Verbal Versus Visual Processing Modes: an Empirical Test of the Cyclical Processing Hypothesis

ABSTRACT - The relationship between verbal and nonverbal modes of processing information during retrieval from memory is investigated. The cyclical processing model suggests that a schema (a verbally oriented representation of a stimulus) should promote the use of mental imagery (a nonverbal means of retrieving information from memory). This proposition is generally confirmed in a study that examined the relationship between the existence of scripts as a type of schema and the usage and vividness of visual imagery as a means of retrieval.

Citation:

Ruth Ann Smith, Michael J. Houston, and Terry L. Childers (1984) ,"Verbal Versus Visual Processing Modes: an Empirical Test of the Cyclical Processing Hypothesis", in NA - Advances in Consumer Research Volume 11, eds. Thomas C. Kinnear, Provo, UT : Association for Consumer Research, Pages: 75-80.

Advances in Consumer Research Volume 11, 1984      Pages 75-80

VERBAL VERSUS VISUAL PROCESSING MODES: AN EMPIRICAL TEST OF THE CYCLICAL PROCESSING HYPOTHESIS

Ruth Ann Smith, University of Maryland

Michael J. Houston, University of Wisconsin-Madison

Terry L. Childers, University of Minnesota

ABSTRACT -

The relationship between verbal and nonverbal modes of processing information during retrieval from memory is investigated. The cyclical processing model suggests that a schema (a verbally oriented representation of a stimulus) should promote the use of mental imagery (a nonverbal means of retrieving information from memory). This proposition is generally confirmed in a study that examined the relationship between the existence of scripts as a type of schema and the usage and vividness of visual imagery as a means of retrieval.

INTRODUCTION

Given the presence of verbal and nonverbal information in the consumer information environment, consumer research that examines verbal and nonverbal information processing strategies is both warranted and necessary. The bulk of consumer research on information processing has, however, assumed or imposed a verbal processing paradigm only. Little attention has been given to nonverbal information processing. Even less attention has been given to the Joint use of verbal and nonverbal processes. The research reported here examines the relationship between verbal and nonverbal processing modes.

Two specific modes of processing are of interest. In the nonverbal realm imagery is a type of processing that seems appropriate for attention in consumer behavior. Moreover, visual imagery is a type of imaginal processing that seems particularly appropriate given the current research attention being paid to it in the consumer behavior literature (e.g., Lutz and Lutz 1977, 1978; Childers and Houston 1983; Rossiter and Percy 1982). In the verbal realm Lord (1980) has maintained that schemas are a verbally oriented approach to processing information, and further maintains that schemata and images are two different modes of processing social information. Scripts represent one specific type of schema receiving current attention in consumer research (e.g., John and Whitney 1982; Smith and Houston 1982). The interest here lies in the degree of independence between visual imagery and scripts and whether they serve as parallel rather than different processing modes. The conceptual features of imagery theory and schema theory are presented. Then a cyclical processing model which posits an interrelationship between imagery and schema is discussed. Finally, the procedures and results of an experiment designed to test the central features of this model are presented.

THEORETICAL FRAMEWORK

Yuille and Catchpole (1977) offer a useful framework for distinguishing between imagery and schema theories. They distinguish between the operative level of thought which is the storehouse of an individual's basic knowledge and the figurative plane which is the servant of operative thought and permits internal representations that support operative activities such as remembering and problem solving. A schema serves its function at the operative level while imagery is relevant to the figurative plane.

Mental Imagery

The mental image is the reconstruction in one's mind of a percept or event that has been previously experienced and stored in memory (Richardson 1969; Childers and Houston 19825. Such images can be visual, tactile, auditory, gustatory, or olfactory. It is the visual image that is of interest here.

Imagery, as a process, is a means of encoding incoming information (in which it has consistently been shown to improve retention of the information) and a means of retrieving the information for operative purposes. At the retrieval stage the visual mental image is a way to process information perhaps in parallel rather than sequential fashion. It thus serves as a means of representing internally our memories and ideas and as such serves as a tool of operative thought.

