Consumer Expertise and the Feature-Positive Effect: Implications For Judgment and Inference

ABSTRACT - Several recent experiments have shown that people tend to overlook nonoccurrences and missing information ("out of sight, out of mind") even when such nonevents have important implications for judgment and choice. This phenomenon is known as the feature-positive effect. The present experiment indicates that consumers are more likely to overlook missing attribute information as expertise decreases. Implications of the results for consumer judgment and inference are discussed.

Citation:

Frank R. Kardes, David M. Sanbonmatsu, and Paul M. Herr (1990) ,"Consumer Expertise and the Feature-Positive Effect: Implications For Judgment and Inference", in NA - Advances in Consumer Research Volume 17, eds. Marvin E. Goldberg, Gerald Gorn, and Richard W. Pollay, Provo, UT : Association for Consumer Research, Pages: 351-354.

Advances in Consumer Research Volume 17, 1990      Pages 351-354

CONSUMER EXPERTISE AND THE FEATURE-POSITIVE EFFECT: IMPLICATIONS FOR JUDGMENT AND INFERENCE

Frank R. Kardes, University of Cincinnati

David M. Sanbonmatsu, University of Utah

Paul M. Herr, Indiana University

ABSTRACT -

Several recent experiments have shown that people tend to overlook nonoccurrences and missing information ("out of sight, out of mind") even when such nonevents have important implications for judgment and choice. This phenomenon is known as the feature-positive effect. The present experiment indicates that consumers are more likely to overlook missing attribute information as expertise decreases. Implications of the results for consumer judgment and inference are discussed.

THE FEATURE-POSITIVE EFFECT

Time constraints and limited resources often force consumers to make product-related judgments and decisions on the basis of incomplete information. Moreover, advertisements, package labels, consumer magazines, and other sources of information typically focus on only a few product attributes. How do consumers deal with partial information? Some studies suggest that consumers form inferences about unmentioned attributes and that these inferences are integrated with judgments of known attributes to arrive at an overall evaluation (Ford and Smith, 1987; Huber and McCann 1982; Johnson and Levin 1985; Meyer 1981), whereas other studies imply that consumers fail to form inferences about unmentioned attributes (Lim, Olshavsky, and Kim 1988). The present experiment focuses on the ability of consumers to detect the absence of information about product attributes. Logically, consumers must first realize that information about an important attribute is missing before they can infer a value for that attribute.

When is missing information overlooked? Jenkins and Sainsbury (1969, 1970) found that discrimination learning in pigeons occurs much faster in the presence than in the absence of a predictive cue, even when the absence of the cue is as informative as its presence. They labeled this phenomenon the feature-positive effect. Newman, Wolff, and Hearst (1980) demonstrated that this effect generalizes to people. In this study subjects were shown a series of cards containing two three-letter strings, a target string and a distractor. In feature-negative conditions, the critical feature was the absence of the letter "T." Performance feedback was presented after each trial, and, as anticipated, the problem was solved in fewer trials in feature-positive than in feature-negative conditions.

Research in areas other than discrimination learning also suggests that people experience difficulty in processing nonoccurrences. For example, when attitudes are weak or nonexistent, people use their own overt behaviors to draw inferences about their attitudes (Bem 1972). However, nonbehaviors (e.g., not laughing at a cartoon) tend to be overlooked (Fazio, Sherman, and Herr 1982; Allison and Messick 1988). Furthermore, in hypothesis testing, people tend to test cases that are expected to possess the property of interest (the positive test strategy) rather than on cases that are expected to lack the property (Klayman and Ha 1987). In covariation judgments, people focus on cases where the expected cause and the expected effect are both present (the positive-positive cell) rather than on cases where one or both events are absent (Bettman, John, and Scott 1986; John, Scott, and Bettman 1986). Finally, production and R&D departments often use checklists or fault trees to troubleshoot a system. Items that are not on the list, however, tend to be overlooked (Fischhoff, Slovic, and Lichtenstein 1978: Hirt and Castellan 1988).

