Influence Processes in Interpersonal Persuasion

Morris Holbrook, Columbia University
John O' Shaughnessy, Columbia University
ABSTRACT - This paper reviews the progress of research on interpersonal persuasion in the salesman-buyer dyad. The nature of the dynamic verbal interactions between salesmen and buyers is identified as a promising focus for research. Methodological suggestions are therefore offered for designing content analyses to represent the unfolding of salesman-buyer interactions over time.
[ to cite ]:
Morris Holbrook and John O' Shaughnessy (1976) ,"Influence Processes in Interpersonal Persuasion", in NA - Advances in Consumer Research Volume 03, eds. Beverlee B. Anderson, Cincinnati, OH : Association for Consumer Research, Pages: 364-369.

Advances in Consumer Research Volume 3, 1976      Pages 364-369


Morris Holbrook, Columbia University

John O' Shaughnessy, Columbia University

[Morris Holbrook gratefully acknowledges the support of Columbia University's Faculty Research Fund.]


This paper reviews the progress of research on interpersonal persuasion in the salesman-buyer dyad. The nature of the dynamic verbal interactions between salesmen and buyers is identified as a promising focus for research. Methodological suggestions are therefore offered for designing content analyses to represent the unfolding of salesman-buyer interactions over time.


The verbal interactions between a salesman and buyer engaged in the selling process are a singularly clear example of interpersonal influence at work. This paper briefly outlines some key distinctions between various ways of studying interpersonal influence in the salesman-buyer dyad. It then employs these distinctions to analyze the development of hypotheses and methodologies directed at studying the selling process. Finally, it draws conclusions concerning the approach of greatest potential use in future studies of interaction between salesmen and buyers.


Robertson and Chase (1968) portrayed the selling process as an open dynamic system in which environmental inputs (e.g., human, technological, and organizational) are transformed into outputs (e.g., cognitive, affective, and behavioral) which are, in turn, fed back into subsequent stages in the salesman-buyer dyad (e.g., as different expectations, feelings, skills, or resources). The heart of the system is the transformation process which converts inputs into outputs and which involves the essence of salesman-buyer interaction: the similarity between participants, their perceptions of one another, how fully they fulfill reciprocal role expectations, and most especially, how they talk and react to each other. It is this interaction process which lends the salesman-buyer dyad its dynamic character.

Unfortunately, the almost self-evident importance of the verbal interaction between salesman and buyer has inspired surprisingly little serious empirical re-search--virtually none of which, we shall argue, deals adequately with the issues involved. Most of the relatively few well-known studies were reviewed by Davis and Silk (1972) and by Capon, Holbrook, and Hulbert (undated). But neither review adopted a critical perspective sharply focused on revealing the most fruitful direction for future research in this area. For example, Davis and Silk distinguished between three types of salesman-effectiveness research: "psychological tests" (e.g., selection studies), "interpersonal influence'' studies (e.g., based on mass-communication theory), and "customer-salesman interaction" studies (e.g., focusing on the similarity or matching between buyer and salesman). There is, however, a conceptually distinct fourth type of selling-process research which these authors neglected, but which may offer the greatest potential for explaining the outcome of selling interactions.


Research on the selling process has been characterized by two key distinctions. [Please note that these distinctions are not synonymous with those introduced by Capon et al. (undated).] First, such studies have typically focused either on static characteristics of the buyer and salesman (e.g., demographic, socioeconomic, or personality variables) or on communication variables (e.g., those represented by the familiar communication-theoretic paradigm--source-message-channel-receiver-effect). An especially important communication variable is the verbal content of the selling interaction. Secondly, research may view the selling process as a two-way interaction (focusing simultaneously on the actions and reactions between buyer and salesman) or may adopt an essentially one-way perspective (dealing separately with the salesman's and buyer's characteristics and/or behavior). These contrasts suggest the typology of selling research shown in Table 1, which assigns each cell a convenient name (and indicates Davis and Silk's terminology parenthetically).




