‘ . “‘ .‘4 0' ' ,, n): vannflu 1%“ ":4 2“. i 4 mm \ .. .--.—'w. i J. 2- 2‘; 51m“ A L? 31?. u.’ uu x. —.v .u. A" .‘<‘ '11.: rL-‘¢’-.K 3 Z:uk'—f€\h£ ‘1‘!- ‘ .<.. . u u .- . u » .: (:5 913% "am ‘53)”.‘1; . . V4. ., .M. ’ qr, \v! 2? game 284% m a, «millilimlll l P01115321?“ ‘_ This is to certify that the dissertation entitled Face Strategies in Negotiation: Conatructing and Validating the goding Scheme for the 1-333. of Fir-nun and Flexibility presented by K11 Ho Kang has been accepted towards fulfillment of the requirements for Ph. D. degree in W101: Major professor Date (1/14/1991 MS U is an Affirmative Action/Eq ual Opportunity Institution 0-12771 - h-_-4- 4 - 4 H... MAD-d. —- *- C A- FACE STRATEGIES IN NEGOTIATION: CONSTRUCTING AND VALIDATING THE CODING SCHEME FOR THE IMAGES OF FIRMNESS AND FLEXIBILITY By Kil Ho Rang A DISSERTATION Sublitted to Michigan State University in partial fulfillaent of the require-eats for the degree of DOCTOR OF PHILOSOPHY Depart-ent of Communication 1991 45:?” 6 71?} ABSTRACT FACE STRATEGIES IN NEGOTIATION: CONSTRUCTING AND VALIDATING THE CODING SCHEME FOR THE IMAGES OF FIRMNESS AND FLEXIBILITY By Kil Ho Kang This study elaborated the conceptual definition of face in negotiation by integrating the ideas of nutual interdependency fro. the negotiation literature and the concept of face from the politeness literature. Face as the images of fir-ness and flexibility was defined as autonomy from both negotiators’ perspectives. Further, the coding scheme for facework in negotiation was developed on the basis of this new conceptualization. Then, this research eapirically tested the construct validity of this coding scheme. This study exanined the validity of the coding scheme by adopting both nomological and multitrait-multimethod approaches. Based on the prior findings about face strategies, task strategies, language intensity, and lexical diversity, face strategies were hypothesized to share moderate, positive associations with task strategies and language intensity, while face strategies were expected to have no correlation with lexical diversity. Coders who were blind to the purpose of this study coded face strategies, task strategies, language intensity, and lexical diversity from two naturalistic negotiations; one from the private sector and the other fro- the public sector. Judges also rated the degree to which these variables were perceived in the data. ii The findings from the MTMN approach showed that the coding scheme for face strategies had convergent and discriminant validity. The perceptual aeasure of face strategies was strongly correlated with the behavioral measure of face strategies, which evidenced the convergent validity of the new coding scheme. The correlations of face strategies between heteromethods (i.e., perceptual and behavioral measures) were larger than most correlations of heteroconstruct-mononethods (NTMM) and heteroconstruct-heteromethods (HTHM). These findings supported the discriainant validity of the new coding scheme. The results fro. the nomological approach also indicated that the coding schene for face strategies. The moderate correlations between face strategies and similar constructs (i.e., task strategies and language intensity) supported the validity of the coding schele for face strategies. Although face strategies were expected not to be correlated with lexical diversity, there were some significant correlations between these two variables. As a result, the overall results fron the nomological approach generally supported the construct validity of the coding ache-e for face strategies. Further, the results showed that the negotiator’s own and opponent’s dimensions are empirically as well as conceptually distinct. This finding inplies that the coding schene for face strategies was well conceptualized. Finally, the ilplications related to the findings and future research were discussed. Further, the general iaplications for the new coding scheme were suggested. iii Dedicated to my parents and wife, Kyung-Shin iv ACKNOWLEDGEMENTS Throughout my graduate education, I have been fortunate enough to have known good scholars and obtained endless support from my falily. The members of my committee, Dr. Steven Wilson, Dr. Willian Donohue, Dr. Jaaes Stiff, and Dr. Nancy Ainsworth are respected teachers as well as researchers. They provided ne with advanced knowledge on communication. They always contributed to the improvement of my knowledge. In addition, the quality of this dissertation was enhanced through their valuable advice and coaments. Special thanks are due then. Dr. Steven Wilson deserves additional mention. He served as ay invaluable adviser, dissertation director, and mentor. His sincere attitude on study and excellent abilities to do research has deeply influenced my graduate education. His numerous intellectual discussions helped me to formulate ay thinking with insightful perspectives. I an very grateful for his contribution to this dissertation and the general quality of my education. I extend my appreciation to ay parents and parents-in-law. They have sacrificed their welfare for Iy education. They have shown their deepest love and faith in ae. A debt of gratitude also is owed to Iy elder brother, Si-Ho and his wife, Hay-Ran. Without their encouragement and financial support, this dissertation could not have been coapleted. My special recognition is owed to my wife, Kyung-Shin. For lore than five years, she has showed endless support. She has always shared ay goal with me. She has also been a source of anxiety sharing. For my education, she has been willing to sacrifice her happiness and goals and take care of my son, Young-Souk almost by herself. This snall work is dedicated to the lost iaportant person, ay wife. Finally, heartfelt thanks are due more than a dozen coders who coded tedious and cumbersome transcripts. They spent much tine in coding during the period of winter when it was severely cold. Without their assistance, I could not have accomplished it. vi TABLE OF CONTENTS LI ST OF TABLES O O O O O O O O O O C O O O O O O O O O 0 CHAPTER I. A NEW CODING SCHEME FOR FACE STRATEGIES IN NEGOTIATION: MANAGING THE IMAGES OF FIRMNESS AND FLEXIBILITY . . . . . . . . . . . . . . . . INTRODUCTION . . . . . . . . . . . . . . . . . LITERATURE REVIEW . . . . . . . . . . . . . . Research on Face: The Negotiation Literature Valued identities in negotiation . . . Earlier studies on face . . . . . . . . Effects of concession-aaking on face . Effects of concern about face on negotiation behaviors . . . . . . . . . Situational factors enhancing concern about face . . . . . . . . . . . . . . Criticisa of negotiation studies . . . Research on Face: The Politeness Literature Earlier studies on face . . . . . . . . Lakoff’s rule-based approach . . . . . Brown and Levinson’s politeness theory Criticism of politeness studies . . . . A REVISED CONCEPTION OF FACE IN NEGOTIATION . FACE STRATEGIES IN NEGOTIATION: A NEW CODING SCHEME . . . . . . . . . . . . . . . . . . . . CONCLUSION . . . . . . . . . . . . . . . . . . II. ASSESSING THE CONSTRUCT VALIDITY: VALIDATION OF THE NEW CODING SCHEME . . . . . . . . . . . . . FRAMEWORK FOR TESTING CONSTRUCT VALIDITY . . . PERCEPTUAL MEASURES OF THE SAME CONSTRUCT . . MEASURES OF SIMILAR CONSTRUCTS . . . . . . . . Task Strategies: Pruitt’s Dual Concern Model Comparison between Task and Face Coding Scheme Orientation . . . Strategies . . . Tactics . . . . . vii Page ix OIOIOOODWH H 0° 10 11 11 12 14 15 16 17 21 22 24 25 25 27 27 29 30 CHAPTER Language Intensity . . . . . . . . . . . MEASURES OF DISSIMILAR CONSTRUCT . . . . . Lexical Diversity . . . . . . . . . . . GENERAL PREDICTIONS FROM THE MTMM APPROACH I I I I METHODS O O O O O 0 O O O O O O O 0 O O 0 0 Data Collection . . . . . . . . . . . . Negotiation from the private sector Negotiation from the public sector Unit of Analysis . . . . . . . . . . . Definition of strategy . . Unitizing reliability . . . Behavioral Measures . . . . . . Images of firmness and flexi Task strategies . . . . . Language intensity . . . Lexical diversity . . . . Perceptual Measures . . . . . l p. p q C O C 0 Images of firmness and f e Task strategies . . . . Language intensity . . Lexical diversity . . . p- o e o H. o a 0 Ho 0 o H. a o o e "a o e o "a e o a: o o o x. o o o p. o o o 0". o o o 0:. o o I v 0 RESULTS 0 O O O O O O O O O O O O 0 O 0 O 0 Descriptive Data . . . . . . . . Research Hypotheses . . . . . . Hypothesis one . . . . . . Hypotheses two and three . Hypotheses four, five, and six Hypotheses seven and eight . . Hypotheses nine and ten . . . . Research Question . . . . . . . . . O O 0 v. DISCUSSIONS O O O O O O O 0 O O O O O O O 0 Findings Regarding Research Hypotheses . Other Findings 0 O O O O O O O 0 O O O O Inplications . . . . . . . . . . . . . . Implications related to the findings General implications for the new coding Bebe-e O O O O O O O O O O O O O O O 0 NOTE 0 O O O O O O O 0 O O O O O O O O O O O O O O O O O APPENDICES O O O O O O O O O O O O O O O O O O O O O O 0 Bl BLIWRAWY O O O O O O O O O O O O O O O O 0 O O O O 0 viii Page 33 35 35 36 38 55 55 61 61 62 71 73 76 80 83 83 90 92 92 95 104 105 131 TABLE 1. 2. 3. 6. 7. 8. 9. 10. 11. LIST OF TABLES Nine Face Strategies Based on the New Conceptualization . . Means and Standard Deviations for the Behavioral and Perceptual Measures of All the Variables Shown in the Hymtheaes O O O O O O O O O O O O O O O O O O O O O O O O 0 Correlation Matrix for the Global-Behavioral and Perceptual Measures of All the Variables Shown in the Hypothesis . . . Correlation Matrix for Single-Unit Behavioral Measures of All the Variables . . . . . . . . . . . . . . . . . . Means and Standard Deviations of Face Strategies for a Negotiator’s Own and the Opponent’s Diaensions at Each Level of Task Strategies . . . . . . . . . . . . Crosstabulation of Face Strategies by Task Strategies Synthetic Correlation Matrix for MTMM . . . . . . . . LISREL Maxi-um Likelihood Estiaates . . . . . . . . . The Coaparison of Chi-Square between Two and Single Factor Model . . . . . . . . . . . . . . . . . . . . . . Sumnary Table of Results supporting the Hypotheses for Validity Test . . . . . . . . . . . . . . . . . . . . . Means and Correlations of Language Intensity and Lexical Diversity for the Nine Face Strategies . . . . . . . . . ix Page 18 58 64 66 67 70 77 82 82 85 100 Chapter One A New Coding Scheae for Face Strategies in Negotiation: Managing the Images of Firsness and Flexibility INTRODUCTION Negotiation is an essential process for resolving disputes at all levels of society. It ranges froa private matters to international affairs. For exaaple, husbands and wives negotiate where to go for summer vacation; automobile salespeople negotiate with their potential custoaers; diplonats negotiate about international probleas, etc. These negotiation situations share several com-on characteristics (see Lewicki & Litterer, 1985; Putnaa, 1989). First, at least two parties are involved in the situations. Second, each party perceives that their goals are inconpatible with the other party’s goals. Third, the parties are interdependent. That is, each party can influence the