THE BEHAVIORAL FACTORS AFFECTING THE FLOW OF INFORMATION IN THE BUYER - SELLER DYAD FOR THE DEGREE OF PH. D. MICHIGAN STATE UNIVERSITY ROBERT WILLIAM SWEITZER LI BRAR Y Michigan State Universi Turf-"‘1 This is to certify that the thesis entitled The Behavioral Factors Affecting the Flow of Information in the Buyer--Seller Dyad presented by Robert William Sweitzer has been accepted towards fulfillment of the requirements for Bh.D. .degree in W Major professor Joseph W. Thompson Dated-016‘ .‘ (71% O~7 639 d, ABSTRACT TUE BEHAVIORAL FACTORS AFFECTING THE FLOW OF INFORMATION IN THE BUYERv-SELLER DYAD BY Robert William Sweitzer A model of buyer-~seller interpersonal communications is develOped. The central construct in the model is the buyer's filtering process. Incoming information is weighted according to its importance and is "filtered" in inverse proportion to its weight. Information from a salesman that exceeds the buyer's importance criterion receives the buyer's attention; otherwise the information is attenuated. The buyer's beliefs about the importance of the salesman’s information are postulated to be a function of the buyer's beliefs about: (a) the "source value" of the salesman, and (b) the relevance of the salesman's message. Three constructs are Operationally defined as measures of the buyer's beliefs about the salesman's source value: empathy, credibility, and preparation. Using a series of likert-type scales two dimensions of the buyer's beliefs about the salesman's empathy are measured: (1) role empathy, the salesman's understanding of the job of the buyer, and (2) task empathy, the salesman's understanding of the specific purchase decision. The buyer's beliefs on three dimensions of the salesman's credibility are measured in a similar manner: (1) expertness, tOpic-bound qualification, (2) trustworthiness, lack of manipulative Robert William Sweitzer intent, and (3) dynamism, a combination of potency and activity. The buyer's belief's about the salesman's preparation and planning prior to the sales call are also measured. It is hypothesized that salesmen believed to possess the above traits are associated with transmissions of relevant messages. The relevance of the salesman's message is operationally defined as the correspondence between the content of the salesman's message and the buyer's purchasing criteria. The relevance of the message is measured by a content analysis of the message transmitted by each salesman in the course of an actual sales call. The buyer's beliefs about the salesman and the relevance of the salesman's message are hypothesized to be associated with the buyer's overall evaluation of the salesman. The buyer's evaluation of the salesman is hypothesized to be associated with the buyer's intention to purchase the salesman's product offering. The data sample was limited to sales calls dealing with "new" products to assure that little or no prior information about the product had been exchanged. Based on thirty-one interviews between buyers and salesmen in the food and dry goods industries, the following results were obtained: 1. The buyer's beliefs about the salesman's: (a) role and task empathy, (b) expertness and nondynamism, and. (c) preparation, show significant, positive associations with the relevance of the salesman's message. 2. The buyer's beliefs about the salesman's: (a) task empathy, (b) expertness and nondynamism, (c) preparation, and (d) the relevance of the salesman's message are significantly, positively associated with the buyer's evaluation of the salesman. Robert William Sweitzer 3. The buyer's evaluation of the salesman is significantly, positively associated with the buyer's intention to purchase the salesman's product offering. 4. The association of the buyer's beliefs about the salesman's task empathy, expertness, nondynamism, and preparation, and the buyer's evaluation of the salesman is not only due to the effects of the relevant message content as an intervening variable. Although the buyer's beliefs about the salesman and the relevance of the salesman's message are significantly associated, the salesman's task empathy, expertness, nondynamism, and preparation still show significant association with the buyer's evaluation of the salesman when the effects of the relevance of the salesman's message are controlled. THE BEHAVIORAL FACTORS AFFECTING THE FLOW OF INFORMATION IN THE BUYER--SELLER DYAD BY Robert William Sweitzer A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Marketing and Transportation Administration 1974 rf’ 0 @ ACKNOWLEDGEMENTS My most sincere and hearty thanks go to Professor Joseph Thompson for creating the opportunity for me to do this research. I am fortunate to have received both his professional and his personal inputs; he gave very generously of both his talents and his time. He contributed to my growth as a would—be scholar and as a human being. For both I am grateful. I would like to thank Professor Gilbert Harrell for his personal support and scholarly guidance. He impressed me with his disciplined and rigorous approach to the research. I trust that this study reflects that impression to some extent. I appreciate his frankness and his friendship. My thanks go to Professor John Allen for his insights into the food industry and his "real world" assessments of this study. A special acknowledgement goes to Mr. Harvey Weisberg, Mr. Dutch Leonard and the purchasing staff at Chatham. They were most generous with their time, their personal insights into buyer--seller relationships, and the data on which the results are based. A special thanks goes to Ms. Krista Lee for editing my stilted prose and making it more readable, and to Ms. Joan McCurdy for typing and retyping and retyping the manuscript. Finally I would like to thank those special people who provided the human inputs, my parents and my then friend and lover and new wife, friend and lover, Terry. They softened the frustration and heightened the elations. ii While many contributed in many ways to improve the quality of this work, the responsibility for the shortcomings in the pages that follow is solely mine. Montreal, Quebec Robert W. Sweitzer February , 1974 TABLE OF CONTENTS ACKNOWLEDGEMENTS . 1. INTRODUCTION . . . 2. A THEORY OF BUYER--SELLER INTERPERSONAL COMMUNICATIONS . The Process of Attention Conceptual Conflict and Attention The Selection of Incoming Information The Buyer's Filtering Process Purchasing Criteria Perceived Traits of the Salesman Evaluation of the Salesman Perceived Marketing Inputs Intention to Purchase Summary of the Model 3. SURVEY OF THE LITERATURE . Theories of Attention Filter Theory Input Selection Theory Weighting of Importance Theory Analysis-by-Synthesis Theories of Cognitive Consistency Symbolic Responses Dissonance Theory Balance Theory Theory of Conceptual Conflict Source Credibility Empathy The Predictive Approach The Situational Approach The Rating Approach The Perceived Approach Dyadic Buyer--Seller Communications Models 4. HYPOTHESES . . . . , 13 15 27 29 3O 32 35 35 36 38 53 SS 55 S6 10 ll 19 20 21 22 . 24 24 34 43 50 57 . 60 5. METHODOLOGY . . Research Instruments D ata Test 6. RESULTS CD‘OWVGU'l-bMNI-i umm Source Credibility Empathy Preparation and Planning Message Content Evaluation Intention to Purchase Sample Statistics Zero—order Correlation Regression Analysis Partial Correlation Multiple Regression Adjustment of Correlation Coefficients Testing Procedure . O I O O O O O O O O Empathy-Message Source Credibility-Message Preparation-Message Empathy-Evaluation Source Credibility-Evaluation Preparation-Evaluation Message-Evaluation Evaluation-Intention Message-Intervening Variable any 7. CONCLUSIONS . Empathy Credibility Preparation Message Content Intention to Purchase Summary Implications 8. APPENDICES. Questionnaires Data Three Buyer--Seller Dialogues Evaluation-Poor Evaluation-Good Evaluation-Excellent 9. REFERENCES. 68 70 73 73 77 78 80 80 81 84 85 86 89 91 95 96 99 103 104 105 106 108 110 111 113 113 115 116 117 119 128 134 135 143 149 .68 68 78 79 .88 . 110 .119 .159 CHAPTER ONE INTRODUCTION This research seeks to answer the following question: Out of all the information that is transmitted by the salesman and that consequently impinges upon the buyer, why is it that some of the infbrmation passes through the buyer's filtering process to become available for decision making and some does not? A substantial amount of research already validates the existence of such a filtering process (Moray, 1969). The purpose here is to describe and explain the effects of the behavioral factors that influence the filtering process in buyer-seller communication situations. An important aspect of buyer-seller communications is the distortion that infermation undergoes between the input of the physical phenomena upon the senses and its subsequent utilization in decision making. Students of human nature have been concerned with this problem as far back as Heraclitus (Nahm, 1964) who refers to the eyes and ears as "evil witnesses." Plato (Lodge, 1956) also comments on the interaction between the processes of nature, the processes of human sense organs and the passage of sensations to the central brain. He notes that sensation is an obstacle to cognition. When Aristotle (Ross, 1949) describes the functions of the "soul", he makes a biological-- psychological division and suggests a relationship between sensations and the reasoning or intellectual parts. Aristotle describes perception as the reorganization of stimuli and the subsequent fermation of images upon which reason is predicated. He notes that the fermation of images does not correspond on a one to one basis to the sensation itself. Aristotle's observations can be explained by a cognitive psychological framework that suggests two basic elements of interpersonal communication: (1) the accumulation and assimilation of incoming information and (2) the subsequent processing of this information leading to overt behavior (Neisser, 1967). This research focuses on the first element. The theoretical model developed for this research draws upon the body of knowledge already in existence regarding interpersonal communications. General communication concepts from research in psychology, sociology and communication arts are considered as well as concepts from research in personal selling and industrial and consumer buyer behavior. Although this research focuses on an industrial buying-selling situation there is justification for drawing on information processing models and consumer models as well. While differences do exist between the two, consumer behavior models nevertheless tend to provide basic frameworks for the industrial buyer models (Webster and Wind, 1972). The major differences between the two models involve the multiple organizational inputs on the industrial buyer not found in the consumer situation and as such have little effect on this research which is concerned with the process of information accumulation and assimilation prior to the actual decision process. In other words, the differences between the industrial and consumer models at the stage with which this research is concerned are slight. At a conceptual level the constructs used to explain industrial and consumer behavior may be identical (Howard and Sheth, 1969). Both industrial buyers and household consumers utilize similar information processing methods. This rationale explains the inclusion of information processing variables from research models in fields other than marketing. The behavioral phenomena of information exchange that is exhibited in buyer-seller interactions is, at a conceptual level, an experience that is common to virtually all the behavioral sciences. Howard (1971) concludes that the buyer takes in that information he expects will be usefu1 in making a purchase decision. In their model of buyer behavior, Howard and Sheth (1969) postulate a concept, "perceptual bias", which functions on the basis of cognitive consistency theory to affect incoming information in a manner consistent with the buyer's frame of reference. Andreasen's (1965) theory of brand choice includes a filtration process where incoming information is selected on the basis of the buyer's attitude structure towards both the message and the source of the message. Howard (1971), however, suggests that the buyer engages in a systematic process of selecting infermation based on a set of rules by which the buyer reduces the incoming information to a manageable quantity that still meets his needs. In a survey conducted by the Supermarket Institute (SMI, 1971) problems most frequently mentioned by grocery buyers were: (1) manufactuers' representatives being unable to supply marketing information and (2) the manufacturer's inability to understand local market situations. Problems most frequently mentioned by sellers were: (1) difficulty in understanding trade demands and requirements, (2) lack of time with buyer and (3) lack of buyer attention. Sellers also complained about inadequate feedback on new product acceptance standards and a lack of feedback on decision criteria used by buying committees. Buyers expressed confusion and frustration growing from the fact that supply representatives were not adequately informed and had a lack of available information. These problems mentioned by buyers and sellers fall within the framework of the behavioral factors affecting the flow of information used in this research. A sample of grocery buyers, salesmen, and their messages is examined to explain why such problems may exist. The theoretical model providing the framework for this research is presented in Chapter TWO. While the sources drawn upon in developing the model are briefly noted there the purposes of Chapter Two is to describe the interrelationships of the constructs. In Chapter Three, The Survey of the Literature, the underpinnings of the constructs are considered more carefully. Alternative approaches to each construct are considered and the rationale is explained for the particular theoretical approach selected for use in the model. The hypotheses tested in this study are presented in Chapter Four and fifteen relationships are set forth based on the theoretical model. These hypotheses are presented in the alternative ferms. The same format is followed in reporting the Results in Chapter Six. Underlying the test of each hypothesis is a null hypothesis of no difference. In each test the calculated level of Type I error (Si) is reported. The reader may develop his own rejection region and thus accept or reject the underlying null hypotheses. However, implicit in the reporting of the results is an .01 criterion. The methodology used to test the research hypotheses is presented in Chapter Five. The hypothetical constructs in the theoretical model are operationalized and examples of the questionnaire items presented. The test statistics used to measure the significance of the association between the independent and dependent variables are also presented. The nature of the sample is discussed. Finally a methodological procedure for testing the causal relationships in the theoretical model is considered. The findings are discussed in Chapter Seven and several conclusions are set forth. The implications of this research are also considered. The research instruments and data are presented in Appendices following Chapter Seven. CHAPTER TWO A THEORY OF BUYER--SELLER INTERPERSONAL COMMUNICATIONS The personal selling function is viewed as essentially a process of interpersonal communications and it follows that the flow of information in the buyer-seller dyad would best be explained by an interpersonal communication model. The theoretical model used in this research contains four basic patterns of relationships which describe and explain the process by which information flows in the buyer—seller dyad may lead to a decision to purchase. These feur sets of interactions are initiated by the salesman's message impinging upon the buyer's filtering mechaniSm. The principal interactions resulting from the input of information consist of: (l) the use of purchasing criteria in the filtering process, (2) the use of the perceived traits of the salesman in the filtering process, (3) the evaluation of the salesman as a message source, and (4) the effects of the salesman and his message upon the buyer's intention to purchase. The model is illustrated in Diagram 1. THE PROCESS OF ATTENTION The central construct in the model, the buyer's filtering process, is predicated on Broadbent's (1958) conceptualization of the process of attention. Not all incoming information impinging upon a buyer is allowed to pass through the filter to become part of the memory and remain available for further use in purchasing decision—making. DIAGRAM 2-1 BUYER--SELLER INTERPERSONAL COMMUNICATION MODEL PURCHASING CRITERIA SALESMAN agggfiéfi BUYER'S LEVALUATION AINTENTION (MESSAGE r:> FILTERING‘ OF THE ‘ TO SOURCE) PROCESS SALESMAN' PURCHASE l ”I 1 m PERCEIVED' PERCEIVED TRAITS OF RELEVANT SALESMAN MESSAGE PERCEIVED ‘ MARKETING [ INPUTS FEEDBACK Broadbent (1958) proposed a filtering mechanism as a hypothetical construct to explain the limitations on the amount and type of information that can be effectively dealt with by an individual. The model also uses the Deutsch and Deutsch (1963) theory to describe the process by which incoming information is limited by the buyer. According to this theory, the buyer weighs incoming information according to its importance. Information that exceeds the individual's importance criterion receives the individual's attention. Information less important than the selection criterion is attenuated. The weighting of importance mechanism processes all incoming information before it becomes available for storage in memory and/or use in purchase decision-making. Neisser (1967) describes a preliminary process which, "picks up portions of the information input that are worth synthesizing and arrives at preliminary identification of these portions". He states that contextual cues and expectations may play a role in the preattentive process. The portion of the phrase, "input information that is worth synthesizing" suggests the weighting of importance construct described by Deutsch and Deutsch (1963). Of the alternative approaches to the process of attention the weighting of importance construct is also consistent with several models of buyer behavior. Howard (1971) concluded that the buyer takes in only that information he expects will be usefu1 in making a purchase decision. Howard and Sheth (1969) suggest perceptual bias operates to limit incoming information, based on cognitive consistency theory, in a manner consistent with the buyer's frame of reference. Thehypothesized input affecting perceptual bias is the buyer's decision mediators and incoming information is weighted in terms of its correspondence to the criteria used to make a purchase decision. Andreasen (1965) has proposed a filtering mechanism which Operates as a source of information. This filtering mechanism is consistent with Allport's (1965) reported effects of affective value and motivational state upon the way in which stimuli are perceived. The filtering process in the model utilizes a weighting of importance construct (Deutsch and Deutsch, 1963) where the selection criteria are a function of both the buyer's puerchasing criteria and the buyer's affective state towards the salesman. CONCEPTUAL CONFLICT AND ATTENTION The underlying motive which guides the buyer's filtering process is the buyer's concern for information which can be used in making purchase decisions. This motive is based on Berlyne's (1960) theory of conceptual conflict. In moving toward a purchase decision the buyer may encounter doubts, mutually exclusive claims, contradictions, incongruities, confusing communications and irrelevant information. Berlyne (1960) indicates that one way of alleviating these types of conceptual Conflict is the acquisition of knowledge through additional information. The drive to reduce conceptual conflict motivates the buyer to pay attention to certain incoming information. Not all information would be equally valuable in alleviating the conceptual conflict surrounding a purchase decision. Therefore not all incoming information would receive the same degree of the buyer's attention. -10- THE SELECTION OF INCOMING INFORMATION For information to be useful to the buyer in the purchase decision process it not only must correspond to the criteria used to evaluate alternative products but must also have "source value". The theoretical model (Diagram 2-1) sets forth two major selection criteria for the buyer's attentive processes. To pass through the buyer's filter, infermation must have two qualities: (1) the incoming information must be perceived as being relevant (have a monotonic correspondence with the buyer's purchasing criteria) and (2) incoming information must be perceived as having originated from a "valuable" source. Generally both these conditions must be met before information will pass through the buyer's filtering process. A message could be seen as relevant but because the buyer doubts its source value it would not serve as a purchasing input. Similarly, information which is irrelevant to the buyer's purchasing criteria but originates from a source perceived as valuable would have diminished utility in decision-making. The case where both qualities are lacking clearly has no utility to the buyer. The beliefs held by the buyer about the relevance of incoming information and the value of the source of information exist in varying degrees. Neither source value nor relevance are seen as dichotomous variables. They are best represented by a continuum of beliefs held by the buyer about the nature of the incoming message and its source. Therefore the filtering process would take place in varying degrees corresponding to the beliefs that serve as inputs to the filtering -11- process. The buyer's filtering process is predicted on 05good's (1957) concept of mediating responses. The patterns of filtering incoming information and the resulting responses are all internal and not directly observable by overt behavior. The mediating responses that take place in the filtering process, while not directly observable, may be inferred by the affective state of the buyer in terms of his perceptions of certain traits of the salesman including his overall evaluation of the salesman as a message source. The buyer's perception of the seller has an overall effect on the quantity of information that passes through the filter. While the effect of the purchasing criteria would be specific in terms of the one aspect of the incoming message to which each particular criterion relates, the perceived traits of the salesman affect the general level of filtering. The inputs to the buyer's filtering process may be specified more exactly. THE BUYER'S FILTERING PROCESS The buyer's weighting of importance function analyzes both the salesman and his message to determine which portions of the incoming information will pass through the filter. The importance weight depends upon the degree to which the information allows the buyer to make the "best" purchase decisions. The buyer's purchasing criteria serve as decision mediators in making his purchase decisions. If a message is to have utility to the buyer the information must relate to the purchase decision mediators. Unrelated infermation would be unimportant -12.. and therefore, attenuated by the filtering process. Thus the buyer's purchase criteria serve as inputs to the weighting of importance function. The buyer's beliefs about the value of the message source also affect the weighting of importance function. The salesman's empathy or "understanding of the internal frame of reference of the buyer and communication of that understanding" (Rogers, 1959) is one element of his value as a message source. The empathy of the salesman determines, in part, his ability to generate a relevant message. Without some ability to "put himself in the other fellow's place" (Allport, 1954) it would be difficult for the salesman to determine the purchasing criteria that will be used to evaluate his product offering. Since it is the buyer's beliefs about the salesman's empathy that effects filtering, the salesman must not only understand but must also communicate that understanding. The multidimensional concept of source credibility (Hovland, Janis and Kelly, 1953) affects the buyer's perceptions of the value of p the salesman as a message source. By being highly regarded or highly credible, a salesman can produce a more highly regarded message which has greater utility to the purchasing agent. Another aspect of the buyer's regard for the salesman is the buyer's perception of the amount of planning and preparation the salesman has done prior to the sales call (Thompson, 1966). The buyer's willingness to listen to the seller's message is affected by his perceptions of the seller's preparation and planning fer that particular sales call. Both the purchasing criteria and the perceived traits of the salesman affect the filtering of incoming information. -13- DIAGRAM 2-2 THE BUYER'S FILTERING PROCESS WEIGHTING OF IMPORTANCE FUNCTION COMPARISON SPECIFIC SALESMAN ACTUAL fl OF CRITERIA AFFECTIVE PERCEIVED (MESSAGE MESSAGE TO ACTUAL H STATES ’- RELEVANT SOURCE) MESSAGE TOWARDS MESSAGE A THE MESSAGE SOURCE Jr PURCHASING PERCEIVED CRITERIA TRAITS OF SALESMAN Purchasing Criteria The motive Of the buyer in seeking out relevant information is based upon his need to receive information on the topic at hand (Berlyne, 1960), infOrmation that enables the buyer to solve his current purchasing problems. The buyer filters incoming information depending upon how important he considers such information in terms of reaching a final purchase decision. Incoming information that does not exceed the importance threshold Of the filter is attenuated. The importance thresholds are determined by the weighting function where the weights are directly proportional to the purchasing criteria. InfOrmation that corresponds to an important purchase criterion is weighted heavily and is not filtered out. InfOrmation that corresponds to an unimportant purchase criterion receives a low weight and is attenuated -14- by the filter. As a result Of the weighting of importance filtering process there will be some discrepancies between the actual message transmitted by the salesman and the message received by the buyer. The utilization Of purchasing criteria to establish a level of importance of incoming information suggests that those aspects of the message which do not correspond to the purchasing criteria will be filtered out. Furthermore some criteria would likely be more important to the buyer than others and therefore, the filtering process takes place by degrees rather than on a dichotomous basis. A "perfect" sales message would be one whose actual content corresponded perfectly to the decision or purchasing criteria Of the buyer. By comparing correspondence of actual message to the so called "perfect message" as seen in the buyer's purchasing criteria one can infer what portion of the actual message was perceived as relevant. By considering the degree Of importance of the particular criteria one may also infer how relevant a particular portion of the message might be to the buyer. If all incoming information is filtered according to its impOrtance, then it is necessary to specify the criteria used to determine importance. In their model of buyer behavior, Howard and Sheth (1969) hypothesized a relationship between decision criteria and perceptual bias. In a similar manner Thompson (1973) discusses the concept Of relevancy where the relevance of incoming information is a function of its correspondence with the buyer's frame Of reference. The specific criteriaused for weighting the importance Of incoming information would be the purchasing criteria which are used to evaluate -15- the alternative product Offerings. In the buyer--seller interpersonal communication model the amount Of relevant message flowing out of the buyer's filtering mechanism is determined in part by the degree to which the actual message impinging upon that filtering mechanism corresponds with the purchasing criteria. However, the perceived traits Of the salesman also influence the filtering process. Perceived Traits of Salesman In his model of buyer behavior, Andreasen (1965) discusses the process Of filtration whereby information inputs are reduced by selection or distorted by individual idiosyncracies in organization, interpretation and judgment Of information. He suggests that the extent of filtration is a function Of pre-existing attitudes, citing a number of studies which show individuals tend to perceive information inputs in terms of their own attitudes. Cox (1962) notes that attitudes about the credibility and usefulness of personal advocate sources, such as salesmen, influence the objectively defined infOrmation input. Andreasen (1965) states that infOrmation about the source Of the message is a primary input into the filtration process of the buyer decision model. -l6- DIAGRAM 2-3 PERCEPTIONS OF SALESMAN'S TRAITS AND THE BUYER'S FILTERING PROCESS PERCEIVED TRAITS -—7 OF THE SALESMAN l. EMPATHY a. Role b. Task 2. SOURCE CREDIBILITY a. Expertness b. Trustworthiness BUYER'S c . Dynamism * FILTERI NG PROCESS 3. SALES CALL OBJECTIVE a. Preparation b. Planning Empathy. The coneptualization Of empathy utilized in the model is drawn from Rogers' (1957) theory which emphasizes the perceptions of the receiver Of the message. This approach stresses not only the salesman's understanding of the buyer's frame of reference but also the communication of that understanding. From Diagram 2-3 it can be seen that two elements of the buyer's frame Of reference are considered. First of all the salesman can display role empathy by communicating his understanding of the job Of the buyer. Perceived role empathy is the feeling of the buyer that the salesman knows what is involved in being a buyer and specifically that the salesman knows the operational mode of the buyer he is facing. -17- Secondly, the salesman can display task empathy. This variable is the salesman's communication of his understanding of the specific purchasing decision facing the buyer in evaluating the salesman's product Offering. It involves the buyer's perception of the salesman's understanding of the purchase criteria that will be used to make a decision regarding his product Offering. Source credibility. The theoretical approach to the source credibility variables lie in the BerlO et a1. (1966) extension of the studies of Hovland, Janis and Kelly (1953). The three factors of: (1) safety or trustworthiness, (2) qualification or topic-bound expertness, and (3) dynamism, all affect the buyer's acceptance of the salesman as a message source. As is the case for the sales call objective and empathy, credibility is also defined in terms of the perceptions of the buyer. Sales call Objective. In addition to the general communication variables of empathy and credibility, another variable specific to the buyer-—seller dyad is considered based on Thompson's (1973), concept of the "sales call Objective". This variable relates to the preparation and planning of the salesman prior to the actual interview with the buyer. From the perspective of the buyer's filtering mechanism the concern is with the buyer's perception of the salesman's degree Of preparation rather than some objective measure of the time spent planning the sales call. Due to the professional role Of the industrial buyer, perceptions of a well planned and prepared salesman have positive inputs on the filtering mechanism and lower the possibility of attenuation of the salesman's message. The presence of a sales call Objective as seen in the buyer's perceptions Of the salesman's -18.. preparation and planning prior to the sales interview is a variable specific to interpersonal communication in the buyer--seller dyad and thus included in the model. The buyer's beliefs about the traits of the salesman provide an input to the weighting of importance function and thus the amount of attention the buyer gives to the salesman. Empathetic, credible, well prepared salesmen are more likely to receive the buyer's attention and have their messages pass through the buyer's filtering process with little loss of information. Questioning the buyer about his perceptions of the salesman along these dimensions, provides an insight into his affective state regarding the salesman as a message source. Given the buyer's commitment to making the "best" purchases for his organization, the buyer is more predisposed towards message sources that are perceived as high in empathy, source credibility and preparation than towards a message source which does not display these characteristics. It has been noted by Andreasen (1965), that in the filtering mechanism, the buyer's attitudes towards the source and the buyer's attitude about incoming infOrmation are clearly interdependent. The buyer's feelings about the source of the message clearly influences his judgments of the relevance of the message. Likewise the message itself has some affect upon the buyer's perception of the value of the message source. Because Of this interrelationship, the "image" of the salesman and the message can be seen to merge in the mind of the buyer. In attempting to operationalize measures of the buyer's beliefs about the salesman and/or his message, this merging effect should be given -19- consideration. The evaluation Of the salesman and his message are not independent of each other. In fact it may be the case that in measuring one an indirect measure of the other is attained. If the buyer does indeed visualize the salesman as primarily a source of information then the salesman and his message will become a similar image in the buyer's mind. EVALUATION OF THE SALESMAN The general affective state of the buyer towards the salesman as a message source is a function Of the preSence of the attributes of the salesman and his message. These attributes are described as inputs to the buyer's filtering mechanism (Diagram 2-1). The message effects are based on the correspondence of the actual message to the buyer's purchasing criteria. The buyer's perceptions of the traits Of the salesman also affect the buyer's combined evaluation of the salesman and his message. The buyer's overall evaluation of the performance of the salesman is determined by the buyer's beliefs about the presence or absence of salient message and message source attributes. This is based on the theory (Fishbein, 1963) that an individual's attitude toward an Object is a function of his beliefs about the Object and the evaluative aspect of those beliefs. The overall evaluation Of the performance of the salesman as a message source is a function of two sets of interrelated, salient beliefs: (1) the traits of the salesman, and (2) the content Of the message. The buyer's evaluation is considered as a general predisposition towards the combined stimulus of salesman and message, -20_ an indicator of the buyer's general affective state towards the salesman's performance. Evaluation serves as a major dependent variable in the research. PERCEIVED MARKETING INPUTS It is certain that there will be other types of marketing information inputs which impinge upon the buyer and influence his intentions. Certain aspects of the product itself, the pricing strategy, promotional activities of the manufacturer and wholesalers other than the salesman himself and the distribution policies of the supplying firm all influence the buyer's intention to purchase. Since the ultimate concern is with explaining purchase behaviors, any model of the buyer--seller dyad which fails to consider such variables is certainly incomplete. While it is realized that the salesman is not the sole input or determinant of the buyer's intentions to purchase, this research is not concerned with explaining the effects of other marketing inputs. No attempt is made to control for the other marketing inputs except to acknowledge their existence. This research is concerned with the question of how much effect the salesman does have upon the buyer's behavioral intentions. Thus the other marketing inputs will be considered as a residual variable explaining that portion of the variation in the buyer's intentions that cannot be explained by the effects of the salesman and his message. _21_ INTENTION TO PURCHASE A precedent for using intention to purchase rather than overt purchase as a dependent variable is found in both Howard and Sheth (1969) and Andreasen (1968). There is a tighter linkage between attitudinal dimensions and intention because the relationship is not so confounded by the situational parameters that may affect overt purchase behavior. In addition to the variety of marketing inputs that may impinge upon the buyer prior to intention, there are inhibitors (Howard and Sheth, 1969) or constraints (Andreasen, 1968) that may impinge between intention and overt behavior. The inhibitors are situational factors such as high prices relative to competitors,offering a temporary ”deal" price, lack of availability of goods, time pressure or financial considerations which may disrupt the buyer's plan Of purchase. The constraints have been described as the obverse of Katona's (1960) "enabling conditions". Such constraints may force postponement or cancellation of a particular purchase intention. Inhibitors and constraints tend to confound further the relationship between predispositions and overt purchase behavior. The buyer's intention to purchase represents his current purchase plan. All other things being equal, it is expected that the buyer will follow through on his expressed intentions with corresponding overt purchase behavior. Using intention as a dependent variable eliminates the additional variation introduced by inhibitors and constraints. Where some amount of planning takes place prior to purchase the conceptualization Of intention is appropriate (Howard and Ostlund, 1973). Thus intention to purchase serves as a major dependent variable in this research. -22- SUMMARY OF THE MODEL The theoretical model that serves as a framework for this research is based on a synthesis of various empirical and theoretical studies. Constructs from various disciplines have been integrated to form a theoretical model that both describes and explains the effects of a number of salient variables upon the flow of information in the dyad. In the introductory chapter it was stated that the personal selling function was viewed as essentially a process Of interpersonal communications. It fellows that the flow Of infOrmation in the buyer-— seller dyad would best be explained by an interpersonal communication model. The model (Diagram 2-1) explains the process of receiving incoming information in terms of a filtering construct. The concept of filtering incoming information is a significant thrust Of studies in the selective aspect of attention (MOray, 1969). The motivation fer the process of attention is the conceptual conflict surrounding purchase decision making. The buyer is motivated to receive infOrmation that reduces conceptual conflict (Berlyne, 1960). Several aspects of the message source and message that may enhance the buyer's expectation Of reducing conceptual conflict are presented as inputs to the filtering process. The expectation of reducing conceptual conflict enhances the attention paid to the message. The survey of the literature documenting the constructs utilized in the model is presented in Chapter Three. Since there are alternative approaches to all