographs, 76(Whole No. 560, 1962).

J. Jacoby, M. C. Olson and R. A. Haddock, "Price, Brand Name, and Product Composition Characteristics as Determinants of Perceived Quality," Journal of Applied Psychology, 55(1971), 570-579.

H. H. Kelley, "Attribution Theory in Social Psychology," Nebraska Symposium on Motivation, 15 (1967), 192-238.

H. H. Kelley, "The Process of Causal Attribution," American Psychologist, 28 (1973), 107-128.

G. A. Kelly, A Theory of Personality: the Psychology of Personal Constructs (New York: Norton, 1955).

R. E. Kirk, Experimental Design: Procedures for the Behavioral Sciences (Belmont, California: Brooks/Cole), 19681

L. A. McArthur, "The How and What of Why: Some Determinants and Consequences of Causal Attribution," Journal of Personality and Social Psychology, 22(1972), 171-193.

M. D. McConnell, "The Price-Quality Relationship in an Experimental Setting," Journal of Marketing Research, 5(1968), 300-303.

K. B. Monroe, "The Influence of Price Differences and Brand Familiarity on Brand Preferences," Journal of Consumer Research, 3(1976), 42-49.

P. Nelson, "Advertising as Information," Journal of Political Economy, 81(1974), 729-754.

J. C. Olson, "Price as an Informational Cue: Effects on Product Evaluations," Working Paper Number 43 (The Pennsylvania State University: College of Business Administration, 1976).

C. E. Osgood, G. J. Suci and P. H. Tannenbaum, The Measurement of Meaning (Urbana: University of Illinois Press, 1957).

H. M. Schroder, M. J. Driver and S. Streufert, Human Information Processing (New York: Holt, Rinehart and Winston, 1967).

E. M. Smith and C. L. Broome, "Experimental Determination of the Effect of Price and Market-Standing Information on Consumer Brand References." In R. Hass (ed.), Science, Technology and Marketing (Chicago: American Marketing Association, 1966), 520-531.

D. S. Tull, R. A. Boring and M. H. Gonsior, "A Note on the Relationship of Price and Imputed Quality," Journal of Business, 37 (1964), 186-191.

E. R. Valenzi and I. R. Andrews, "Effect of Price Information on Product Quality Ratings," Journal of Applied Psychology, 55(1971), 87-91.

J. S. Vannoy, "Generality of Cognitive Complexity-Simplicity as a Personality Constructs," Journal of Personality and Social Psychology, 2(1965), 385-396.

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