other’s outcoaes and their own outcoaes can, in turn, be influenced by the other. Fourth, in order to resolve their affairs, they exchange positions and proposals. Thus, negotiation can be defined as a process in which two or lore interdependent parties resolve conflicting goals through the exchange of arguaents and proposals (Putnaa, 1989). Since the 1970s, the study of negotiation has emerged as a concern of coamunication scholars (for reviews, see Donohue, Dies, a Stahle, 1983; Putnan & Jones, 1982; Rubin, 1983). One of the prilary topics for communication scholars has been the relationship between negotiation interaction and outcomes. Research has exaained the effects of bargaining strategies (Putnam & Wilson, 1988), argumentation (Putnam & Geist, 1985; Putnam, Wilson, Walt-an, & Turner, 1986), and inforaation 2 management strategies (Donohue & Diez, 1985) on the outcomes of negotiation. Recently, the importance of face strategies has been emphasized (Pruitt & Smith, 1981). Negotiation is a process in which participants pursue identity as well as instrumental and relational goals (Wilson & Putnam, 1990). Regarding the former objectives, negotiators show such concern about how they maintain their "face" or protect against loss of face. Hence, strategies for maintaining or saving face can affect the instrumental outcomes of negotiation. "In some instances, protecting against loss of face "swamps" the importance of the tangible issues at stake and generates intense conflicts that can impede progress toward agreement and increase substantially the costs of conflict resolution" (Brown, 1977, p.275). Accordingly, scholars often acknowledge the importance of face strategies within negotiation (for example, see Deutsch, 1969; Donahue a Diez, 1985; Rubin & Brown, 1975; Stevens, 1963). Previous studies, however, contain conceptual ambiguity in defining face. Naturally, this conceptual ambiguity has limited the develop-eat of systematic coding schemes for face strategies in negotiation. For these reasons, little research on the use of face strategies has been done. The purpose of this study is to clarify the concept of face in negotiation and to develop the conceptual rationale underlying a coding scheme for face strategies in negotiation. In order to achieve this purpose, a new conceptualization of face in negotiation will be developed by integrating the concept of face from the literatures on 3 negotiation and politeness. Based on the new conceptualization, a coding scheme for face strategies will be developed. In the second chapter, methods for evaluating the construct validity of the new coding scheme will be proposed. LITERATURE REVIEW e c e: e e t' ti e t e Negotiation studies on face have focused on several issues: what images negotiators attempt to project, how concession-making affects their images, what behaviors are associated with concern about face, and what situational factors increase concern about face. Valued identities in neggtiatigp. Generally, face refers to positive social values that persons claim for their public selves (Goffman, 1967). This broad definition of face has two inplications. First, face is an image of self that persons want to be known to others (Brown & Levinson, 1978; Lim, 1990). That is, face is a public image that persons desire to be approved by others. Second, face is composed of numerous identities or images. The specific qualities which are perceived as socially desirable vary across situations (Tjosvold, 1977; Tracy, 1990). For example, I might lose face if I appeared incompetent to my students or disloyal to my wife. Face is socially situated on the basis of what are desirable images in the specific situation. In negotiation, the images of firmness and flexibility are two of the most dominant or important appearances that participants try to maintain (Pruitt & Smith, 1981). The images of firmness and flexibility have been defined as the degree to which negotiators are willing to make concessions. Put differently, the image of firmness is the appearance 4 of being reluctant to make concessions, while the image of flexibility is the appearance of being willing to make concessions (Pruitt & Smith, 1981). Thus, the former underlines competitive moves or distributive bargaining, while the latter cooperative moves or integrative bargaining (Walton & McKersie, 1965). Negotiators attempt to manage the images of firmness and flexibility, since they are directly related to the outcomes of negotiation. In order to maximize their outcomes, negotiators maintain the inage of firmness, that is, of the reluctance to give ground. At the same time, however, they should show the image of flexibility, that is, of the willingness to cooperate, in order to prevent a deadlock which would result in losses for both negotiators. Thus, the images of firmness and flexibility are desirable images that negotiators seek to cultivate. Some negotiation scholars also emphasize the importance of looking strong and tough (Brown, 1968, 1977; Tjosvold, 1977) and appearing trustworthy (Wilson a Putnam, 1990). The images of firmness and flexibility, however, are closely related with these images. Looking tough is conceptually equivalent with the image of firmness. Appearing trustworthy is tied to the image of flexibility, since negotiators need to be perceived as sincere in order to appear genuinely ready to make concessions and to bargain in good faith (i.e., to appear flexible). In sum, the images of firmness and flexibility are two of the most important appearances that negotiators attempt to manage. In the remainder of this section, various findings related to negotiator’s images of firmness and flexibility will be reviewed. 5 £§§11_§§gd1g§_gn_1§gg. Students of negotiation have recognized that face plays an important role in negotiation (Donahue & Diez, 1985; Rubin, 1983; Tjosvold, 1983; Wilson, Meischke, & Kim, 1990). Early studies (e.g., Deutsch a Krauss, 1962, Hornstein, 1965; Benton & Druckman, 1973; Druckman & Bonoma, 1976) usually drew on the concept of face post hoc in order to interpret their findings about negotiators’ behaviors. For example, Deutsch (1961) explained negotiators’ frequent choice of competitive tactics in terms of their face-maintenance motive. Although these interpretations are very suggestive, the studies do not provide direct empirical evidence that concern about face influences the process of negotiation. EIfggtg_g1_ggaggggign;ggkigg_gn_1gg§. A few studies have attempted to directly investigate the relationship between movement on instrumental issues and perceptions of firmness/flexibility. For example, Hiltrop and Rubin (1981) investigated the relationship between position and image losses. In this study, subjects were asked to read a story describing a dyadic bargaining episode and to take the role of one of the disputants. They then rated the weak or strong images associated with each of a series of offers. As expected, position loss and image loss were found to be positively correlated. That is, the more favorable a particular offer was to the buyer, the greater was the buyer’s rated appearance of strength. However, beyond a certain point, continued surrender of position was not associated with continued loss of image. These results suggest that subjects perceived a threshold point, that is, a point at which they had no more "face" to lose. Pruitt and Johnson (1970) also found that bargainers expected to 6 associate their own increased concessions with increased loss of face. In this negotiation experiment, subjects who played a role as buyers were given more noney when they made an agreement than when they did not. Confederates who were always sellers made concessions by preprogrammed fashion. Then, various time pressure was imposed on subjects in order to manipulate the degree to which they needed to make a concession. In the high time pressure condition, the subjects were told that they had very little tine left to reach agreement and a loudly ticking timer was set in motion. In the low time pressure condition, they were told that they had plenty of time left to reach agreement. At the end of the simulated negotiation, the subjects were asked to rate their personal strength. The results showed that the subjects felt weaker under high time pressure condition where more and larger concessions were made than under low time pressure where fewer and smaller concessions were produced. In sum, the findings on the relationship between concession-making and face indicate that increasing concession-making is associated with increasing loss of face. Although these studies directly examine the relationship between negotiator’s face and task-related behaviors, they do not specify what behaviors are enacted when negotiators attempt to maintain/regain face. The first effort to observe negotiators’ face-saving and face-restoring behaviors was made by Brown (1977). Arguing that face-maintenance motives . frequently induce negotiators to save or restore their face, Brown enumerated verbal and nonverbal expressions which he believed were related to face-saving and face-restoration. These behaviors are 7 examples of what Goffman (1955) labelled facework: actions described to help one or both parties save face. According to Brown, face-saving behaviors reflect actions designed to hide or soften an appearance of weakness, whereas face-restoring behaviors reflect actions designed to reestablish negotiators’ strength after the negotiators feel that they have been damaged. For example, ’avoiding discussion of specific issues,’ ’credentialing disclaimers (e.g., I’m no expert, but ...),’ and ’withholding certain information’ are regarded as face-saving behavior, since these behaviors ward off anticipated damage to negotiator’s personal image or reputation. ’Giving warnings of a future resistance or noncooperativeness’ and ’threatening an opponent’ are examples of face-restoring behavior. These behaviors purport to reassert negotiators’ strength. Although his analysis deserves much attention in the sense that it directly identifies negotiators’ face-related behaviors, it does not address why negotiators save or restore their face. Some studies (Brown, 1968; Carnevale et als., 1979; Tjosvold, 1977) have been designed to investigate situational factors which motivate negotiators to maintain their face. These studies typically observe behaviors which imply a concern for saving face such as threats or refusals to make concessions/compromises under the presence and absence of certain situational conditions. For example, Carnevale, Pruitt, and Britton (1979) found that negotiators take tougher positions when their constituents monitor their behaviors than when their constituents do not. 8 Other studies report that negotiators become sore concerned about their face when they believe that they: (a) have lost their face by looking foolish to their audience (Brown, 1968), (b) have been affronted by an opponent of lower status (Tjosvold, 1977), (c) are perceived as weak by their opponents (Tjosvold & Huston, 1978), (d) represent ingroup members rather than outgroup members (Breaugh a Klimoski, 1981), or (e) are highly accountable