Experiences With Script Elicitation Within Consumer Decision Making Contexts

ABSTRACT - Considerable interest in script theory has developed in the areas of artificial intelligence, cognitive science, cognitive, social, developmental and clinical psychology. In this paper we report on a series of exploratory studies designed to address issues related to script elicitation and script organization within a consumer decision context. Findings for both conceptual and methodological issues are discussed and implications for future research are noted

Citation:

Thomas W. Leigh and Arno J. Rethans (1983) ,"Experiences With Script Elicitation Within Consumer Decision Making Contexts", in NA - Advances in Consumer Research Volume 10, eds. Richard P. Bagozzi and Alice M. Tybout, Ann Abor, MI : Association for Consumer Research, Pages: 667-672.

Advances in Consumer Research Volume 10, 1983      Pages 667-672

EXPERIENCES WITH SCRIPT ELICITATION WITHIN CONSUMER DECISION MAKING CONTEXTS

Thomas W. Leigh, The Pennsylvania State University

Arno J. Rethans, The Pennsylvania State University

ABSTRACT -

Considerable interest in script theory has developed in the areas of artificial intelligence, cognitive science, cognitive, social, developmental and clinical psychology. In this paper we report on a series of exploratory studies designed to address issues related to script elicitation and script organization within a consumer decision context. Findings for both conceptual and methodological issues are discussed and implications for future research are noted

INTRODUCTION

Recently, considerable interest has developed in abstract knowledge structures or schemata (Bobrow and Norman 19;s, Ortony 1978, Rumelhart 1980, Schank and Abelson 1977, Taylor and Crocker 1980). These schemata are perceived to be the building blocks of cognition. They are seen as being employed in the process of interpreting sensory data, of retrieving information from memory, of organizing actions, of determining goals and subgoals, and generally in guiding the flow of information in the system (Rumelhart 1980).

One simple form of schema is the script, which embodies knowledge of stereotyped event sequences (Abelson 1980). More precisely, a script is a schematic "structure that describes appropriate sequences of events in a particular context" (Schank and Abelson 1977). The script concept is suggested to have broad application and has been applied in such areas as artificial intelligence (Abelson 1981), cognitive science (Schank and Abelson 1977), and cognitive, social, developmental, and clinical psychology (Rumelhart 1980, Abelson 1976, den Uyl and van Oostendorp 1980, Tomkins 1978).

Despite the popularity of the script concept in other areas, there has been little attention to script theory in consumer contexts. Although references to scripts have been made in the marketing literature (Bozinoff 1981, Calder 1978), no empirical use of scripts in the study of consumer decision making has been reported. Therefore, we set out to explore the use of script theory within the context of consumer decision making. What follows is a preliminary report on four exploratory studies designed to address issues related to script elicitation and script organization within one consumer decision context

THEORETICAL FRAMEWORK

In 1976, Abelson proposed the cognitive script as the basis for an alternative theory of social cognition to the then predominant views. He defined script to mean "a coherent sequence of events expected by an individual, involving him either as a participant or as an observer" (Abelson 1976, p. 33). More recent definitions have incLuded "[a script is] a structure that describes appropriate sequences of events in a particular context" (Schank and Abelson 1977) and "[a script is] a hypothesized cognitive structure that when activated organizes comprehension of event-based situations" (Abelson 1981).

The theory postulates that upon activation, scripts which have been stored in memory are used to direct behavior. This activation is believed to occur automatically as a function of the situational context. For example, upon entering a restaurant, a RESTAURANT script would be activated and begin to guide behavior. The script would contain a standard sequence of typical activities in a restaurant from the point of view of the customer. It may include such activities as talking to the hostess, being shown to a table, reading the menu, reading the wine list and so forth. The script would also include standard roles, objects, ordinary conditions for entering upon the activity, a standard sequence of scenes or actions wherein one action enables the next, and some normal results from performing the activity successfully. All this information contained in the script aids the restaurant goer in understanding, planning, and executing conventional activities.