Schemata

The concept of a memory schema has, for some time, been found quite useful in explaining how knowledge is represented in memory. As noted by Rumelhart and Ortony (1977), "While originating from the senses, knowledge is not a blind record of sensory inputs. Normal people.. seem to process and reprocess information, imposing on it and producing from it knowledge which has structure. Some of this knowledge seems to be in the form of specific memories of particular events which we have experienced; some of it seems ta be in the form of more general abstractions no longer tied to any particular time, place, or source. (P. 99)

How these general abstractions are represented in memory, and how the representations influence subsequent information processing, are central concerns to schema theory. As defined by Taylor and Crocker (1981), a schema is wa cognitive structure that consists in part of the representation of some stimulus domain. The schema contains general knowledge about that domain including a specification of the relationships among its attributes as well as specific examples or instances of the stimulus domain. The schema provides hypotheses about incoming stimuli, which include plans for interpreting and gathering schema-related information. It may also provide a basis for activating actual behavior sequences or expectations of specific behavior sequences" (p. 91). Thus, a schema is just a mental representation of an individual's generic knowledge about some domain, abstracted through repeated experiences with various stimulus configurations within that domain.

Schemata are of several types and may be classified according to the stimulus domain which acts as the organizing theme. One such category, person schemata, is organized around the characteristics of an individual (including oneself), or prototypical individuals such as extraverts or introverts (Cantor and Mischel 1977). Role schemata are a second category, and have as an organizing theme a particular occupation, social role or social group. Thus, an individual may possess role schemata which define the typical characteristics and behaviors of people occupying various roles such as professor, father, or athletic club member. A third category, object schemata, is organized around the typical attributes and functions associated with groups of related objects, such as stereo equipment or automobiles. In addition to the uses and physical traits, these schemata may also specify the typical patterns of motor responses utilized with the objects.

The final category of schemata, event schemata, is often referred to as scripts, and will form the focus of the investigation to be reported. Scripts are organized around "a predetermined, stereotyped sequence of actions that define a well-known situation" (Schank and Abelson 1977, p. 41). Abbott and Black (1980) define scripts more precisely as follows: "A script for a commonplace event consists of the standard actions, characters and objects involved in the event. Scripts are intended to represent knowledge about events which are so well practiced that their performance is stereotyped' (p. 5-).

Scripts are distinguished from other types of schemata in several ways. Most importantly, the actions which define a scripted event are associated by a temporal sequence which is causal in nature. That is, prior actions in the event must first occur to enable subsequent actions to take place. Additionally, the actions which compose this sequence differ in their importance, or centrality, to the event. For instance, when dining in a restaurant, a farsighted person would likely don reading glasses prior to ordering a meal. However, the action of ordering is clearly more central to the main event than is putting on one's glasses. The term "base actions" will be employed in the remainder of this paper to refer to those actions which are most central to an event.

Similar to all categories of schemata, scripts perform a variety of functions which facilitate the processing of information. One of these is that a script acts to facilitate the retrieval of stored information. Findings by Anderson and Pichert (1978) and Hasher and Griffen (1978) suggest that scripts are so effective in this capacity that they may be used to retrieve information from memory even if some other schema was activated when the information was encoded. Furthermore, the finding of Alba et al. (1981) that the recall advantage of subjects with an activated schema over those without disappears in a recognition task implies strongly that a schema affects retrieval, if not encoding, of stored information. An unanswered question, however, pertains to whether the facilitating effect of a schema, or script, on retrieval works in concert with, or independently of, alternative retrieval strategies such as mental visual imagery.