the constructs utilized in the model, each construct is -23- considered from a number of viewpoints. The rationale for the conceptualizations found in the model is also presented in Chapter Three. CHAPTER THREE SURVEY OF THE LITERATURE This survey of the literature examines selling as a process of 51 :14 interpersonal communication in light of theoretical and empirical findings from many disciplines. Interpersonal communication provides the rationale on which selection is based. An attempt has been made to identify variables that describe and explain a variety of dyadic communication situations with particular emphasis on buyer—seller dyads. The survey first considers several approaches to the concept of attention which provides alternative explanations of the process of limiting incoming information. The motivation for the process of attention is considered under the heading of conceptual conflict. I Theories of cognitive consistency provide an insight into why an individual pays attention to some information and not to others. Finally, dyadic communications are considered in the context of attention and the motivation for attention. Attributes Of message sources that may enhance the likelihood of receiving attention are discussed. THEORIES OF ATTENTION The concept of attention refers to the limitations on the amount and type of information with which an individual can deal effectively. Attention has varied meanings. Moray (1969) suggests at least seven different interpretations of the concept: mental concentration, -25- vigilance, selective attention, search, activation, set, and analysis— by—synthesis. In reviewing the literature and the various definitions Posner and Boies (1971) suggest that there are three major topics under which I various studies of attention might be grouped: l) alertness 2) ability to select information 3) limited processing capacity Alterness refers to the individual's ability to develop and maintain an optimum level of sensitivity to external stimulation after he has received some cue to prepare for incoming information, (Mackworth, 1970). Attentiveness and vigilence studies have tried to answer the basic question Of how much attention is given to a particular stimuli. Berlyne (1970) has referred to this as the intensive aspect of attention. Another intensive aspect of attention is the degree of mental concentration involved in the intake of information. Wittenborn (1943) examined the individual's ability to concentrate on some particular task such as mental arithmetic or trying to seclude an incoming stimuli which might interfere with the performance of the specified task. The ability to select information from a source of one kind or another is the major category in the existing experimental work on attention. Moray (1969) suggests that a listener is able to exercise considerable voluntary control over what he will hear on the basis of the voluntary use of specific criteria. Broadbent (1958) has suggested that some aspects of selection are performed by filtering mechanisms which block out or attenuate input. Treisman (1960) has suggested a two-stage process in which signals are first sorted along various preselected dimensions and are then differentiated in terms of threshold settings of -26— an internal recognition network. Deutsch and Deutsch (1963) have proposed an explanation of selective attention based on a weighting of importance mechanism. Both the intensive and selective aspects of attention are considered; how much attention buyer and seller pay to each other in the process of exchanging information and how attention is distributed among the various stimuli included in this process. Particular emphasis is placed on the latter. Moray (1969) suggests that the relation between selective listening and arousal is that arousal acts as a parameter which alters the overall efficiency of selection and rejection. The nature of the buyer or purchasing agent's job dictates some overall level of arousal. On this basis it appears that the pertinent aspect of attention is how a buyer or purchasing agent selects information from the many information inputs he receives in a given period of time. The third approach to attention relates to the concept of limited centural processing capacity. It has been widely Observed that individuals have difficulty in handling two tasks simultaneously even when they wish to do so. This phenomenon has been explained in terms of a single channel limitation (Welford, 1967). Any two operations requiring the use of the limited capacity mechanism will interfere with each other. In studying the buyer-—seller dyad, the question of simultaneous processing of stimuli from multiple sources seems to have little application. The unit of investigation is one buyer interacting with one seller. If more than one seller is present during the transaction rarely do both salesmen attempt to speak at the same time. Experiments used to study each of the three components of attention have been different and there has been little attempt to develop a -27- framework which allows the components to be interrelated (Posner and Boies, 1971). In buyer-~seller communications the concept of attention will be used in the selective context. The level of arousal or alertness will be considered but the fOcus in the process of attention is on the selective aspect. Filter Theory One Of the first models for attention was Broadbent's (1958) filter theory. According to the filter theory, information enters the system through a number of sensory channels which pass through a filter construct. If these inputs arrive in a parallel form the filter performs a transformation so that infOrmation is passed on to the subsequent mechanisms in serial form. Information theory suggests a limited capacity channel the capacity of which is much smaller than the total capacity of the initial parallel inputs. Broadbent suggests a short term memory at the end of the parallel inputs followed by the filter. The function of the filter is to select one of the inputs and give information access to the limited capacity channel. Input information not selected will hold in short-term memory storage for a period of only a very few seconds, and thus one can expect a loss of information between the total inputs and the latter information flows in the system. The hypothetical construct of the filter accounts for this loss of information. The transfer of information through the filter from the parallel to the serial stage involves an encoding process based on the filtering criteria. Overloading of the limited capacity channel is prevented by serializing the input to it. Broadbent does not specify _23_ Figure 3—1 Broadbent's Filter Theory Short Limited Motor -———e>m _ . i ._>. Term , Filter Capac1ty .——_§. Response . .___e, Memory - Channel System the explicit pattern of analysis of the filter; however it is clear that the filtering process provides a method of distinguishing between relevant and irrelevant stimuli (Moray, 1969). The limited capacity channel provides hypothesized linkages to and from the long-term store of learned associations or "long-term memory” as well as to the motor response system. The theory also has feedback provision for signals which have passed through the limited capacity channel to enter short term memory again. A central question addressed by Broadbent's filter theory, involves the conditions under which the filter will switch channels. Broadbent (Broadbent and Gregory, 1964) considers the channels which the filter might select and suggests two kinds of selective attention in line with his filter theory; 1) stimulus set and 2) response set. In response set irrelevant stimuli are not appropriate to any one of the allowable responses. With stimulus set irrelevant stimuli are appropriate but are distinguished by some other feature which in itself has no connection with the class of allowed responses (Broadbent, 1970). -29_ Input Selection Theory The general formulation of the Treisman model is based on Broadbent's filter theory but makes much more explicit the selection of rules governing the actions of the filter, and the problem of identifying particular signals. According to Treisman information flows to the individual through a number of parallel channels. The term filter describes the process whereby signals are selected along various dimensions based on particular criteria. The filter acts to attenuate the signal strength of the output based on this analysis. Thus according to Treisman's theory the filtering process weakens the signal output of all but that particular message which scores highly on the particular filter criteria. Figure 3-2 - TreiSman's Input Selection Theory _ 5 AI Selected Meqqnop r -———-———§;~ Pattern Filter __ ;. . . _ _____ Attenuatgd_ _ _ _ _ .- _ _ - -> Recognltion ————>- ""‘7;’L_ _ ~.._M§§sage§ _______ . _._€>_ Network Weakened messages and the one unweakened message reach a ”pattern recognizer" where conscious perception of the message takes place. This conscious perception is the result of the functioning of various "dictionary units”. The dictionary units have different and variable thresholds. Variations in thresholds are the result of criteria of the receiver. Unimportant signals will not be able to -30- trigger the appropriate dictionary unit since their signal strength has been attenuated by the filter, but unattenuated signals will be able to do so and hence the message will be heard. The occurance of a particular signal will, if it triggers a dictionary unit, lower the threshold for other signals. Treisman's theory states that attention is a two-stage process: (1) filtering on the basis of channel characteristics, and (2) analysis by the threshold settings Of the dictionary units. Weighting of Importance Theory Deutsch and Deutsch (1963) have proposed an importance-weighting construct to explain selective attention. They criticize the Treisman model as being redundant and suggest that by suitably altering the properties of the dictionary, the lower level filter is made unnecessary. As a flow diagram, this theory might be represented by the same diagram used for Treisman's model except that the filter would be omitted and the input lines run without interruption directly into the recognition network. The activity of the network is described in Figure 3—3. The output of the recognition network is determined by the importance of the mesSage to the individual, where the more important the message is, the more likely it is to pass out of the recognition network. The importance weighting is a function of past experience and expected outcomes. All incoming information is weighted as to its importance and the output is inspected. The most important information is allowed to be transmitted to further stages of the mechanism, and thus the most important stimulus captures the attention. -31- Figure 3-3 The Deutsch and Deutsch Importance Weighting Theory A F‘ , Selection P—-——1-I——————-—------—- . . Criterion o o a p F“ H o o. E H 'a_ b c d e9 Messages If another information input is more important the first will be displaced. The most important signal both captures the system and sets the criterion for any other signal to surpass. As the level of arousal increases, more signals will have access to the response side Of the system providing that they are the most important at the moment they occur. Changes in importance may be brought about by long term differences in learning. Pattern recognition precedes conscious perception and is not identical with it. Rather, perception is a response to the output of the pattern recognition network. The Deutsch and Deutsch model selects incoming signals in the same sense which Treisman's model does, since all inputs are analyzed with filtering taking place subsequent to the analysis. -32- Analysis—by-Synthesis Neisser (1967) has dismissed much of the experimental work covered in the previous theories as "pre-attentional selection". Neisser (1967) postulates two stages of attention: (1) the preattentive or preliminary analysis of incoming information by filtering or attenuation of input, and (2) analysis-by-synthesis. The attention process acts as if making the "best bet" over the interpretation of certain kinds Of incoming information. The process that selects a message from incoming information includes transitory information inventory. Neisser calls this an "echoic memory" which occurs prior to the post-perceptual short-term memory store. He also argues for the brain acting as a general purpose processor that redirects search processes and allocates different input, output and transformation tasks. It is during the process of analysis-by-synthesis that interpretation of filtered information takes place. Neisser (1967) proposes that filtered information goes through a second process which he calls analysis—by-synthesis. This process consists of three stages: 1) an hypothesis about the original message 2) application of rules to determine what the information input would be like if the hypothesis were true and 3) checking to see if the infOrmation input is really comparable to the original hypothesis. In the buyer—-seller interpersonal communication model Neisser's analysis- by-synthesis takes place on that relevant information that has passed through the buyer's filtering process. This analysis provides a transformation of the relevant information so that it is available for storage in memory and/or use in making actual purchase decisions. Howard and Sheth (1969) draw upon Neisser's (1967) formulation of _33- Figure 3—4 Neisser's Analysis-by—Synthesis Theory Preattentive Processes HEESZee Analysis Passive Echoic g by L—9 Memory Synthesis Messages analysis-by-synthesis to discuss the transfer of information from their perceptual sub-system to their learning sub-system. Howard and Sheth infer that only screened information is available for use by learning variables in making the purchase decision. In this respect Howard and Sheth use Egeth's (1967) terminology of "stimulus-as-coded" to represent information that has been screened and passed through Neisser's analysis—by-synthesis process. I In his review of the literature Egeth (1967) discusses the‘ reception of stimuli in terms of the "stimulus-as—coded”. He suggests that the understanding of attention lies in the discovery of the coding and decoding rules which are applied from moment to moment by the individual who is being confronted with the information. This parallels Neisser's approach. This formulation of attention is the application of such coding rules arranged in an hierarchy through which data is transferred and transformed until finally recognition and responses are obtained. The application of information flow models seems dependent upon a deeper -34_ understanding of the mechanisms which control the flow of information through the individual's system. Moray (1969) suggests that considering the communication channel to be an entirely passive transmission line is inadequate. Egeth (1967) suggests that the task of future research is to specify the coding responses which enable subjects to respond selectively to portions of conflicting stimuli. The various approaches to selection attention explain how an individual limits incoming information but not why. In their comprehensive theory of buyer behavior Howard and Sheth (1969) draw upon Berlyne's theory of conceptual conflict. Berlyne (1960) suggests that the idea of conceptual conflict may have implications for attention. Selective attention has been regarded as a phenomenon made up of two parts: (1) the intensification of the process on which attention is being concentrated, and (2) the holding in check of other distracting processes. Whether there is intensification or inhibition depends on what is being compared with what. For the buyer the comparison is between what information is required for making a purchase decision and what information is available. The comparison operates on the basis of cognitive consistency theory. THEORIES OF COGNITIVE CONSISTENCY Theories of cognitive consistency deal with the relationships brtween hypothetical constructs that are not directly observable by overt behavior. The effects of the interaction of these constructs are intcrnal responses. For example, Berlyne (1960) suggests that the acquisition of knowledge is associated with symbolic responses. -35- Symbolic Responses The notion of a symbolic response has been considered by Osgood (1957), in his formulation of mediating reactions. As a result of the stimulus being presented in the form of a ”sign” it evokes some but not all of the effects that would be produced by the stimulus object that it stands for. A sign evokes a mediating reaction which is the "meaning" of the sign. This results in a self-stimulation and finally an overt behavior evoked by the self stimulation. In this way the notion of conceptual conflict is considered in terms of mediating responses to various signs. Figure 3—5 Osgood's Theory of Symbolic Response Stimulus Overt Cue Behavior S——->— rm ———-——->- sm ——-—-———9- Rx Mediating Self- Reaction stimulation Dissonance Theory A number of theories have been proposed to deal with the incompatibilities between symbolic responses and the conceptual conflict PrIDduced by them. Festinger's "dissonance" (1957) is a relationship between two cognitive elements, (i.e. beliefs, evaluations, perceptions) 0T between a cognitive element and an overt action that has been eXecuted. He focuses predominently on dissonance between evaluations rather than between factual beliefs. -36- When dissonance is present Festinger indicates that the strength of the drive toward its reduction depends upon the importance of the areas between which the conflict occurs. Dissonance for example can be reduced by changing evaluations of conflicting elements, by reducing the importance attached to them or by seeking social support from others. Festinger predicts that dissonance will also affect which stimuli are sought, favoring those which are likely to moderate dissonance and avoiding those which are likely to aggravate it. Because Festinger focuses on the relationship between a COgnitive element and overt behavior a major application of his theory to marketing has occurred relative to the post-transaction function. Schewe (1973) reviewed the implications of cognitive dissonance theory for marketing and found the major area is post decision infOrmation seeking. Dissonance studies may be confounded by measuring a variable other than dissonance (i.e. anxiety) or by concluding that predicted and observed behavior is the result Of an attempt to reduce dissonance when the behavior was caused by something else (Oshikawa, 1972). Thus other cognitive consistency theories may have more implications for information gathering prior to purchase. Balance Theory Another theory, "Cognitive Imbalance" (Ableson and Rosenberg, 1958) focuses on discrepancies among evaluations rather than among factual beliefs. Cognitive elements are valued positively, negatively or neutrally. Between any two of these elements there can be an associative relation or a disassociative relation. Imbalance exists when two positively or two negatively valued elements are disassociatively -37_ linked or when a positively valued and a negatively valued element are associatively linked. In either of these cases, there is postulated to be a ”pressure towards the attainment of cognitive balance". Abelson suggests that imbalance can be reduced by reorganizing attitudes and beliefs in any of the following ways: 1) Denial of the evaluation of one of the elements involved, 2) Bolstering of one of the elements by linking it with other ideas that are associated with strong attitudes, and thus outweighing the opposing belief or evaluation, 3) Differentiation within one of the conflicting elements by making a distinction between some aspect of it that is valued positively and another which is valued negatively, 4) Transcendence of the conflicting elements by combining them into some larger unit which is collectively viewed with favor or disfavor. Day (1972) cites a number of marketing studies using attitude model based on consistency theory that explain consumer attitude structure. Abelson's theory has been used to consider the effect of two—sided immunization messages on consumer attitudes (Bither, Dolich, and Nell, 1971), as well as to predict the effects of disconfirmation of expectations of produce performance (Olshavsky and Miller, 1972). Bither and Wright (1973) utilize a balance theory framework to examine the relationship of consumer self—confidence and attitude change in response to advertising. The same framework is used to explain the defensive information—processing strategy adopted by consumers in environments that prohibit thorough information analysis (Wright, 1973). -38- Theory of Conceptual Conflict Berlyne (1960) considers conceptual conflict that arises primarily out of the dennotative content rather than the affective or connotative aspect of beliefs or thoughts and the relationship between such conflicts in the pursuit of knowledge. Individuals have a tendency to associate incoming information with other concepts. The individual has already learned that properties of various stimuli are not independent of one another; some inputs occur together more often and some less often. Berlyne states that an ”acquired mutual inhibitory relation" may arise between concepts and lists the following six types of conceptual conflict which may result: 1) Doubt. The tendency to both believe and disbelieve the same statement is known as doubt. 