Thus, scripts are perceived as playing a dual role. First, scripts are presumed to play an important role in comprehension and inference-making (Gibbs and Tenney 1980, McCartney and Nelson 1981, Nottenberg and Shoben 1980). Second, since they are formed on the basis of repeated episodes of a particular event sequence, scripts guide behavior. Empirical evidence supporting both the inference function and the structuring function is growing steadily (Bower, Black and Turner 1979, den Uyl and van Oostendorp 1980, Nottenburg and Shoben 1980).

OBJECTIVES OF THE STUDIES

Abelson (1981) noted that the first requirement-for scripts to play their dual role is that the individual must have a stable cognitive representation of a particular script. The first requirement then of a research program designed to examine script theory in consumer decision making is to establish the existence of script norms. That is, it must be shown that consumers agree on the characters, props, and actions used to describe decision making activities (Bower, Black and Turner 1979). The purpose of the four studies presented here was to investigate whether or not script norms could be developed in the context of an automobile purchase. This decision context has been one of the more frequently studied in marketing (cf. Agerwal and Ratchford 1980).

The objective of Study 1 was to establish the existence of script norms. Empirical experience with this study raised a number of methodological issues which were further explored in Studies 2, 3 and 4. The second study explored the possibility that the paper-and-pencil elicitation procedure employed in Study 1 might have resulted in more detailed reporting of activities. Study 2 therefore utilized an oral elicitation procedure.

By varying entering and role conditions through manipulation of instructions to the subjects, Studies 3 and 4 examined the issue of response sensitivity to task delineation. Although again addressing a methodological issue, these studies simultaneously explore the conceptual issue of script organization. This conceptual issue centers around the question of whether or not information surrounding script roles, props and action is stored at varying levels of abstraction.

In combination, then, these studies were designed to explore some fundamental methodological and conceptual issues in the application of script theory in a consumer behavior context.

STUDY 1: SCRIPT ELICITATION

Thirty consumers were asked to generate a list of events or actions involved in an automobile purchase decision from the perspective of the consumer. The specific instructions to the consumers for generating the scripts were adapted from Bower, Black and Turner (1979) and read as follows:

"Please write a list of actions describing what people generally do when they bus a new car. We are interested in the common actions of a 'buying a new car' stereotype. Start the list with 'decides to buy a new car' and end it with 'drive the car out of the showroom.' Include about 20 actions or events and put them in the order in which they occur."

The subjects, volunteers and members of civic groups in an Eastern town, were predominantly married females with children with a median age in the 30-39 category. The majority of subjects had completed some college coursework. They considered themselves to be familiar with the car purchase situation and found the instructions to be clear and the task to be easy, not too fatiguing, not too artificial, although a bit long.

The script presented in Exhibit 1, as well as each o, the scripts which follow, was edited and tabulated according to the procedures described in Bower, et al. (1979).

EXHIBIT 1

SCRIPT NORMS FOR AUTOMOBILE DECISION

Specifically, we designated the group script according to frequency of citation of specific actions or events. The criterion percentages selected, and their form of presentation in the script, were respectively: 75% as the minimum inclusion criteria (in lower case); 40% (in italics); 50% (in capital letters); and, 75% (in capital letters followed by an asterisk). The 75% level was added because of the relatively large number of such common actions noted. The actions are listed in the modal serial order in which they were reported. Since these procedures, though reflective of current practices in cognitive psychology, are somewhat arbitrary (Bower et al. 1979), the following statistics are provided: the number of mentions for each action or event included; the mean, median, mode, and range of actions mentioned; and the split-half reliability among the subjects

What is interesting about this first script is the degree of agreement in the "basic action" language that consumers use to describe a potentially complex decision process. We find actions described in both capital and italics letters indicating the relative popularity of these actions. This uniformity is also reflected in the low incidence of unique actions or events. Only two completely unique events (mentioned by a single consumer) were identified. The events were "bargain on options" and "look under hood." Furthermore, there is a high reliability in the frequency with which particular actions and events were mentioned in the script. A split half reliability coefficient of 0.71 resulted from correlating the frequency of mentioning specific actions by two randomly determined halves. Thus, we conclude that script norms do exist within a consumer context and that these norms can be reliablY elicited.

