Comment on Classically Conditioning Human Consumers

ABSTRACT - The assumptions underlying recent articles advocating attention by consumer behavior scholars to classical conditioning theory are examined. The advocacy is viewed as unjustified because it relies heavily on animal studies that use organisms whose cognitive processes differ from normal humans and because conditioning theorists have yet to address adequately the thorny problem of awareness. Conditioning research should routinely assess awareness of human subjects. To date studies observing this criterion have not supported classical conditioning theory as useful in consumer research.

Citation:

Lynn R. Kahle, Sharon E. Beatty, and Patricia Kennedy (1987) ,"Comment on Classically Conditioning Human Consumers", in NA - Advances in Consumer Research Volume 14, eds. Melanie Wallendorf and Paul Anderson, Provo, UT : Association for Consumer Research, Pages: 411-414.

Advances in Consumer Research Volume 14, 1987      Pages 411-414

COMMENT ON CLASSICALLY CONDITIONING HUMAN CONSUMERS

Lynn R. Kahle, University of Oregon

Sharon E. Beatty, University of Alabama

Patricia Kennedy, University of Oregon

ABSTRACT -

The assumptions underlying recent articles advocating attention by consumer behavior scholars to classical conditioning theory are examined. The advocacy is viewed as unjustified because it relies heavily on animal studies that use organisms whose cognitive processes differ from normal humans and because conditioning theorists have yet to address adequately the thorny problem of awareness. Conditioning research should routinely assess awareness of human subjects. To date studies observing this criterion have not supported classical conditioning theory as useful in consumer research.

INTRODUCTION

In 1974 Brewer suggested that there is no convincing evidence of the classical conditioning of normal human adults. He believed that no experiment had to date unambiguously demonstrated that normal human adults are classically conditioned without awareness, as conditioning theory requires. Although in marketing some authors question the importance of conditioning theory (e.g., Assael 1984), one recent review of classical conditioning written for marketers (McSweeney and Bierley 1984), which implied that normal human adults can be conditioned, presented an especially sophisticated analysis of the controversy over the viability of classical conditioning from the perspective of a conditioning proponent. Because these authors clearly had a better understanding and explanation of the issues involved than some conditioning proponents, a comment from an alternative perspective may be warranted.

Awareness in the conditioning paradigm implies that research participants grasp the nature of the hypotheses of the study in which they are participating. Aware subjects frequently behave in such a way as to comply with these demand characteristics of the experiment in order to make themselves and the experimenter look good.

Proponents of classical conditioning theory frequently insist on discounting or trivializing the controversy over the role of awareness and demand characteristics in the classical conditioning experiment. They often argue that we ought to ignore this "counterproductive" controversy, in order to use our time more efficiently on worthwhile tasks, such as increasing our understanding of other aspects of classical conditioning (cf. McSweeney and Bierley 1984). Yet the controversy over classical conditioning and demand characteristics is neither trivial nor theoretically unimportant, and it ought to be resolved, if possible, before any further attempts to elaborate conditioning theories proceed.

HISTORIC CONTEXT OF CLASSICAL CONDITIONING

A few notes on history may help to put the current controversy into an interpretable context. Classical conditioning theory grew from the work of Pavlov and especially J. B. Watson, in part as a response to the excesses of introspectionists in the early twentieth century. Many of the introspectionist propositions were not amenable to scientific investigation, and the conditioning theorists sought to rectify this weakness by insisting on the use of behavior, which is readily observable, as the fundamental fact of social science. At that time the objectification of social science was an important and valuable contribution of conditioning theorists. Conditioning theory and behaviorism grew together, inextricably intertwined as a methodological and theoretical approach to the objectification of social science.

When Staats and Staats (1957) articulated the relevance of classical conditioning for attitudes and human social behavior in their classic article on the topic, it was no accident that they insisted, "The meaning of stimuli may be learned without awareness--without cognition" (Staats and Staats 1957, p. 76). Awareness and cognition by definition nave no role in the science of behavior in this view. By insisting on the irrelevance of cognition and awareness, Staats and Staats were making a pointed statement to their colleagues in social psychology, who at the time seemed infatuated with cognitive approaches, such as balance theory (Heider 1958), cognitive dissonance theory (Festinger 1957), and a host of similar explications of the function of cognition by such social scientists as Cartwright and Harary, Newcomb, and Osgood and Tannenbaum (cf. Kahle 1984). To study anything besides behavior, in Staats and Staats' view, is to compromise science in much the same way as the earlier introspectionists. The Staats and Staats theory of conditioning may be the most useful to marketing because it carefully considers the nature of higher order conditioning and the relationship between attitudes and conditioning.