2) Perplexity. This type of conceptual conflict occurs when various factors lead the subject toward a number of mutually exclusive beliefs, e.g. when conflicting facts are each supported by some evidence and the subject has no way of knowing which fact is true. This notion is very similar to the information theory measure of uncertainty. Perplexity encompasses doubt since there are factors simultaneously supporting and rejecting each of the alternative beliefs if taken separately. 3) Contradiction. The term contradiction refers to symbolically illogical sequences. The avoidance of contradiction seems to be a learned quality that may be a function of education and/or conceptual abilities. Because one values logic, an overt expression of thinking that violates the laws of systematic thought tends to lead to disapproval by peers. Individuals will tend to avoid illogical _39_ sequences provided that they recognized that a contradiction exists. 4) Conceptual incongruity. This type of conceptual conflict occurs when an individual has learned to believe that property A is unlikely to be found together with property B and yet sources of information indicate that a certain object or event has both A and B. 5) Confusion. Stimulus patterns that are ambiguous or can be confused with one another give rise to conflicting symbolic responses in much the same way as they arouse conflicting identifying or overt responses. Confusion is frequently found when the person communicating has not succeeded in conveying his intent, or because the symbolic stimulus pattern undergoes distortion between him and the receiver's receptors. 6) Irrelevance. The learning process involved here may be the same as that responsible for the clustering of words belonging to a common category. Individuals learn to connect thoughts with certain supraordinate concepts or topics and tend to reject ideas on a different topic until all available ideas pertaining to the topic at hand have been exhausted. Individuals learn to distinguish the ways in which thoughts condusive to successful thinking are related to the problem at hand. In attempting to follow an individual's conversation, frustration and arousal are apt to result when efforts to connect one message element with another meet with difficulty. Discomfort arises when the elements or thoughts are not clearly linked. Berlyne (1960) suggests that the concept of conceptual conflict may have implications for the concept of attention. Selective attention has been regarded as a phenomenon with two parts: 1) intensifying the process on which the attention is being concentrated, and 2) holding in -40_ check other distracting processes. Whether there is intensification or inhibition depends on what is being compared with what. One means of alleviating this conceptual conflict is the acquisition of knowledge through additional information. A reduction in conflict can be accomplished by: l) concilliation —- making the competing response tendencies less compatible 2) swamping -— introducing a new response tendency that is stronger than those which are in competition, or 3) disequalization _— strengthening or weakening one or more of the competing response tendencies and, thus, rendering the conflict unequal (Berlyne, 1960). Regardless of the process used to reduce conceptual conflict, the crucial element is information. A receiver is more likely to pay attention to message sources which provide information he considers necessary than to message sources which ignore such information. Berlyne (1960) contends that conflict among symbolic response tendencies provides the motivation for intrinsic epistemic behavior and supplements extrinsic motivation when information is sought for practical needs. The latter case seems especially relevant in purchasing decisions. In search of the information required to reduce the conceptual conflict naturally arising out of the purchasing situation, the buyer may be more highly motivated to attend the information provided by highly credible and/or highly empathic message sources. Allport (1955) cites a number of studies that show the effects of affective value and motivational state on the way in which a stimulus is perceived. When a motive is aroused the tendency is to perceive stimuli relevant to that motive more readily than irrelevant stimuli. When an individual is motivated it is expected that information which é -41- facilitates satisfaction Of the aroused motive would prevail over other incoming information. The affective value of incoming information is determined by the association with rewarding or punishing situations, either through overt behavior or through internally mediated, symbolic responses. Berlyne (1960) considers several experiments on the effects Of affective valued stimuli on attention and concludes that there is a significant tendency fOr stimuli associated with rewards to be more readily perceived. Stimuli that accompany opportunities for satisfaction are also much more likely to be recalled. In the buyer--seller relationship the buyer's satisfaction may arise from the opportunity to make "correct" purchase decisions. It fellows that a buyer would have positive affective values towards those sellers who facilitate the making of "correct" purchases by providing information which enhances the buyer's ability to make "correct" decisions. Berlyne's theory has been used to explain consumers' affective feelings toward a specific characteristic of a new brand even when that characteristic is linked to an unfamiliar and highly ambiguous brand name (Miller, Mazis, and Wright, 1971). Howard and Sheth (1969) have utilized Berlyne's work to explain consumer information seeking behavior by building a link between conceptual conflict and attention. The information requirements necessary to relieve different types of conceptual conflict may be tempered by the affective inhibitions of the receiver caused by the receiver's perceptions of the source of the information. Message sources and their messages must have certain qualities in order to reduce conceptual conflict through information input. In order for the buyer to place a higher value on one particular -42- belief or set of beliefs, the message source must necessarily be credible or believable. Thus the notions of doubt, perplexity and contradiction seem inextricably linked to the notion of credibility. In order for the message to be consistent with the buyer's frame of reference and for the message to contain some information about that framework, the message source necessarily must seem empathetic to the buyer. Thus, reduction of conceptual conflict due to confusion and irrelevance may be dependent upon the empathy of the source of the message. The buyer's "filter" may be attuned to incoming information from only certain types of message sources. The drive towards the reduction of conceptual conflict may explain why the receiver is more inclined towards message sources high in credibility and/or empathy. The receiver is more likely to believe the message content from a highly credible message source. The highly empathic message source is more likely to be received since this message is apt to be comprised of words that are easily understood by the receiver and that pertain to the receiver's current problem solving situation. Contradiction, perplexity, and doubt thus seem more likely to be relieved by a highly credible message source. The message content from such a source will seem more believable. The buyer may be able to relieve some the mutual inhibitory relationships between concepts by placing greater weight on the information provided by highly credible salesmen. The satisfaction resulting from a reduction of conflict would serve to reinforce the receipt of information in the future from the same highly credible message sources. The motivational cycle would tend to produce a drive to attend salesmen seen as having high credibility when the buyer faces situations of conceptual conflicts in making purchase decisions. -43_ Irrelevance, confusion, and conceptual incongruity seem to result from situations where the stimuli has little or no relationship to the patterns expected by the buyer. In these situations the motivational cycle leading to relief Of conceptual conflict seems susceptable to message sources high in empathy. By having empathy the message source is likely to encode the message in such a way that the content will correspond to the buyer's frame of reference. By so doing, the likelihood of reducing the conceptual conflict due to irrelevance, confusion and conceptual incongruity should be greatly enhanced. The buyer would be motivated to attend empathetic message sources to reduce the conceptual .9“ conflict arising out of the product purchase decisions. The drive to reduce the conceptual conflict arising in the purchasing situation may explain why credible and empathetic message sources are more attractive to buyers. As receivers of information, the buyers would be motivated to attend the message content from credible and/or empathic salesmen. There is considerable empirical evidence that message sources rated as possessing high credibility or high empathy are effective communicators. The specific definitions and dimensions of these two variables bear further consideration. SOURCE CREDIBILITY Communications research consistently supports the conclusion that a highly regarded message source is an effective transmitter of information. The influence of the receiver's evaluation of the message source on the effectiveness of the communication generally falls under the concept of source credibility. The available literature in this area indicates that a wide range of characteristics forms the basis for a highly -44- regarded or highly credible source. Mausner (1953) found that the receiver's tendency to agree with statements made by the message source is enhanced when the statements are contributed to prestigeful or authoritative sources. The following studies have suggested a wide variety of personal characteristics which, when present in the message source, lead to credibility and thus effective communication: perceived objectivity (Hovland and Mandell, 1962); lack of manipulative intent (Wallster and Festinger, 1962); competence (Hollander, 1960); the source's prestige (Adams, 1960); or trustworthiness (Hovland and Weiss, 1951). These studies all attempted to measure credibility as an index on a single dimension of the source's projected attributes. The index is produced by single linear rating scales that judge the receiver's perception of one trait of the communicator. Regardless of the dimension of the personal characteristic of the message source, the conclusion reached by all these studies is that ”high credible" sources are more effective than "low credible" ones. In their communication studies Hovland, Janis and Kelley (1953) suggest the use of a multi-dimensional model of source credibility. They conclude that two dimensions of the message source are especially relevant to communication effectiveness - (1) perceived expertness and (2) perceived trustworthiness. The researchers state that: "An individual's tendency to accept a conclusion advocated by a given communicator will depend upon ... 1) the extent to which the communicator is perceived to be a source of valid assertions, his "expertness", and 2) the degree of confidence in the communicator's intent to communicate the assertion he considers most valid, his "trustworthiness". In certain matters an individual is likely to feel the person with status, value, interest and needs similar to his own sees -45- things as he does and judges them from the same point of view". (1953, pp. 21-22). A number of studies following along the lines of the research done by Hovland, et. al., work to determine the specific dimensions of the communicator which make him acceptable as a source of information and to provide reliable instruments with which to measure the receiver's perception of these dimensions. Studies by Schweitzer and Ginsberg, (1966); McCroskey, (1966); and Berlo, Lemert and Mertz, (1966); have all investigated these dimensions using similar approaches. Message sources of predetermined acceptability or unacceptability are rated on semantic differential or Likert-type scales by large numbers of individuals. The scores are then factor-analyzed and scale items are selected on the basis of the strength of association and the purity of the factor loadings which represent each of the principle factors. Despite the differences in methods of obtaining the original scales, types of message sources, subjects used and factor analytic methods, all these studies yeilded relatively similar results. All these studies support the existence of the Hovland, et. al., factors of expertness and trustworthiness as the principle dimensions of evaluation of a message source which affect the source's acceptance and influence potential. Schweitzer and Ginsberg, (1966) assessed the adequacy of the Hovland, et. al., model of communicator credibility. They found evidence to support the dimensions of perceived expertness and perceived trustworthiness and suggest the existence of other dimensions as well. Berlo, et. al., (1966) found receivers evaluate a message source according to the source's perceived "safety" (his manipulative intent and Predictability), his "qualification" (his tOpic-bound expertness as well -46- as his perceived general ability and intelligence), and his "dynamism" (a combination of potency and activity dimensions). McCroskey, (1966) described a factor as "authoritativeness" which accounted for forty—seven per cent of the variance in source evaluation and a second factor which he called "character" which accounted for twenty-nine per cent of the variance. These factors also correspond with the qualification and safety factors reported by Berlo, et. a1. While McCroskey does not report the "dynamism" factor as appearing in his studies he also points out that this is not necessarily an indication that it does not exist. Mertz (1966) further examined the factor analytic, multi- dimensional model of source evaluation originally developed by Berlo and Lemert (1961). The principal research hypothesis considered receiver judgments of the message source on three independent, evaluative dimensions (safety, qualification and dynamism) as the principle determinants of influence acceptance. A relational conceptualization of credibility and the source-receiver relationships implied by credibility judgments were included in the second major area of Mertz's research. Multiple regressions were computed to determine the relationship of pre- post topic attitude change scores and (1) measures of perceived source, safety qualification, and dynamism and (2) measures of perceived self- source discrepancy on the three credibility dimensions. Zero order correlations between topic attitude change and measures of absolute source evaluation and self-source discrepancy were computed separately for each of the dimensions. Mertz (1966) reports a significant positive relationship between influence acceptance and perceived safety, qualification, and dynamism of the message source. However, this relationship was significant -47- only for those subjects who perceived the message as uniformly high or low on all three of the evaluative dimensions. NO conclusive results were found fOr subjects who saw the message source as high on one dimension and relatively low on others. The multiple regression on the three dimensions of source evaluation attempting to predict influence and acceptability failed to produce significant results. The predicted positive relationship between influence acceptance and the relative perceptions of source safety was not fOund. Mertz (1966) could not support the hypothesis that subjects who perceived the source as higher than themselves on the safety dimensions would show a positive relationship between influence acceptance and perceptions of relative source dynamism. The hypothesis that subjects who perceived the source as lower than themselves on the safety dimension would show a negative relationship between influence acceptance and perceptions of relative source dynamism was not supported. The hypotheses related to the direction and magnitude of perceived self-source discrepancies on the three dimensions were not supported. The predicted relationship between influence acceptability and perceptions of relative source qualification was supported and this variable was the best single predictor of influence acceptance. Mertz proposes that his data suggests that further credibility measures should be designed to define the source-receiver relationship which is expressed by credibility judgments and to determine the impact of source evaluation in different influence situations. One difficulty may lie in the use of pre-post topic attitude change as a -48- dependent variable for measuring influence acceptability. A question arises whether effective communication necessarily results in attitude change. Howard and Sheth (1969) discuss the dynamics of attitude change and the confounding relationship of feedback mechanisms over time. If the concern is with effective communication rather than with attitude change, Mertz's unsupported hypotheses need not be considered as negative results. Talbott (1968) investigated the generality of dimensions of source credibility reported by Berlo, et. al., (1966) across subjects in relation to different kinds of sources of information. In doing the factor analysis of the bipolar adjective scales Berlo, et. al., summed across both subjects and sources. In this way the factor structure could have been a result of the variance attributable to subjects, sources or a combination of both. Talbott (1968) developed two sets Of source evaluation data from each subject. One set, belief sources, was judged to have beliefs and ideas similar to the subject's own. The other set, disbelief sources, was judged to have beliefs and ideas different from the subjects. The dimensions of source evaluation and their correspondence with