Selection Studies

Studies of salesman selection (e.g., Mayer & Greenberg, 1964) have been ably reviewed by Cotham (1970) and will not concern us here, except to point out that sometimes studies which purport to deal with salesman-buyer interaction may actually confine their attention to a rather limited, one-way analysis of static salesman characteristics. Albaum (1967), for example, predicted the sales penetration of service stations using independent variables like the dealer's business attitudes, number of services provided, and customer waiting-time. But clearly, such factors do not adequately represent the complex, interactive verbal communication between buyer and salesman.

Communication Model

The effects of both the sales message and its source have received attention from researchers approaching selling research from a communication-theoretic perspective.

Message effects. In an early study of communication in selling, Pace (1962) found a significant tendency for higher-performing door-to-door saleswomen to earn higher interview-based ratings on their use of emotional appeals, dramatization, and language (but not on several other communication characteristics). More directly, Levitt's (1965) well-known experimental study of one-way persuasive communication compared the effects of contrasting sales messages themselves, demonstrating the significant superiority of "good" over "poor" filmed presentations in persuading subjects to react favorably to a new paint product.

Such clear-cut results, however, have often failed to appear in subsequent studies of message effects. Farley and Swinth (1967), for example, found that a "product pitch" (describing the features of a roll-up yardstick) and a "personal pitch" (relating the yardstick to the subject's role as student or wife) exerted significantly different effects on attitude, but not on choice behavior. Similarly, though Reizenstein's (1971) hard- and soft-sell messages were equally ineffective in causing reversal of preference between first- and sixth-ranked patterns of crystal stemware, the soft-sell did produce significantly greater decision reversal between the first- and second-ranked alternatives, a difference which Reizenstein explained by invoking the theory of cognitive dissonance.

Even more confusing findings have emerged from a recent series of communication-theoretic studies comparing "canned" presentations with "interactive" deliveries in which the salesman encourages the prospect to talk. Jolson's (1975) interactive presentations of a new reference book (in which a classroom audience was permitted to ask questions) generated greater product comprehension, but his canned version (with heavy use of audiovisual aids) produced a higher rate of purchase 'intention. It appears dangerous, however, to generalize even this equivocal finding from Jolson's classroom setting to a selling situation involving one-to-one interaction.

Capon (1975) avoided this methodological weakness by comparing the effectiveness of canned and interactive sales presentations in the telephone selling of magazine subscriptions. Using six messages in a complex quasi-factorial design, he found no significant main effect of his canned/interactive manipulation on attitudes or purchase behavior, though significant interaction effects did suggest that certain salesmen were most successful with certain messages.

Another possible situational determinant of communication effectiveness is the business environment characterizing the seller's product. Pasold (1975) hypothesized that selling strategies which he called "preactive" (salesman dominated), "interactive" (mutually directed), and "proactive" (focusing on the alignment of long-term objectives) might, respectively, be most effective in markets characterized by imperfect competition, oligopoly, and turbulence. A correlational study of stocking sales found beta coefficients with the predicted relative magnitudes for the contrasting selling strategies. But the tiny t values of these coefficients and the failure of their relative magnitudes to be consistently upheld in a validation sample raise questions concerning the statistical significance of Pasold's finding.

Source effects. Levitt's (1965) pathbreaking study also extended the Hovlandian tradition of one-way communication-theoretic research to an examination of the effects of source credibility in selling. He claimed to find both a source effect (more favorable response to the salesman identified with the more credible parent firm) and a sleeper effect (the erosion of this differential over time), but Capon, Holbrook, and Hulbert (1972) have questioned the significance and consistency of these findings, without serious challenge from Levitt (1972).

Conclusions on the communication model. Perhaps the safest conclusion to be derived from the one-way communication-theoretic studies is that the nature of source or message effects tends to depend upon the situation--on the dissonance of the customer or his social setting or his profession, on the salesman's personal characteristics, on market conditions--or, in short, on the two-way process of interaction between buyer and salesman. Apparently, a research paradigm which views the selling process as one-way communication is inadequate to the task of generating consistent findings. Let us turn, then, to studies which have viewed selling as a two-way process of interaction.