to their constituents (Roloff a Campion, 1987). In contrast, third party intervention can help negotiators reach agreement, when they face tension between a need to make concessions and a need to save face (Donohue, Allen, a Burrell, 1988; Pruitt & Johnson, 1970). In sum, scholars of negotiation have examined the effects of concession-making on negotiators’ face, the effects of concern about face on negotiators’ behaviors, and situational factors-which activate negotiators to maintain their face. However, there is little research on how negotiators employ message strategies to maintain or save their own or their opponent’s face. This may be because the concept of face in negotiation contains some limitations. C i ' ° ' t ' . Studies on the role of face in negotiation contain a number of conceptual and methodological problems. Conceptually, the way in which face has been defined is problematic. As argued above, face in negotiation can be regarded as the images of firmness and flexibility (Pruitt a Smith, 1981). These authors, however, fail to distinguish between qualitatively different ways of being firm or flexible. They imply that if a negotiator is not willing to concede his/her own proposal or prior position, then the negotiator’s 9 image is firm. According to their view, the image of firmness and flexibility is a unidimensional concept which lies on the continuum of reluctance-willingness to make concessions. Since negotiators are highly interdependent on each other to achieve their goals, however, each negotiator’s image can affect the opponent’s image. Since interdependence between negotiators creates a need to manage both negotiators’ inages of firmness and flexibility, face strategies in negotiation contain information about both negotiators’ face, that is, the images of ’I’ and ’YOU’. For example, negotiators can stand firm eithgz by arguing that they won’t concede their prior position 9; by forcing their opponents to concede their prior position. Additionally, a negotiator can be "firm" about both negotiators’ images by introducing new alternatives which satisfy both negotiators’ demands. In sum, while negotiators may enhance their images of firmness or flexibility, they can also influence their opponent’s images of firmness and flexibility. In this sense, the images of firmness and flexibility are two dimensional. Thus, face strategies in negotiation should take such qualitative differences into account. Methodologically, prior studies have not employed detailed measures of facework. One limitation is that scholars have not developed comprehensive coding scheaes for coding facework present in all negotiators’ behaviors. For example, studies of situational features have focused on only a limited range of behaviors, such as threats or refusals to make concessions. Brown (1977) also addresses only a few face-saving and face-restoring behaviors. Accordingly, the face-related 10 behaviors which appear in every turn cannot be coded with his analysis. In addition, Brown’s categorizations of negotiator’s face-related behaviors are not exhaustive. His scheme does not contain face-lowering behaviors such as making concessions for avoiding deadlock. Finally, Brown’s analysis does not consider how negotiators’ behaviors for managing their own face influence their opponent’s face. A second methodological limitation is that negotiation research has not considered the relationship between the images of firmness and flexibility and linguistic style. That is, studies have employed coding schemes which analyze gross categories of talk without examining linguistic variations in the range of messages which are members of those categories (Jackson & Jacobs, 1983). For example, Longabaugh (1963) has coded negotiators’ talk into ’accept,’ ’reject,’ ’seek,’ and ’offer’ categories. However, although a negotiator does not make any concession, he/she still can be somewhat flexible by altering the wording of his/her refusal (e.g., "I know what you mean, but I can’t accept your proposal"). Longabaugh’s system fails to distinguish various forms of rejections (or other categories), which does not provide a detailed description of how negotiators can be firm and flexible. A more detailed conceptualization of face in negotiation, i.e., the images of firmness and flexibility, can be drawn from the literature on politeness theory. e e: e ' r ture Politeness studies have focused on several issues. Specifically, early studies examined the relationship between politeness and expressions with various syntactic and lexical features and syntactic 11 rules of politeness employed by language users. Recent studies have developed a more comprehensive theory of politeness by taking a functional approach. r st ie n ce. Traditionally, politeness studies have tried to identify a hierarchical structure among expressions with different syntactic and lexical features in terms of their degree of politeness. For example, Jespersen (1964) and Quirk and Greenbaum (1973) found that past tense of the modal was more polite than present tense in requests. That is, the request "would you give me a book?” is more polite than the request "will you give me a book?" Brown and Gilman (1960) showed that an utterance with a minimal length was a less polite request structure than an utterance with a more than the necessary number of words. But, these studies also revealed difficulties in explaining politeness phenomena, since they were not =based on any theoretic perspective. ’ e-b h. A major effort to explain politeness phenomena systematically was made by Lakoff (1973, 1975, 1977). She tried to explain politeness in requests in terms of rules which dictated to the speaker how to act towards the bearer. She proposed three rules of politeness: "don’t impose," "give options," and "make the listener feel good." These were based on the assumption that increasing the addressee’s freedom to refuse the request would correlate with an increasing degree of politeness. Lakoff applied her rules of politeness to differentiate the degree of politeness in different syntactic strategies. For example, in making requests imperatives are less polite than declaratives, since 12 imperatives are more imposing than declaratives. And questions are more polite than declaratives, since they offer more options. Even though Lakoff’s rules of politeness successfully identify the various degrees of politeness in requests in terms of syntactic features, they do not seem to include the pragmatic aspects of politeness in communicative interaction, i.e., how the rules relate to communicative goals or intentions in interactive situations. Thus, while Lakoff’s efforts to establish a theoretical perspective on the politeness of request deserve much attention, the rules of politeness are so syntax-oriented that they cannot comprehensively explicate the complex phenomena of politeness. BMW! Brown and Levinson’s (1978, 1987) work on politeness overcame many of these limitations. The authors treat face as needs or goals that individuals attempt to attain and maintain during interaction. Brown and Levinson (1978) suggest that there are two types of face that can be found in many different situations: approval and autonomy. Approval as face, which they call positive face, refers to an individual’s desire for significant others to recognize and positively evaluate his/her abilities or attributes. Autonomy as face, which Brown and Levinson call negative face, represents an individual’s desire to have others recognize that he/she has rights to carry out behaviors or decisions free from interference. Scholars have attempted to construct coding schemes with which the degree of politeness contained in discourse can be effectively identified (Brown a Levinson, 1978, 1987; Goffman, 1955, 1967; Shimanoff, 1987; for a review, see Spradlin & Bhargava, 1989). For instance, Goffman (1967) categorized politeness strategies into face- 13 loss or ”wrong face," face-saving, face-attack or "aggressive facework," and face-giving strategies. Face-loss occurs when "an information is brought forth in some way about an individual’s social worth which cannot be integrated into the line that is being sustained for him" (Goffman, 1967, p. 8). Face-saving refers to "the process by which the person sustains an impression for others that he has not lost face" (Goffman, 1967, p. 9). Face-attack is used to ”introduce information favorable to himself and unfavorable to the others" (Goffman, 1967, p. 25). Finally, face-giving is to "arrange for another to take a better line than he might otherwise have been able to take" (Goffman, 1967, p. 9). Brown and Levinson (1978, 1987), building upon Goffaan’s idea of facework, proposed a more comprehensive and detailed coding schele. They assume that all people are concerned about their face, i.e., the self-image they present to others, and that people recognize that other individuals have similar face needs. Further, they argue that many communicative acts inherently threaten the face needs of one or both participants, especially the addressee. Thus, when a speaker performs a face-threatening act (FTA), the speaker needs politeness strategies to compensate for the hearer’s loss of face. In this sense, politeness strategies are the elements of messages which provide hearers’ face needs. Brown and Levinson divide politeness strategies into five superstrategies: don’t-do-the-FTA, do the FTA off-the-record, do the FTA with negative politeness, do the FTA with positive politeness, and do the FTA baldly, on-record. Each superstrategy is instantiated by numerous detailed "output strategies." For example, ”be pessimistic," 14 "ainimizing the inposition,’ and "apologizing" are output strategies for negative face, while "notice, attend to hearers," "intensifying interest ’ and "avoiding disagreement" are those for positive face. to hearers,’ Several recent studies (e.g., Baxter, 1984; Lim, 1989; Shimanoff, 1987) have applied Brown and Levinson’s ideas in particular contexts. For example, Baxter (1984) employed Brown and Levinson’s coding scheme in a compliance-gaining context. She found that subjects use both positive and negative face strategies in situations where they convince others to perform desired actions. Cri i i ' es ies. The politeness literature helps address many of the conceptual and methodological lilits of the negotiation studies, but this literature is not without its own limitations. Although Brown and Levinson and others acknowledge that a speaker’s and the hearer’s face is interdependent, this insight is not reflected in their coding schemes. That is, their categorical systems of face strategies are mainly developed in terms of a hearer’s face. Hence, these systems ignore how statements uttered by a speaker affect the speaker’s own face. One typical coding scheme showing this shortcoming is Shimanoff’s (1987) system which was constructed by applying Brown and Levinson’s ideas. Shimanoff (1937) codes conversations between married couples into face-honoring, face- compensating, face-neutral, and face-threatening categories. The underlying dimension of her coding scheme is the pleasant-unpleasant