Berlo's dimensions for sources judged to have beliefs similar to the subject's own beliefs differ from the dimensions for sources judged to have beliefs not similar to the subject's own. On an overall basis the dimensions for belief sources correspond more to the Berlo dimensions than do the dimensions for disbelief sources. On a factor to factor basis Talbott found no differences for the safety factor. However, he did find differences for the qualification and dynamism factors. Belief sources exhibited a greater tendency for these factors to be present than did disbelief -49- sources. Talbott also Observed a merging of the Berlo qualification and dynamism factors into a single dimension. The tendency to do so was much greater for the disbelief sources than for the belief sources. Regarding the generality of the Berlo, et. al., dimensions of source credibility Talbott (1968) reached the following conclusions: 1) the generality of the safety dimension is high and is acceptable for most conceivable purposes, 2) the generality of the qualification and dynamism dimensions approaches that Of the safety dimension for the belief sources only, 3) for the qualification and dynamism dimensions generality is lower for disbelief sources, 4) generality is lowest if representatives Of all three dimensions are required among the individual factors, and 5) conclusion four is true for both kinds of sources except that for disbelief sources it is less likely that all three dimensions will be represented in the individual three factor solutions. Using each dimension one at a time, Talbott found the generality across subjects for both kinds of sources was highest for the safety dimension (78%). Generality for belief sources was 63% and 69% respectively for the other two dimensions, qualification and dynamism. For disbelief sources the generality of across subjects dropped to 45% for qualification and 39% for dynamism. While the overall factor analysis on his data questions whether this overall factor structure represents a very meaningful average of the factor analyses done separately on the data obtained from each subject, Talbott (1968) concludes that the results of all analyses were not interpretively different from the Berlo analysis. -50- EMPATHY Allport (1954) found the idea of ”putting yourself in the other fellow’s place" common to many definitions of empathy. This notion stresses the imitative characteristic which suggests an identity between the empathizer and his subject. Sensitivity to the affect of another thus depends on analogy, identification or a real similarity. Another approach stresses the detachment and the objectivity of the empathizer. English and English (1958) suggest that empathy is: ”Comprehension of the state Of mind of another person without feeling (as in sympathy) what the other feels. While the empathic process is primarily intellectual, emotion is not precluded, but it is not the same emotion as that of the person with whom one empathizes. The attitude in empathy is one of acceptance and understanding on an implicit I see how you feel.” Rogers (1959) describes the state of being empathic as one of perceiving: "...the internal frame of reference of another with accuracy and with the emotional components and meaning which pertain thereto, as if one were the other person but without ever losing the ”as if" condition. If this "as if" quality is lost then the state is one of identification." The ObjeCtiVG, detached—but—concerned attitude is what differentiates empathy from sympathy. Fenichel (1945) states that empathy: "...consists of two acts: 1) an identification with the other person and 2) an awareness of one's own feelings after the identification, and in this way an awareness of the object's feelings." Chessick (1965) cites this definition in his own work and -51- views empathy as an intellectual and detached process. Dymond (1948) stressed the role-taking aspect of the empathic process in the following definition: "the imaginative transposing of one's self into the thinking, feeling and acting of another and so structuring the world as he does." Speroff (1953) defined empathy as: "the ability to put yourself in the other person's position, establish rapport and anticipate his feelings, reactions and behavior. Empathy and role reversal are mutually complimentary." Murray (1938) describes the interactive component of the empathic process which he refers to as "reciapathy". Stewart (1954) defined empathy in terms of "mutual transference" and thus implies that the empathic process is an interactive one. Describing empathy and role reversal and complimentary processes, Speroff (1953) implies a process involving interaction. Buchheimer (1963) stressed the interactive component of the empathic process and referred to it by using the term "confluence". Buchheimer, et. al., (1964) defined empathy as "the ability to structure the world as another person sees it; i.e., the counsellor's task is to feel to react and to interpret the counsellee's world as he sees it without the counsellor's inactment of these perceptions as if he were the counsellee." Rogers (1951) included interaction variables in his definition of empathy: "...assuming so far as he is able the internal frame of reference Of the client,...and to communicate something of this empathic understanding to the client." In these proceding definitions of empathy not only is the ability to put yourself in another's position present but also there is the dimension of interaction. Rogers stresses the importance of communication of understanding or empathy as an ..52- aspect of the empathic process. This definition seems to capture the essence of the dyadic notion of the communication-—empathy relation. Rogers and Bhowmik (1970) indicate the importance of understanding the dyadic nature of the communication relationship when considering the effects of empathy on communication effectiveness. From the various definitions of empathy, there has arisen a wide variety of approaches to the measurement of the concept. Astin (1957) distinguishes between the ”predictive test” of empathy and the ”situation test" of empathy. The predictive test is one which involves prediction of the behavior of another. The predictor is either acquainted with the subject or is given data about him. The predictor's task is to predict the subject's performance on test items or responses to attitude or personality inventory items. The situation test of empathy provides the subject with some standardized situation to which he responds. The situation may be video tapes, films, role playing sessions and/or typed script materials. Kurtz (1970) classifies empathy measures as: (l) predictive approach, (2) situation approach, (3) perceived approach, (4) rating approach. In the perceived approach the counsellor is asked to rate his empathy in a particular interview after the completion of an interview and the client can be asked to report how empathic a counsellor was during the same interview. Barret-Lennard (1962) labelled this as "experienced" or "perceived" empathy. Chessick (1965) defined the same process as the "retrospective approach". In the rating approach trained judges listened to observed interactions between persons on the basis of certain -53- predetermined criteria. The judges rate the amount of empathy displayed by the counsellor in the interactions. Katz (1962) compared the judged and predictive approaches with what he termed "behavioral empathy” which was measured by having trained judges listen to or observe interactions and then rate the amount of empathy displayed. The Predictive Approach Many predictive tests of empathy have followed the procedure outlined by Dymond (1948). Empathy is measured by assessing the degree of similarity between individual's ratings of others known to them through some interaction and other's actual self—ratings. Ratings are usually done on some personality characteristics using one of several attitude and/or personality tests. If individuals A and B were to be tested for their empathy for each other using Dymond's procedure it would be as follows: 1) A rates A 2) B rates B 3) A rates B 4) B rates A 5) A rates B as he thought B rates himself 6) B rates A as he thought A rates himself 7) A rates A as he thought B rated him 8) B rates B as he thought A would have rated him. A group pencil and paper device called the "empathy test" (Kerr and Speroff, 1954) was designed for use in industrial selection and was intended to measure the "ability to put yourself in another -54- person's position, establish rapport, anticipate his feelings, reaction, and behavior". R.L. Thorndike commented on this test as follows: "There appears to be no inherent validity in the Operations called for in this test. This test cannot be recommended as either a useful practical device or a contribution to the description and understanding of an individual" (Buros, 1965). Wallace Hall (Buros, 1965) comments on the same test: "There appears little to recommend this test for the purpose stated by its authors". Both the Dymond ”specific other" predictive test and the Kerr and Speroff "generalized other" predictive test have been severely criticized. Hall and Bell (1953) find that there is no empirical relationship between the two predictive tests even though they reported to be measuring the same construct. Questions have been raised whether or not these tests or any other such predictive tests were actually measuring empathy or whether they were measuring or at least confounded by certain other variables such as projection (Norman and Leiding, 1956) or identification (Jackson and Carr, 1955). Lindgren and Robinson (1953) evaluated Dymond's test and found the subjects making predictions tended to conform to social norms, that the reliability of the test was too low for predictive purposes and that the validity of the test could be questioned. Borgatta (1960) found another bias in the Dymond "specific other" predictive test that he called "self-image error" which was the tendency of a person rating himself to consider whether he was high or low in certain qualities without paying any attention to the qualities of others to whom he was being compared. Campbell (1967) notes that since most predictive tests are variations of Dymond's approach, they all contain these shortcomings to some extent. The Situational Approach Kagan and Krathwohl (1967) developed a situational test using video tape segments of actual counselling interviews. The test does not propose to measure empathy but rather an important component of empathy, "affective sensitivity" which is the ability to perceive and .n-.n:'m_-'_--4 4. 4 identify the affective states of others. Analyzing the situational approach it would be hypothesized that an individual scoring high on a situation test would show high empathy in actual interviews. In the counselling situation Kurtz (1970) found that the trait of affective sensitivity scale did not correlate with the counselling process or outcome. He also found that predicted empathy was not consistent, I did not correlate with other empathy measures and was not related to the counselling process and outcome. The Rating Approach Cheminsky and Rappaport (1970) have raised questions about just what is being measured by this method. When trained judges listened to or observed interactions between persons, they found that the rating approach measures such variables as "language style , communicated commitment, or general positive affect”. Caracena and Vicory (1969) studying correlates of subject's perceived and judges' rating methods of empathy, found that the judges' ratings were related to verbal comments and the quantity of words used while clarity of expression was related to the subject's perceptions. It was also found that the expression of interest or involvement related to both measures. These studies suggest something other than empathy is being treated by judges using this method. Truax and Carkoff (1967) —56— suggest that judge rated empathy was more predictive of outcomes of therapy than subject perceived empathy. The Perceived Approach The perceived approach to the measurement of empathy involves having the subject rate how empathic he perceived his counsellor had been during the previous interview. This procedure is consistent with Rogers theory (1957) which emphasizes that the subject's perceptions of being understood are crucial for personality change. Barret-Lennard (1962) provided an empirical test of Roger's theory and developed a questionnaire yielding an empathy index for an individual in a particular situation. In comparing clients' and therapists' perceptions of empathy he finds "...evidence of the primary relevance to therapeutic change of the client's perception of the relationship rather than the therapist's." The amount of empathy experienced by the client seems to be related to the outcome of the therapy since the more changed groups of client perceived significantly higher levels of empathic understanding. Rogers, et. al., (1967) fOund low correlations between patient's perception of therapist empathy and the therapist's perceptions of his own empathy. It was also found the patient's perceptions were more predictive of the therapy process and outcOme than were the therapist's perceptions. Kurtz (1970) suggests that counsellor perceived empathy was not significantly correlated with client perceptions and was also consistently unrelated to both the counselling process and outcome measures. He found client perceived empathy, based on the Barret- Lennard relationship inventory, was the best predictor of all indices -57- of counselling outcome. This again is consistent with Rogers' (1957) theory which says that the client's experience of the therapist's response is the primary focus of influence in the relationship. While personality change is not in any way seen as a necessary element of the buyer--se11er interaction process, the behavioral factors described in Rogers' client-centered therapy is consistent with the concept of receiver-centered marketing communications. Rogers and Bhowmik (1970) hypothesized that the source having a high degree of empathy with the receiver will be positively associated with effective communication. This hypothesis was developed from a source—-receiver relational framework. The dyadic approach is utilized in a number of studies in buyer——seller communications. DYADIC BUYER—~SELLER COMMUNICATIONS MODELS An early contribution to buyer--seller communications research is the dyadic interaction model developed by Evans (1963). This study stresses the perceived similarities of buyers and sellers. These perceptions are specific to each buyer—-seller situation and therefore focus on the dyadic interaction of buyer and seller as the unit of investigation. Rather than use some "objective" traits of buyers and sellers Evans (1963) utilizes the buyer's perceptions of the seller as the major independent variable. Tbsi (1966) also suggests that buyer-— seller interactions are central to understanding personal selling as a communication process. Tosi (1966) uses a dyadic interaction model and SuPports Evans conclusion that the outcome of a sales call cannot be IxPedicted by looking at either the buyer or seller alone. In this .I ~58- context it can be predicted that any "qualities" a salesman may possess affect the buyer only to the extent of the buyer's ability and/or willingness to perceive those qualities. Stiden (1967) explains personal selling in terms of interpersonal communication or ”intercommunication," a continually evolving specific relationship between each sender and each receiver. He concludes that the effectiveness of the salesman is a function ofthe ability to tailor messages that are specifically related to individual buyers. This approach also utilizes a relational model to explain buyer and seller communications. Levitt (1967) examined the "source effect" in industrial selling and found that the perception of who communicates is Often more important than what is communicated. He reports that the effects of a particular message ("presentation effect") cannot explain the variations in sales outcomes independently of the perceived source. The concept of credibility has been treated as a dimension or set of dimensions by which the receiver evaluates the source of the communication. Dyadic conceptualizations of credibility emphasize the direct measurement of perceptions of the message source. Mertz (1966) points out that these measures should be interpreted as expressing a source--receiver relationship. Birdwhistle (1968) suggests that the ideal model for the communication process is based on the dyad. Rogers and Bomick (1970) present a convincing case for the use of the relationship between the source and receiver as the basic unit of study for both empathy and credibility in human communications. Mertz (1966) conceptualized the dyadic notion by stressing the source-- receiver relationship which is implied by credibility judgements. Talbot (1968) emphasized the multidimensionality of source evaluation -59_ and support the argument that credibility must be defined in terms of the perceptions of the receiver rather than by any innately determined characteristics of the source. Thompson (1966) has explained the communication effectiveness Of the industrial salesman in terms of three behavioral science concepts: readiness, empathy and source credibility. Integrating these three concepts into an interactive buyer—~seller framework Thompson (1969) accounts for several phenomena described in previously mentioned studies. The concept of source credibility is described by Levitt (1967) as a "source effect". Thompson (1966) analyzes the concept of source credibility in terms of the Berlo et. al., (1966) three—factor solution of competence, trustworthiness and dynamism. The perceptions of similarity of the seller and the buyer (Evans, 1963; Tosi, 1966) are central to the concept of empathy. The mutual understanding and the intercommunication of same is a major element of Stiden's (1967) relational model. Thompson (1966) utilizes the perceived approach to empathy in his framework. In addition to empathy and credibility Thompson (1966) presents the concept of readiness, the ability and desire to communicate in a buyer——seller context. Generating readiness in the buyer—-seller dyad depends upon the buyer's perception of the salesman's preparation and planning or what Thompson (1973) calls the "sales call objective". The three concepts — readiness, empathy, and credibility - are presented in a interpersonal communication framework (Thompson 1966, 1969, and 1973) that treats the selling function from an integrated, behavioral science vieWpoint. The buyer--seller communications model presented in Chapter 2 is based on this integrated framework. CHAPTER FOUR RESEARCH HYPOTHESES This research examines the relationship between the buyer's perceptions Of the salesman's empathy, credibility and preparation and the buyer's intention to purchase the salesman's product. This relationship is believed to exist due to two intervening variables: the salesman's message content and the buyer's overall evaluation of the salesman. Furthermore it is expected that a relationship exists between the salesman's message content and the buyer's evaluation of the salesman. These relationships are illustrated in Figure 5-1. FIGURE 5-1 Hypothesized Relationships Empathy Credibility .