Similarity or Matching Studies

In search of similarity effects. Attempts to relate sales success to the similarity or matching of salesman and prospect on static characteristics have proven inconclusive, almost without exception. Evans' (1963) classic study in this area claimed such relationships in the sale of life insurance, but when reanalyzed by Capon, Holbrook, and Hulbert (undated), his data did not pass conventional tests of statistical significance. Gadel (1964) showed that life-insurance buyers tended to resemble salesmen in age, but because she looked only at purchasers, no inference can be made concerning the effects of age similarity on sales success. Capon (1975) found no significant relationship between the perceived similarity of a magazine salesman and attitudes or intention to subscribe. Mathews, Wilson, and Moncky (1972) manipulated the perceived similarity/dissimilarity of students engaged in a selling simulation and found the expected difference between treatment groups in cooperative and competitive behavior, but interpretation of these results was beclouded by a still higher incidence of cooperation and conflict in a control group than in the respective treatment groups. Finally, Tosi (1966) adopted a slightly different view of the similarity hypothesis, focusing on the matching of role expectations between buyers, salesmen, and sales managers; but with various measures of sales success, only one of his seven hypotheses received statistically significant support.

Comparisons of similarity and source effects. Two ambitious studies have pitted the already shaky source and similarity hypotheses against one another, again with mixed results. Brock's (1965) two confederates tried to switch paint customers to a different-priced brand using either a similarity or an expertise appeal (claiming recently to have purchased about the same amount of paint or twenty times as much). The results supported the similarity hypothesis at the expense of the source effect, but their interpretation was complicated by the questionable believability of the expertise manipulation, the salesman's apparent complicity in the experimental design, and the failure of that design to permit the simultaneous support of both hypotheses. Indeed, when Woodside and Davenport (1974) created a factorial manipulation of both perceived similarity (a record-store saleswoman's claimed preference for the same or a different type of music) and perceived expertise (her apparent ability to explain the proper use of a tape-recorder cleaning device),both treatment variables exerted significant positive main effects on purchase of the cleaning gadget, with a significant difference in favor of the source effect. Perhaps these results differed from Brock's because the salesgirl convincingly demonstrated her expertise by explaining the product's use whereas the paint salesman only claimed to have bought an almost preposterously large amount of paint.

Conclusions on source and similarity effects. Clearly, the debate between advocates of source and similarity effects could be continued, probably without ever resolving the issue definitively. It appears, however, that a clearer understanding of selling might emerge from a reorientation of research on the salesman-buyer dyad toward a more deliberate focus both on the two-way interaction between buyer and salesman and on the verbal content of that interpersonal communication process. This is the approach we call the study of "interaction synchronization" and to which we now turn.

The Study of the Synchronization of Verbal Interaction

Unfortunately, we cannot point to examples of research which has successfully viewed selling as a dynamic process of verbal interaction. We can, however, cite several studies which have helped to clear a path to examining the synchronization of verbal behavior.

Willett and Pennington (1966) were the first to focus systematically on a content analysis of the verbal components of buyer-salesman interaction. They applied Bales ' (1950) Interaction Process Analysis to protocols of the transactions between 14 appliance salesmen and their customers and found, for example, that successful sales outcomes were associated with a high incidence of giving suggestions and asking for opinions and with a relatively higher involvement of salesmen in asking task-oriented questions. Pennington's (1968) own content analysis of the same data viewed the salesman-buyer dyad as an example of bargaining behavior. A discriminant analysis successfully predicted about 80% of the purchase outcomes using the frequencies of verbal acts falling into various bargaining categories (direct offers, attempts to change concession limits, commitment to concession limits, etc.). A further re-analysis of the same protocols by Olshavsky (1973) counted the number of verbal acts referring to various product-related attributes during succeeding stages of the sales interview and highlighted, for example, the customer's predominant concern for price during what he called the "evaluation" phase and the relegation of concern for nonproduct attributes like delivery or service to the concluding "consummation" phase. Thus, these three studies have quite appropriately focused attention on various facets of the verbal content of selling interactions. They have, however, shared a common tendency to treat these verbal interactions in an aggregative fashion (e.g., by summing the frequency of verbal acts in each category across customers and/or salesmen) with no explicit attempt to characterize the interaction through time of the content of the verbal acts within a specific customer-salesman pair. Because of this failure to investigate the unfolding interaction of the salesman's and buyer's verbal behavior, these studies tell us little about how a salesman might adjust the verbal content of his presentation to the responses of the buyer in order best to achieve sales success.