The Cyclical Processing Model

The potential relationship between imagery and schema theory has recently been formalized by Hampson and Morris (1979) in their cyclical processing model. Central to this model is the cycle as the basic unit of imagery production and schemata as the cognitive structures needed to engage this activity. Schemata serve two purposes in the model: (1) the schema serves as a format in that it accepts only information in an appropriate form, and (2) schemata initiate the exploration of information sources and thus serve as plans (Hampson and Morris, p. 12). Kant (1787), as quoted in Rumelhart and Ortony (1977, p. 101), provides a similar view of this interrelationship:

The image is a product of the empirical faculty of reproductive imagination; the schema of sensible concepts, such as figures in space, is a product and, as it were, a monogram of pure a priori imagination, through which, and in accordance with which, images can be connected with the concept only by means of the schema to which they belong.

Within the Hampson and Morris framework, an imagery cycle consists of three components: the schema, a data base in abstract propositional form, and the surface or mental image. As shown in Figure 1, the three components interact in processing internally as well as externally generated stimuli. Image production thus involves the selection and activation of an underlying schema from which a surface image is constructed (see Palmer 1975 for a description of the process) and the use of a set of expectations derived from the schema to anticipate the product of the construction process. The product is a dynamic image which is modified and manipulated through the cyclical process of continual information pickup and expectation refinement guided by the underlying schema.

FIGURE 1

THE IMAGERY CYCLE ILLUSTRATING (1), THE IMAGE SCHEMA WHICH CONTROLS (2), THE ABSTRACT - DATA BASE, WHICH PRODUCES (3), THE SURFACE IMAGE, WHICH IS PICKED UP BY (1), THE IMAGE SCHEMA.  FROM HAMPSON AND MORRIS (1979, P. 15)

The cyclical processing model postulates a close relationship between the organization of information in memory and the manner in which it is internally represented as an image. Recent research by Rethans and Hastak (1981) provides some corroborating evidence on this relationship. They found that 25-30% of the thoughts generated regarding product hazard knowledge structures were imaginally based. They, in turn, suggest that research on information representation will almost always be incomplete if it focuses on linguistic information only. It is this relationship between knowledge structures (scripts) and imaginal processing that is the focus of the current study. Within the cyclical processing framework, imagery should play a significant role in the retrieval of consumer knowledge when consumers have established an existing knowledge structure. For a scripted activity, perhaps related to the sequence of events relating to the purchase of a product, certain events would most likely be considered base actions (e.g., go to the store, pay for the merchandise, etc.) while, for more experienced consumers, scripts would be expected to be enriched by idiosyncratic actions (e.g., examined brand B and compared it to brand D). As a retrieval strategy, imagery would probably play a greater role in the representation of the more detailed idiosyncratic actions as compared to commonly held base actions.

HYPOTHESES

The key notions of the cyclical processing model suggest that we must distinguish between individuals who possess a well-formed knowledge structure (i.e., schema) in a particular domain of behavior from those who do not, and then determine if visual imagery occurs to a greater degree by those possessing a schema. From here on, those possessing a knowledge structure for the event of interest are referred to as "schematics" and those not possessing a schema as "aschematics." The following hypotheses are offered for testing:

H1: Schematics will use visual imagery as a retrieval process in recalling an event to a greater extent than aschematics

H2: Schematics using visual imagery will generate more vivid images than aschematics who use visual imagery.

The first hypothesis is the basic proposition of the cyclical processing model: the existence of a schema for an event will promote the use of imagery as a means of recalling that event. The second hypothesis recognizes that imagery can be used in the absence of a schema but that the existence of a schema will generate more vivid images of a previously experienced event. Confirmation of this hypothesis would provide further evidence of the imagery-eliciting role of schemata.

METHODOLOGY

Context

The hypotheses were tested in a laboratory experiment conducted over a month-long interval. The event which formed the context for the investigation was the process of interviewing through the Placement Office operated by the School of Business of a large midwestern university. This event was selected for two reasons. First, it is highly routinized and repetitive, with the procedures followed to arrange, attend, and follow up on an interview being quite invariant for every individual using the service and for every interview. Under such conditions, script-formation would be facilitated.

