Achieving Correspondence Among Cognitive Processes and Physiological Measures

ABSTRACT - The empirical bias inherent in buyer behavior research will hinder the theoretical usefulness of psychophysiological measures. This situation will arise primarily from the confusion of relationships among independent and dependent variables and levels of abstraction. A return to some basics of theory construction together with more emphasis on conceptual analysis may alleviate this problem.

Citation:

Michael J Ryan (1982) ,"Achieving Correspondence Among Cognitive Processes and Physiological Measures", in NA - Advances in Consumer Research Volume 09, eds. Andrew Mitchell, Ann Abor, MI : Association for Consumer Research, Pages: 170-172.

Advances in Consumer Research Volume 9, 1982      Pages 170-172

ACHIEVING CORRESPONDENCE AMONG COGNITIVE PROCESSES AND PHYSIOLOGICAL MEASURES

Michael J Ryan, The University of Michigan

ABSTRACT -

The empirical bias inherent in buyer behavior research will hinder the theoretical usefulness of psychophysiological measures. This situation will arise primarily from the confusion of relationships among independent and dependent variables and levels of abstraction. A return to some basics of theory construction together with more emphasis on conceptual analysis may alleviate this problem.

INTRODUCTION

In order to develop the central thesis, a common ground must first be established concerning the notions of science, explanation, the role of concepts, and the relationships among abstractions and phenomena The applied orientation of buyer behavior has led to some confusion concerning science and technology Whereas fuller treatments are available elsewhere (Calder, Phillips, and Tybout 1981; O'Shaughnessy and Ryan 1979), a fundamental distinction must be made since this paper teals only with theory building which is in the realm of science. Science seeks explanation through abstraction whereas technology seeks prediction and control of everyday phenomena. Thus, there is a basic distinction in their objectives that dictates different methodological approaches.

LEVELS OF ABSTRACT - ION

Science seeks networks of concepts in which each concept accounts for a number of phenomena. Its objective, the explanation of the largest number of phenomena with the fewest number of concepts, is gained by overlaying networks of variables at more abstract levels (see Ryan and O'Shaughnessy 1980) Torgerson's illustration, shown in the Figure, captures this notion quite well. [The description that follows is based heavily on Kaplan's (1964) exposition. Slightly different viewpoints will be found in Baggozzi (1980) and Zaltman, et al (1973).]

At the extreme left are concepts which have no existence in the empirical world. Each concept is itself a mental process by which theoreticians account for concepts at lower levels of abstraction. For example, the concept cognitive dissonance may be useful in accounting for anger frustration and complaint intentions. What is important to note is that by their very nature concepts have no direct connection to the observable world they are general analogies useful for explaining "as if" they existed and their validity is judged by their usefulness If, for example, the "consolidation theory of memory" is replaced with a "depth of processing" theory that does not contain "short term memory this would imply that the new theory accounts for observable phenomena in a way more useful for the purpose at hand than did the old theory The existence of short term memory" is not an issue since its imaginary life in the theorists mint is quite apart from observable phenomena.

A concept depends upon theory for its meaning, not observable phenomena, as it has only systemic (Kaplan 1964) meaning. [The position taken here is based on the philosophical school known as instrumentalism. For a more detailed exposition and juxtaposition with other schools of thought see Messick (1981).] For example, cognitive dissonance derives its meaning from its relationships (the solid lines in the Figure) with other concepts contained in balance theory. Furthermore, meanings change depending on theoretical orientation. Thus, for example, the leaning of attitude depends on whether one subscribes to single component or tripartite attitude theory (Peter 1981).

FIGURE

RELATIONSHIPS AMONG ABSTRACT - AND EMPIRICAL VARIABLES

Constructs are inventions whereby we link concepts to the empirical world. They contain both systemic and observational meaning. Thus, they are linked to concepts through laws or propositions (the solid lines in the Figure) and to phenomena through rules of correspondence (the dashed lines in the Figure). Thus, "sensory input, short term memory," and long term memory may be viewed as constructs having systemic meaning within consolidation memory and observational meaning through rules of correspondence linking them to appropriate phenomena. Whether a term is a concept or construct will, of course, depend on the theory that contains it.