- Evaluation an. Intention :5 Message Content The salesman's perceived empathy, perceived credibility and perceived preparation are believed to be associated with the buyer's evaluation of the salesman and associated with the salesman's message content. The salesman's message content and the buyer's evaluation of —61— : the salesman are expected to be associated. The buyer's evaluation of the salesman is hypothesized to be associated with the buyer's intention to purchase. Hypotheses one through three consider the association of: (l) empathy, (2) credibility, (3) preparation and the message content. Hypotheses four through seven consider the association of: (4) empathy, (5) credibility, (6) preparation, (7) message content and evaluation. Hypothesis eight proposes a relationship between evaluation and intention to purchase. In addition to these eight hypotheses alternative relationships between the variables are considered. Hypothesis nine examines the possibility that the relationships between empathy, credibility, preparation and evaluation exist due to the relationships of these variables with the message content. In a similar manner the viability of evaluation as an intervening variable is considered. Hypothesis ten considers the possibility that empathy, credibility, preparation and message content are directly associated with the buyer's intention to purchase rather than indirectly due to evaluation. Hypothesis One: The buyer's beliefs regarding the salesman's role and task empathy are associated with the salesman's message content score. Thompson's (1973) concept of relevancy suggests that the salesman is more likely to encode a relevant message if he has some understanding of the buyer's frame of reference. Applying Rogers' (1957) theory, this understanding is a function of the salesman's empathy. This hypothesis utilizes the perceived approach to empathy found in Barret—Lennard's (1962) operational definition of Rogers' (1957) client—centered counseling theory. The salesman's capacity of -62- producing a relevant message is hypothesized as being dependent upon his empathy for the buyer. The relevance of a message is a function of its correspondence with the buyer's frame of reference (Thompson, 1973). Die frame of reference of the buyer is operationally defined as the decision criteria the buyer uses in making a purchase decision. Comparison of the content of the actual message with the buyer's decision criteria yields the message content score. The relevance of the salesman's message is operationally measured by the message content score. Salesman perceived as empathic are expected to transmit relevant messages. Thus the buyer's belief of the salesman's empathy is hypothesized as associated with the salesman's message content score. Hypothesis Two: The buyer's beliefs regarding the salesman's source credibility on the dimensions of: a) qualification b) trustworthiness and c) dynamism are associated with the salesman's message content score. These hypotheses follow from Thompson's (1966) application of the three dimensional model of source credibility set forth by Berlo and Lemert (1961). It is expected that the salesman's source credibility is reflected in the content of the salesman's message. The Berlo, et al. (1966) model presents three aspects of the communicator that affect his acceptance as an information source. The multiple dimensions follow from Mertz's (1966) report on inconclusive results when subjects rated a message source as high on one dimension and relatively low on others. The acceptability of the salesman should be reflected in his message. It is expected that some combination of 5 -63- the three dimensions of source credibility is associated with the salesman's message content score. Hypothesis Three: The buyer's belief regarding the salesman's preparation and planning is associated with the salesman's message content score. Thompson (1966) proposes the need for preparation and planning grows out of the need for the salesman to stimulate the buyer's readiness and enhance the buyer's willingness and ability to communicate. This hypothesis follows from (Thompson, 1969) the application Of the concept of readiness to buyer--seller communications. The salesman's ability to positively affect the buyer's willingness to communicate depends, in part, upon the salesman planning and preparation prior tO the buyer--seller interaction. Thompson (1973) explains this with the concept of a "sales call Objective". The need for the salesman's planning exists so that the salesman's message contains only those elements relevant to the buyer's frame of reference. The salesman believed to be well planned is expected to present a relevant message. Therefore the buyer's belief regarding the salesman's preparation and planning is hypothesized as associated with the salesman's message content score. Hypothesis Four: The buyer's beliefs regarding the salesman's role and task empathy are associated with the buyer's evaluation of the salesman. In a source--receiver relational communications framework Rogers and Bhowmik (1970) hypothesize that a receiver's perception of the source having a high degree of empathy with the receiver is positively associated with effective communication. The perceived measure of empathy (Barret-Lennard, 1962) shows the degree of understanding the -64- salesman has communicated to the buyer. This understanding should lead to more effective communication between salesman and buyer (Thompson, 1969). If the salesman were an effective communicator he would be highly valued by the buyer. Thus it is hypothesized that the buyer's beliefs regarding the salesman's empathy are associated with the buyer's evaluation of the salesman. Hypothesis Five: The buyer's beliefs regarding the salesman's source credibility on the dimensions of: a) expertness b) trustworthiness and c) dynamism are associated with the buyer's evaluation of the salesman. Hovland, Janis and Kelley (1953) propose that the perceived credibility Of a message source affects the communication effectiveness Of that source. Berlo and Lemert (1961) identified three dimensions of the communicator that affect his acceptance as a source of information: qualification (expertness), safety (trustworthiness), and dynamism. Hypothesis five examines the Berlo, et a1. (1966) dimensions of source credibility and the salesman's communication effectiveness. Thompson (1966) and Levitt (1967) have both applied the concept of source credibility to buyer--se11er communications. It is expected that the effectiveness of the salesman's communication will affect the buyer's evaluation Of the salesman. Communication effectiveness is affected by the perceptions of the credibility of the message source (Mertz, 1966). Thus the buyer's beliefs regarding the salesman's source credibility are hypothesized as associated with the buyer's evaluation of the salesman. ~65- Hypothesis Six: The buyer's belief of the salesman's preparation and planning is associated with the buyer's evaluation of the salesman. The buyer's belief regarding the salesman's preparation and planning is presented (Diagram 2-3) as one of the inputs to the buyer's filtering process. The buyer's degree of belief regarding the salesman's preparation and planning affects the buyer's predisposition towards receiving the salesman's information (Thompson, 1969). The well planned salesman is expected to be held in higher regard by the buyer than the poorly planned salesman. Therefore, it is hypothesized that the buyer's belief regarding the salesman's preparation and planning is associated with the buyer's evaluation of the salesman. Hypothesis Seven: The salesman's message content score is associated with the buyer's evaluation of the salesman. This hypothesis follows from the Howard and Sheth (1969) buyer behavior theory where perceptual bias is based on the buyer's decision criteria. The buyer takes in that information he expects will be useful in amking a purchase decision (Howard, 1971). The usefulness Of information is a function of the correspondence of the information to the buyer's decision criteria. A salesman providing useful information is expected to be highly evaluated by the buyer. This process can also be explained by Thompson's (1973) concept of relevancy. Relevance is operationally defined in terms of the buyer's decision criteria. Information corresponding to the buyer's decision criteria is, by definition, relevant. The message content score reflects the degree of correspondence of the salesman's message and the buyer's decision criteria. Buyers are expected to value relevance. Therefore it is hypothesized that the salesman's message content score —66- is associated with the buyer's evaluation Of the message source. Hypothesis Eight: The buyer's evaluation Of the salesman is associated with the buyer's intention to purchase. This hypothesis tests the relationship between the buyer's evaluation of the salesman and the buyer's intention to purchase. Intention to purchase may be influenced by other marketing functions (i.e., price, product, distribution, etc.). However, it is expected that there is a relationship between the salesman's performance as perceived by the buyer and the buyer's intention to purchase the salesman's product. This hypothesis considers if the buyer's intention is associated with the buyer's evaluation of the salesman. Hypothesis Nine: The buyer's evaluation of the salesman is associated with: a) empathy b) source credibility c) preparation due to the association with message content. Hypotheses one, two and three propose an association between empathy, credibility, preparation and the message content. Hypotheses four, five and six propose an association between empathy, credibility, preparation and evaluation. Hypothesis seven suggests a relationship between the message content and evaluation. Thus it might be the case that empathy, credibility and preparation are associated with evaluation because: (a) empathy, credibility, and preparation are associated with the message content and (b) the message content is associated with evaluation. Hypothesis nine considers if a relationship between empathy, credibility, preparation and evaluation exists because all are associated with the intervening variable, the message content. -67— Hypothesis Ten: The buyer's intention to purchase is associated with: a) empathy b) source credibility c) preparation due to the association with the buyer's evaluation of the salesman. This hypothesis considers the relationship between empathy, credibility, preparation and intention to purchase in terms of evaluation as an intervening variable. Hypothesis eight considers the relationship of evaluation and intention to purchase. Hypotheses four, five and six suggest an association between empathy, credibility, preparation and evaluation. Hypothesis ten considers if an association between empathy, credibility, preparation and intention to purchase is due to an association of each with evaluation, or if empathy, credibility and preparation are directly associated with intention to purchase. CHAPTER FIVE METHODOLOGY RES EARCII I NS TRUMEN'IS Source Credibility The conceptual background for the operational definitions of credibility lie in the multi-dimensional model put forward by Hovland, Janis and Kelly (1953). A study of Schweitzer and Ginsberg (1966) empirically confirmed the existence of the two factors, perceived expertness and perceived trustworthiness, and suggested the existence of a third factor, dynamism, that is described by Berlo and Lemert (1961). Mertz (1966) and Talbot (1968) support the validity of Berlo's (1961) model. McCroskey (1966) investigated the three factors and confirmed the existence of an "authoritativeness" factor and a "character" factor. These two factors correspond with the expertness and trustworthiness factors identified by Hovland, et. a1. (1953). McCroskey (1966) did not confirm or deny the existence of the dynamism factor. The dimensions of source credibility are identified in the above studies through the use of factor analytic techniques. Semantic differential or Likert-type scale items are selected on the basis of the strength of association and the purity of the factor loadings. McCroskey (1966) demonstrated a highly significant correlation between Likert—type items and the semantic differential items. The Likert-type scale items utilized in this study were selected by taking a consensus 459— of the scales used in the Schweitzer and Ginsberg (1966), McCroskey (1966), and Berlo, et. a1. (1966) studies. The selections were based on the criteria that the scales must have demonstrated good explanatory value in previous research. The wording of the items was modified slightly to pertain to the buyer——seller situation. The scales selected for the three credibility factors are as follows: PERCEIVED EXPERTNESS 4. He is very well trained for his job. 1 2 3 4 5 6 7 minimum maximum 13. He seems to be very well informed on all aspects of this new product. 1 2 3 4 5 6 7 minimum maximum 25. He seems to be highly qualified for his job. 1 2 3 4 5 6 7 minimum maximum 31. He is highly skilled at his job. 1 2 3 4 5 6 7 minimum maximum PERCEIVED TRUSTWORTHINESS 2. He is a person that I can trust to tell the truth. 1 2 3 4 5 6 7 minimum maximum 12. He seems to be a just and fair person. 1 2 3 4 5 6 7 minimum maximum 21. He seems to be honest and ethical. l 2 3 4 5 6 7 minimum maximum -70_ 29. He seems to be a dangerous person to do business with. l 2 3 4 5 6 7 minimum maximum PERCEIVED DYNAMISM 9. He is aggressive and forceful. l 2 3 4 5 6 7 minimum maximum 17. He is bold and emphatic. l 2 3 4 5 6 7 minimum maximum 27. He seems to be meek and reserved. 1 2 3 4 5 6 7 minimum maximum 32. He is active and energetic. l 2 3 4 5 6 __j minimum maximum Across the studies considered, the above items were most reliable and had greatest explanatory value in accounting for the variation of the evaluation of a message source. The buyers beliefs of the seller on three dimensions of source credibility are presented in Chapter Two as an input to the buyer's filtering process. These beliefs are Operationally defined by the preceeding Likert-type scale items. Empathy The conceptual basis for the empathy items in the questionnaire is found in Rogers' (1957), client—centred theory. The operational definitions were drawn from Barret-Lennard's (1962) empirical test of Rogers' theory. He developed a questionnaire yielding a empathy index -71- for therapy situations. The following Likert—type items were selected from the Barret-Lennard Relationship Inventory based on their predictive value in counselling situations. They provide an Operational measure of the seller's empathy towards the role of the buyer. 1. He tries to see things through my eyes. 1 2 3 4 5 6 7 minimum maximum 6. He seems interested in knowing what my experiences as a buyer mean to me . l 2 3 4 5 6 7 minimum maximum 7. He nearly always knew exactly what I meant. l 2 3 4 5 6 7 minimum maximum 14. Sometimes he seemed to think that I felt a certain way because he felt that way. l 2 3 4 5 6 7 minimum maximum 16. He seems to understand the way I do my job as a buyer. 1 2 3 4 5 6 7 minimum maximum 19. His own attitudes toward some of the things I say, or do, stop him from really understanding my job as a buyer. l 2 3 4 5 6 7 minimum maximum 24. He appreciates my own feelings about my job as a buyer. 1 2 3 4 5 6 7 minimum maximum 26. He did not realize how strongly I felt about some of the things we discussed. I 2 3 4 5 6 7 minimum maximum _72- 28. He usually understands all of what I say to him. 1 2 3 4 5 6 7 minimum maximum 30. He responded to me mechanically. l 2 3 4 5 6 7 minimum maximum The previous scale items provide a measure of the buyer's beliefs about the seller's understanding of his role as a buyer. These scale items are used to measure the variable called role empathy in this study. In addition to role empathy an attempt was made to operationalize Regers' theory in a buyer--seller specific sense. As well as understanding the buyer as a person, it is important for the seller to understand the buying task specific to each buyer--seller situation. These items are based conceptually in the Howard and Sheth (1969) model of buyer behavior. The buyer's decision process is described as information processing leading to some comprehension of the brands in the product class. These comprehensions are then compared to certain choice criteria specific to the product class. The following four items measure the buyer's perceptions of the seller's understanding of this purchase decision—making process. These scales measure the variable referred to as task empathy in this study: 8. He understands the criteria that are used to make purchase decisions. 1 2 3 4 5 6 7 minimum maximum 23. He seems to fully understand if his product will fit our needs. 1 2 3 4 5 6 7 minimum maximum 15. The things he talked about are important to know in making a purchase decision. l 2 3 4 5 6 7 minimum maximum 20. He provided me with the information that is necessary to make the best purchase decision. 1 2 3 4 5 6 7 minimum maximum Preparation and Planning Thompson's (1973) concept of a sales call objective is operationally defined as the buyer's perceptions of the preparation and planning of the salesman. Four Likert-type scale items are used to measure the buyer's belief about the seller's preparation and planning prior to the buyer-—seller encounter. The following items measure the buyer's belief about the presence of the salesman's sales call objective: 5. He seems to have done a great deal in preparation for this appointment. 1 2 3 4 S 6 7 minimum maximum 11. His sales presentation was well planned. 1 2 3 4 S 6 7 minimum maximum 18. He seems to have put things in order specifically for this appointment. 