Ironically, the oft-forgotten studies by Chapple (Chapple & Donald, 1947; Norman, 1954) displayed just the opposite virtue, focusing resolutely on the temporal relationship between the speaking (or gesturing) of a salesman and that of his interlocutor. Chapple's Interaction Chronograph generated such indices as the salesman's activity level (the average length of his verbal and nonverbal acts) and his initiative-dominance level (his relative tendency to break a silence and to keep talking when interrupted). Work with such measures found good fits between the sales-performance rank of salesmen and the degree to which their interaction-timing indices resembled those of highest-ranking performers. Yet, while elegantly handling the temporal pattern of communication behavior, Chapple's method almost completely ignored the verbal content of salesman-buyer interactions. The Interaction Chronograph might help us determine when a salesman should talk, but it cannot tell us what he should say.


An Illustration of the Content Analysis of Interaction Synchronization

Apparently what we need, then, is research relating selling success to the way in which the salesman adjusts both the timing and the content of his verbal behavior to that of his customer. The raw materials for such a methodology would include, first, some content-analytic scheme theoretically derived from the nature of interpersonal influence and, secondly, some way of analyzing the temporal unfolding of verbal behavior.

Let us begin by presenting a rather idealized illustration of such an approach to the content analysis of interaction synchronization. Elaborating upon O=Shaughnessy's (1971/72, 1974) analysis of interpersonal influence, the illustrative content-analytic scheme shown in Table 2 was developed by arraying Kelman's (1961) types of influence (compliance, identification, internalization) against two of Etzioni's (1964) means to power (material or tangible and symbolic or intangible).



As suggested by Bales (1950), each verbal act by the buyer (B) or salesman (SM) might involve one or more instances of seeking (S) or offering (0) a particular type of content. The protocol of an illustrative salesman-buyer dialogue might then be coded interaction-by-interaction as follows:

1. SM-O-1b, SM-S-2b, SM-O-3a (salesman offers approval, seeks information concerning the buyer's values, offers factual information concerning a tangible feature of the product)

2. B-O-2b, B-S-3b (buyer answers question about values and seeks evaluative information about an intangible feature)

3. SM-O-2a, SM-O-3a (salesman grabs buyer by the arm and, instead, tells him something about a tangible feature )

4. B-O-1b, B-S-3b (buyer offers approval and repeats his question)

5. * * * * * * * * * * * * * * * * * * * * * * *

Such a content-analytic coding of the interactions between buyer and salesman could generate a measure of the degree of match or synchronization between their verbal behavior--that is, the extent to which each participant offers the type of verbal acts which the other seeks. Thus, for example, interaction synchronization might be defined as the relative frequency with which verbal acts of type B-S-i are followed by verbal acts of type SM-0-i (i = 1a, ..., 3b). Alternatively, the concept might be operationally defined as the average proportion of content categories represented by each verbal act which have corresponding content components in the preceding verbal act. The latter measure would ignore whether information was sought or offered, but would represent focusing on the same topic. Obviously, a myriad of possible content-matching measures exists, given the basic content data, with little a priori basis for choosing between them. The crucial point is that such measures would depict some facet of interaction synchronization as a basis for predicting sales success.