Secondly, within the population of Business School students, there are individuals with virtually no experience with the service, as well as a group with substantial experience. The former group consists primarily of students who are not yet at the- point in their careers where a sustained job search is appropriate. The latter group, however, is composed mainly of students in the final stages of their academic programs, and who are actively engaged in searching for professional positions. Consequently, the Placement Office event permitted comparisons of the imagery of schematics and aschematics whose between-group differences were minimized.

Schematics vs. Aschematics

Preliminary to the main experimental procedures, it was first necessary to categorize the subjects as schematic or aschematic with respect to the event of interest. This was accomplished on the basis of scores on an experience survey. This multiple item measure was designed to evaluate knowledge about the event of interest which had been obtained either through direct personal use of the service as well as from a variety of vicarious sources such as conversations with peers. [A copy of this instrument may be obtained upon request from the authors.]

Responses to this measure were obtained from 106 students, all of whom had been formally admitted to the School of Business. Half of these subjects were registered with the Placement Office and had completed at least two interviews. The remainder were not registered with the Placement Office, and therefore had no interviewing experience through that service. The internal consistency of the measure was evaluated using the responses of this group to compute coefficient alpha, which was determined to be .97, indicating a reliable instrument.

The scores on the survey were computed, revealed a distribution with a mean of 84.9, a standard deviation of 37.5, and a range of 3 to 133 on a measure with a maximum possible score of 148. The median score of 94.5 was used to categorize subjects, and those whose scores exceeded this level were considered to be schematic with respect to the interviewing event, while the remainder were categorized as aschematic. From each group of 53 subjects, a sample of twenty was randomly selected to participate in the experiment. Six subjects (one aschematic and five schematics) were subsequently eliminated from the sample either for noncompletion of some part of the experiment or due to the presence of a high probability of change in the mental representation of the event during the course of the study. The final sample of 34 subjects was composed of 19 aschematics and 15 schematics.

Measures

The experimental stimulus consisted of a probe to obtain a written protocol describing the event of concern from each subject. The probe was modeled after that used by Bower, Black and Turner (1979), and was worded as follows:

We are interested in knowing how the Business Placement Office interviewing process operates. Please write a list of actions that describes what usually happens when a student who is properly registered with Business Placement arranges an interview with a campus recruiter. Include all the actions that would either be required to get an interview, or that would be done voluntarily to prepare for or to follow up on the interview. Begin your list with the first action involved in signing up for the interview and conclude it with the last follow up action after the interview is over. Include about 15 actions in your list and put them in the order in which they would normally occur.

Two dependent measures were employed to evaluate the relationship between the presence of a script for the event and the amount and vividness of visual imagery. The first of these was an item analysis of the protocol which was used to compare the frequency and vividness of imagery related to the scripted event across the two group i. This measure required that each subject rate the individual actions in their protocol on two separate seven-point scales. On the first scale, subjects rated each action in terms of whether they had visualized or verbalized it. Those actions which were visualized were then rated on the second scale, which evaluated whether the image was vivid (clear), or pale (indistinct).

I In order to be able to attribute any between-group differences on the protocol item analysis to the presence or absence of a script, rather than to group differences in imaging abilities, a second dependent measure was employed. The Vividness of Visual Imagery Questionnaire (Marks 1973) was used to evaluate the imagery abilities of the two groups. This 16-item scale measures the "clarity" of voluntarily formed images and has been found to possess desirable reliability and validity qualities by White, Sheehan and Ashton (1977) and Childers (1982).

Procedures

The experimental procedure involved meeting with subjects on two different occasions, separated by a one-month interval. Subjects were run in two groups, composed of either schematics or aschematics as categorized using the previously described experience questionnaire. All experimental tasks were performed in a classroom setting under the supervision of one of the investigators.

