On the Use of Verbal Protocols in Cross-Cultural and Cross-National Consumer Research

Susan P. Douglas, New York University
ABSTRACT - Verbal protocols have recently received considerable attention as a methodology for examining consumer information acquisition and processing. They also have potential as a methodology for identifying constructs to be examined in cross-cultural consumer research. Some illustrations of ways in which verbal protocols can provide helpful insights are provided based on data from a study of consumer information-processing currently being conducted in France.
[ to cite ]:
Susan P. Douglas (1980) ,"On the Use of Verbal Protocols in Cross-Cultural and Cross-National Consumer Research", in NA - Advances in Consumer Research Volume 07, eds. Jerry C. Olson, Ann Abor, MI : Association for Consumer Research, Pages: 684-687.

Advances in Consumer Research Volume 7, 1980     Pages 684-687

ON THE USE OF VERBAL PROTOCOLS IN CROSS-CULTURAL AND CROSS-NATIONAL CONSUMER RESEARCH

Susan P. Douglas, New York University

[The author wishes to acknowledge the helpful comments of Professors Elizabeth Hirschman, John Rossiter and Yoram Wind, as also of three anonymous reviewers, on an earlier version of this paper.]

[Department of International Business, 90 Trinity Place, New York, NY 10086.]

ABSTRACT -

Verbal protocols have recently received considerable attention as a methodology for examining consumer information acquisition and processing. They also have potential as a methodology for identifying constructs to be examined in cross-cultural consumer research. Some illustrations of ways in which verbal protocols can provide helpful insights are provided based on data from a study of consumer information-processing currently being conducted in France.

INTRODUCTION

One of the more innovative trends in consumer research over the past several years has been the use of unstructured data collection techniques such as protocols, to examine consumer cognitive processes and decision-making procedures (Bettman 1979 , Russo 1978). These developments have been stimulated in large measure by increased interest in understanding how consumers actually arrive at, and use information in making decisions, as well as in the impact of situational variables such as the retail store environment on these decisions.

Traditional data collection procedures are largely inadequate for such purposes, since they focus on decision outcomes or end results, rather than tapping the internal workings of these processes. Survey data, for example, rely on a highly structured model of consumer response, analyzing final decisions, and even when using open-ended responses, require specification of a purchase or choice situation. Motivation or depth interviews provide generalized introspective insights into decision-making processes, but primarily reflect consumer ex post rationalizations and recollections of these events.

Protocol methodologies are specifically designed to avoid the imposition of a prespecified choice model on the respondent allowing him to respond freely in his own terms in relation to an actual choice task or decision situation. They, thus. provide information relative to situational effects and the purchase environment, typically missed by other data collection techniques. Yet, such factors are often critical in influencing the choices which are finally made.

To date, protocol techniques have mostly been used to examine the information utilized by consumers in making decisions, focusing primarily on the use of information in the retail store environment (Alexis et. al. 1968, Bettman 1970, Payne & Ragsdale 1978, Svensen 1974, Wright 1970 ). Some protocols have also been collected monitoring thoughts generated by advertising and product stimuli (Wright 1974, Olson, Toy and Dover 1978), though these typically require modification of the 'classic' protocol techniques. [These protocols" are typically not collected concurrently with the performance of a decision task, but are "thoughts" related to advertising and product stimuli, and hence more closely resemble free-response techniques (Payne and Ragsdale 1978).] This orientation appears to be largely the result of the interest shown in government and consumerist circles in the regulation of information provided to consumers, and the suitability of these techniques for analyzing consumer information needs and processing behavior.

Protocols, nonetheless, have considerably wider potential application. In particular, they are highly appropriate for use in cross-cultural and cross-national consumer research. Since no formal structure or conceptual framework is utilized, the imposition of constructs, terms, or measures developed in relation to one culture, which may not be appropriate in another, is avoided and constructs or terms relevant to each individual culture identified.