emotional effects of messages on hearers. If a speaker’s discourse arouses pleasant emotion in hearers, it is coded into face-honoring. If a speaker’s conversation induces unpleasant emotion in hearers, it is 15 regarded as face-threatening. Face-compensating and face-neutral are located between face-honoring and face-threatening. Consequently, her coding scheme does not consider the effects of a speaker’s discourse on his/her own face. In sum, since the coding systems from the politeness literature implicitly assume that a speaker’s discourse contains inforlation only about the hearer’s face, and not the speaker’s own face, they ignore the interdependence between interactants which is dominant in negotiation situations. By integrating concepts from the literatures on politeness theory and negotiation, we can overcome the current problems of an ambiguous conceptualization of the inages of firmness and flexibility in the negotiation literature, as well as a failure to explicitly recognize interdependence between both parties’ facework in the politeness literature. A REVISED CONCEPTION OF FACE IN NEGOTIATION The images of firmness and flexibility from work on face in negotiation can be conceptualized in terms of Brown and Levinson’s (1978) concept of autonomy. The image of firmness in negotiation results from attempts to maintain or protect the rights to behave independently and be unimpeded in pursuing a position, proposal, or course of action. The image of flexibility is just the opposite. That is, by giving up autonomy, negotiators cultivate the image of flexibility. Since negotiators are highly interdependent, they can do this in two ways. Specifically, the image of flexibility can be defined as the degree to which negotiators concede their own autonomy and protect their opponent’s autonomy. The image of firmness reflects the 16 degree to which negotiators maintain their own autonomy and constrain their opponent’s autonomy. In sun, the images of firmness and flexibility can be managed through negotiator’s autonomy. The firsness (flexibility) for a negotiator’s own dimension refers to greater (less) attempts to maintain his/her own autonomy. On the other hand, the firmness (flexibility) for the opponent’s dimension represents greater (less) attempts by the negotiator to aaintain the opponent’s autonomy. Accordingly, in this study, the images of firmness and flexibility will be used interchangebly with the term, "autonomy,” as a same construct. This definition merits some attention. First, the new definition emphasizes the interdependent aspect of negotiation. The definition highlights that a negotiator’s face can be aanaged through his/her opponent’s as well as his/her own behavior. Second, this definition is broad enough to reflect the usual sense of the images of firmness and flexibility. Third, the new conceptualization encourages a focus on how negotiators constantly create and sustain the images of firmness and flexibility through the details of language use, since it adopts the concept of autonomy froa the literature on politeness. Based on this conceptualization, a new coding scheme of face strategies in negotiation is proposed. FACE STRATEGIES IN NEGOTIATION: A NEW CODING SCHEME In the context of negotiation, face strategies can be categorized into three broad classes: face-saving, face-neutral, and face- threatening. Face-saving strategies enhance autonomy, face-neutral strategies sustain current levels of autonomy, and face-threatening strategies lower autonomy. These three categories can exist along two 17 dimensions which represent both negotiators’ perspectives. By crossing these two dimensions, we can create nine face strategies (see Table I). As shown in this table, the coding scheme indicates that negotiators can manage their own as well as their opponent’s face]. Specifically, while increasing their own firmness, negotiators also may enhance, neutralize, or lower their opponent’s face. Those three face strategies are named as integrating, defending, and distributing strategies, respectively. Likewise, without altering their current level of autonomy, negotiators also enhance, neutralize, or lower their opponent’s face which are called respecting, sidestepping, and attacking strategies. Finally, while negotiators lower their own autonomy so as to show their flexibility to their opponent, they can also enhance, neutralise, or lower their opponent’s face. Those three strategies are labelled as upgrading, disclosing, and discounting. Further, for each strategy, several tactics were developed. These tactics were partially elaborated from Brown and Levison’s (1978) and Brown’s (1977) coding schemes for face strategies. As argued above, however, these prior coding schemes are not comprehensive and exhaustive, since they have conceptual problems. In order to overtake these limitations, some tactics were constructed through actual observations on naturalistic and simulated negotiations. Appendix A summarizes the nine face strategies and tactics of each strategy. CONCLUSION As argued above, current research on face in negotiation has conceptual and methodological limitations. Conceptually, negotiation studies have not considered that interdependency between negotiators 18 Table 1 Nine Eaga Stratagiea Baaeg an the New Canceptaalizatton low constraint to self high Qppgnent’s Perspacttve Nagotiator’s Perapactive Firmness Neutral Flexibility low Firmness integrating respecting upgrading constraint Neutral defending sidestepping disclosing to other Flexibility distributing attacking discounting high Nata. The terms employed in this new coding scheme are technical terms only for this coding scheme. Put differently, these terms are different in meaning from the terms of everyday language. 19 creates a need to manage both negotiators’ images of firmness and flexibility. Methodologically, they have not developed comprehensive schemes for coding facework present in all of a negotiator’s behaviors. Further, they also ignore the relationship between the images of firmness and flexibility and linguistic style. Although studies on face from the politeness literature help address many of these limitations, their coding systems do not consider the effects of a speaker’s discourse on his/her own face, but only on the hearer’s face. In order to overcome these limitations, a new conceptualization of the images of firmness and flexibility was attempted. The images of firmness and flexibility were defined as autonomy from both negotiators’ perspectives. This definition resolves the prior conceptual problems by emphasizing the interdependent aspects of negotiation. A new coding scheme for facework in negotiation also was developed on the basis of the new conceptualization. The coding scheme reflects the interdependent aspects of negotiation; hence, it captures both negotiators’ perspective. In addition, the coding scheme is comprehensive in coding every turn in the negotiation. Since the new coding scheme supplements the prior limitations shown in previous coding systems, the coding scheme is likely to suggest new avenues for research on negotiation. For example, interaction patterns in negotiation can be investigated through the new coding scheme in terms of the images of firmness and flexibility. What patterns of facework strategies negotiators employ in order to maximize their task outcomes also can be examined. In addition, this coding scheme could be used to investigate how negotiators’ discourse contributes their 20 interpersonal relationship. In sum, the new coding scheme seems to have several implications for future research. However, prior to examining these topics as future research, it seems to be necessary that the new coding scheme should be validated. Thus, in the next chapter, a study investigating the construct validity of the new coding scheme will be presented. Chapter Two ASSESSING THE CONSTRUCT VALIDITY: VALIDATION OF THE NEW CODING SCHEME In the previous chapter, the images of firmness and flexibility were defined as a negotiator’s desire to keep his/her autonomy. This definition integrated the concepts of mutual interdependency from the negotiation literature and facework from the politeness literature. Based on this new conceptualization, a new coding scheme which consists of nine face strategies was presented. Each strategy is instantiated by various tactics. In considering this coding scheme, a key question must be raised: does this new coding scheme measure accurately what it intends to measure? This question of validity is central because its answer determines the grounds on which one assigns codes to a message or makes revisions in coding categories (Folger, Hewes, & Poole, 1984). Construct validity is one of the most useful and significant validity tests. Construct validity assesses the extent to which a purported measure of a construct is associated with other measures deduced from a theoretical framework. In other words, the validity of the measure is assessed via other measures which tap theoretically related or unrelated constructs (Nunally, 1968). The purpose of this section is to test the construct validity of the revised coding scheme of face strategies in negotiation. The constructs of task strategy, language intensity, and lexical diversity will be used to evaluate the construct validity of the new coding scheme. Specifically, those constructs will be theoretically and conceptually compared with the negotiator’s images of firmness and 21 22 flexibility. Then, the relationships between the images of firmness and flexibility and the other constructs will be empirically tested. Both perceptual and behavioral measures of each construct will be employed. Prior to predicting the relationships between the images of firmness and flexibility and other constructs, the framework for testing construct validity will be addressed. FRAMEWORK FOR TESTING CONSTRUCT VALIDITY As argued above, construct validity attempts to test the extent to which the measure is associated with other measures deduced from a theoretical framework. In general, there are two approaches to which construct validity can be tested: the nomological approach and the multi-trait multi-method (MTMM) approach. In this study, both approaches will be combined to test the construct validity of facework coding scheme. According to the first approach, nomological network should be considered in order to evaluate construct validity (Cronbach & Meehl, 1955). A noaological network refers to the predicted pattern of relationships that would permit naming a construct. In other words, a construct is defined implicitly by a network of associations. Accordingly, the validity of a proposed measure of that construct can be inferred from whether or not it produces the predicted relationships with measures of other constructs within the nomological network (Cook & Campbell, 1979; Cronbach a Meehl, 1955). If the relationships between the construct of interest and other constructs which are predicted on the basis of theory are empirically confirmed, then the proposed measure of the construct is likely to be valid. Thus, the nomological approach 23 demands theoretical