THE ROLE OF EXPERTISE

When are consumers unlikely to overlook missing information? One factor that is likely to influence sensitivity to omissions is expertise. In their comprehensive review, Alba and Hutchinson (1987) defined expertise as "the ability to perform product-related tasks successfully" (p. 411) and they discussed how experts and novices differ on five important epistemological dimensions: cognitive effort, cognitive structure, analysis, elaboration, and memory. Analysis, or the extent to which attention is directed towards all and only important information, is particularly relevant to the present study. If sensitivity to important information increases as expertise increases (Brucks 1985; Furse, Punj, and Stewart 1984; Johnson and Russo 1984), then sensitivity to omitted important information should also increase with expertise.

THE PRESENT STUDY

Subjects received an incomplete description of a fictitious ten-speed bicycle. Information about two important attributes, the weight of the bicycle and the strength of the frame, was deliberately omitted. Following Sujan (1985), a subjective measure and a ten-item multiple-choice questionnaire was developed to assess expertise about bicycles. Subjects were divided into low, moderate, and high expertise groups on the basis of their scores on these measures. In addition, half of the subjects received a prompt to consider the weight of the bicycle and frame strength when evaluating the bicycle and half received no such prompt.

Evaluative judgments and confidence judgments served as the primary dependent measures. It was reasoned that if subjects spontaneously detected the absence of information about weight and frame strength that the prompt manipulation would have no effect on these judgments. Thus, if experts have a greater ability to recognize the absence of important attribute information, their product evaluations and their confidence ratings should not differ as a function of the prompt manipulation. Moreover, because of the uncertainty generated by the lack of information about these important attributes, experts should exhibit less extreme evaluations and lower confidence ratings, relative to less expert subjects.

The prompt manipulation should have a strong impact on the judgments of moderately knowledgeable subjects. If these subjects are insensitive to important omissions, incomplete information would seem relatively complete. When people treat incomplete information as if it were complete, they are likely to form polarized evaluations and they are likely to hold these extreme judgments with a high degree of confidence. By contrast, when a prompt forces them to consider information that otherwise would have been overlooked, less extreme and less confidently-held judgments- should be formed.

Finally, novices are less able to identify and use important information (Alba and Hutchinson 1987). Thus, even if a cue highlighting the absence of important information is presented, they are unlikely to see the evaluative implications of this cue. Hence, the prompt manipulation should not influence the judgments of novices. Relatively extreme and confidently-held judgments should be formed whether or not the prompt is presented.

METHOD

Procedure

One hundred thirty-one male undergraduates were asked to participate in a study concerning how people evaluate products. They were asked to read a description of a ten-speed bicycle (Brand A) that was based on a "study conducted by a popular magazine for bicycle enthusiasts." They were given two minutes to read the description and were instructed to form an impression of the bicycle. Afterwards, the description was taken away and a questionnaire was administered.

Stimuli and Measures

The description of the Brand A ten-speed bicycle contained information pertaining to six important product attributes: braking, shifting, cornering, wheels, styling, and comfort. Information about two important attributes, the weight of the bicycle and the strength of the frame, was omitted. The presented attributes were described very favorably and care was taken to present attributes that were subjectively unrelated to the other presented attributes and to the unmentioned attributes. All subjects received the exact same product description.

Product evaluations were measured on three nine-point semantic differential scales anchored at 4 and +4, with endpoints labeled bad/good, unsatisfactory/satisfactory, and unfavorable/ favorable. These ratings were averaged to form a single product evaluation index. Subjects indicated their confidence in their evaluations on a scale ranging from 1 (Not at all confident) to 9 (Highly confident).

RESULTS

Product Evaluations

Alerting subjects to the fact that important attribute information is missing should affect their evaluations only if they failed to detect the absence of this information spontaneously, and if they are knowledgeable enough to understand the implications of the missing attributes. Thus, prompting should influence only moderately knowledgeable subjects. As predicted, moderately expert subjects formed more favorable evaluations in no prompt that in prompt conditions. In contrast, prompting did not influence the evaluations of low or high expertise subjects.

Confidence Judgments

It was predicted that the prompt manipulation would influence confidence judgments as well as evaluations. Again, prompting should influence only moderately knowledgeable subjects. Consistent with this prediction, moderately knowledgeable subjects had greater confidence in their evaluations in no prompt than in prompt conditions. In contrast, prompting did not influence the confidence judgments of low or high expertise subjects. Hence, considered together, the product evaluation and confidence judgment data indicate that experts are sensitive to omissions, whereas less knowledgeable individuals overlook important missing information.