Methodological Problems in Content Analysis

The most obvious objection to the kind of content-analytic methodology suggested above is its enormous difficulty of execution. Whatever the content-analytic classification dictated by the researcher's theory of interpersonal influence, perplexing methodological issues arise, first, in defining the content unit to be categorized (a word, a phrase, a sentence, a complete thought, the verbal act, or even the entire transaction) and, secondly, in formulating operational definitions of each content category which are clear, tight, and valid enough to permit some degree of inter-coder reliability. Certainly, there are other tough methodological issues to be solved in research on verbal interaction in selling, but exploratory attempts in this direction at Columbia have convinced us that they are dwarfed by the problems of the content analysis itself (O'Shaughnessy, 1971/72). The difficulty of a research technique, however, has never been a persuasive argument against its potential usefulness, and we remain convinced that refinement of the investigation of personal influence in selling depends upon our ability to deal meaningfully with the interaction between the content of the verbal behavior of buyer and salesman. Even though we may be using rough tools, they appear to stand a better chance of reaching the heart of the matter.

Simplifying the Content-Analytic Scheme

Accordingly, we have been searching for ways to simplify the rather idealized content analysis outlined above. One solution is based on the assumption, clearly articulated by Winch (1958) that human action is rule-governed behavior and that these rules are reflected in the concepts used by the individual to describe his actions. If the salesman is to understand how a buyer is likely to act, he must familiarize himself with the concepts in terms of which the buyer is viewing the situation and recognize that these concepts reflect the rules being used to guide buying action. If these concepts change (e.g., through salesman/buyer interaction), so do the buyer's perceptions of the situation; his actions change as a result. Of course, salesmen are intuitively aware that words reflect attitudes, as when they speak of the buyer making an "investment" rather than a "payment" or when their product is described as "cheap" instead of "inexpensive." However, less obvious is the claim that concepts mirror the set of rules that direct a buyer on what action to take.

Since the notion of following a rule is inseparably linked to the possibility of making a mistake, Winch (like Wittgenstein) has defined behavior as rule-following whenever it makes sense to distinguish between a right and wrong way to do something. It follows that one practical guide to a buyer's rules is to ask him what errors he particularly seeks to avoid. Although this viewpoint has been the subject of much discussion and criticism, its utility for our purposes has not been seriously under, ed (Gellner, 1971; MacIntyre, 1971). The immediate problem, however, is to relate this perspective to our own focus on interaction synchronization.

Hypotheses on Rule-Following in the Salesman-Buyer Dyad

Peters (1958) has argued that any satisfactory explanation of human action must constitute a "sufficient reason" in the sense that the action must be perceived as both an efficient and a socially appropriate means to goal attainment. From our viewpoint, then, a buyer would have (a) rules reflecting beliefs as to what constitutes an efficient means (offering) for attaining the goals stemming from his functional requirements and/or role responsibilities and (b) rules reflecting what he considers socially appropriate behavior on the part of the salesman.

If buying action is rule-following, tentative hypotheses are that the verbal interactions of more successful salesmen (1) will seek to identify the buyer's rules; (2) will have content that is relevant to a rule that has just been reflected in the concepts used by the buyer; (3) will have content that satisfies rules regarding socially appropriate salesman behavior; (4) will have content (if possible) that satisfies product-related rules or (if not possible) that provides the buyer with additional concepts to achieve modification of these rules.

Revised Categories for Content Analysis

The two types of rules suggested above (efficiency and social appropriateness) provide the broad outline for a content-analytic scheme congruent with the above discussion of the interpersonal influence process, except that it ignores what Kelman (1961) referred to as "compliance'' as being of little relevance in selling. Such a classification scheme might include more detailed categories, as listed below.

1. Rules reflecting expectations of socially appropriate behavior on the part of salesmen. If the salesman adheres to the buyer's rules for socially appropriate behavior, the buyer might tend to accept influence via Kelman's process of identification, dependent upon the salesman's "attractiveness." Such rules are associated with what we might call "the four I's": (a) Image Projected--the image the buyer wishes to project (e.g., a verbal focus on "saving money" might indicate the rule that "communications from the salesman must support an image of myself as a hardheaded businessman"); (b) Integrity Sought--the professional and personal norms he values (e.g., a statement that "firms who divulge a supplier's competitive bid to the competition are just not playing fair" might reflect the rule that "communications from the salesman must not contain inquiries about competitive bids and preferably should show respect for my objections to such breaches of confidence''); (c) Interpersonal Relations--how the salesman fits into the buyer's overall role set (e.g., remarking on "the need to confer with everyone affected by a purchase" might indicate the rule that "communications from the salesman must show a sensitive regard for the fact that I need to consult with others before giving an order"); (d) Innovation Risk--fear of the possible consequences of a departure from customary buying patterns (e.g., a reminder that "the firm has been using the same material for twenty years" might reflect the rule that "communications from the salesman must address the problems I face in acting as a change agent within my own organization").