The purpose of the present paper is to draw attention to possible uses of protocol techniques in cross-cultural and cross-national consumer research. First, protocol methods are briefly reviewed. Their advantages and limitations in relation to the specific problems of cross-cultural and cross-national research are next examined. Some illustrative examples, taken from an ongoing study of consumer information processing in France are then discussed, and some other possible applications indicated.

PROTOCOL METHODS

The use of "protocols" to study consumer choice behavior has been largely inspired by methods used in studying generalized problem-solving behavior. Here, individuals were asked to think out loud when solving a problem. The objective of these studies was to assess the procedures used by "expert" problem solvers, in chess games, theorem proving, or musical composition, in order to develop routine procedures or schemes for solving such problems (Newell and Simon 1972). A protocol has thus been defined as "a transcript of the verbalized thought and actions of a subject, when instructed to think or problem-solve aloud. The transcript is a record of the subject's thought processes, while engaged in making a decision." (Clarkson 1962).

One of the earliest studies using protocols to examine buying decisions was conducted by Clarkson to assess the rules used by trust investment officers to select stock portfolios (Clarkson 1962). Here, the officers were asked to recall the procedures and decision-rules used to reach a number of stock selection decisions. In another early study of consumer decision-making, two female subjects were sent on trips to buy clothing, accompanied by an interviewer who tape-recorded their verbalizations about how they were selecting clothing (Alexis et. al. 1968).

The first consumer protocols were thus typically collected on actual shopping trips (King 1969, Bettman 1970), and subjects were often recruited among friends, relatives and acquaintances. The interviewer held the microphone for the shopper and sometimes, helped to push the shopping cart. [An alternative procedure, though generally less successful, is to provide shoppers with a small tape recorder hung around the neck or attached to the lapel, and request them to record their thoughts as they go around the store making purchases.] He, therefore, provided an excuse or target for verbalization, and helped to alleviate the embarrassment that many shoppers feel in talking alone into a microphone.

These protocols provided a richness of detail about the shopping environment and consumer thought processes not revealed by other data collection procedures. No prior structure was imposed on the shopper's response pattern, and hence, the subject supplied spontaneously the information, relating, for example, to the product attributes, or situational variables, of most concern to him. The impact of store environmental variables such as store layout, shelf-facing, store promotions or product labeling on consumer choice was, for example, often clearly apparent.

On the other hand, the unstructured character of the early protocols meant that the data were often partial in character, and difficult to analyze. As a result, "prompted" or retrospective protocols subsequent to the performance of the decision tests have been collected, often, in conjunction with other data collection procedures, as, for example, information display boards, eye fixation or other laboratory tasks (Arch Bettman and Kakkar 1978, Payne 1976, Russo and Rosen 1975). The task provides a specific focus for the protocol, giving it a more structured character, and thus considerably facilitates the analysis.

USE OF PROTOCOL DATA IN CROSS-CULTURAL AND CROSS-NATIONAL RESEARCH

Although to the author's knowledge, protocol techniques have been little utilized in the context of cross-cultural and cross-national research, they are particularly appropriate in this regard. Here, a key issue, especially in the initial stages of research, is to identify relevant concepts and constructs to be examined in different cultural settings (Anderson 1967, Berry 1969, Sears 1961, Straus 1969). In this context, avoidance of a 'pseudo-etic' bias, i.e. the use of constructs, or measures developed in relation to a specific cultural environment is an important consideration. These may not be appropriate or well-adapted to other cultural environments and so fail to capture some important aspects of the problem studied (Ember 1977, Gordon 1968, Triandis 1972).

Protocol methods, particularly the classic "free response" protocols are well-suited to dealing with such issues. In contrast to more traditional data collection procedures, the researcher does not define or specify the form, nor in "classic" protocols, the particular stimuli to which the subject should respond, and hence does not impose his own cultural frame of reference on the respondent. Each subject identifies, of his own accord, the factors of importance to him.