Phenomena refer to things recognizable by one or more of the five senses. By an appropriate phenomena is meant something that falls within the domain of the construct (Nunnally 1967). A specific type of recall, for example, may evidence a phenomenon only explainable by long term memory or a particular written statement may indicate dissatisfaction.

Operational definitions (the sets of parallel lines in the Figure) state how certain characteristics of the phenomena are quantified. For example, "strength" may be a characteristic of intentions which we may quantify through use of a probability scale or "duration" and "detail" may be characteristics of "recall" also susceptible to paper and pencil testing. The important point here is that we never directly measure any construct no matter how low its level of abstraction. We only quantify certain characteristics that we attribute to it (Jones 1971). We may, for example, measure the width, height, weight, light reflecting property, etc. of a table. Yet, table is an arbitrary category since these characteristics could just as well refer to "cube" or "box." In the same way, a galvanic skin response measure could be accounted for by a number of different abstractions such as sexual arousal, anger, frustration, etc. Thus, phenomena depend on abstractions for their meaning.

The above discussion suggests that it is a long way from concept, on the extreme left of the Figure, to observable, on the extreme right of the Figure. Furthermore, the interplay among the different levels of abstraction makes it impossible to view certain levels in isolation from others. Thus, measurement of any type cannot proceed without conceptual analysis.

CONSUMER RESEARCH ORIENTATION

The field as a whole is continually recognized for its deficiency in conceptual analysis and lack of theoretical frameworks (Ferber 1979; Jacoby 1978; Kollat, et al. 1970; Sheth 1967). Areas of research have been abandoned to conceptual confusion arising primarily from inadequate measurement procedures (e.g., Kassarjian 1971; Wilkie and Pessemier 1973) and the general problem of reification has been pointed out (Fishbein 1976).

Consider the following quote from a recent article on voice graph analysis (Nighswonger and Martin 1981): "Attitudes are constructs which exist in the mints of individuals" (p. 350, 1st line). [The notion that constructs actually exist is consistent with the philosophical school known as realism which flourished in the 18th and 19th century (see Kaplan, 1964. especially Chapter 2 and 8).] The only individuals in whose minds attitudes exist are the researchers not the subjects. Attitude may provide a conceptual scheme useful in interpreting voice pitch and voice pitch analysis may help document the usefulness but not the existence of attitude.

ORIENTATION OF PSYCHOPHYSIOLOGY

Due to tremendous technological limitations, psychophysiology has been preoccupied with the development of measurement apparatus. This orientation, together with the primary objective of understanding physiological processes, has produced a large body of literature primarily concerned with relationships among empirical phenomena (e.g., Thatcher and John 1977). For example, a tree sways back and forth in the breeze and a corresponding pattern of electrical impulses are detected in the brain. When the physiologist talks about electrical impulses as a representation of tree movement he is not referring to this representation as a mental construct. As Jacobson (1973) notes: "in electrophysiology, of course, we record signalization, not meaning" (p. 4). And "...when a physiologist speaks about mental activity to the experienced psychologist, he is talking in a foreign language' (p. 5). To the physiologist, then, a mental representation would consist of coordinated neural activity shifting and changing in an orderly position as the tree sways (Thatcher and John 1977, p. 135) me internal representation does not, of course, resemble a tree On the other hand, the behavioral theoretician would view the internal representation as the ideal type tree which, of course, does resemble a particular swaying tree, and is an imaginary concept useful for organizing and interpreting the world

Psychophysiologists do use mental concepts to organize physiological phenomena However, they are careful to make the distinction noted above between abstractions and phenomena Thus, for example, Cacioppo and Petty (1981) refer to electromyograms as responses which may be accounted for by "information processing, an abstraction. More to the point, Cacioppo and Petty clearly delineate levels of abstractions (e.g , models of memory + linguistic elaboration + self reference vs orthographic task) in a way that their theoretical framework provides meaning for observed EMG activity

PROBLEM PRONENESS

There are always problems not understood by the inventors and the borrowers concerning the magnitude of errors or inaccuracies inherent in the measurement process The empirical bias of consumer researchers suggests that they will focus on the complex technological limitations of physiological measures and overlook the conceptual limitations known to psychophysiologists. There are already indications that this bias will manifest itself in two ways; first, a misunderstanding of which concepts may account for psychophysiological phenomena, second, a lack of recognition of the role of artifacts and intervening variables.