consideration about whether the construct of interest should be associated with other constructs. According to the MTMM approach, construct validity can be subdivided into convergent and discriminant validity (Campbell a Fiske, 1959). Convergent validity is examined when we devise several different measures of the same construct. A valid measure should be related to other measures tapping the construct of interest. Discriminant validity is exaained when we compare the measure with measures of other unrelated constructs. A valid measure should be unrelated to measures of dissimilar constructs. To establish construct validity, both convergent and discriminant validity are required. The MTMM approach reconmends a validation process utilizing a matrix of intercorrelations among tests representing at least two traits, each measured by at least two methods. For evaluating convergent validity, the correlations between measures of the same -construct should be significantly different from zero. For evidence of discriminant validity, measures of the same construct should correlate higher with each other than they do with measures of different constructs which employ the same method (e.g., self report questionnaires). In this study, the nomological and MTMM approaches will be integrated in order to evaluate the construct validity of the new coding scheme. The underlying construct of the new coding scheme are the images of firmness and flexibility. For the nomological approach, three other constructs will be chosen: task-oriented behavior, language intensity, and lexical diversity. Then, the relationships between the 24 images of firmness and flexibility and these three constructs will be theoretically predicted. Further, since the images of firmness and flexibility are a two-dimensional construct, predictions will be made for each dimension. For the MTMM approach, each construct will be measured by two methods. That is, perceptual and behavior measures of each construct will be employed. PERCEPTUAL MEASURES OF THE SAME CONSTRUCT The new coding scheme attempts to tap the images of firmness and flexibility. In order to test convergent validity, the degree to which the images of firmness and flexibility are perceived in negotiator’s behaviors will be assessed. The behavioral and perceptual measures of facework are expected to be highly correlated, since both measures attempt to tap the same construct, i.e., the images of firmness and flexibility. While the new coding scheme intends to measure negotiators’ behaviors for managing the images of firmness and flexibility embedded in their messages, the perceptual leasure intends to identify the degree to which those images are perceived in negotiator’ behaviors. Since both aeasures attempt to tap the same construct, the results froa both measures are expected to be similar. Prior studies of conflict styles have found positive associates between behavioral and perceptual measures. For example, in order to assess the construct validity of Sillars’ (1980) coding scheme for classifying verbal behaviors in interpersonal conflict, Sillars, Coletti, Parry, and Rogers (1982) had subjects watch discussions in which disputants dominantly used one particular tactic of Sillars’ coding scheme. Then, subjects were asked to rate how they perceived the discussions along 25 theoretical underlying dimensions of the coding scheme. The results showed that observers’ perceptions were highly correlated with the coding scheme. Theoretically, negotiation behaviors simultaneously have relevance for both the negotiator’s and the opponent’s images of firmness and flexibility. Each tactic of the new coding scheme integrates these two dimensions. In order to test convergent validity, the perception of the images of firmness and flexibility needs to be rated along these two dimensions. If the new coding scheme is valid, the tactics of the new coding scheme will also show high correlations with perceptual measures of both the negotiator’s and his/her opponent’s autonomy. Based on the above rationale, the following hypothesis is proposed. HYPOTHESIS 1: Behavioral and perceptual measures of face strategies will share a strong, positive association for both the negotiator’s and opponent’s dimension. MEASURES OF SIMILAR CONSTRUCTS t t ' ' 'tt’ e1 Pruitt (1984) has developed a dual concern model for explaining negotiators’ task-related behaviors. According to Pruitt, four basic task strategies are available to negotiators: (a) problem solving, which involves an effort to find an alternative that is acceptable to both parties; (b) contending, which involves an effort to force one’s will on the other party; (c) yielding, which involves a reduction in one’s basic aspirations; and (d) inaction, which involves doing as little as possible in the negotiation. Further, he argues that these strategies are incompatible because they require different psychological orientations. 26 Pruitt’s dual concern model assumes that concern about both one’s own and the opponent’s outcomes affects each negotiator’s strategic behaviors. The dual concern model predicts that the: (1) problem- salving strategy will be used when concern about both one’s own and the opponent’s outcomes is high, (2) yielding strategy will be used when concern about only the opponent’s outcomes is high, (3) contending strategy will be used when concern about only one’s own outcomes is high, and (4) inaction strategy will occur when concern about both one’s own and the opponent’s outcomes is low. As his dual concern model shows, Pruitt’s concern lies in what motivational factors influence strategic choices. He does not, however, specify what tactics can instantiate each strategy. Because of this, none of his research testing the dual concern model has directly analyzed negotiator’s tactics or verbal interactions. He only suggests some examples of tactics for each strategy. If these examples can be regarded as a coding scheae for his strategies, then his coding scheme may be summarized as follows (Pruitt, 1984): (a) problen-solving strategy - conceding with the expectation of receiving a return concession; mentioning possible compromises as talking points; revealing one’s interest (i.e., one’s goals and values); sending disavowable intermediaries; talking in back channels; communicating through a mediator; retracting a proposal in the face of resistance; expressing a willingness to negotiate or compromise; promising reward if the opponent behaves in a stated manner; and searching for a mutually beneficial proposal. (b) contending strategy - making demands that far exceed what is 27 actually acceptable (i.e., one’s resistant point); making commitments to unalterable positions; making persuasive arguments aimed at convincing the other that concessions are in his/her best interests; using threats; demonstrating that there is more time pressure on the other than on oneself; retracting a previously nade concession so as to make the position clearly less agreeable to the opponent; rejecting the opponent’s proposal/position; demanding that the opponent sakes accommodations or concessions. (c) yielding strategy: making non-reciprocal concessions; agreeing to the opponent’s proposal/position. (d) inaction strategy: taking no action (i.e., not having enough motivation to negotiate. In the next section, Pruitt’s coding scheme and the new coding scheae will be compared in terms of orientation, strategies, and tactics. These two coding schemes can be compared with regard to orientation, that is, the goals which motivate the behaviors assessed by each coding scheme. Due to their differences in orientation, the negotiation strategies and tactics of both coding schemes also are contrasted. inantatiga. The underlying construct which Pruitt’s (1984) coding scheme taps is negotiators’ strategic behaviors for accomplishing task goals. Pruitt’s coding scheme is concerned with how negotiators resolve conflict between their own and their opponent’s task goals with regard to the outcones of negotiation. Thus, his coding scheme is oriented to 28 identify conflicting-resolving strategies for task issues. In contrast, the new coding scheme taps the images of firsness and flexibility, which are important components of each negotiator’s face wants. The new coding scheme attempts to identify how negotiators’ strategic behaviors continually influence their own as well as their opponent’s autonomy. In sum, whereas Pruitt’s coding scheme identifies strategic behaviors used to resolve conflicting taat_gaata, the new coding scheae examines strategic behaviors regarding the images of firmness and flexibility, one of the most important negotiators’ {aaa seala- These two goals should be moderately related. That is, task and face goals can be consistent or inconsistent. Task goals are tied to particular issues and future state of affairs that negotiators try to attain. Face goals are present in varying degrees throughout the negotiation of all task issues. Generally, since maintaining face goals is necessary to attain future task goals, they are positively associated with each other. For example, negotiators usually attempt to maintain their face partly due to the structure of negotiation, which forces participants to have mixed motives (Kelley, 1966; Podell & Knapp, 1969; Stevens, 1963). Their primary goal is to maximize their outcomes in the negotiation. In order to maximize their outcomes (task goal), they lust give a firm image to their opponent (face goal). However, to reach an agreement with their opponent (task goal), they also have to show a flexible image (face goal). As a result, to be effective, they must be firm without appearing too strong and flexible without appearing too conciliatory. In this sense, face goals seem to be related to task 29 goals. Under certain conditions, however, face goals may be pursued at the expense of task goals (Brown, 1968, 1977; Folger & Poole, 1984; Tjosvold, 1977). Face goals can be inconsistent with task goals when negotiators become highly aotivated to maintain their face. For exaaple, negotiators may refuse a proposal that meets their resistant point (task goal) when they feel their opponent has been aggressive in making demands (Tjosvold, 1977), when they are being monitored by their constituents or a third party (Brown, 1968; Pruitt et als., 1986; Roloff a Campion, 1987), when they are representatives of their group members (Breaugh & Klimoski, 1981), or when their own prior actions jeopardize their face (Tjosvold a Huston, 1978). In sum, face goals generally are related to task goals, but face goals can conflict with task goals when negotiators feel weak or are motivated to maintain their face. Sttatagiaa. Negotiation strategies refer to a game plan, or a set of tactics which negotiators employ to achieve their goals. As argued above, Pruitt’s coding scheme identifies four strategies for resolving conflict between participants’ task goals: contending, problem-solving, yielding, and inaction strategies. According to Pruitt, prior to or early in the negotiation, motivational and contextual factors lead negotiators to employ one specific strategy out of these four