Protocols supplied by a variety of subjects in each cultural context also provide indications as to relevant concepts to be examined in each context and how these may most appropriately be defined. Based on this analysis some initial measures adapted to each context can be developed. For example, protocol data could be used to identify relevant product attributes, attitudinal statements, or life-style items, to be examined.

Protocol data also enable the researcher to identify the terminology used by consumers when thinking about products, or in relation to purchase situations, or the store environment. This can be particularly useful when conducting research in different linguistic environments, where the researcher, although familiar with the formal language, may not be conversant with specific consumer "speech terms."

In addition, since protocols are collected in a real or simulated purchase environment, they are helpful in bringing to light the role of situational variables, such as the type of retail outlet, or the degree of interaction with store personnel. Often such factors differ from one national or cultural environment to another, and may underlie or account for, many of the differences in behavior observed between countries or cultures.

While protocol methods offer a number of advantages in cross-cultural research, they suffer from some serious limitations, which also need to be taken into consideration. These stem primarily from their unstructured character.

In the first place, interpretation of cross-cultural protocol data is highly subjective. Since there is no prespecified conceptual model to provide a framework for analysis, the onus of defining what constructs are relevant, and determining where there are similarities and differences between countries is placed on the researcher. Consequently, any inferences drawn from the data are open to criticisms of subjective, cultural bias on the part of the researcher.

This is best resolved by the use of multiple Judges, each with different cultural self-referents, to analyze the data, and develop relevant coding categories. For example, in a two country/culture study, one monolingual Judge from each culture or country could be used, and two bilingual Judges - each a native speaker from different countries. An iterative, back-checking process similar to that used in questionnaire translation could then be applied to ensure "decentring" of concept interpretation or coding, and elimination of a single culture referent bias (Werner and Campbell 1973).

Protocol methods also assume that subjects express via their recorded thoughts, factors relevant in reaching a decision. This has been questioned in relation to research conducted in the U.S. (Nisbett and Wilson 1977) where respondents have relatively high educational levels. It may pose additional problems in other countries, where respondents are less educated and less well-informed.

The validity and ease of collecting protocol data also depends on respondent willingness to cooperate and to talk out loud while making choice decisions (Wright 1976). Again, difficulties may be encountered in cross-cultural and cross-national research. Cultures differ with regard to their loquaciousness (Mitchell 1965), and willingness to respond to questions (Brislin, Lonner and Thorndike 1973) and this may have an important impact on their capacity to verbalize.

Protocol data also tends to be partial in character. It consists predominantly of fragmented pieces of information and phrases, and information stored in secondary memory, relating, for example, to routine purchases and frequently bought brands may not be tapped.

Consequently, the extent to which it provides a comprehensive picture of all relevant elements of the decision process is not always clear, and certain aspects relevant to cross-national comparisons maybe under-emphasized. Such information can, however, be obtained through complementary data collection procedures, as for example, supplementary questionnaires or diary data.

While protocols provide a wealth of detail and a breadth of coverage not furnished by structured data collection techniques, they are essentially qualitative in character. They are, therefore, best suited for use in the exploratory phases of research, and for generating relevant cross-cultural research hypotheses to be examined in subsequent phases of research, rather than for testing explicit research hypotheses.

SOME POTENTIAL AREAS OF APPLICATION

[These illustrations are drawn predominantly from an ongoing study on consumer information processing, currently being conducted at the Centre d'Enseignement Superieur des Affaires, Jouy-en-Joses, France, with financial support from the Delegation Generals de Recherche Scientifique. The author wishes to acknowledge assistance from her colleagues at CESA in data collection.]

In using protocol methods in cross-national and cross-cultural research, a number of potential areas of application may be identified. Among these, are projects where:

1. attention is centered on examining cues used in making consumer purchase decisions in different cultural and national environments;

2. product class boundaries, and product usage may differ in each national and cultural environment;

3. interaction at point of purchase is likely to have a significant impact on consumer decisions, and the nature of retail store environment varies from country to country.