Pupillometrics provides an apt example of the first problem. Reviewers of this literature report that, among other things, pupil dilation-contraction may be accounted for by affect, sexual arousal, discrimination of auditory pitch, cognitive processing, task difficulty, and memory Loading (e.g , Hess 1972; Watson and Gatchel 1979) Watson and Gatchel and Blackwell, et al (1970) conclude there is doubt as to which psychological processes underlie pupil response Yet, Hansen (1981), writing in the Journal of Consumer Research, reports that what is being measured is most frequently believed to be involvement, activity, or engagement in the issue Without a conceptual explanation this interpretation of selected empirical findings is simply misleading. In fact, Hess (1972) cites a number of empirical studies to conclude that pupillometrics is valuable because it reflects many different types of nervous system functioning and mirrors ongoing neurological activity in all parts of the brain He does not address the more interesting question as to how this general level of neurological activity can indicate distinct mental constructs

An excellent example of the intervening variable problem has been provided by Edelbert (1972) in regard to electrodermal responses (EDR). He states that EDR is an indicator of autonomic (primarily sympathetic) nervous activity and, more specifically, reflects the sympathetic inflow to the cutaneous area under observation

it is clear that more often than not the investigation in a psychophysiological experiment conceptually bypasses this step and equates electrodermal activity either to the level of arousal or to emotional activity. It should be apparent that this is an abstraction based on an assumption that there is a direct relation between sympathetic activity and these behavioral correlates, an assumption which, in view of the neuro-physiological subsrate of the electrodermal reflex may be unwarranted.

Writing in the Journal of Consumer Research, Kroeber-Riel (1979, 1980) does equate EDR to level of arousal In an analogous situation he also defines information as only the sensory input reaching short term memory. Yet, eye fixations are used to indicate information This interpretation is based on the assumption that information within the fixated area is necessarily processed Unfortunately, the selective and distorted perceptions that attend information processing mitigate against this assumption (Ryan 1980)

CONCLUSION

Psychophysiology may hold promise in the development of consumer behavior theory, especially in areas where verbal responses are inappropriate (cf. Nisbett and Wilson 1977) However, beliefs that physiological measures provide direct access to mental processes are based on outmoded views of science. Indeed, the reliability and validity problems of psychophysiological measures say be more awesome than those we face with verbal, paper and pencil, or more traditional observational measures In order to avoid the problems of the past it is suggested that those using physiological measures to build theory incorporate the following suggestions First, there should be a clear delineation of levels of abstraction with physiological response meaning embedded in a theory Second, it should be clear which parts of the conceptual analysis involve primarily deductive and/or inductive reasoning (cf. Marx 1963).

REFERENCES

Bagozzi, Richard P. (1980), Causal Models in Marketing, New York Wiley

Blackwell, Roger D , Hensel, James S., and Sternthal, Brian (1970), "Pupil Dilation: What Does it Measure?" Journal of Advertising Research, 19, No 4, pp. 15-18.

Calder, Bobby J., Phillips, Lynn W., and Tybout, Alice M. (1981), "Designing Research for Application," Journal of Consumer Research, 8 pp. 197-207.

Edelberg, Robert (1972), "Electrical Activity of the Skin Its Measurements and Uses in Psychophysiology," in Handbook of Psychophysiology, (eds.)," S. Greenfield and R. A Sternbach, New York, Holt, Rinehart, and Winston, pp. 267-418.