strategies. In this sense, each strategy is concerned with global behaviors which are performed consistently during the negotiation. It seems to be difficult to examine negotiators’ behaviors as a process with this perspective. On the other hand, the new coding scheme is designed to code 30 negotiators’ concrete behaviors regarding their images of flexibility and firmness embedded in their interaction. The new coding scheme divides negotiators’ inage-related behaviors into nine categories according to the degree to which negotiators show concern about their own and their opponent’s autonomy. In contrast with Pruitt’s strategies, these nine strategies are assumed to vary during negotiation. Thus, negotiators’ behaviors can be examined as a process. Iaatiga. Tactics refer to the communicative behaviors that operationalize strategies. Since Pruitt intends to code negotiators’ global behaviors, he pays little attention to tactics. Pruitt just suggests some examples of tactics of each strategy. For example, demands that far exceed resistant point and commitments to unalterable positions are tactics for the contending strategy, since those behaviors are directly related to maxiaizing one’s own task goals at the expense of the opponent’s task goals. Tactics from the new coding scheme focus on negotiators’ communicative behaviors which are related to their own and their opponent’s autonomy. For example, apologizing to an opponent, or revealing one’s own weakness (without making a concession) are tactics for disclosing, since those expressions directly lower a negotiator’s own autonomy without altering the opponent’s autonomy. When one compares tactics for the two coding schemes, it becomes apparent that they are quite different. Most of Pruitt’s tactics do not correspond directly to the tactics of the new coding scheme. That is, depending on how or in what context information is communicated, Pruitt’s tactics can be coded into various tactics from the new coding scheme. For example, threat, a tactic in a contending strategy from 31 Pruitt’s coding scheme, can be coded as a distributing strategy (e.g., "If you don’t accept my proposal, we will strike"), or as an attacking strategy (e.g., "If you don’t accept my proposal, you will have to take losses by closing the conpany for the present time"). Persuasive arguments, another tactic in Pruitt’s contending strategy, can be coded as a distributing strategy (e.g., "The price I suggest is the bottom price. You cannot find this product with this price anywhere"), or as a depending strategy (e.g., "The price I suggest is the bottom price. Actually, I have many customers who are looking for this product at this price"). Pruitt’s other tactics such as revealing one’s interest, demands that far exceed resistent points, and making concessions also do not equate with any single face tactics. Depending on how to be expressed, Pruitt’s tactics can be coded into various tactics from the new coding scheme. A few tactics from the new coding scheme, however, have moderately positive relationships with Pruitt’s tactics. For example, making concessions in Pruitt’s yielding strategy seems to be part of face strategies which show a negotiator’s own flexible image, since making concession is a behavior that lessens a negotiator’s autonomy by giving up his/her prior position. Although concession-making might be coded into strategies such as respecting, upgrading, or disclosing from the new coding scheme, that tactic would not be coded into distributing, attacking, or defending categories. Similarly, problem-solving strategy can be coded into integrating or discounting strategies, but not into distributing or upgrading categories. In sum, no one-to-one relationships exist between tactics froa Pruitt’s coding scheme and 32 those from the new coding schene, but moderate patterns of relationships may occur. Thus, two coding schemes nay share a moderate relationship. As suggested above, perceptual measure of task strategies also will be employed in order to evaluate validity through the MTMM approach. Since the perceptual measure of task strategies taps the same construct as Pruitt’s behavioral coding scheme, these two measures are expected to be highly correlated. Further, the perceptual measure also is expected to be moderately correlated with perceptions of face strategies. Theoretically, task behaviors simultaneously have relevance for concern about both the negotiator’s and the opponent’s outcomes. Each tactic of the Pruitt’s coding scheme integrates these two dimensions. In order to test convergent validity, the perception of negotiators’ task behaviors needs to be rated along these two dimensions. If the Pruitt’s coding scheme is valid, the tactics of the Pruitt’s coding scheme will also show high correlations with perceptual measures of concern about both the negotiator’s and his/her opponent’s outcomes. Based on this rationale, the following hypotheses are forwarded. HYPOTHESIS 2: Behavioral and perceptual measures of task strategies will share a strong, positive association for both negotiator’s and opponent’s dimensions. HYPOTHESIS 3: Behavioral measures of face and task strategies will share a moderate, positive association. Perceptual measures of these strategies also show a moderate, positive association. Specifically, H3a: Pruitt’s concession strategy will be moderately correlated with respecting, upgrading, and disclosing strategies from the new coding scheme. H3b: Pruitt’s problem-solving strategy will be moderately correlated with integrating and discounting strategies from the new coding scheme. 33 H3c: Pruitt’s contending strategy will be moderately correlated with defending, distributing, and attacking strategies from the new coding scheme. Languag§_lnisn§iix Language intensity refers to the quality of language which indicates the degree to which the speaker’s attitude toward a concept deviates from neutrality (Bowers, 1963). If a speaker chooses language which expresses more extreme positions towards an object, his/her linguistic expressions are regarded as more intense. For example, the statement, "It is very nice" is more intense than the statement, "It is nice." Although little research has examined the relationship between the images of firmness and flexibility and language intensity, studies which have examined the relationship between language intensity and aggressive behaviors indirectly suggest that a negotiator’s firm image will be positively associated with language intensity. A negotiator may foster his/her own firm image by stating his/her position more aggressively, advocating extreme positions without considering the opponent’s position, or deviating from the opponent’s positions/arguments. Indeed, prior studies (Greenberg, 1976; McEwen & Greenberg, 1970) indicate that when speakers state their attitudinal positions, speakers who adopt more intense language are perceived to be sore aggressive than those who use less intense language. Similarly, Donahue (in press) also found that couples reaching agreement during divorce mediation significantly reduced their language intensity over time, while nonagreement couples became much more intense from the start and middle parts of the interaction. Perhaps, agreement couples fostered their flexible image 34 over time by using less intense language. Thus, negotiators’ own image of firmness should be positively associated with their language intensity. In addition, a negotiator who attacks the opponent’s firm image in order to lower the opponent’s face might use more intense language than a negotiator who attempts to enhance the opponent’s face. Negotiators can attack the opponent by disagreeing with the opponent’s positions, criticizing the opponent’s arguments/positions with prejudice, or deviating from what the opponent has said. Accordingly, the image of "firmness" in upholding the opponent’s autonomy expected to be negatively associated with language intensity. As suggested above, perceptual measure of language intensity will also be employed in order to evaluate validity through the MTMM approach. Since the perceptual measure of language intensity taps the same construct as the behavioral measure, the two measures should be highly correlated. Further, the perceptual measure of language intensity also is expected to be moderately correlated with perceptions of face strategies. Based on the above rationale, the following hypotheses will be established. HYPOTHESIS 4: Behavioral and perceptual measures of language intensity will share a strong, positive association. HYPOTHESIS 5: Behavioral measures of a negotiator’s own firmness will share a moderate, positive association with the negotiator’s language intensity. HYPOTHESIS 6: Behavioral measures of a negotiator’s firmness towards the opponent will share a moderate, negative association with the negotiator’s language intensity. 35 MEASURES OF DISSIMILAR CONSTRUCT W For evaluating discriminant validity, lexical diversity is chosen. Lexical diversity refers to the manifest range of a source’s vocabulary (Bradac, Bowers, a Courtright, 1979). This range is quantified in the form of a type-token ratio: the number of different words in a message (types) divided by the total number of words (tokens). Usually, lexical diversity is operationalized as a mean segmental type-token ratio, for instance, the average ratio of types to tokens in samples of 100 words. Although both the new coding scheme and lexical diversity are language variables, the nine face strategies from the new coding schene are conceptually and operationally different from lexical diversity. The nine face strategies tap the degree to which the images of firmness and flexibility are reflected in a message, whereas lexical diversity taps a negotiator’s ability to command diverse vocabulary in a message. In order to measure these underlying constructs, the nine face strategies focus on both surface (i.e., linguistic expressions) and deep structure (i.e., content) in a message, while lexical diversity centers only on one linguistic component, vocabulary. The nine face strategies should not be correlated with the lexical diversity due to these conceptual and operational differences. Firm or flexible images can be communicated regardless of the range of vocabulary employed in a message. For example, in order to enhance a negotiator’s own face, the negotiator can simply say, "I won’t negotiate this," or ”I will argue vociferously until this proposal is accepted.” Thus, the nine face strategies are expected to have little relationship 36 with the lexical diversity. As suggested above, a perceptual measure of lexical diversity will also be employed in order to evaluate validity through the MTMM approach. Again, the perceptual measure of lexical diversity is expected to be highly correlated with the behavioral measure. Further, that measure is also expected to be unrelated to perceptions of face strategies. Thus: HYPOTHESIS 7: Behavioral and perceptual measures of lexical diversity will share a strong, positive association. HYPOTHESIS 8: Behavioral measures of face strategies and lexical diversity will be unrelated. Perceptual measures