Cues Used in Selecting Products: In examining cues used in brand selection, protocol data can be useful in identifying the reactions of consumers to, instore variables, such as shelf-displays, promotions, as well as product packaging and labeling, and their perception and interpretation of such stimuli. A protocol collected in a French supermarket, for example showed that a subject thought that the weight of a promotional gift was included in the weight of the package. This influenced her choice of promotional gift. Such a reaction is unlikely to be anticipated by a researcher, and would probably not be detected by other data collection techniques such as in-depth interviews or surveys.

Product Class Boundaries: Protocol data can also be useful in providing insights into product categorization and in determining appropriate definitions of product classes or cultural settings. Protocol data collected in France showed, for example, that in purchasing butter and fruit, different product variants were purchased for different end uses. One type of butter was purchased for cooking, and another for table use. This implied that butter was not a homogeneous product category. Similarly, in purchasing oranges, one variety was bought for making Juice, and another for eating. This suggests that in examining fruit purchases, the relevant product class should not only include fruit but also substitutes for fresh Juice, such as frozen, canned or bottled Juice.

Interaction With the Retail Store Environment: Protocols can also aid in understanding how store personnel, store layout and other store environmental variables influence consumer choice decisions. Again, protocol data collected in traditional service establishments in France clearly demonstrates the influence of store clerks on purchase decisions. Suggestions are offered about products to fit specific needs, for example, purchases of gifts, clothing suitable for a particular occasion, or menu-planning. Information is also often provided about different products and brands, and influences decision, sometimes to the point of causing brand switching.

While these illustrations provide some examples applying protocol methods in cross-cultural and national research, it should be emphasized that these are limited in scope, due to the nature of the study from which they are drawn. This was not intended as a comparative study of information processing, but rather to test the feasibility of using protocol techniques in a different cultural and retail environment. Protocol techniques can also be appropriate in a number of other situations. More generally they can be used to:

- examine the consumer choice models, i.e. lexicographic, disjunctive, used in different cultural environments;

- identify and compare relative parent or child involvement in purchase decisions in different cultures and countries;

- explore consumer thoughts and attention to advertisements or other promotional stimuli, in non-store environments-such as the home, the street and public transport.

In addition, protocol techniques could be used in conjunction with other methods of data collection. For example, videotapes of actual shopping trips, or eye fixation data could be collected as well as instore protocols, in order to compare the use of visual vs. verbal data in cross-cultural and national research.

CONCLUSIONS

In brief, while protocol analysis has primarily received attention as an appropriate technique for examining consumer information seeking and processing behavior, it has much wider potential application in cross-cultural and cross-national research. The absence of a preimposed conceptual framework or verbal stimuli which might generate a culture-specific or pseudo-etic bias, make protocols a particularly suitable methodology in this context, especially in the initial stages of research. Further exploration of their use could thus constitute an important step in developing improved methodologies in an all too rarely and poorly investigated area of research.

REFERENCES

Alexis, Marcus, George H. Haines and Leonard Simon. (1968), "Consumer Information Processing: the Case of Woman's Clothing." in Robert L. King (ed.) Marketing and the New Science of Planning, Chicago: American Marketing Association.

Anderson, R., (1967), "On the Comparability of Meaningful Stimuli in Cross-Cultural Research," Sociometry, 30, 124-136.

Arch, David C., James R. Bettman and Pardeep Kakkar, (1978), "Subjects' Information Processing in Information Board Studies," in Advances in Consumer Research, Vol. V, Chicago: Association for Consumer Research.

Berry, John W., (1969), "On Cross-Cultural Comparability", International Journal of Psychology, 6, 2, 119-128.

Bettman, James R. (1970), "Information Processing Models of Consumer Behavior," Journal of Marketing Research, 7,370- 376.

Bettman, James R. (1977),"Data Collection and Analysis Approaches to Studying Consumers Information Processing," in Advances in Consumer Research, Vol. IV, Atlanta: Association for Consumer Research, 342-3~8.