Interestingly, the script seems to provide some evidence for the recent "consumer decision making: fact or fiction" controversy (Olshavsky and Granbois 1979, 1980, Ursic 1980). The list of events and actions is in accordance with the five-step (problem recognition, search, alternative evaluation, choice and outcomes) conceptual model of consumer decision making. The script also reflects the previous literature on information seeking (Newman 1977).

In all, we found the results of this initial attempt at script elicitation within a consumer context sufficiently encouraging to pursue some critical issues about script theory that surfaced in this initial experience. Specifically, issues of script elicitation and script hierarchy were addressed.

STUDY 2: ALTERNATIVE ELICITATION

In the second study the possible effect of elicitation procedure on script composition was explored. Seventeen volunteers, members of a civic organization in an Eastern town, were directed to verbalize their scripts for the new car purchase situation. That is, upon receiving instructions and an exemplary written script identical to those used in the first study, subjects were asked to orally state a list of activities and/or events describing what people do when they purchase a new car. Their verbal statements were unobtrusively tape recorded for later analysis. Most subjects indicated that they found the task to be easy, just right in terms of length, and natural rather than artificial. As with the subjects in the first group, they considered themselves familiar with the car purchase decision process.

Exhibit 2 presents the group script, edited and tabulated according to the procedures previously described. Consistent with script theory, this script seems to overlap on at least the most salient aspects with that of the first group. The second group demonstrated less conscious "out of store" information seeking, but the remainder of the scriPt is remarkably similar.

EXHIBIT 2

SCRIPT NORMS FOR AUTOMOBILE DECISION

At the individual level, the scripts elicited orally tended to be shorter in length as indicated by the measures of central tendency and tended to have a lower percentage of less common activities. A subjective analysis of the scripts suggests that this variation may be due in part to scene selection and the equifinality of some of the actions as suggested by Abelson (1981). More importantly, however, we sense that verbal elicitation may not tap as efficiently activities which border on being "automatic" or "mindless" (Langer, Blank and Chanowitz 1978). It appears that paper-and-pencil elicitation induces a more thoughtful processing of the script which results in somewhat more detailed reporting of actions This finding is particularly interesting in light of the "verbal reports as data" controversy (Nisbett and Wilson 1977. Ericsson and Simon 1980).

In addition, verbal elicitation seems to result in a higher incidence of episodic script rather than categorical and/or hypothetical activities (Abelson 1976). Consumers seem to access scripts at these three levels at differential rates not only as a function of individual differences but also as a function of elicitation procedure.

Overall, the results of the second study suggest some convergent validity of the results obtained in the first study. On the other hand, the data also suggest the need for a more definitive analysis of the possible existence of a method factor.

STUDY 3: SUBACTIONS AND ROLE EXPECTATIONS

A key concept of the script model is that scripts are hierarchically organized with varying levels of abstraction. The entering conditions define the level of abstraction appropriate for the situation. For example, the entering conditions in Study 1 were broadly defined shopping for a car. The actions elicited therefore represented large "chunks" of activity: talking to friends or salesperson, test driving, or negotiating price. These actions max represent main conceptualizations, each of which has attached to it a hierarchy of related subordinate actions (Shank and Abelson 1977).

One purpose of this third study, therefore, was to explore the effect of different entering conditions on the degree of abstraction elicited. The situation defined was a subset of the script in Study 1, beginning with "getting out of your car on the dealer's lot" and ending with "getting back into your car on the dealer's lot."

A second issue investigated in this study was the ability of the participants to report the expected role actions of the interacting partner. Role expectations are considered important in buyer-seller interaction (cf. Sheth 1975, Spiro et al. 1977). Script theory, moreover, suggests that the act of reporting one's own actions involves an implicit review of intersecting roles (Shank and Abelson 1977). This study, therefore, asked that subjects report the actions of the salesperson rather than their own.