of these constructs also will be unrelated. GENERAL PREDICTIONS FROM THE MTMM APPROACH The relationships between the images of firmness and flexibility and three other variables have been addressed from the nomological perspective. Further, the high predicted correlations between the behavioral and perceptual measures of the same variables reflect the MTMM perspective. As argued above, the MTMM approach also demands discriminant validity in testing construct validity. For discriminant validity, monoconstruct-heteromethod correlations should be higher than heterocanstruct-heteromethod correlations, as well as heteroconstruct- monomethod correlations. Based on these conditions for discriminant validity from the MTMM approach, two additional hypotheses are proposed. HYPOTHESIS 9: Monoconstruct-heteromethod correlations, such as correlations between the perceptual and behavioral measures of the images of firmness and flexibility, will be significantly larger than heteroconstruct- heteromethod correlations, such as the perceptual measure of the images of firmness and flexibility 37 and the behavioral measure of lexical diversity. HYPOTHESIS 10: Monoconstruct-heteromethod correlations, such as the perceptual and behavioral measures of the images of firmness and flexibility, should be significantly larger than heteroconstruct-monomethod correlations, such as the perceptual measures of the images of firmness and flexibility and of lexical diversity. Aside from these hypotheses, we also can inquire about the relationships between being firm or flexible towards oneself versus one’s opponent. There is little evidence about whether these two dimensions, i.e., a negotiator’s and the opponent’s dimensions are orthogonal or oblique. Put differently, there is no empirical evidence about whether these two dimensions are correlated. In order to answer this question, the following research question will be raised. RESEARCH QUESTION 1: Are a negotiator’s and the opponent’s dimensions of face strategies correlated? Chapter Three METHODS The purpose of this study is to evaluate the construct validity of the new coding scheme for negotiator’s facework. The relationships between behavioral and perceptual measures of facework, Pruitt’s task strategies, language intensity, and lexical diversity were discussed in the previous chapter. In this chapter, methods for testing the construct validity of the new coding scheme will be addressed. Sources of data used and the unit of analysis for coding the data will be described. Then, behavioral and perceptual neasures of negotiator’s facework, Pruitt’s task strategies, language intensity, and lexical diversity will be explicated. t e t' n Data were collected on naturalistic negotiations from various published sources (e.g., negotiation books, articles, etc.). Specifically, two negotiations were chosen: one from the public sector and another from the private sector. As a matter of fact, the negotiation from the public sector was a simulated negotiation from expert negotiators, whereas the negotiation from the private sector was a naturalistic negotiation. Since the simulated negotiation from the public sector was conducted by the expert negotiators with several issues which they usually encountered in real negotiations, the negotiation was similar to a naturalistic negotiation. Bath negotiations cannot be necessarily regarded as a representation of negotiations. However, given that both negotiations were naturalistic and contained several topics which occurred in natural negotiations, 38 39 they seem to be typical enough to validate the new coding scheme for facework. In this sense, the data chosen have high external validity. The transcript from the public sector has 501 turns and 488 observational units, while that of the private sector contains 632 turns and 425 observational units. Consequently, a total of 913 observational unit were used as the data for this study. To assist in understanding the nature of these two negotiation situations, the following brief background information is provided. Nagatiattan_t;aa_tha_ptigata_aaatgg. The interaction from the private sector is an informal grievance negotiation in a co-ownership company (Morley a Stephenson, 1977). Three electricians participated in this negotiation with three management representatives. One representative from management had a six-year working relationship with one of three electricians. The negotiation occurred according to the electricians’ request. The main topic of this negotiation was the callout procedure for electrical fitters. Management’s position was that the electricians should be on stand by during bank holidays in preparation for emergency situations, while the electricians’ position was that they did not want to be on stand by for callout during bank holidays. The parties reached an agreement through this debate that management would increase the number of electrical staff by one and rotate responsibility for electrical coverage during bank holidays between the available staff. Nagatiatian_1:9a_tha_patlig_§agtgz. The interaction fro. the public sector was a simulated negotiation conducted by professional representatives of a Michigan school board for practice (Diez, 1983). 40 This negotiation was a complete contract negotiation. The professional negotiators were paired into school board and teacher negotiator teams. Then, they practiced this negotiation. The siaulation contained twenty unresolved issues which included binding arbitration of grievance, workload for teachers, evaluation of teachers’ ability, and salary. After heated debate, both parties reached agreements on a new contract. U 'ts o n s e 'n' 'o trate . A strategy were employed as the unit of analysis. A strategy is an abstract behavioral unit and provides general guidelines for action (Berger, 1985). A strategy refers to a sequence of actions or to a family of related actions (Wheeless, Barraclough, & Stewart, 1983). For example, the following statement is coded as a single strategy, since this statement consists of a sequence of sentences confirming what was just said. I’m not asking you if you’re a doctor. You just told as if a doctor agrees that David Kass is sick that you being management would agree to it. Is that not what you just said. The statement, "Well, I recognize that." also is another example of a strategy, since this statement can be a complete action of understanding what was just said. As shown in these examples, a strategy can vary in number of sentences. The unit of strategy also can be distinguished from turn and arguaent. A turn occurs each time that negotiators switch speaking roles. A strategy is a subjective unit in the sense that it is an 41 abstract behavioral unit that expresses intentions regarding face, while a turn is an objective unit in the sense that it is determined without reference to message content. Most often a single turn contains one and only one strategy. Sometimes, however, longer turns are likely to contain more than one strategy, since negotiators communicate various intentions within one turn. For example, in the following example a labor negotiator uses two strategies in a single turn, An article 26 employee who is granted a sick leave shall be paid for holidays following within the first 30 calendar days of his approved sick leave. 26 was read in such a way that its intent was to be interpreted. An article 20 is a second one. To be eligible for a holiday pay, an employee must work his last schedule day prior to his rescheduled work day after a holiday. This paragraph was written as a protection clause for the company. In other words, to keep employees from taking undue advantage of the company and prevent an excessive absenteeism around a holiday time. This turn contains two strategies, i.e., one strategy regarding article ZG and another strategy regarding article 2C. In this study, one constraint in coding units was established in order to avoid missing strategy units in longer turns. That is, turns which were longer than five sentences were automatically coded as containing two or more strategies. A ten sentence turn, for example, could be coded as containing two different instances of a discounting strategy, or one discounting and one yielding strategy. In contrast, multiple turns sometimes can include only one 42 strategy, even if they are spoken by different people on the same side. This would occur if one negotiator makes an argument to the opponent and then, another negotiator from the same side supports his/her colleague’s claim, as shown in the following of statements by two representatives for the teachers at the negotiation table. A: If you did ten percent across the board I think that would have done it. B: Ten percent across the board would have sold it. Speakers from opposing sides, however, cannot jointly produce one strategy. Most often a single argument consists of one strategy, since it usually expresses one intention. When negotiators attempt to transfer their intentions, their face strategy contains an explicit or implicit claim (i.e., what these negotiators believe) and evidence which supports the claim. Accordingly, one negotiator’s argument (i.e.,a claim and its supporting evidence) is regarded as one strategy. Negotiators, however, sometimes can manage their face strategies without providing clear, supportive evidence for their arguments. For example, negotiators may simply make an assertion without explicit reason giving ( e.g., a negotiator simply says, "I don’t want it"). Thus, a single argument or assertion is usually correspondent to one strategy. In sum, the coding scheme intends to code the extent to which negotiators’ message content contains firmness and flexibility. Since negotiators manage their images of firmness and flexibility through strategies, a strategy seems to be appropriate for coding the extent to which negotiators intend to show these images. In the next section, the 43 coding procedure for unitizing the data and unitizing reliability is addressed. itizi e i b' ' . For the behavioral measures, the author and one coder who was familiar with the concept of strategy independently unitized the entire data set. To enhance reliability, a codebook for unitizing was developed, as shown in Appendix A. The codebook presents the definition of a strategy, general rules for unitizing, comparisons of other units to a strategy unit, rules for overlaps (i.e., simultaneous talk) and specific cues for unitizing. Two training sessions were held before unitizing reliability was satisfied. In the first training session, the content of the codebook for unitizing was explained through examples. For both training sessions, the coders practiced coding units with about fifty turns of the actual data employed in this study. Following this training, the coders independently unitized the entire data set (i.e., both transcripts). To check unitizing reliability, Guetzkow’s U and unit-by- unit agreement were calculated. Guetzkow’s (1950) U, which indicates the degree to which two coders identify the same numbers of strategies in the entire negotiation, was .03. Unit-by-unit agreement estimates the