This view of Staats and Staats on awareness was not an extreme view among conditioning theorists. It was rather an orthodoxy inseparable from the most fundamental tenants of the theory. Thus, when authors such as Cohen (1964) and Page (1969) proposed that awareness and experimental demand characteristics provide an alternative explanation to the results of conditioning experiments, classical conditioning proponents quickly rallied to attempt to disprove and discount what in their view was an apparent throwback to earlier theoretical days. Awareness presents a potential methodological problem for all types of deception research conducted on adult human subjects, but it is a key issue in classical conditioning studies because awareness is a theoretical issue as well.

The issue addressed here is perhaps the most fundamental question of social science--namely, what is social science to study. Is the fundamental fact of social science behavior or cognition? Can future actions be best predicted from previous behavior, as behaviorists maintain, or does some cognitive factor, such as intention (cf. Fishbein and Ajzen 1975), better illuminate future actions? If behavior is not a sufficient and complete topic for investigation, then classical conditioning theory is inadequate in its account of social behavior because the most basic premise of classical behaviorism is undermined. "Cognitive classical conditioning" is an inherently contradictory phrase because what distinguishes classical conditioning theory from numerous other theoretical accounts of the consequences of pairing events (e.g., Kahle 1984) is its insistence on behavior as the fundamental fact of social science.

THE EMPIRICAL EVIDENCE FROM PSYCHOLOGY

Numerous studies have demonstrated that among normal human adults the empirical phenomenon of classical conditioning only occurs when subjects are cognitively aware (e.g., Cohen 1965; Page 1969). Typically, these studies measure awareness immediately after measuring evidence of conditioning, and it is demonstrated that "conditioned" subjects knew the hypothesis of the experiment and complied with it. Aware subjects often view these experiments as problem-solving studies, in which their task is to figure out that the experimenter wants them to like certain words .

Proponents of conditioning have criticized these experiments because the measures of awareness are often relatively long, "funnel-type" questionnaires, but these questionnaires have been shown to be more valid (e.g., Page 1971, 1973) than the shorter, two-item questionnaires proponents of conditioning utilize. The two-item questionnaires ("Write down anything you thought during the experiment" and "What was the purpose of this experiment?") often generate weak or irrelevant data (e.g., "I thought about the girl sitting next to me and about my test this afternoon" and "The purpose was for me to get research participation credit and for the experimenter to get a publication"). Some sophisticated subjects furthermore know that they are not "supposed to" give the true purpose of the experiment on the post-experimental questionnaire, although they do know the purpose. Orne (1962) has called this the "pact of ignorance" between the experimenter and the sophisticated subject.. Hence, it is necessary to have the longer, funnel-type questionnaire that starts with abstract questions ("write down anything you thought during the experiment") and uses a series of questions that lead to a specific inquiry about the purpose of this study.

Proponents of conditioning have also criticized these experiments as "correlational," pointing out that in some instances correlation is not causation. This argument is quite weak as evidence for classical conditioning, however, because conditioning theorists predict no correlation between awareness and behavior where one is observed. Correlational studies such as these provide necessary but not sufficient evidence for the cognitive interpretation of conditioning experiments and may thus be criticized. But they clearly provide necessary and sufficient falsification for the conditioning interpretation; hence, their "noncausal" nature hardly supports conditioning theory.

Several true experiments also have addressed the issues of classical conditioning and demand awareness using evidence that should be acceptable to behaviorists. Page (1974), for example, "conditioned" subjects and then, just prior to measuring attitudes, told subjects to do the opposite of what they were supposed to do if they knew of the experimenter's hypothesis. Compared to a normal conditioning experiment, Page found a mirror image of the normal bi-modal distribution of subjects. Instead of a mode of aware subjects with scores in the direction of the experimental conditioning, Page found a mode of subjects with scores in the opposite direction of the conditioning theory prediction. (The second mode consists of unconditioned or unaware subjects.) Evidently, some subjects did know what they were supposed to do and did it, while other subjects were neither aware nor "conditioned." The mere existence of a bi-modal distribution in conditioning experiments is difficult to explain from conditioning theory, but the mobility of the "conditioned" mode due to a minor experimental manipulation is clearly contrary to conditioning theory.