1 2 3 4 5 6 7 minimum maximum Message Content The message content is defined as the degree to which the “fan 15% . -74_ salesman's message corresponds to the purchasing criteria of the buyer. Three concepts provide the operational measure of message content: (1) buyer's purchasing criteria, (2) salesman's actual message, and (3) content analysis of salesman's message in terms of buyer's purchasing criteria. Purchasing criteria. The buyer's purchasing criteria were determined through a two—stage interview. The first stage consisted of an open- ended inquiry into the attributes of a new product offering that were considered by each of the buyers. From the first stage interviews eighteen different attributes or criteria were presented to the buyers in questionnaire form. In the second stage each buyer was asked to rank the importance of the eighteen criteria on a nine point Likert—type scale. This method utilizes two successive direct questioning approaches to determine the buyer's criteria or those attributes of the product offering which determine preference and purchase behavior. The direct questioning approach has been shown to be an acceptable method for determining criteria or determinant attributes, (Alpert, 1971). Salesman's message. Each salesman—~buyer dialog was tape recorded and later transcribed. The buyer's consent to tape record the interaction was requested and in every case obtained. The salesman did not specifically know they were being tape recorded but were aware some record was being made due to the researcher's presence during the interview. The salesman's permission was requested and received in every case. The transcribed tape recording serves as the buyer's actual message. This process causes the "loss" of certain information I -75- since the nonverbal and paralinguistic activities were not recorded. It has been shown that in cases where words, tone of voice and facial expression are in conflict that the nonverbal or facial expression accounts for the significant portion of the variance (Bugental, Kaswan and Love, 1970). In messages not having such conflict it is not clear how much of the variance is accounted for by the verbal and nonverbal components (Duncan, 1969). Therefore this study assumes that a significant portion of the information transmitted is found in the verbal content. Content analysis. The message content score was calculated by content analyzing the transcript of the salesman's message. Content analysis is typically concerned with the second phase of a three step communication process where: (1) various motives produced (2) a message that (3) intended to express these motives and/or produce various effects on the audience (Mitchell, 1967). In Lasswell's (1946) classic sentence identifying the process of communication as "who says what to whom, how and with what effect,” the communication content is the what. Content analysis deals with what is said in the communication process (Berelson, 1952). Rather than asking the buyers or sellers to respond to the message content in terms of a series of scales, content analysis takes the communication that has been produced and inquires into the nature of the message. Recently the trend in content analysis is towards hypothesis testing as opposed to purely descriptive research (Budd, Thorp and Donohew, 1967). This extension allows the use of content analysis in combination with other data gathering techniques. This allows an analysis of the relationship between the content of the :.fil:‘h§.'f3' V". IL -76— message and other variables surrounding the communications process. There are two basic units of measurement involved in content analysis: (1) the coding unit and (2) the context unit (Budd, Thorp and Donohew, 1967). The coding unit is the smallest segment of content ."m'. .. counted and scored. An assertion is the coding unit used in this 5'! research. An assertion could be a sentence, a part of a sentence or perhaps even a single word. For example the sentence, ”Green beans come packed 24 to the case," represents a single assertion. However, in the sentence "The price for green beans is $2.12 in cases of 24 with terms 3 net 30,” there are three separate assertions: (1) green beans come packed 24 to a case, (2) the price of green beans is $2.12, and (3) the credit terms on green beans are 3 net 30. The reason for using the assertion as a coding unit relates from the concept of comparing the buyer's purchasing criteria to the content of the message. Ultimately . this will involve an analysis of the correspondence between criteria and message content. In this sense it is possible to have a number of correspondences in a given sentence where each assertion might relate ,i to a different one of the buyer's criteria. The context unit is the body material surrounding the coding unit. In this research, the context unit will be each statement by the buyer and the seller. Statement begins and ends with the change of the flow of conversation from buyer to seller or vice versa. Each change in flow of conversation from buyer to seller marks the end of a statement by one and the beginning of a statement by another. The message content score was calculated by content analyzing the frequency and value of the assertions in the salesman's message. If the salesman mentioned something pertaining to a purchasing criteria which -77- the buyer indicated as having an importance of 7 on a scale of 1 to 9, then the salesman received 7 points for that particular assertion. If the salesman made an assertion which pertained to none of the purchasing criteria rated by the buyer, he received no points at all for that assertion. The message content score is the sum of the number of points received for each assertion made during the sales interview. The message content score is a function of both the frequency of assertions pertaining to purchasing criteria and the relative importance of each criterion to each buyer. The importance scores of the buyer involved in the dialog were utilized in the calculation of the message content score. Therefore it reflects the relevance of a particular message to a particular buyer. Evaluation One of the variables in this research is the buyer's evaluation of the salesman. The following scale was used to measure evaluation: A. How would you rate the overall performance of the salesman you have just seen? 1 2 3 4 S 6 7 minimum maximum The above scale stems from the concept of the salesman as a message source. This item was worded to minimize the contaminating effects of the buyer's attitude towards the salesman as a person. The respondent is asked to evaluate only the salesman's performance. Thus the focus is intended to be on the communication activitiy of the salesman and not other, general characteristics. -73- Intention to Purchase The dependent variable in this study is the intention of the buyer to purchase the product presented by the salesman. The following scale was used to measure this variable: B. Short of making the final decision, how would you rate your current intention to purchase the new product just presented? 1 2 3 4 5 6 7 minimum maximum DATA SAMPLE The sample of buyers in this study came from the merchandising department of a large midwestern food and dry goods retail chain. The sample of salesmen came from the population of salesmen who called on these buyers. A total of 93 sales calls were recorded. The collection of data was done with the cooperation of buyer and salesman. At no time was there any request that any data be withheld from the study. The data sample was further restricted by the buyer's degree of prior exposure to the salesman's offering. For the message content scores to be valid they have to include all the information that the seller has transmitted to the buyer about that product. If the buyer and seller had discussed the offering prior to the interview being recorded the data was excluded from this study. Of the total of 93 sales calls with ten different buyers, 31 sales calls with seven buyers were used in this study. This research focuses on buyer-~seller interaction about "new" offerings, The product and/or the distribution, pricing and .4.“ ...79- promotion were "new" to the buyer. Thus the sample of thirty—one sales calls consists solely of new offerings. By limiting the sample to communications about "new” products and/or "new" marketing mixes the content of the message should reflect the entire information exchange between buyer and seller on that topic. Otherwise the message content scores would not provide a valid measure of the information flow between buyer and seller. The restricted data sample may limit the applicability of the results. However, the wide scale of operations of the retail chain buyers suggests that they may be representative of other buyers in the food and dry goods industries. The sample of salesmen range from manufacturers' salesmen to brokers, manufacturer's agents and merchant wholesalers. Thus the sample of salesmen may be representative of the ' food/dry goods industry and perhpas many other business settings as well. Given the similarity of this retail buying operation to others in the U.S. and Canada and the range of salesmen included in the sample, the data gathered in this study may be representative of a variety of buyer-—seller situations. TEST STATISTICS The relationships in the theoretical model suggest that high amounts of credibility, empathy and preparation and planning result in high message content scores, evaluation and intention to purchase. One way to measure the linear dependence of these variables is correlation analysis (Ezekiel and Fox, 1959). A related technique, regression analysis, uses a correlation coefficient to predict the value of the dependent variables from the value of the independent variables. The /—‘ -80.. ability to make such a prediction provides an insight into the underlying causal links between the variables (Johnston, 1963). Zero-Order Correlation The zero-order Pearson product-moment correlation coefficient measures the degree of association between two variables (Blalock, 1972). The significance of the size of the correlation coefficient can be measured by calculating the "t" statistic and comparing it to the Students' t distribution with N-2 degrees of freedom. Regression Analysis The statistical techniques utilized must test the theoretical model's ability to explain the variation in the dependent variables in terms of the variation of hypothesized independent variables or explanatory variables. For analyzing dependence regression analysis is a commonly used technique (Massey, 1962). An underlying assumption of regression analysis requires a single, interval-scale, dependent variable to be explained by a set of independent variables also assumed to be interval scale. Other assumptions underlying the regression model are summarized below (Massey, 1962): l. The regression error term must be randomly distributed. 2. The error terms must be homoscedastic., 3. The error terms must be statistically independent of one another. 4. The error terms must be uncorrelated with the independent variable. 5. The relationship between variables or transformed variables must be strictly linear. One possible source of error arises from violation of the independence assumption. The theoretical model postulates the effects —81- of several multi—dimensional variables. Empathy (Rogers, 1957) consists of understanding and communication of understanding of both the role and the task of the buyer. Credibility is considered in terms of a three—variable model consisting of perceived qualification, trustworthiness and dynamism (Berlo, et. al., 1966). While the dimensions of these variables are conceptually separate, there is evidence in the case of credibility, for example, that the dimensions may not occur independently of one another (Mertz, 1966). Therefore it cannot be stated with certainty that, for the phenomenon under consideration, the explanatory variables can be found independently. IA salesman with high credibility on one dimension may be more likely to be credible on another dimension as well. This presents the methodological problem of isolating the effects of an individual independent variable from the possible contaminating effects of other related variables. Partial Correlation The task of determining the correlation between two variables with the influence of a third eliminated can be accomplished by correlation of the deviations from two regression lines instead of from the array means (McNemar, 1969). The correlation of these deviation scores, which is designated by the symbol r (the 12.3 correlation between x1 and x2 with x3 held constant), is known as the partial correlation coefficient. The partial correlation coefficient is calculated by constructing "new” independent and dependent variables with the effect of the control variable removed. If x3 is unrelated to x and x2 the partial r will equal r 2, and if either r or r is 12.3 13 23 negative and r12 positive, ”partialling out" x3 will raise the l l correlation (McNemar, 1969). -32- Partial correlation can be used to identify three types of problems encountered in correlation analysis: (1) spurious correlation, (2) intervening variables, and (3) "hidden" relationships. Spurious correlation is a relationship between two variables (x1 and x2) where xl's correlation to x2 is solely the result of the fact that x1 varies with some other variable, x3, which is the true predictor of x If the 2. effects of x are controlled x2 no longer varies with x The vanishing 3 1' of the partial correlation is evidence of the spuriousness of the zero- order correlation (Simon, 1957). Partial correlation can also be used to identify intervening variables as is the case in the following situation: x1 leads to x2 which influences x3. In this situation the hypothesis is that a major proportion of the correlation between x1 and x3 is due to the impact of x1 upon the intervening variable x2 which is in turn a major predictor of x3 (Blalock, 1972). The third case involves the possibility that some variable acts to hide or suppress a relationship (Massey, 1962). This can be seen in the following Situatlon: x1 1 related to x3, which is in turn positively related to x2. correlation of x1 and x2 controlling for the effects of x3 would indicate that x1 is positively related to x2. In summary, a partial correlation can be used to: (1) locate a spurious correlation, shows no relationship to x2 because x is negatively The partial (2) determine the importance of an intervening variable, and (3) uncover a relationship where none appears to exist (Nie, Bent and Hull, 1970). McNemar (1969) cites a number of issues of importance when partial correlation techniques are used in determining the direction of causation. The partial correlation coefficient, r provides a 12.3’ -83- degree of correlation between x and x2 which would exist provided 1 variation in x3 were controlled. The interpretation of the partial r depends upon an understanding of the influence of x in producing 3 variation in x1 and x2. It would not make sense to interpret r12 3 as the correlation between x1 and x2 with the influence of x3 removed if it were not known that x3 could influence both x1 and x2. The relationships controlled for should be derived from the conceptual framework so that the partial correlation method is not simply an exercise in multivariate manipulation. The theoretical model should indicate that x3 can produce or contribute towards a correlation between x1 and x2 by way of x3 producing variations in both x1 and x2. The partial correlation coefficient based on a small sample can be tested for significance by the t—test. The calculated t—ratio is then compared with the standard Students' t distribution. When there is a directional nature to the alternative hypothesis, a one—tailed t-test is appropriate. The t-test has been shown to be powerful even in those situations where the assumptions are violated (Boneau, 1960). Although it is not clear that the phenomenon being considered in this study conforms to normal shape parameters this violation of assumption has been shown not to effect the t statistic drastically (Boneau, 1960). One assumption that may effect the t statistic, homoscedasticity, seems to exist in this situation. Table 5-1 shows that the assumption of equal variances of the error terms (homoscedasticity) seems to held since the distributions of the variables are relatively similar shapes. —84— TABLE 5-1 Variable 55321 Variance Skewness Qualification 4.51 2.50 —.641 Trustworthiness 4.67 .91 -.421 Dynamism 4.47 1.24 .032 Preparation 4.24 2.73 —.500 Role Empathy 4.32 .88 —l.ll7 Task Empathy 4.43 1.63 —.582 Message Content 45.23 239.31 .227 Evaluation 4.29 2.08 —.589 Intention 3.61 3.05 .305 Multiple Regression An extension of the Pearson product~moment correlation is multiple regression (Ezekial and Fox, 1959). The correlation coefficient provides a measure of the linear relationship between one independent variable and a dependent variable. Multiple regression allows the study of the linear relationship between a set of independent variables and the dependent variable, taking into account the interrelationships among the independent variables. A powerful variation of multiple regression is step-wise regression which provides a method of choosing independent variables which yield the best prediction of the dependent variable. In the first step the independent variable which is the best predictor of the dependent variable is chosen. The second independent variable added to the regression equation provides the best prediction -35_ in conjunction with the first variable. The algorithm procedes in this recursive fashion adding variables step by step until no other variables will make a significant contribution to the prediction equation. The significance of predictive value of adding an additional independent variable is done by considering the size of the regression coefficient associated with that independent variable. The significance of the size of the normalized regression coefficient (8 ) is measured by the F statistic. Adjustment of Correlation Coefficients There are two factors that may cause the correlation coefficient to be overstated: the sample size and the number of explanatory variables. A small number of observations and/or a large number of explanatory variables may cause the correlation coefficient and thus its significance to be overstated. Therefore adjusted zero-order and multiple correlation coefficients are utilized. Sample Size. If only two observations are taken from any population the resulting regression line will have a perfect fit and r2 = 1. Subsequent observations may or may not fall on this straight line. In general small numbers of observations tend to overstate the proportion of the variance in the dependent variable (Y) that is associatedwith the independent variable (X) in the universe from which the sample was drawn (Ezekiel and Fox, 1959). A correction for this can be obtained by adjusting the estimates of variance for the dependent variable, 5;, and the residual variance, the variance not accounted for by X, 82 (Ezekiel and Fox, 1959). y.x -86- Number of Explanatory Variables. The larger the number of independent variables considered, the greater will be the tendency the observed standard error to understate the true correlation in the universe (Ezekiel and Fox, 1959). It is possible to obtain a multiple correlation coefficient equal to one for n observations if there are n-l independent variables. Following from the procedure used to adjust for small samples, it is appropriate to compute the adjusted R2 based on degrees of freedom (Green and Tull, 1970). Testing Procedure The classical procedure of testing a hypothesized relationship begins with postulating a null hypothesis of no difference. An appropriate statistical test is then applied to determine if the hypothesized relationship could have occured by chance and if so, the null hypothesis is not rejected. If the null hypothesis is rejected there is some associated level of type I error (alpha), the likelihood of rejecting a true null hypothesis. Thus the classical statistical prOcedure consists of: (1) specification of some acceptable level of type I error, commonly .05 or .01, (2) calculation of an appropriate test statistic and, (3) acceptance or rejection of the null hypothesis depending upon whether or not the statistical test produced an & (alpha-hat) greater or less than the specified acceptable level of type I error. The research hypotheses in this study are stated in their alternative forms rather than as the null hypothesis of no difference. Application of the appropriate test statistic, resulting in rejection of the null hypothesis, causes acceptance of the alternative hypothesis as -37- stated. This manner of presentation is used for purposes of clarity. It will be the procedure in this study to report the alpha-hat (‘3) associated with each application of a test statistic. The alpha-hat (:3) statistic represents a degree of belief rather than a dichotomous accept-reject variable (Cohen, 1960). Implicit in the reporting of the results is an .01 criterion. However, the calculated alpha,-d (alpha- hat), allows the reader to judge the significance of the relationship reported. The results are presented in the following chapter. CHAPTER SIX RESULTS The linkages in the theoretical model are operationally defined in Chapter Four. Nine hypotheses are set forth as a test of the model. Figure 6-1 illustrates the relationships under consideration. Figure 6-1 Summary of Research Hypotheses Role Empathy Task Empathy Expertness Trustworthiness ——’- Evaluation—4 Intention Dynamism ‘l Preparation / J Message Content . 