At the first meeting, subjects provided written protocols in response to the stimulus probe. These responses were used to identify the base actions which are central to the event of concern, a procedure necessary to compare the imagery on idiosyncratic versus base actions in the protocols. The base actions were identified in the following manner: All the nonrepetitive and nonmutually exclusive actions included in any of the twenty schematic subjects' protocols were combined into a prose narrative. This narrative was then presented to a convenience sample of three schematics, none of whom were experimental subjects. These judges were instructed to select only those actions in the narrative which they would usually expect to happen, or that they usually performed, during a typical occurrence of the event. Those actions which were selected by at least two of the three judges were retained as the base actions central to the event. Any other action which appeared in a description of the event was operationalized as an idiosyncratic action. A total of twenty base actions were identified through this procedure, which is adaptation of that used by Bower, Black and Turner (1979). The sequential order of these actions was then established using a paired comparison technique developed by John and Whitney (1982).

This set of responses was also used as a check on the ability of the protocol measure to discriminate between known groups of schematics and aschematics. This was accomplished by assigning a score to each response which represented the level of agreement between the contents of the protocol and the sequenced base actions of the script. Scores were determined as follows: a group of three judges, all of whom were naive both with respect to the event of interest and the experimental hypotheses, coded the protocols to identify the base actions present. A base action was considered to be included in a response if at least two of the three judges coded it. One point was assigned for every base action present in a protocol, and an additional point was given for every pair of base actions in the script-defined order. The score was the total of all assigned points. As expected, the mean score of schematic subjects (15.5) was significantly greater than that of aschematics (9.9) with t=3.42 and p<.001.

At the second meeting, subjects first completed a two-item measure indicating the amount and type of involvement with the Placement Office event during the interval since the first meeting. The purpose of this measure was to identify any subject for whom a high probability of change in the mental representation of the event existed. One schematic subject reported having had "unusual" interviewing experiences, and was eliminated from the sample. The remaining subjects reported either no- experiences, or only "ordinary" experiences.

Each subject then generated a second written protocol describing the event of interest in response to the same stimulus probe used at the first meeting. It was assumed that the month-long interval between the meetings would minimize any memory effects in the second set of responses. Scores on these responses were computed as before, and the measure again discriminated between the two groups, with the mean score of schematics (15.4) being significantly higher than that of aschematics (8.7), with t=5.11, p<.001.

Subjects then completed the protocol item analysis measure, rating every action in their protocol with respect to whether they visualized or verbalized it, and the vividness or paleness of visualized actions. For each subject, a mean rating was computed across all protocol actions on each of the two scales, as well as for just base actions and just idiosyncratic actions. Mean ratings were preferred in order to take into account the different number of actions in the protocols of various subjects. Finally, subjects completed the 16-item Vividness of Visual Imagery Questionnaire, concluding the experimental tasks.

RESULTS

Preliminary to testing the two hypotheses advanced in this investigation, the responses of schematic and aschematic subjects on the Vividness of Visual Imagery Questionnaire were compared. The mean score of schematic (;7.2) was not significantly different from that of aschematics (39.6) with t=-.61 and p>.05. Consequently, any differences in the amount and vividness of visual imagery which are found may more confidently be attributed to the presence or absence of a script, rather than to any inherent differences in the imagery ability of the two groups.

The first hypothesis predicts that schematics will use mental imagery as a retrieval process to a greater extent than aschematics. This prediction, which is consistent with the cyclical processing model of Hampson and Morris (1979), was tested by comparing the group means on Scale A of the Protocol Item Analysis measure. On this scale, subjects had rated each action in their recall protocol with respect to whether they had visualized or verbalized the actions. As expected, for all actions rated, the grand mean of schematic subjects was significantly greater than that of aschematics, as shown in Table 1.