Kahle and Page (1976) also carried out a true experiment that provides evidence interpretable within the cognitive but not conditioning framework. They replicated a study by Staats, Minke, Martin, and Higa (1972), which had been offered as evidence of the "physiological" (i.e., noncognitive) nature of classically conditioning attitudes. Staats et al. showed nonsense syllables paired with food words to subjects who either were or were not hungry. Hungry subjects liked the nonsense syllables better than did full subjects, and Staats et al. interpreted these data as evidence of the physiological basis of attitude conditioning. Kahle and Page held physiology, the crucial conditioning variable, constant by using only food-satiated subjects. They preserved the demand characteristics of the original study by telling half of the subjects to pretend to be hungry. "Conditioning" was just as effective among subjects who were exposed to the demand characteristic of the Staats et al. study but lacked the crucial-for-conditioning physiological difference. Page and Kahle (1976) also demonstrated that the Staats et al. paradigm only works with subjects who have already learned about classical conditioning in their psychology classes. The cognitive interpretation of both of these studies is that knowledge of the experimenter's goals facilitates "conditioning" of normal human adults, but the conditioning interpretation is not apparent.

Much of the more recent literature in psychology has built on the base of assuming that awareness undermines classical conditioning explanations of the empirical phenomenon under consideration here (e.g., Baumrind 1985; Cahoon and Edmonds 1984; Carlopio, Adair, Lindsay and Spinner 1983; Dawson, Schell, Beers and Kelly 1982; Dulany, Carlson, and Dewey 1985; Perruchet 1985; Reardon and Katz 1983).

THE EMPIRICAL EVIDENCE FROM THE ANIMAL KINGDOM

The question of what subjects to use also enters into this debate. Do animals represent a legitimate topic of study in consumer behavior? Certainly for some topics where animAls do influence consumer choices (e.g., dog food), animals should be studied. And animals should be studied when problems and processes between animals and humans are quite similar and practical or ethical considerations restrict human research (e.g., research regarding a product's carcinogenic potential). Furthermore, animal research is interesting and valuable in its own right, regardless of its implications for consumer behavior. For example, we would support and encourage a biologist who wants to know why pigeons peck or how whales communicate.

But studying animals to learn about human consumer behavior per se necessarily relies on analogy. Animals differ from humans in many respects: Animals cannot imagine, symbolize, or plan in the same manner as humans. Humans can cognitively override even basic physiological processes, as in anorexia nervosa. Humans differ from animals in culture, customs, and, of course, genetics. The basic anatomy of information processing in humans is unique. Thus, the burden of proving the analogy from animal behavior to human consumer behavior is always on the researcher, and in the area of classical conditioning ample evidence shows that the analogy is invalid.

McSweeney and Bierley (1984) find it not "intuitively appealing" to assume that the behavior of animals and humans differ, but we disagree. To us it seems, on the contrary, intuitively obvious that the consumer behaviors of humans and, say, goldfish (cf. Bitterman 1964, a study McSweeney and Bierley cite) do differ. Philosophers of science even have a name, anthropomorphism, for the logical fallacy of assuming that nonhumans have human characteristics. In a discipline where research resources are precious and where some critics question the appropriateness of research on even the venerable college sophomore, the intuitive appeal of consumer research on goldfish, dogs, sheep, rats, rabbits, goats, and pigeons may not be universally obvious. Yet virtually all of the empirical studies cited by McSweeney and Bierley rely on these nonhuman organisms, with a strong preference for rats and pigeons.

THE EMPIRICAL EVIDENCE IN CONSUMER BEHAVIOR

The most cited study of classically conditioning normal human consumers (Gorn 1982) paired music with a slide of a pen once and then found that people selected the pen more if the music had been pleasant than unpleasant. This study failed to employ the "funnel-type" post-experimental questionnaire (Page 1973) that has been shown to be more valid at assessing demand awareness.