DISCUSSION

The ability to recognize the absence of important attribute information plays a key role in consumer judgment and inference. In the present study, novices and even moderately knowledgeable individuals consistently overlooked important omitted information. When important information is overlooked, extreme, confidently-held judgments are formed. Experts, on the other hand, spontaneously detect the absence of important information and form more moderate judgments.

Research in several paradigms has shown that people often overlook missing information. These paradigms include discrimination learning (Jenkins and Sainsbury 1969, 1970), concept identification (Newman et al. 1980), self-perception (Fazio et al. 1982), false consensus (Allison and Messick 1988), hypothesis testing (Klayman and Ha 1987), covariation assessment (Bettman et al. 1986, John et al. 1986), and problem solving (Fischhoff et al. 1978, Hirt and Castellan 1988).

The results of the present study suggest that expertise should moderate sensitivity to omissions within each of these paradigms. This proposition should be tested directly because other moderating variables are also likely to be implicated. For example, some product attributes may be too important to be overlooked (e.g., price). Further, more extensive analytic processing is likely as time pressures decrease and as concern about the consequences of inferential errors increases (Kruglanski 1989, Kruglanski and Freund 1983).

Overconfidence

People are frequently overconfident about the validity of their judgments (Einhorn and Hogarth 1978; Fischhoff and Beyth-Marom 1983; Lichtenstein, Fischhoff, and Phillips 1982). Several explanations for overconfidence have been offered, including (a) people often receive incomplete feedback about the consequences of their decisions (e.g., after purchasing one brand from a set of alternatives, little is known about the performance of alternatives not chosen), (b) people tend to seek information that confirm their expectations, (c) confirmatory evidence is more accessible from memory, and (d) judgments influence and are influenced by outcomes (e.g., self-fulfilling prophecy). However, consumers are often unaware of the reciprocal relation between judgments and outcomes. Finally, our results suggest (e) that insensitivity to missing information leads consumers to treat incomplete information as if it were complete. When consumers believe they know more than they actually do, overconfidence results.

Inference Formation

The results of the present study also have implications for an apparent inconsistency in the consumer inference literature. Huber and McCann (1982) found that when explicitly presented attributes are subjectively related to omitted attributes, consumers spontaneously form inferences about omitted attributes. However, Lim et al. (1988) recently failed to replicate Huber and McCann's results. Lim et al. found that inferences were formed when consumers were explicitly asked to do so, but no evidence for spontaneous inference formation was found. The present results indicate that consumers are more likely to notice omissions as expertise increases. Because consumers must first detect omissions before they can draw inferences about these omissions, the likelihood of spontaneous inference formation should increase as expertise increases. If Huber and McCann's student subjects were more knowledgeable about their target product (beer) than Lim's et al. were about theirs (automobiles), spontaneous inference formation would be more likely in the Huber and McCann study.

To summarize, the results of the present study indicate that the ability to detect the absence of important information increases as expertise increases. Consumers who are sensitive to omissions form less extreme judgments and are less likely to exhibit overconfidence. Furthermore, the ability to detect and identify missing information is likely to serve as an important precondition for spontaneous inference formation. Consumers must notice the gaps before they can fill them.

REFERENCES

Alba, Joseph W.and J. Wesley Hutchinson (1987),"Dimensions of Consumer Expertise," Journal of Consumer Research, 13 (March), 411454.

Allison, Scott T.and David M. Messick (1988), "The Feature-Positive Effect, Attitude Strength, and Degree of Perceived Consensus," Personality and Social Psychology Bulletin, 14 (June), 231-241.

Bem, Daryl J. (1972), "Self-Perception Theory," in Advances in Experimental Social Psychology, Vol. 6, ed. Leonard Berkowitz, New York: Academic Press, 1-62.

Bettman, James R., Deborah Roedder John, Carol A. Scott (1986), "Covariation Assessment by Consumers," Journal of Consumer Research, 13, (December) 316-326.

Brucks, Merrie (1985), 'The Effects of Product Class Knowledge on Information Search Behavior," Journal of Consumer Research, 12 (June), 1-16.

Einhorn, Hillel J. and Robin M. Hogarth (1978), "Confidence in Judgement: Persistence in the Illusion of Validity, Psychological Review, 85, 395-416.