4.. + -39- Hypothesis One: The buyer's beliefs regarding the salesman's role and task empathy are associated with the salesman's message content score. Table 6-1 displays the zero-order Pearson product-moment correlation coefficients (adjusted for sample size, N = 31) between role empathy, task empathy and message content. The adjusted correlation coefficients (5) provide a better estimate of the "true" correlation in the population when the sample size is small (Ezekiel and Fox, 1959). TABLE 6-1 CORRELATION COEFFICIENTS ADJUSTED FOR SAMPLE SIZE (N=31) AND PROBABILITIES OF TYPE I ERROR 5(a) Role Task Message Empathy Empathy Content 1. Role Empathy 1.00 .8655 .6298 (.001) (.001) 2. Task Empathy 1.00 .6437 (.001) 3. Message Content i 1.00 Both role empathy and task empathy are significantly associated with the message content; however, significant correlation exists between the dimensions of empathy. The associations with message content‘ may exist due to a spurious correlation (Simon, 1957) or one of the dimensions of empathy may be an intervening variable between the other dimension and the message content (Massey, 1962). It might also be the case that both role empathy and task empathy are measures of the same construct. Partial correlation can be used to identify these types of -90- problems encountered in correlation analysis. Table 6-2 shows the partial correlation coefficients between each dimension and the message content score while holding the other dimension constant. TABLE 6-2 PARTIAL CORRELATION COEFFICIENTS ADJUSTED FOR SAMPLE SIZE (N=31) Variables Partial Correlation Probability of Type Coefficients I Error (9“ j 1. Role Empathy £13 2 = .1896 5.158 2. Task Empathy £23 1 = .2534 5.089 3. Message Content The partial correlation coefficients do not suggest the existence of a spurious relationship nor do they support an intervening relationship between one dimension of empathy and the message content score while controlling for the other. Table 6-1 shows that both role empathy and task empathy dimensions have significant associations with the message content score. However, role empathy and task empathy are more highly correlated with each other than with the message content score. The multicollinearity between the two dimensions of empathy may obscure the relationship with message content. In a multiple regression role empathy may be partialed out of the relationship because it is highly correlated with task empathy which has a slightly higher correlation with message content. The multicollinearity and the not significant partials suggest that role empathy and task empathy may measure the same thing relative -91- to the message content. The null hypothesis of no association is rejected for both role empathy and task empathy, and message content. However, role and task empathy are not independent and should not both be used in a predictive equation. Since task empathy is a better predictor of message content, it will receive the emphasis. Figure 6-2 represents the pattern of the relationships. FIGURE 6-2 CORRELATION COEFFICIENTS(§) AND PROBABILITY OF TYPE I ERROR(3£) (ADJUSTED FOR SAMPLE SIZE, N=31) .8655 Role Empathy ( a). Task Empathy (.001) .6437 .6298 (.001) in- Message Content (.001) Hypothesis TWO: The buyer's beliefs regarding the salesman's source credibility on the dimensions of: a) qualification b) trustworthiness and c) dynamism are associated with the salesman's message content score. Table 6-4 shows the zero—order Pearson product-moment correlation coefficients adjusted for sample size (N=31). The correction is obtained by increasing the estimate of the residual variance (Ezekiel and Fox, 1959). In some cases (i.e. trustworthiness and dynamism, -92- dynamism and message content) the correction increased the estimate of the residual variance to a value greater than one. In these cases the . . 2 . correlation was taken as equal to zero Since r cannot be negative. TABLE 6—4 CORRELATION COEFFICIENTS ADJUSTED FOR SAMPLE SIZE (N=31) AND PROBABILITTES OF TYPE I ERROR ' 5(a) Expertness Trustworthiness Dynamism Message Content Expertness 1.00 .6036 .4097 .7049 (.001) (.012) (.001) Trustworthiness 1.00 .0000 .5370 (.500) (.001) Dynamism 1.00 .0000 (.500) Message Content 1.00 It appears that expertness and trustworthiness are associated with the message content and that no relationship exists between dynamism and the message content. However the former relationship may be spurious (Simon, 1957) and the latter may be a suppressed relationship (Massey, 1962). Because of the intercorrelations between the dimensions of credibility may confound the zero-order correlations, partial correlation coefficients are calculated. -93- TABLE 6—5 PARTIAL CORRELATION COEFFICIENTS ADJUSTED FOR SAMPLE SIZE (N=31) Variables Partial Correlation Probability of Type Coefficients I Error (2,) = 5 1. Expertness r14.23 .6600 .001 ' : 5 2. Trustworthiness r24.13 .0539 .391 ‘ :_ 5 3. Dynamism r34.12 .4115 .014 4. Message Content Controlling for the effects of perceived trustworthiness and perceived dynamism, perceived expertness still shows a significant association with the message content score. Based on the not significant partial (r24.13) the association of perceived trustworthiness and message content seems due to the correlation of both with perceived expertness. The null hypothesis of no association is rejected for perceived expertness and message content score but is not rejected for perceived trustworthiness and message content score. When the effects of perceived expertness are controlled, perceived dynamism has a significant, negative association with the message content score. The hypothesis of association between dynamism and message content score is accepted, given that the dimension of expertness is present. For a given level of expertness higher message content scores are associated with lower levels of dynamism. The message content score is positively associated with a lack of dynamism, i.e., nondynamism. This relationship can be seen in the regression equation (Table 6—6). _94- TABLE 6—6 REGRESSION EQUATION ADJUSTED FOR SAMPLE SIZE (N=31) Equation i2 F & Message Content = 8.2918(Expertness) 41.51 5.001 -4.8296(Dynamism) 6.97 5.01 +29.4225 .5972 20.75 5.001 The regression equation accounts for over 59% of the variance in the message content score. The overall F-ratio is significant at.the .001 level. The F-ratios for the coefficient of expertness and dynamism are significant at the .001 level and the .01 level respectively. The null hypothesis between a combination of the dimensions of credibility and the message content is rejected. Figure 6-3 illustrates the relationship. FIGURE 6-3 CORRELATION COEFFICIENT (E) AND PROBABILITY OF TYPE I ERROR (a3) (ADJUSTED FOR SAMPLE SIZE, N =31) N (l) Expertness 1 + ’7728 .4,- Message Content (.001) (2) Nondynamist -95- Hypothesis Three: The buyer's belief regarding the salesman's preparation and planning is associated with the salesman's message content score. Table 6-7 shows the regression equation between the message content scores and the buyer's belief of the salesman's preparation and planning. TABLE 6-7 REGRESSION EQUATION ADJUSTED FOR SAMPLE SIZE (N=31) Equation r2 F 6 Message Content = 6.2163(Preparation) .4415 22.92 5.001 +18.8887 The above equation accounts for 44% of the variance in message content scores using preparation and planning as a predictor variable. The equation is significant at the .001 level. The null hypothesis of'no association between preparation and message content score is rejected. Figure 6-4 illustrates the relationship. FIGURE 6-4 CORRELATION COEFFICIENT (E) AND PROBABILITY OF TYPE I ERROR ( a) (ADJUSTED FOR SAMPLE SIZE, N=3l) .6644 Preparation ear- Message Content (.001) -96.. Hypothesis Four: The buyer's beliefs regarding the salesman's role and task empathy are associated with the buyer's evaluation of the salesman. Table 6-8 shows the matrix of adjusted zero-order Pearson product- moment correlations. Since the sample (N=31) is small, the adjusted coefficients (r) provide a better indication of the degree of association in the population. TABLE 6—8 CORRELATION COEFFICIENTS ADJUSTED FOR SAMPLE SIZE (N 31) AND PROBABILITIES OF TYPE I ERROR Eca) Role Task Empathy Empathy Evaluation Role Empathy 1.00 .8655 .7897 (.001) (.001) Task Empathy 1.00 .8800 i (.001) Evaluation 1.00 The adjusted zero-order Pearson product—moment correlation coefficients between role empathy and evaluation, and between task empathy and evaluation are both significant at the .001 level; however, so is the correlation coefficient between role empathy and task empathy. Partial correlation coefficients are computed (Table 6-9) to test if the relationships are spurious or if an intervening relationship exists. -97.. TABLE 6-9 PARTIAL CORRELATION COEFFICIENTS ADJUSTED FOR SAMPLE SIZE (N=31) Variables Partial Correlation Probability of Type I Coefficients Error (61L 1. Role Empathy 513 2 = .1179 5.268 2. Task Empathy £23 1 = .6359 ‘.001 3. Evaluation Table 6-9 shows the partial of (l) role empathy and (3) evaluation, while holding (2) task empathy constant, is not significant. The partial Of (2) task empathy and (3) evaluation, holding (l) role empathy constant, is significant at the .001 level. It appears that task empathy is an intervening variable between role empathy and evaluation. The intervening relationship is shown in the regression equations (Table 6-10). TABLE 6-10 REGRESSION EQUATIONS ADJUSTED FOR SAMPLE SIZE (N=31) . —2 , Equation r F g 1. Evaluation = 1.2157(Role Empathy)- .9606 .6236 48.05 5.0013 2. Evaluation = .9927(Task Empathy)- .1051 .7744 99.55 $.001 3. Evaluation = .8836(Task Empathy) 19.37 5.001 .1721(Role Empathy) .39 5.27 — .3651 .7744 48.93 5.001 The first equation explains 62% of the variance in evaluation and the —98- second explains 77%. Both regression equations have F ratios which are significant. However, the multiple regression equation using both task empathy and role empathy does not explain any more of the variation in evaluation than does the regression using task empathy alone. In the third equation the F ratio of the coefficient of role empathy is not significant. Role empathy and evaluation have a significant zero—order correlation coefficient but this seems to be due to the effects of the significant correlation between role empathy and task empathy. The empathy specific to the buyer--seller situation (task empathy) is" significantly associated with generalized empathy between seller and buyer (role empathy). The not significant partial between role empathy and evaluation suggests that task empathy may be an intervening variable between the two. The effect of role empathy upon evaluation may exist because role empathy is associated with task empathy, which, in turn, is associated with evaluation. On the strength of the correlation coefficients the null hypothesis of no association between task empathy and evaluation is rejected. Due to the not significant partial correlation coefficient and not significant coefficient in the regression equation, the null hypothesis of direct association between role empathy and evaluation is not rejected. The association between role empathy and evaluation exists because both are associated with task empathy. Figure 6-5 illustrates the relationship. -99- FIGURE 6-5 CORRELATION COEFFICIENTS (I) AND PROBABILITY OF TYPE I ERROR (& ) (ADJUSTED FOR SAMPLE SIZE, N=31) .8655 Role Empathy .v4.. Task Empathy (.001) .8800 (.001) Evaluation Hypothesis Five: The buyer's beliefs regarding the salesman's source credibility on the dimensions of: a) expertness b) trustworthiness and c) dynamism are associated with the buyer's evaluation of the salesman. Table 6-11 shows the zero-order Pearson product-moment correlations adjusted for sample size between evaluation, expertness, trustworthiness and dynamism. The zero-order correlations suggest significant assOciation with evaluation for expertness and trustworthiness, but not for dynamism. However, significant correlation also exists between the dimensions Of credibility. -100- TABLE 6-11 ZERO-ORDER CORRELATIONS ADJUSTED FOR SAMPLE SIZE (N 31) AND PROBABILITIES OF TYPE I ERROR r(i&) Expertness Trustworthiness Dynamism Evaluation Expertness 1.00 .6036 .4098 .8833 (.001) (.012) (.001) Trustworthiness 1.00 .0000 .6448 (.500) (.001) Dynamism 1.00 .1028 (.292) Evaluation 1.00 While dynamism and trustworthiness are not significantly associated, expertness shows a significant correlation with trustworthiness and dynamism. 'Because of this intercorrelation, partial correlation coefficients are calculated (Table 6-12) between each dimension of credibility and evaluation while holding the other two dimensions constant . -101- TABLE 6—12 PARTIAL CORRELATION COEFFICIENTS ADJUSTED FOR SAMPLE SIZE (N=31) Variables - Partial Correlation Probability of Type I Coefficients Error {&1 _ _ = ‘ 1. Expertness r14.23 .8725 .001 2. Trustworthiness r24.13 = .1238 . -.262 3. Dynamism r34.12 =-.5620 ‘.001 4. Evaluation The partial correlation coefficients indicate that the association of trustworthiness and evaluation is due to the association of both with expertness. The association of expertness and evaluation is significant. Perceived dynamism has a significant negative association with evaluation when the effects of perceived expertness and perceived trustworthiness are held constant. For a given level of expertness and trustworthiness increasing dynamism is associated with decreasing evaluation of the salesman. These associations produce the following regression equations (Table 6-13). - TABLE 6-13 MULTIPLE REGRESSION EQUATION ADJUSTED FOR SAMPLE SIZE (N=31) -2 Equation r F & Evaluation = .9223(Expertness) 171.19 s.001 - .4039(Dynamism) _16.24 2.01 +1.9367 .8609 56.71 2.001 —102— The addition of trustworthiness to the multiple regression equation increases the f2 by only .0021. Despite the significant zero- order correlation the expected association between trustworthiness and evaluation is not supported. The not significant partial and the not significant addition to the regression equation suggest no direct relationship exists. The adjusted zero-order, partial, and multiple correlation coefficients all show a significant association between perceived expertness and evaluation. The expected hypothesis is accepted. The multiple regression equation shows dynamism with a significant negative coefficient, which follows from the degree of association seen in the partial correlation coefficient (Table 6-12). The effect of dynamism runs in the opposite direction to expertness. In and of itself, dynamism explains but 1% of the variation in evaluation. However for a given level Of expertness dynamism depresses the buyer's evaluation of the salesman. Dynamism shows a Significant negative association with evaluation when the effect of expertness is considered. The null hypothesis is rejected for this association of evaluation and the combination of expertness and nondynamism. Figure 6—6 illustrates the relationship. -103— FIGURE 6-6 CORRELATION COEFFICIENT (r) AND PROBABILITY OF TYPE I ERROR (‘1) (ADJUSTED FOR SAMPLE SIZE, N=31) Expertness + ' 9279 > Evaluation (.001) Nondynamismj Hypothesis Six: The buyer's belief of the salesman's preparation and planning is associated with the buyer's evaluation of the salesman. Table 6-14 shows the regression equation between the buyer's belief of the seller's preparation and planning and the buyer's evaluation of the salesman. TABLE 6-14 REGRESSION EQUATION ADJUSTED FOR SAMPLE SIZE (N=31) Equation f2 F 6 Evaluation = .7625(Preparation) .7643 94.04 $.001 +1.0599 The null hypothesis between the buyer's evaluation of the salesman and the buyer's belief of the salesman's preparation is rejected. The regression equation shows a significant association. Figure 6—7 illustrates the relationship. -104- FIGURE 6-7 CORRELATION COEFFICIENT (E) AND PROBABILITY OF TYPE 1 ERROR (a) (ADJUSTED FOR SAMPLE SIZE, N.31) . .8472 Preparation *9- Evaluation (.001) Hypothesis Seven: The salesman's message content score is associated with the buyer's evaluation of the salesman. Table 6-15 shows the regression equation between the message content score and the buyer's evaluation of the salesman. TABLE 6-15 REGRESSION EQUATION ADJUSTED FOR SAMPLE SIZE (N=31) -2 Equation r .F 8, Evaluation = .0678(Message Content) .5291 32.57 2.001 +1.2237 The null hypothesis between the buyer's evaluation of the salesman and the salesman's message content score is rejected. The F-ratio of the regression equation is significant. Figure 6-8 illustrates the relationship. ~105- FIGURE 6-8 CORRELATION COEFFICIENT (E) AND PROBABILITY OF TYPE I ERROR (a ) (ADJUSTED FOR SAMPLE SIZE, N=31) .7273 Message Content -_4.. Evaluation (.001) Hypothesis Eight: The buyer's evaluation of the salesman is associated with the buyer's intention to purchase. The zero-order correlation coefficient between evaluation of the salesman as a message source and intention to purchase is .5495, which 'is significant at the .001 level. Adjusting for the sample size (N=31) the correlation coefficient is reduced to .5271, which is significant at the .002 level. The resulting regression equation accounts for almost 28% of the variance in intention to purchase. TABLE 6-16 REGRESSION EQUATION ADJUSTED FOR SAMPLE SIZE (N=31) Equation . r2 F & Intention = .6379(Evaluation) + .8763 .2779 11.16 .01 Analysis of variance yields an F ratio which is significant at the .01 level. The expected hypOthesis is accepted. While there may be a number of other factors influencing the buyer's intention to purchase, there is, nevertheless, a significant association between the buyer's evaluation of the salesman and the buyer's intention to -106- purchase. The relationship is illustrated in Figure 6-9. FIGURE 6-9 CORRELATION COEFFICIENT (I) AND PROBABILITY OF TYPE I ERROR (& ) (ADJUSTED FOR SAMPLE SIZE, N=31) .5271 Evaluation S» Intention (.002) Hypothesis Nine: The buyer's evaluation Of the salesman is associated with: a) empathy b) source credibility c) preparation due to the association with message content. Figure 6-10 shows the hypothesized relationships based on the theoretical model (Diagram 2-1). Hypothesis nine reconsiders the viability of hypotheses four, five and six in terms of an alternative explanation of the relationships. It may be the case that the direct associations of task empathy, expertness, nondynamism and preparation with evaluation do not exist. These assoCiations may be due solely to the relationships with the message content scores. Partial correlation coefficients are calculated (Table 6—17) to test this hypothesis. Table 6-17 shows that controlling for the effects of the message content does not decrease the association of task empathy, expertness, nondynamism or preparation with evaluation. This result suggests two types of association with evaluation. There are significant associations with evaluation through the intervening variable, message content. A pampcou memmoz 4/ T3 1 / \ 3V coaumhwmopm E f Emfiewcxpcoz coaumsfim>m.nml Illllllg coaucoucHnA + nmv mmocpHOQXm 8v 3 E - xcummam xmfl. A HIUHm OH mzo mmmmzfiom>z ”mmHmmZOHH44mm no >m<223m OHIO mmDuHm anummEm OHom -108— However, as an intervening variable message content does not account for all of the association of task empathy, expertness, dynamism and preparation with the buyer's evaluation of the salesman. There are significant direct associations in addition to the indirect associations. The relationships identified in Hypotheses Four, Five and Six cannot be explained solely in terms of an intervening variable. Hypothesis Nine is not accepted. TABLE 6-17 PARTIAL CORRELATION COEFFICIENTS ADJUSTED FOR SAMPLE SIZE (N=31) Variables Partial Correlation Probability of Type Coefficients I Error (5;) 1. Role Empathy r18.27 = .0385 5.422 2. Task Empathy r28.17 = .6104 5.001 3. Expertness r38.457= .7941 5.001 4. Trustworthiness r48.357= .1208 5.271 5. Dynamism r58.347=-.5126 5.003 6. Preparation r68.7 = .7623 $.001 7. Message Content 8. Evaluation Summary Of the nine research hypotheses eight were supported by the data. Figure 6—11 summarizes the significant relationships. Onoucou Owe-Two: in maoo.v goo. / Coo; nose-Hammad _ mnmn. “Hoo.v me. . 4 _ mmoo.v flaoo.v Emaeaaseeoz “Hoo.v 53:35 Al :oflmgmapm A . + $43. Humm. memo mmocuaoaxm maoo.c m a . . Hoe U ooww / zap-692m 632. A Vanessa-2m mfiom mmow. Aamuz .mNHm mgaz