Another human study (Milliman 1982) has also been interpreted as providing evidence of classical conditioning, although its author made no reference to classical conditioning. In that study, fast versus slow background music related to total store sales, although a lack of music did not differ from either of the other conditions. Milliman just barely detected a significant difference (p = .05) using all data from all shoppers in a store between January 28 and March 31 and using an optimally powerful statistical procedure (ANOVA). With a less powerful statistic (Chi Square) and a tiny subset of subjects, the tendency toward awareness, measured with only one item, failed to attain statistical significance. We do not know whether a more comprehensive sampling technique, a more valid measure of awareness, and a more powerful data gathering and statistical procedure would have changed the inference that awareness did not influence the outcome of this study, but these features would have been necessary had the author sought to prove classical conditioning. This study is interesting and contributes to the consumer behavior literature, but it does not provide compelling evidence for classical conditioning (nor does Milliman make such a claim). A variety of theoretical accounts can explain the roles of mood and emotion in human behavior (cf. Dienstbier 1979). For example, social adaptation theory (Kahle 1984) maintains that situational and cognitive factors interact in determining behavior. Behavior is more related to adaptation than to conditioning in this view.

Allen and Madden (1985) clearly did use an appropriate funnel-type postexperimental questionnaire in an attempted conceptual replication of Gorn (1982). This conceptual replication failed to replicate Gorn on parallel dependent variables. Interestingly, Allen and Madden report that on a unique dependent variable--subjects' willingness to sell back a pen they had been "conditioned" to select-conditioning without awareness did occur. Unfortunately, conditioning did not occur for pen selection, calling into question the mediation process. The authors report the probability level associated with their inference but do not detail how they computed it. In any event, the authors proceed to articulate an interesting position that is quite different from traditional classical conditioning theory. At this time insufficient data testing their theory have been presented either to endorse or reject it.

CONCLUSION

In conclusion, it appears that the data have not yet come forth to contradict Brewer's (1974) claim that there is no convincing evidence of classical conditioning in normal human adults. The impact of classical conditioning theory has been particularly important in consumer behavior, perhaps because when J. B. Watson left psychology for the field of advertising, he brought with him his psychological and conditioning expertise. But now empirical studies on conditioning adult humans seem less useful because other theoretical approaches, which view human behavior as motivationally and cognitively more complex, appear more fruitful (e.g., Kahle 1984). Dawson, Schell, seers, and Kelly (1982) take essentially the same perspective as the present review--that awareness is a persisting problem in classical conditioning research. If nearly a century of research has not produced the convincing evidence of classical conditioning sought by Brewer, perhaps researchers should study new topics.

REFERENCES

Adair, John G., Terrance W. Dushenko, and R. C. L. Lindsay (1985), "Ethical Regulations and Their Impact on Research Practice." American Psychologist, 40 (January), 59-72.

Adair, John G. and Barry Spinner (1981), "Subjects' Access to Cognitive Processes: Demand Characteristics and Verbal Reports," Journal for the Theory of Social Behavior, 11 (March), 31-52.

Allen, Chris T., and Thomas J. Madden (1985), "A Closer Look at Classical Conditioning," Journal of Consumer Research, 12 (December), 301-315.

Assael, Henry (1984), Consumer Behavior and Marketing Action, Second ed. Boston: Kent Publishing.

Baumrind, Diane (1985), "Research Using Intentional Deception: Ethical Issues Revisited," American Psychologist, 40 (February), 165-174.

Bierley, Calvin, Frances K. McSweeney, and Renee Vannieuwkerk (1985), "Classical Conditioning of Preferences for Stimuli," Journal of Consumer Research, 12 (December), 316-323.

Bitterman, M. E. (1964), "Classical Conditioning in the Goldfish as a Function of the CS-US Interval," Journal of ComParatiVe and Physiological Psychology, 58 (December), 359-366.

Brewer, William F. (1974), "There is No Convincing Evidence for Operant or Classical Conditioning in Humans," in Cognition and Symbolic Processes, eds. W. R. Weiner and D. S. Palermo, Hillsdale, NJ: Lawrence Erlbaum, 1-35.

Cahoon, Delwin D. and Ed M. Edmonds (1984), "Guns/No Guns and the Expression of Social Hostility," Bulletin of the Psychonomic Society, 22 (July), 305-308.