TABLE 1

SCALE A: VISUALIZE VERSUS VERBALIZE PROTOCOL ACTIONS

Comparisons were also made of the grand means of the two groups on Scale A for just the base actions and just the idiosyncratic actions in the recall protocols. Base actions were defined to be any of the 20 actions identified from the first set of protocols as central to the event (see Methods section), and coded by at least two of the three judges. Idiosyncratic actions were defined as all other actions. These comparisons revealed that schematics are significantly more prone to visualize base actions than are aschematics. However, the difference between the two groups on idiosyncratic actions was not significant, although it is in the predicted direction (see Table 1). These findings are generally supportive of the first hypothesis in that schematic processing and mental imagery appear to be complementary retrieval processes.

The second hypothesis predicts that schematics using mental imagery will generate more vivid images than aschematics. This hypothesis was tested by comparing the grand means of the two groups on Scale B of the Protocol Item Analysis measure, on which subjects rated each visualized action in the protocol in terms of the vividness of the image formed. As shown in Table 2, the ratings of schematics indicated that their images were significantly more vivid than those of aschematics.

TABLE 2

SCALE B: VIVIDNESS OF IMAGES FORMED

The vividness of visualized base actions and of visualized idiosyncratic actions was also compared between the two groups. Schematics were found to generate significantly more vivid images of base actions than aschematics. Furthermore, although on Scale A, schematics and aschematics did not differ significantly in their tendency to visualize idiosyncratic actions, the images generated by schematics for such actions were found to be significantly more vivid (see Table 2). These findings may be interpreted as supportive of the second hypothesis.

IMPLICATIONS

Based on the results of this investigation, it may be concluded that the use of visual imagery and cognitive scripts as strategies to retrieve from memory information about previously experienced events are not independent. Consistent with the cyclical processing model, individuals with well-formed knowledge structures in the form of scripts reported a greater use of a visual means of retrieval and greater vividness in the images formed. These findings have several implications for the use of imagery instructions as a retrieval strategy, and for the pictorial content of advertising messages as well.

Although numerous studies report beneficial effects of imagery instructions on both recall and evaluations of information (Lutz and Lutz 1978), the present findings suggest that differences would be expected in the magnitude of those effects for individuals with and without a relevant schema. Instructions to generate images about events or concepts with which one is unfamiliar would not be expected to produce facilitative effects on either cognitive or evaluative responses Methodologically, therefore, it would appear to be inappropriate to attempt an examination of the effects of this imagery eliciting strategy without controlling for the presence or absence of related schemata.

Concerning the design of advertising messages, there is again substantial empirical evidence supporting the positive influence of certain kinds of images, especially interactive images, on recall and recognition of brand names and product classes. However, the success of the images so employed may be limited should there exist inconsistencies with previously -formed knowledge structures among the target audience. Thus, ensuring congruence of pictorial advertising content to pre-existing cognitive schemata may be a critical prerequisite for messages which effectively assist in the subsequent retrieval of presented information.

CONCLUSIONS

This investigation examined the relationship between cognitive scripts and mental visual imagery as strategies to retrieve information about an event. Rather than operating independently, it was found that individuals with a script generated more images about the event, and that those images were more vivid, compared to individuals with na script. These findings are consistent with the cyclical processing hypothesis, and may provide some useful guidelines in identifying the precise conditions under which imagery will favorably affect the recall and evaluation of information.

REFERENCES

Abbott, V. and Black, J. 8. (1980), "The Representation of Scripts in Memory," Cognitive Science Technical Report #5, New Haven: Yale University Press.

Alba, J. W., Alexander, S. G., Rasher, L. and Caniglia, K. (1981), "The Role of Context in the Encoding of Information," Journal of Experimental Psychology: Human Learning and Memory, 7, 283-292.

Anderson, R. C. and Pichert, 3. W. (1978), "Recall of Previously Unrecallable Information Following a Shift in Perspective," Journal of Verbal Learning and Verbal Behavior, 17, 1-12.