2. Rules covering beliefs as to efficient means of achieving functional requirements and/or role responsibilities. If a salesman adheres to (or succeeds in modifying) the rules regarding beliefs about efficient means, the buyer might move toward influence acceptance via Kelman' s "internalization," as dependent on the salesman's "credibility." Such rules relate to considerations which we might term "the four S's". (a) Suitable Performance/Quality Characteristics of the product and/or vendor (e.g., verbal reference to "the supplier's quality control" might indicate the rule that "communications from the salesman must satisfy me that his firm has the experience to manufacture at the quality level required"); (b) Sources of Supply--the likely availability of the vendor's product (e.g., statements foreboding "the ill effects of late delivery" might reflect the rule that "communications from the salesman must demonstrate his firm's ability to meet its promised delivery date"); (c) Size of Delivered Cost--as opposed to quoted price (e.g., reference to "the costs of handling a raw material" might indicate the rule that "communications from the salesman must show that total use costs are at least competitive"); (d) Services Expected (e.g., mention of "installation and training problems" might reflect the rule that "communications from the salesman must explain the extent of these problems and the help we would get from his firm").

A Simplified Measure of Interaction Synchronization

To repeat, the simplified content analysis which we envision would not attempt to categorize all verbal acts between buyer and salesman, but rather would focus only on that verbal behavior which can readily be classified into those rule-related categories outlined above. More specifically, each implicit rule expression by the buyer would be identified and the salesman's verbal behavior scored on whether or not it responds directly to the rule represented by the buyer's most recent verbal act--either by providing information congruent with the rule or by attempting to modify the rule itself by questioning the buyer in a way that encourages him to re-examine his assumptions while providing him with additional concepts. In other words, the protocol of a salesman-buyer transaction would be scored on the degree to which the verbal behavior of the buyer and salesman is linked by their common focus

on an implicit set of buying rules. In that sense, the measure would represent the salesman's responsiveness in offering the information which the buyer seeks and would therefore serve as an index of interaction synchronization as defined above. It is our hope that such a measure would predict sales success better than other currently available techniques.


Gerald Albaum, "Exploring Interaction in a Marketing Situation," Journal of Marketing Research, 4 (May 1967), 168-72.

Robert F. Bales, Interaction Process Analysis (Cambridge, Mass.: Addison-Wesley, 1950).

Timothy C. Brock, "Communicator-Recipient Similarity and Decision Change," Journal of Personality and Social Psychology, 1 (June 1965), 650-54.

Noel Capon, "Persuasive Effects of Sales Messages Developed from Interaction Process Analysis," Journal of Applied Psychology, 60 (1975), 238-44.

Noel Capon, Morris B. Holbrook, and James Hulbert, "Industrial Purchasing Behavior: A Reappraisal," Journal of Business Administration, 4 (Fall 1972), 69-77.

Noel Capon, Morris Holbrook, and James Hulbert, "The Selling Process: A Review of Research," unpublished paper, Columbia University, undated.

Eliot D. Chapple and Gordon Donald, "An Evaluation of Department Store Sales People by the Interaction Chronograph,'' Journal of Marketing, 12 (October 1947), 173-85.

James C. Cotham III, "Selecting Salesmen: Approaches and Problems," MSU Business Topics, 18 (Winter 1970), 64-72.

Harry L. Davis and Alvin J. Silk, "Interaction and Influence Processes in Personal Selling," Sloan Management Review, 13 (Winter 1972), 54-76.