Third, university seniors with diverse majors wore recruited through a newspaper advertisement. Each was paid for participating. Of these thirty participants, fourteen reported direct experience with shopping or purchasing a new car. The mean rating on a 7-point familiarity scale for shopping for a new car was 4, with 16 participants reporting a 5 or better. The probability of buying a new car in the next few years was almost universally high. The instructions provided were identical to those of Study 1 except that the actions requested are those of the salesperson. The subjects reported the instructions to be clear and the task easy, not fatiguing, or artificial.

EXHIBIT 3

SCRIPT NORMS FOR AUTOMOBILE SALESPERSON

The resulting script, presented in Exhibit 3, demonstrates a high degree of agreement in the basic actions. Seven actions were mentioned by more than fifty percent of the participants; 16 actions were mentioned by at least 40%. This degree of agreement is encouraging given the greater specificity of the actions solicited in this study. The ratio of unique actions to total mentions for the nonpurchase script, 12/488, is also better than was anticipated. Examples of unique actions included asks my name, compliments my driving, fixes tie, uses breath deodorant, kicks tires, and works to keep up my enthusiasm. The results are consistent with the idea that scripts are hierarchically organized and that the entering conditions define the level of abstraction -appropriate in a situation. The subjects had little apparent difficulty filling in the dealer visit portion of the Study 1 script with more concrete or basic action units. Further decomposition of such actions "introduces self" or "takes you to his office" seems unnecessary. These basic units are especially prominent in the opening/closing portions of the script. The data also suggest that subJects had little difficulty reporting on the expected role enactment of the salesperson. The lists of actions generated and the modal ordering seem highly plausible and social interaction norms, as well as typical selling procedures, seem appropriately represented.

A final encouraging outcome is that the basic actions occurring in the original new car shopping script are considerably enhanced while the basic ordering is maintained. For example, the dealer visit portion of the original script included: talk to salespersons, check on color, check on mileage, get price estimates, and test drive. The more detailed script in this study adds the approach, greatly decomposes the "talk to salesperson" action, and maintains the order of the test drive, price negotiation, and discussion of trade-in and financing.

Several problems, however, were noted with the data collection process. Redundancy and looping occurred in several of the individual scripts. There was a tendency for individual actions to be repeated at different points in the script (as when the consumer sequentially inspects several cars). In addition, some scripts included a summary repeat statement almost like a "do loop" in computer programming. A second problem was the tendency for some participants to interject their own actions into the script. This fact principally accounts for the reduced number of mentions elicited in this study These problems plus the desire to further explore the interactive dimensions relevant to the automobile shopping process led to the design of a new script elicitation method and Study 4.

STUDY 4 ROLE DECOMPOSITION USING AN INTERACTIVE METHODOLOGY

The automobile shopping situation used in the previous study was modified to allow the respondent to list actions or behaviors for both the consumer and the salesperson. The specific instructions read as follows:

"You have recently decided to shop for a new car. We are interested in the common actions or behaviors of sales representatives and their customers as they discuss the new car purchase. Please write a list of your own personal actions or behaviors and those you would expect from a new car salesperson. Start the list with 'getting out of your car on the lot' and end it with 'getting back into the car.' For each action, circle 'C' if it is your own (consumer) action or 'S' if it is a salesperson action. Include about 30 actions or behaviors in the order in which they occur. Please note that several actions for either the salesperson or yourself may occur in a row."

The expectation was that these instructions would generate a summary script which would better reflect the temporal and interactive nature of the dealer visitation script, would allow highly specific concrete actions for both participants, and would provide some insight into the origin of initiation of the various actions. Since the data were collected from the consumer perspective, the expectation was that the consumer-initiated actions would predominate.

Twenty-four university seniors with diverse majors were recruited through a newspaper advertisement. They were screened for prior participation or knowledge of the study. Each was paid for participating. Thirteen participants reported direct prior experience with shopping for or purchasing a new car. The mean rating on the familiarity scale for shopping for a new car was 4.1, with 11 respondents reporting a 5 or better. The reported probability of buying a new car in the next few years was universally high. The instructions were reported to be clear, moderately easy, not fatiguing, and natural, though perhaps a little too long.