Ferber, Robert (1979), "How Not to Write a Prize-Winning Article," Journal of Consumer Research, 5, pp. 303-305.

Fishbein, Martin (1976), "Extending the Extended Model Some Comments," in Advances in Consumer Research, Vol. 3, (ed.), B. B. Anderson. Cincinnati The Association for Consumer Research, pp. 491-497.

Hansen, Flemming (1981), "Hemispherical Lateralization: Implications for Understanding Consumer Behavior," Journal of Consumer Research, 8, pp 23-36.

Hess, Eckhard H. (1972), "Pupillometrics A Method of Studying Mental Emotional, and Sensory Processes," in Handbook of Psychophysiology, (eds.), N. S Greenfield and R A Sternbach, new York: Holt, Rinehart, and Winston, pp. 491-531.

Jacobson, Edmund (1973), "Electrophysiology of Mental Activities and Introduction to the Psychological Process of Thinking," in The Psychophysiology of Thinking, (eds ), F. J. McGuigan and R A Shoonover, New York Academic Press, pp 3-31.

Jacoby, Jacob (1978), "Consumer Research: A State of the Art Review," Journal of Marketing, 42, pp. 87-96.

Jones, Lyle V (1971), "The Nature of Measurement," in Educational Measurement, Second Edition, (ed ) R L Thorndike, Washington: American Council on Education, pp. 335-355.

Kaplan, Abraham (1964), The Conduct of Inquiry, San Francisco: Chandler.

Kassarjian, H. H. (1971), "Personality and Consumer Behavior: A Review," Journal of Marketing Research, 8, pp. 409-419.

Kollat, David T , Engel, James R , and Blackwell, Roger D (1970), "Current Problems in Consumer Behavior Research," Journal of Marketing Research, 7, pp 327-332.

Marx Melvin H. (1963), "The General Nature of Theory Construction," in Theories of Contemporary Psychology, (ed.). M. H. Marx, London: The MacMillan Company, pp. 3-46.

Messick, Samuel (1981), "Constructs and Their Vicissitudes in Educational and Psychological Measurement," Psychological Bulletin, 89, pp 575-588.

Nighswaonger, Nancy J., and Martin, Claude R. Jr. (1981), "On Using Voice Analysis in Marketing Research," Journal of Marketing Research, 18, pp. 350-355.

Nunnally, Jum C (1967), "Psychometric Theory, New York McGraw-Hill.

O'Shaughnessy, John and Ryan, Michael J (1979), "Marketing, Science, and Technology," in Conceptual and Theoretical DeveloPments in Marketing, (eds.), O. C Ferrell, Stephen Brown, and Charles Lamb, Chicago American Marketing Association, pp. 577-585.

Peter, J. Paul (1981), "Construct Validity A Review of Basic Issues and Marketing Practices," Journal of Marketing Research, 18, pp. 133-145.

Ryan, Michael J , and O'Shaughnessy, John (1980), "Theory Development: The Need to Distinguish Levels of Abstraction," in Theoretical Developments in Marketing, (eds.), C. W. Lamb Jr., and P. Dunne, Chicago American Marketing Association, pp. 47-50.

Sheth, Jagdish N. (1967) 5 "A Review of Buyer Behavior," Management Science, 13, B718-B756.

Thatcher, Robert W., and John, E Roy (1977), Foundations of Cognitive Processes, Hillsdale, NJ Lawrence Erlbaum Associates.

Torgerson, Warren Sc (1958), Theory and Methods of Scaling, New York Wiley.

Watson, Paul J , and Gatchel, Robert J. (1979), "Autonomic Measures of Advertising," Journal of Advertising Research, 19, No. 3, pp. 15-26.

Wilkie, W L., and Pessemier, E A (1973), "Issues tn Marketing's Use of Multi-Attribute Attitude Models," Journal of Marketing Research 10, pp 428-441.

Zaltman, Gerald, Pinson, Christian, R. A , and Angelmar, Reinhard (1973), Metatheory and Consumer Research, New York Holt, Rinehart, and Winston

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