Amitai Etzioni, Modern Organizations (Englewood Cliffs, N.J.: Prentice-Hall, 1964).

Franklin B. Evans, "Selling as a Dyadic Relationship: A New Approach," American Behavioral Scientist, 6 (May 1963), 76-9.

John U. Farley and Robert L. Swinth, "Effects of Choice and Sales Message on Customer-Salesman Interaction," Journal of Applied Psychology, 51 (April 1967), 107-10.

M. S. Gadel, "Concentration by Salesmen on Congenial Prospects," Journal of Marketing, 28 (April 1964), 64-6.

Ernest Gellner, "Concepts and Society," in Bryan R. Wilson, ed., Rationality (New York: Harper Torchbook, 1971).

Marvin A. Jolson, "The Underestimated Potential of the Canned Sales Presentation," Journal of Marketing, 39 (January 1975), 75-8.

Herbert C. Kelman, "Processes of Opinion Change," Public Opinion Quarterly, 25 (Spring 1961), 57-78.

Theodore Levitt, Industrial Purchasing Behavior (Boston: Harvard University Graduate School of Business Administration, 1965).

Theodore Levitt, "Industrial Purchasing Behavior: A Bayesian Reanalysis," Journal of Business Administration, 4 (Fall 1972), 79-8I.

Alasdair MacIntyre, "The Idea of a Social Science," in Bryan R. Wilson, ed., Rationality (New York: Harper Torchbook, 1971).

H. Lee Mathews, David T. Wilson, and John F. Moncky, Jr., "Bargaining Behavior in a Buyer-Seller Dyad," Journal of Marketing Research, 9 (February 1972), 103-5.

David Mayer and H. M. Greenberg, "What Makes a Good Salesman," Harvard Business Review, 42 (July-August 1964), 119-25.

J. T. Norman, "Now They Stand Up Under Pressure," Sales Management, 43 (August 1, 1954), 36-7, 84-5.

Richard W. Olshavsky, "Customer-Salesman Interaction in Appliance Retailing," Journal of Marketing Research, 10 (May 1973), 208-12.

John O'Shaughnessy, "Selling As an Interpersonal Process," Journal of Retailing, 47 (Winter 1971-1972), 32-46.

John O=Shaughnessy, "The Influence of Personal Communication on a Buyer's Choice Criteria in the Salesman/Buyer Dyad," in Jagdish N. Sheth, ed., Models of Buyer Behavior (New York: Harper & Row, 1974).

R. Wayne Pace, "Oral Communication and Sales Effectiveness," Journal of Applied Psychology, 46 (October 1962), 321-4.

Peter W. Pasold, "The Effectiveness of Various Modes of Sales Behavior in Different Markets," Journal of Marketing Research, 12 (May !975), 171-6.

Allan L. Pennington, "Customer-Salesman Bargaining Behavior in Retail Transactions," Journal of Marketing Research, 5 (August 1968), 255-62.

R. S. Peters, The Concept of Motivation (London: Routledge and Kegan Paul, 1958).

Richard C. Reizenstein, "A Dissonance Approach to Measuring the Effectiveness of Two Personal Selling Techniques Through Decision Reversal," Proceedings, American Marketing Association, Fall Conference, 1971, 176-80.

Thomas S. Robertson and Richard B. Chase, "The Sales Process: An Open Systems Approach," MSU Business Topics, 16 (Autumn 1968), 45-52.

Henry L. Tosi, "The Effects of Expectation Levels and Role Consensus on the Buyer-Seller Dyad," Journal of Business, 29 (October 1966), 516-29.

Ronald P. Willett and Allan Pennington, "Customer and Salesman: The Anatomy of Choice in a Retail Setting," Proceedings, American Marketing Association, Fall Conference, 1966, 598-616.

P. C. Winch, The Idea of a Social Science (London: Routledge and Kegan Paul, 1958).

Arch G. Woodside and J. William Davenport, Jr., "The Effect of Salesman Similarity and Expertise on Consumer Purchasing Behavior," Journal of Marketing Research, 11 (May 1974), 198-202.