Bower, G. H., Black, J. B. and Turner, T. J. (1979), "Scripts in Memory for Text," Cognitive Psychology, 11, 177-220.

Cantor, N. and Mischel, W. (1977), "Traits as Prototypes: Effects on Recognition Memory," Journal of Personality and Social Psychology, 35, 38-48.

Childers, T. C. (1982), "Sensory and Semantic Bases of Interactive Imagery," Unpublished Doctoral Dissertation, University of Wisconsin-Madison.

Childers, T. C. and Houston, M. J. (1982), "Imagery Paradigms for Consumer Research: Alternative Perspectives from Cognitive Psychology," in R. Bagozzi and A. Tybout (eds.), Advances in Consumer Research, Vol. X, Ann Arbor: Association for Consumer Research.

Hampson, P. and Morris, P. (1979), "Cyclical Processing: A Framework for Imagery Research," Journal of Mental Imagery, 3, 11-22.

Hasher, L. and Griffen, M. (1978), "Reconstructive and Reproductive Processes in Memory," Journal of Experimental Psychology, 4, 318-330.

John, G. and Whitney, J. C. (1982), "An Empirical Investigation of the Serial Structure of Scripts," in B. J. Walker (ed.), An Assessment of Marketing Thought and Practice, Chicago: American Marketing Association.

Kant, K. (1787), Critique to Pure Reason (1st ed. 1781, 2nd ed. 1787, translated by N. Kemp Smith, 1963), London: MacMillan.

Lord, C. (1980), "Schemas and Images as Memory Aids: Two Modes of Processing Social Information," Journal of Personality and Social Psychology, 38, 257-269.

Lutz, K. and Lutz, R. (1977), "Effects of Interactive Imagery on Learning: Application to Advertising," Journal of Applied Psychology, 62, 493-98.

Lutz, K. (1978), "Imagery Eliciting Strategies: Review and Implications of Research, in H. K. Hunt (ed.), Advances in Consumer Research, Vol. V, 611-20.

Palmer, S. F. (1975), "Visual Perception and world Knowledge: Notes on a Model of Sensory-Cognitive Interaction," in D. A. Norman, D. E. Rumelhart, and the LNR Research Group, Explorations in Cognition, San Francisco: Freeman.

Percy, L. and Rossiter, J. (1982), Effects of Picture Size and Color on Brand Attitude Responses in Print Advertising," in R. Bagozzi and A. Tybout (eds.), Advances in Consumer Research, Vol. X, Ann Arbor: Association for Consumer Research.

Rethans, A. and Hastak, M. (1980), Representation of Product Hazards in Consumer Memory," in A. Mitchell (ed.), Advances in Consumer Research, Vol. IX, Ann Arbor: Association for Consumer Research, 487-93.

Richardson, A. (1969), Mental Imagery, London: Routledge and Kegan Paul.

Rumelhart, D. and Ortony, A. (1977), "The Representation of Knowledge in Memory," in R. Anderson, R. Spiro, and W. Montague (eds.), Schooling and the Acquisition of Knowledge, New Jersey: Lawrence Erlbaum Associates.

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

Smith, R. A. and Houston, M. J. (1983), "Script-Based Evaluations of Satisfaction with Services," in L. Berry et al. (eds.), Imerging Perspectives on Services Marketing, Chicago: American Marketing Association.

Taylor, S. E. and Crocker, J. (1981), "Schematic Bases of Social Information Processing," in E. T. Higgins, C. D. Herman and M. P. Zanna (eds.), Social Cognition: The Ontario Symposium, Hillsdale, NJ: Lawrence Erlbaum Associates.

White, R., Sheehan, P. W. and Ashton, R. (1977), "Imagery Assessment: A Survey of Self-Report Measures," Journal of Mental Imagery, 1, 145-70.

Yuille, J. C. and Catchpole, M. J. (1977), "The Role of Imagery in Models of Cognition, Journal of Mental Imagery, 1, 171-80.

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