Exhibit S presents the script, edited and tabulated as described in Study 1. Consumer and salesperson initiated actions were tabulated separately in order to reflect the origin of initiation for the action. The range of actions listed in this study was tightly concentrated around the maximum number of spaces provided (only one subject provided less than 28 responses). The close correspondence between the number of actions provided and requested may reflect many factors. The interactive nature of the script allowed two actions, one for each participant, to describe an event. For example, a considerable number of action combinations such as "asks about features" and "responds/discusses features" were noted. Additionally, the interactive nature of the methodology allowed for specific inclusion and counting of repeated activities and more detailed decomposition of the negotiation/dickering process. The number of actions requested may have been insufficient in light of these considerations. Finally, since the problem of incorrect role interjections was eliminated, there were fewer losses of data due to such errors.

EXHIBIT 4

NORMS FOR THE ROLE DECOMPOSITION SCRIPT

The origination of actions varied considerably among respondents. Consumer actions ranged from 10 to 78% of the total actions listed. Respondents occasionally listed an activity, such as "walked to the car" as Jointly initiated. The mean number of consumer actions was 48.6%. Exactly one-half of the respondents listed more than 50% consumer actions. Yet the pattern of responses exhibited relatively long strings of either consumer or salesperson role actions. The data, therefore, do not seem to reflect either a consistent perspective effect or a highly oversimplified interaction pattern.

The degree of agreement in the basic actions remained fairly high. A total of 38 actions were mentioned by at least 25% of the respondents. Sixteen actions were mentioned by at least 40%. However, only seven actions were mentioned by at least 50%, and only "showing various models" achieved the 75% level. The reduced commonality of various actions is at least partly due to the fact that consumer and salesperson actions were separately listed. Several actions, for example "inquires about type of car," "shows various models," "inquires about price," "suggests test drive," and "negotiates/dickers" would have had higher scores if roles were ignored. A second factor tenuating commonness was the fact that location of a statement in the script is more explicitly respected in the interactive approach. For example, "asks about features" occurs at three separate locations in the normative script. Features were brought up at ten distinct locations in the scriPt.

The ratio of unique actions to total mentions was 39/694. Examples of unique actions included: be seated, hands keys to the customer, adjusts seat, checks blue book value, asks about customizing, acts disinterested, hugs the wife, appears to be thinking, asks father's opinion, calls bank to clear credit, daydream about yourself in the car, and says "you look good in there." In addition to the increased number of unique acts, a considerable number of actions were mentioned by only 2 or 3 respondents. Finally, the total number of actions mentioned at least once totalled 161. The fact that the technique allows for differential roles and locations at least partially accounts for this large number of actions. The method may also encourage more detailed visualization and therefore a larger number of idiosyncratic actions. Many of the less common actions are due to the fact that individuals differed in the outcome of the interaction: some decided to buy, some to think it over, and some to negotiate (with a variety of strategies).

In spite of the preceding problems, the script is quite consistent with the scripts developed in the previous studies. In comparison with Study 3, several observations are interesting. Several consumer actions were added prior to the initiation of the interaction with the salesperson's approach. The salesperson's approach is essentially identical. And the portion of the script oriented toward determining consumer interests and showing various models is essentially the same The key difference is that the roles are more specifically decomposed. In addition, the script more specifically delineates the process of narrowing consumer interests to a particular car model, the test drive, and the financing discussions

In sums the interactive procedure allowed more specific role decomposition than the earlier methods. The script structure was essentially maintained in terms of both the actions listed and their temporal ordering. The problems of redundancy, looping, and role interjection were apparently handled. Lastly, the procedure offered the opportunity to investigate the consumer's view of the relative degree of control of the interaction exerted by the consumer and salesperson.

SUMMARY AND CONCLUSIONS

Cognitive scripts have been proposed as a cognitive structure composed of a coherent sequence of events expected by an individual, including his/her own actions, actions of others, standard objects, ordinary entering conditions, and expected outcomes. Scripts are hypothesized to play an important role in comprehension, inference-making, and hence behavior in social situations. The script concept would seemingly be useful in describing and explaining consumer decision-making processes.