Bettman, James R. (1979), An Information Processing Theory of Consumer Choice, Reading, Mass: Addison Wesley.

Brislin, R., W. Lonner and R. Thorndike.(1973), Cross-Cultural Research Methods. New York: Wiley.

Clarkson, Geoffrey P. E. (1962), Portfolio Selection: A Simulation of Trust Investment, Englewood Cliffs, New Jersey: Prentice Hall.

Ember, Carol R. (1977), "Cross-Cultural Cognitive Studies" Annual Review of Anthropology, 6, 33-56.

Gordon, L. V. (1968), "Comments on Cross-Cultural Equivalence of Personality Measures," Journal of Social Psychology, 75, 11-19.

King, Robert H. (!969), "A Study of the Problem of Building a Model to Simulate the Cognitive Process of a Shopper in a Supermarket," in George H. Haines, Consumer Behavior: Learning Models of Purchasing, New York: Free Press, 22-67.

Mitchell, Robert E. (1965), "Survey Materials Collected in the Developing Countries: Sampling, Measurement and Interviewing Obstacles to Intra and International Comparisons," International Social Science Journal, 17, 4.

Newell, Allen and Herbert A. Simon (1972), Human Problem-Solving, Englewood Cliffs, New Jersey: Prentice Hall.

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

Olson, Jerry C., Daniel R. Toy and Phillip Al Dover, (1978), "Mediating Effects of Cognitive Responses to Advertising on Cognitive Structure," in Advances in Consumer Research, Vol. V, Chicago: Association for Consumer Research, 72-78.

Payne, John W. (1976), "Task Complexity and Contingent Processing in Decision-Making: An Information Search and Protocol Analysis," Organizational Behavior and Human Performance, 16, 3, 366-387.

Payne, John W., and E. K. Easton Ragsdale. (1978),"Verbal Protocols and Direct Observation of Supermarket Shopping Behavior: Some Findings and a Discussion of Methods," in H. Keith Hunt (ed.) Advances in Consumer Research, Vol. V, Chicago: Association for Consumer Research, 571-577.

Pike, Kenneth. (1966), Language in Relation to a Unified Theory of the Structure of Human Behavior, The Hague: Mouton.

Russo, Edward J. and Larry D. Rosen. (1975), "An Eye Fixation Analysis of Multi-Alternative Choice," Memory and Cognition, 3n, 267-276.

Russo, Edward J. (1978), "Eye Fixations Can Save the World: A Critical Evaluation and a Comparison Between Eye Fixations and Other Information Processing Methodologies," in H. Keith Hunt (ed.) Advances in Consumer Research, Vol. V, Chicago: Association for Consumer Research, 571-577.

Sears, Robert R. (1961), "Trans-Cultural Variables and Conceptual Equivalence," in Bert Kaplan (ed.), Studying Personality Cross-Culturally, Evanston, Ill.: Row, Peterson and Co., 445-455.

Straus, Murray A. (1969), "Phenomenal Identity and Conceptual Equivalence in Cross-National Comparative Research," Journal of Marriage and the Family.

Svenson, Ola (1975), "Coded Think Aloud Protocols Obtained When Making a Choice to Purchase One of Seven Hypothetically Offered Houses: Some Examples," Unpublished paper. University of Stockholm,

Triandis, Harry C. (1972), The Analysis of Subjective Culture, New York: Wiley.

Werner. O. and D. Campbell. (1973), "Translating, Working Through Interpreters and the Problems of Decentring," in Raoul Naroll and Ronald Cohen (eds.), A Handbook of Method in Cultural Anthropology, New York: Columbia University Press, 398-420.

Wright, Peter L. (1974), "On The Direct Monitoring of Cognitive Response to Advertising," in C. David Hughes and Michael L. Ray (eds.), Buyer/Consumer Information Processing, Chapel Hill: University of North Carolina, 220-248.

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