Fazio, Russell H., Steven J.Sherman, and Paul M. Herr (1982),"The Feature-Positive Effect in the Self-Perception Process: Does Not Doing Matter as Much as Doing?" Journal of Personality and Social Psychology, 42 (March), 404411.

Fischhoff, Baruch, Paul Slovic, and Sarah Lichtenstein (1978),"Fault Trees: Sensitivity of Estimated Failure Probabilities to Problem Representation," Journal of Experimental Psychology: Human Perception and Performance, 4, 330-334.

Fischhoff, Baruch and Ruth Beyth-Marom (1983), "Hypothesis Evaluation From a Bayesian Perspective," Psychological Review, 90 (July), 239-260.

Ford, Gary T. and Ruth Ann Smith (1987), "Inferential Beliefs in Consumer Evaluations: An Assessment of Alternative Processing Strategies," Journal of Consumer Research, 14 (December), 363-371.

Purse, David H., Girish N. Punj, and David W. Stewart (1984), "A Typology of Individual Search Strategies Among Purchasers of New Automobiles," Journal of Consumer Research, 10 (March), 417431.

Hirt, Edward R. and N. John Castellan (1988), "Probability and Category Redefinition in the Fault Tree Paradigm," Journal of Experimental Psychology: Human Perception and Performance, 14 (February), 122-131.

Huber, Joel and John McCann (1982), "The Impact of Inferential Beliefs on Product Evaluations," Journal of Marketing Research, 10 (August), 324-333.

Jenkins, H. M. and R. S. Sainesbury (1969), 'The Development of Stimulus Control Through Differential Reinforcement," in Fundamental Issues in Associative Learning, eds. N. J. Mackintosh and W. K. Honig, Halifax, Nova Scotia, Canada: Dalhousie University Press.

Jenkins, H. M. and R. S. Sainesbury (1970), "Discrimination Learning with the Distinctive Feature on Positive or Negative Trials," in Attention: Contemporary Theory and Analysis, ed. D. Mostofsky, New York: Appleton-Century-Crofts.

John, Deborah Roedder, Carol A. Scott, James R. Bettman (1986), "Sampling Data for Covariation Assessment: The Effects of Prior Beliefs on Search Patterns," Journal of Consumer Research, 13 (June), 38-47.

Johnson, Eric J. and J. Edward Russo (1984), "Product Familiarity and Learning New Information," Journal of Consumer Research, 11 (June), 542-550.

Johnson, Richard D. and Irwin P, Levin (1985), "More Than Meets the Eye: The Effect of Missing Information on Purchase Evaluations," Journal of Consumer Research, 12 (September), 169-177.

Klayman, Joshua and Young-Won Ha (1987), "Confirmation, Disconfirmation, and Information in Hypothesis Testing," Psychological Review, 94 (April), 211-228.

Kruglanski, Arie W. (1989), Lay Epistemics and Human Knowledge: Cognitive and Motivational Bases, New York: Plenum.

Kruglanski, Arie W. and Tallie Frenud (1983), "The Freezing and Unfreezing of Lay Inferences: Effects on Impressional Primacy, Ethnic Stereotyping, and Numerical Anchoring," Journal of Experimental Social Psychology, 19 (September), 448468.

Lichtenstein, Sarah, Baruch Fischhoff, and Lawrence D. Phillips (1982), "Calibration of Probabilities: The State of the Art to 1980," in Judgement Under Uncertainty: Heuristics and Biases, eds. Daniel Kahneman, Paul Slovic, and Amos Tversky, New York: Cambridge University Press, 306-334.

Lim, Jeen-Su, Richard Olshavsky, and John Kim (1988), "The Impact of Inferences on Product Evaluations: Replication and Extension," Journal of Marketing Research, 25 (August), 308316.

Meyer, Robert J. (1981), "A Model of Multiattribute Judgments Under Attribute Uncertainty and Informational Constraint," Journal of Marketing Research, 18 (November), 428441.

Newman, Joseph, William T. Wolff, and Eliot Hearst (1980), "The Feature-Positive Effect in Adult Human Subjects," Journal of Experimental Psychology: Human Learning and Memory, 6 (September), 630-650.

Sujan, Mita (1985), "Consumer Knowledge: Effects on Evaluation Strategies Mediating Consumer Judgments," Journal of Consumer Research, 12 (June), 31-46.

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