The utility of script concepts in consumer contexts requires investigations in three areas: structure and content, i.e., what belongs or is contained in such scripts and how it is organized; procedure, i.e., how do such scripts operate; methodology, i.e., how is such information best elicited (den Uyl and van Oostendorp 1980, Bower et al. 1979). The studies described in this paper were designed to investigate several structural and methodological issues within the context of an automobile purchase decision.

With respect to the structure of scripts in this consumer context, several conclusions seem noteworthy. The studies suggest that there may be considerable agreement in the basic action language used to describe the automobile shopping/purchase experience. Each of the four studies found considerable agreement on basic actions, behaviors, and events. In addition, the structure of the scripts elicited was quite consistent, both in the actions elicited and their temporal ordering. Furthermore, the studies are consistent with the advocated position that scripts are hierarchically organized with varying levels of abstraction. These scripts were seemingly decomposable into subactions, role expectations, and interaction sequences. The studies are thus encouraging from a structural perspective.

With respect to methodological issues, the studies suggest that instructions and data collection methods are important determinants of the abstractness, length, and focus of the scripts provided. Further investigation of the effect of evoking context, entering conditions, role perspective, length of script requested, and data collection method would seem to be worthwhile. These studies indicate certainly that people can and will provide script-related information under a variety of methodological guises. Several other methodological issues were noted in these studies. First, the validity of establishing norms by strict reference to a count of common actions recalled needs assessment. Given the availability of a normative script, recognition, ordering, and fill in the missing actions tasks might be used to validate the results (Bower et al. 1979, Kemper 1982). Properly designed, these approaches would enhance confidence in serial ordering and consistency (or reliability). Second, the sensitivity of the norming procedure to the base rate of actions collected needs investigation. Third, statistical alternatives to modal serial ordering should be investigated (John and Whitney 1982). Lastly, the observation that the actions provided might be readily structured into "scenes" seems worthy of further investigation. The existence of scenes or other simplifying structural factors, may simplify and improve the data collection process. These and a variety of other traditional reliability and validity issues need investigation if such concepts are to be used to investigate how scripts operate in consumer contexts.

Investigation of the procedural issues of scripts, of course, promises to be the most interesting aspect. A variety of script-theoretic concepts would seem to be potentially useful in understanding both individual and group consumer processes. These concepts include such extra-script concepts as life themes, life perspectives, plans, and goals as well as intra-script dimensions such as props, roles, scenes, obstacles, errors, and distractions (Schank and Abelson 1980. Bower et al. 1979).

The script-theoretic approach offers an opportunity to explicate the "expectations" component of consumer behavior models --a component that has been tossed around for years with only limited empirical study. Investigations into this component can focus on the effect of existing scripts on individual and group behaviors, the effect of deviations from normative expectations on such processes, and the acquisition and modification processes involved in updating scripts to fit special situations or novel contexts.

REFERENCES

Abelson, Robert P. (1981), "Psychological Status of the Script Concept," American PsYchologist, 36, 7, 715-29.

Abelson, Robert P. (1976), "Script Processing in Attitude Formation and Decision Making," in J. Carroll and J. Payne (eds.), Cognition and Social Behavior, Hillsdale, NJ: Erlbaum.

Agerwal, Manoj K. and Brian T. Ratchford (1980), "Estimating Demand Functions for Product Characteristics: The Case of Automobiles," Journal of Consumer Research, 7, 3, 949-62.

Bobrow, Daniel C. and Donald A. Norman (1975), "Some Principles of Memory Schemata," in D. G. Bobrow and A. M. Collings, eds., Representation and Understanding: Studies in Cognitive Science, New York: Academic Press.

Bower, Gordon H., John B. Black and Terrence J. Turner (1975), "Scripts in Memory for Text," Cognitive Psychology, 11, 177-220.

Bozinoff, Lorne (1982), "A Script Theoretic Approach to Information Processing," in A. .Mitchell (ed.), Advances in Consumer Research, Vol. 9, Ann Arbor, MI: Association for Consumer Research (forthcoming).