Carlopio, James, John G. Adair, R. C. L. Lindsay, and Barry Spinner (1983), "Avoiding Artifact in the Search For Bias: The Importance of Assessing Subjects' Perceptions of the Experiment," Journal of Personality and Social Psychology, 44 (April), 693-701.

Cohen, Burton H. (1964), "Role of Awareness in Meaning Established by Classical Conditioning," Journal of Experimental Psychology, 67 (April), 373-378.

Dawson, Michael E., Anne M. Schell, James R. Beers, and Andrew Kelly (1982), "Allocation Of Cognitive Processing Capacity During Human Autonomic Classical Conditioning," Journal of ExPerimental Psychology: General, 111 (September), 273-295.

Dienstbier, Richard A., ed. (1978), Nebraska Symposium on Motivation, Vol. 26. Lincoln, NE: University of Nebraska Press.

Dulany, Don E., Richard A. Carlson, and Gerald I. Dewey (1985), "On Consciousness in Syntactic Learning and Judgment," Journal of ExPerimental Psychology General, 114, 25-32.

Festinger, Leon (1957), A Theory of Cognitive Dissonance, Stanford, CA: Stanford University Press.

Fishbein, Martin and Icek Ajzen (1975), Belief, Attitude, Intention, and Behavior: An Introduction to Theory and Research, Reading, MA: Addison-Wesley.

Gorn, Gerald J. (1982), "The Effects of Music in Advertising on Choice Behavior: A Classical Conditioning Approach," Journal of Marketing, 46 (Winter), 94-101.

Heider, Fritz (1958), The Psychology of Interpersonal Relations, New York: John Wiley.

Kahle, Lynn R. (1984), Attitudes and Social AdaPtation: A Person-Situation Interaction Approach, Oxford, U.R.: Pergamon.

Kahle, Lynn R. and Monte M. Page (1976), "The Deprivation Satiation Effect in Attitude Conditioning Without Deprivation but with Demand Characteristics," Personality and Social Psychology Bulletin, 2 (Fall), 470-473.

McSweeney, Frances K. and Calvin Bierley (1984), "Recent Developments in Classical Conditioning," Journal of Consumer Research, 11 (September), 619-631.

Milliman, Ronald E. (1982), "Using Background Music to Affect the Behavior of Supermarket Shoppers," Journal of Marketing, 46 (Summer), 86-91.

Orne, Martin T. (1962), "On the Social Psychology of the Psychological Experiment," American Psychologist, 17 (November), 776-783.

Page, Monte H. (1969), "Social Psychology of a Classical Conditioning of Attitudes Experiment," Journal of Personality and Social Psychology, 11 (February), 177-186.

Page, Monte M. (1971), "Postexperimental Assessment of Awareness in Attitude Conditioning," Educational and Psychological Measurement, 31 (Autumn-Winter), 891-906.

Page, Monte M. (1973), "On Detecting Demand Awareness by Postexperimental Questionnaire," Journal of Social Psychology, 91 (December), 305-322.

Page, Monte M. (1974), "Demand Characteristics and the Classical Conditioning of Attitudes Experiment," Journal of Personality and Social Psychology, 30 (October), 468-476.

Page, Monte M. and Lynn R. Kahle (1976), "Demand Characteristics in the Satiation-Deprivation Effect on Attitude Conditioning," Journal of Personality and Social Psychology, 33 (May), 553-562.

Perruchet, Pierre (1985), "Expectancy for Airpuff and Conditioned Eyeblinks in Humans," Acta Psychologica, 58 (January), 31-44.

Reardon, Richard and Stuart Katz (1983), "Comprehending Ambiguity in the Sentence-Verification Paradigm: Basic Processes of Problem Solving?" Bulletin of the Psychonomic Society, 21 (September ), 373-376.

Staats, Arthur W., K. A. Minke, C. H. Martin, and W. R. Higa (1972), "Deprivation-Satiation and Strength of Attitude Conditioning: A Test of Attitude-Reinforcer-Discriminative Theory," Journal of Personality and Social Psychology, 24 (November), 178-185.

Staats, Carolyn K. and Arthur W. Staats (1957), "Meaning Established by Classical Conditioning," Journal of Experimental Psychology, 54 (July), 74-80.

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