Special Session Summary Teasing Processes Apart in Consumer Research: Novel Experimental Methodologies

Citation:

Michel Tuan Pham and Patti Williams (1999) ,"Special Session Summary Teasing Processes Apart in Consumer Research: Novel Experimental Methodologies", in NA - Advances in Consumer Research Volume 26, eds. Eric J. Arnould and Linda M. Scott, Provo, UT : Association for Consumer Research, Pages: 372.

Advances in Consumer Research Volume 26, 1999      Page 372

SPECIAL SESSION SUMMARY

TEASING PROCESSES APART IN CONSUMER RESEARCH: NOVEL EXPERIMENTAL METHODOLOGIES

Michel Tuan Pham, Columbia University

Patti Williams, New York University

One of the principal challenges of consumer researchCand of research in cognitive and social psychologyCis to document the processes underlying the effects of interest.

There are a number of well-established methodological approaches for identifying underlying processes. These include: (1) process measures (e.g., thought listings, eye tracking, pattern of information-search); (2) manipulation and confounding checks; (3) complex experimental designs with intricate patterns of predictions; (4) mediation analyses; and (5) multiple operationalizations (triangulation) across studies. However, these approaches often become impractical when several processes are at work (e.g., the multiple processes underlying the compromise effect) and when they operate in an intimately related way (e.g., familiarity and liking). The purpose of this special session was to introduce a new class of methods that can be referred to as process decomposition.

The principles of process decomposition were originally developed in the cognitive psychology literature (Jacoby 1991) as a means of disentangling automatic versus controlled processes of memory. Pham and Johar (1997) have recently shown that similar principles can be many applied to a variety of other problems. The three papers presented in the session all capitalize on the principles of process decomposition, albeit in slightly different manners, addressing a variety of consumer behavior issues.

Separating automatic and conscious memory processes has been the focus of a large literature in psychology and is an emerging topic in consumer research. One approach, the task-dissociation methodology, compares performance on tests of explicit (e.g., recall or recognition) and implicit memory (e.g., stem-completion) to infer whether advertising exposure results in nonconscious effects. Krishnan and Shapiro compare the task-dissociation methodology with Jacoby’s (1991) process-dissociation methodology to examine advertising effects. Results show that the test of implicit memory provided a biased estimate of the effects of ad clutter and delay on automatic ad influences, whereas the process-dissociation methodology yielded more accurate estimates.

Fitzsimons and Williams examine the impact of automatic versus effortful processes upon the "mere measurement effect." Research on this effect has shown that when asked an intent question, consumers not only repot biased assessments of their likelihood to perform a behavior, but also become more likely to perform in accordance with the biased report of intention. In other words, simply by asking consumers to form and report a purchase intention, marketing researchers are changing consumers’ actual purchase behavior. Results replicate the mere-measurement effect and show that by virtue of being asked to form and report an intention to purchase a new candy bar, subjects become more likely to purchase in a particular pattern. More interestingly, the results of this research demonstrate that the mere-measurement effect is largely driven by automatic processes. In fact, the magnitude of the mere-measurement effect that can be attributed to an automatic mechanism is more than three times greater than the component that may be attributed to intentional processing.

Johar and Pham use a process decomposition model to estimate the relative magnitudes of four different processes consumers use to identify sponsors of eventsCdirect retrieval, relatedness between the event and the sponsor’s product category, market prominence of the sponsor, and pure guessing. The method relies on (1) stimuli-based a priori specification of conditional probabilities of providing a certain response (e.g., identifying a related and prominent brand as the sponsor), given that one has used a certain process (e.g., pure guessing), and (2) the observed response frequencies. Process decomposition estimates on data collected in an experiment manipulating relatedness and prominence indicate that reliance on relatedness and prominence exceeds reliance on direct retrieval and pure guessing and that use of relatedness dominates use of prominence. Sensitivity analyses varying the conditional probability specifications for each process enhances the estimated contribution of that process at the expense of pure guessing. A major limitation of the model concerns its assumption that any one process is singly used on any response and that the four processes are exhaustive of the family of processes used in sponsor identification. Relaxing this assumption requires specifying conditional probabilities for combination processes and increasing the number of response levels to equal the posited number of individual and combination processes.

At the end of the session Michel Pham offered a review of the major assumptions and limitations of the process decomposition approach, as well as suggestions on how to address these limitations. In summary, we believe that process decomposition approaches: (1) have broad applicability in consumer research, (2) can be invaluable complements of traditional approaches (e.g., ANOVA), (3) are worthy of further methodological development. Interested readers are invited to contact the researchers directly. Their electronic addresses are shown below:

Gavan J. Fitzsimons: gavan@mktgmail.wharton.upenn.edu

Shankar Krishnan: skrishna@indiana.edu

Gita V. Johar: gvj1@columbia.edu

Michel Tuan Pham: tdp4@columbia.edu

Stewart Shapiro: sshapiro@udel.edu

Patti Williams: pwilliam@stern.nyu.edu

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