Calder, Bobby J. (1978), "Cognitive Response, Imagery and Scripts: What is the Cognitive Basis of Attitude?", in H. Keith Hunt (ed.), Advances in Consumer Research, Vol. 5, Ann Arbor, MI: Association for Consumer Research, 630-34

den Uyl, Martijn and Herre van Oostendorp (1980), "The Use of Scripts in Text Comprehension," Poetics, 9, 275-94.

Ericsson, K. Anders and Herbert A. Simon (1980), "Verbal Reports as Data," Psychological Review, 87, 3, 215-251.

Gibbs, Raymond W. and Yvette J. Tenney (1980), "The Concept of Scripts in Understanding Stories," Journal of Psycholinguistic Research, 9, 3, 275-84.

John, George and John C. Whitney (1982), "An Empirical Investigation of the Serial Structure of Scripts." in An Assessment of Marketing Thought and Practice, Educators Proceedings, Bruce J. Walker et al., eds., Chicago: American Marketing Association. 75-79.

Kemper, Susan (1982), "Filling in the Missing Links," Journal of Verbal Learning and Verbal Behavior, 21, 99-107.

Langer, Ellen J., Arthur Blank and B. Chanowitz (1978), "The Mindlessness of Ostensibly Thoughtful Action," Journal of Personality and Social Psychology, 36, 6, 635-42.

McCartney, Kathleen and Katherine Nelson (1981), "Children's Use of Scripts in Story Recall," Discourse Processes, 4, 59-70.

Newman, Joseph W. (1977), "Consumer External Search: Amount and Determinants," in Arch G. Woodside et al., (eds.), Consumer and Industrial Buying Behavior, New York: North Holland Publishing Company.

Nisbett, Richard and Timothy Wilson (1977), "Telling More Than We Know: Verbal Reports on Mental Processes," Psychological Review, 84, 3, 231-259.

Nottenburg, Gail and Edward J Shoben (1980), "Scripts as Linear Order," Journal of Experimental Social Psychology, 16, 329-47.

Olshavsky, Richard W. and Donald H. Granbois (1980), "Consumer Decision Making-Fact or Fiction: Rejoinder," Journal of Consumer Research, 7, 3, 333-34.

Olshavsky, Richard W. and (1979), "Consumer Decision Making-Fact or Fiction?", Journal of Consumer Research, 6, 2, 93-100.

Ortany, Andrew (1978), "Remembering, Understanding, and Representation," Cognitive Science, 2, 53-69.

Rumelhart, David E. (1980), "Schemata: The Building Blocks of Cognition," in R. J. Spiro et al., (eds.), Theoretical Issues in Reading Comprehension: Perspectives from Cognitive Psychology Linguistics, Artificial Intelligence and Education, Hillsdale, NJ: Lawrence Erlbaum Associates.

Schank, Roger C. and Robert P Abelson (1977), Scripts, Plans, Goals and Understanding, Hillsdale, NJ: Lawrence Erlbaum Associates, Publishers.

Sheth, J. N. (1975), "Buyer-Seller Interaction: A Framework," Proceedings of the Association for Consumer Research (Winter). 382-86.

Spiro, R. L., W D. Perreault, Jr. and F. D. Reynolds (1976), "The Personal Selling Process: A Model and Critical review," Industrial Marketing Management, 5, 6, 351-63.

Taylor, Shelly E. and Jennifer Crocker (1980), "Schematic Basis of Social Information Processing," in E. T. Higgins et al., (eds.), Social Cognition, The Ontario Symposium, Hillsdale, NJ: Erlbaum Associates. Publishers.

Tomkins, Silvan S. (1978), "Script Theory: Differential Magnification of Effects," in Nebraska Symposium on Motivation 1978, Lincoln: University of Nebraska Press, 1978, 201-236.

Ursic, Michael (1980), "Consumer Decision Making-Fact or Fiction: A Comment," Journal or Consumer Research, 7, 3, 331-33.

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