degree of agreement between two coders in identifying specific segments of the entire negotiation. For calculating unit-by—unit agreement, sentences were used as an objective fixed unit. If both coders agreed that a strategy occurred or that a strategy did not occurred at the beginning or end of a sentence, this was counted as an agreement. If they disagreed, that was regarded as a disagreement. Following this procedure of calculation, the two coders showed 89.1 X of 44 unit-by-unit agreement. When disagreement occurred, the number of units was determined by discussion and consensus. These codings were used as the unit of analysis for all of the variables in this study. For the perceptual measures, a more global level of analysis was employed. Specifically, perceptions of the images of firmness and flexibility, task outcomes, language intensity, and lexical diversity were rated by judges for every 20 strategy units. It would be very difficult for judges to rate the degree of firmness, flexibility, intensity, etc. for every unit, since this work is boring and tedious. Further, such work may lower the reliabilities of the perceptual measures. It appears to be reasonable to use a more global unit of analysis for the perceptual measures than each strategy. Thus, 20 strategy units were employed as the global unit for the perceptual measures in this study. When the cutting point of the global units occurred within the same negotiator’s turn, the whole turn was included in the end of current global unit and in the first part of next global unit. Since the private sector negotiation contained 425 strategy units, there were 22 global units in this sector. On the other hand, the public sector negotiation contained 488 strategies and 25 global units. As a result, the total number of global units in the data was 47 units. In the next section, codings of the behavioral measures and ratings of the perceptual measures will be addressed. Wags Iaagaa_at_£1:aaaaa_aag_11agtatlity. For coding the images of firmness and flexibility from the unitized data, two different coders 45 who were blind to the purpose of this study were trained with the new coding scheme (see Appendix A). Eight training sessions were held before reliability was satisfied. In the first training session, the definition of the images of firmness and flexibility which underlie the new coding scheme was explained. In addition, the nine face strategies also were explicated through examples. For each training session, each coder practiced coding face strategies with about 50 strategy units of the actual data used in this study. The coding results which were not consistent between coders were discussed in order to clearly understand the nine face strategies. Training sessions in which coders practiced codings were continued until their coding results were reliable. Then, both coders coded the entire data set (i.e., 913 strategy units). After finishing codings, both coders’ results for 913 strategy units were compared to check reliability. Categorizing reliability, as checked with Cohen’s (1960) kappa, was .83. When disagreement occurred, the final codings were determined by discussion and consensus. To compute the global behavioral indices, the values of one, two, and three were assigned to lowering, neutralizing, and enhancing face categories, respectively for the self and opponent dimensions (see Table 1). For example, the strategy of "disclosing" was scored as "1" on the self dimension and "2" on the opponent dimension. Then all of the values assigned to tactics were averaged for every 20 strategy units which are one global unit. These average values were used as a behavioral index of more global images of firmness and flexibility. 46 Iaat_atgatagtaa. Pruitt’s coding scheme consists of four task- strategies. One of his strategies, the inaction strategy, cannot be employed in coding interaction since it refers to strategically avoiding negotiation with the opponent. In other words, the inaction strategy is literally a "no-interaction" strategy. Verbal interactions between negotiators already surpass the inaction strategy. Hence, it was not analyzed in this study. Instead, an "other" category was included in Pruitt’s task coding scheme, since acts such as procedural statements and questions of clarification can be irrelevant to task strategies. The codings of the "other" category were eliminated from statistical analyses which contain task strategies, since this category did not include any information about tasks. Because Pruitt’s coding scheme does not clearly specify an exhaustive set of tactics for each strategy, it had to be elaborated for the present research. For clarifying this coding scheme, exanples of tactics which Pruitt has used to illustrate each strategy were collected and relevant tactics from other coding schemes were added. Specifically, categories related to task goals from Putnam and Jones’ (1982), Sillars’ (1980), and Putnam and Wilson’s (1988) coding schemes were chosen to supple-ent Pruitt’s coding scheme. The modified coding scheme is shown in Appendix 8. Two additional coders were employed to analyze the data with this elaborated version of Pruitt’s coding scheme. For increasing the reliability of coding, four training sessions were held. For each training session, both coders practiced coding task strategies with about 50 strategy units of the actual data employed in this study. 47 Categorizing reliability, as calculated by Cohen’s kappa, was .87. When disagreement occurred, final codings were determined by two coders’ discussion and consensus. Two global behavioral indices were created separately along two dimensions which underlie Pruitt’s coding scheme: that is, concern about a negotiator’s own outcome and concern about the opponent’s outcome. One, two, and two were assigned to yielding, problem-solving, and contending strategies for the negotiator’s own dimension, respectively. Two, Two, and one were assigned to yielding, problem-solving, and contending strategies for the opponent’s dimension. The dual concern model suggests that the choice of these three strategies is determined under consideration of two underlying dimensions, that is, the degree to which a negotiator is concerned about the negotiator’s own and the opponent’s outcomes. When negotiators are highly concerned about their own outcome either contending or problem-solving strategies are chosen, while the yielding strategy is selected when the negotiators are less concerned about their own outcomes. Similarly, when negotiators are highly concerned about their opponent’s outcomes they choose yielding or problem-solving strategies, while when negotiators are less concerned about their opponent’s outcomes they choose the contending strategy. Thus, these weighted values differently assigned for negotiator’s own and the opponent’s perspective seem to be meaningful. For each dimension, all the values assigned to tactics were averaged for every 20 units, i.e., a global unit. These average values for each dimension were used as a global index of task behaviors. 48 Langaaga inteasity. In this study, a language intensity coding scheme was borrowed from the measure recently employed in Rogan’s (1990) study of naturalistic hostage negotiation and Donohue’s (in press) work on divorce mediation. The measure extends Bower’s (1964) typology of intensity correlates. Using 482 sets of alternative words or phrases, Bowers’ coding scheme sought to identify possible correlates of language intensity. Judges rated each word or phrase of these sets as high or low intense language by imagining that each tern (i.e., word or phrase) was to be inserted into a sentence. Mean intensity of each term was correlated with various features such as nuaber of syllables, presence of qualifiers, and metaphorical quality possessed by each term. His results showed five reliable correlates of intensity, including obscure language, metaphors, qualifying adjectives, sex-based metaphors, and death-based metaphors. Correlations of intensity for the five predictors with the 482 terms were .59, .83, .89, 1.0, and 1.0 respectively. Donahue (in press) used a slightly modified version of these correlates by adding "profanity" as a predictor. In producing a composite aeasure of language intensity for each utterance, he weighted the frequency of each predictor within each utterance by Bower’s correlations of intensity, summed across the six predictor categories, and then divided by the number of words within the utterance. In his recent study, Rogan (1990) argued that Bower’s definition of qualifiers is inconsistent with his definition of language intensity. Although Bower’s scheme focuses on the intensity of message, it focuses on individual words as the unit of analysis. According to Bowers, "for 49 a term to be classified as qualified, it was required that the qualification be embodied in a separate word. "Greater height," for example, is a qualified term; "higher" is not." (Bowers, 1964, p.352). Using Bowers’ definition of qualifiers to code the statements, "That is a very nice sweater" and ”No, I don’t want you to do that" would reveal a lack of sensitivity to intensity in Bowers’ definition of what constitutes a qualifier. For the first statement, "very" would receive a code. For the second statement, there would be no qualifiers coded. Yet, the second statement clearly expresses an attitude which deviates from a position of neutrality. In his elaborated version of language intensity, Rogan (1990) includes qualifying adjectives, adverbs, affirmation, negation, auxiliary verbs, and contractions as qualifiers. This definition solves Bower’s problems of qualifiers by extending Bower’s definition of qualifiers. For example, according to Ragan’s definition, the statement, "No, I don’t want you to do that" has more qualifiers coded and consequently a higher intensity level than has the statement "That’s a nice sweater." He also combines profanity and sex-based statements into one single category for empirical usefulness. In this study, Rogan’s coding scheme was adopted, since it turned out to be useful in negotiation situations. A copy of the general coding procedure is presented in Appendix C. As Rogan did, when producing a composite measure of language intensity for a unit, the frequency of each predictor within each unit was weighted by Bower’s correlations of intensity and summed across the six predictor categories. However, the summed score was not divided by the number of 50 words within the unit in this study. If the composite score is divided by the total number of words, that measure still seems to be inconsistent with the definition of language intensity. For example, according to Rogan’s method, the statement, "No" has the same intensity with the statement, ”No, no." Similarly, the statement, "No" is more intense than the statement, "No, I don’t want to do that." In both examples, however, the first statement which repeats the same words twice or clearly expresses a negative intention is more intense than the second statement. Further, those who deviate froa neutrality towards a concept and are expected to use more intense messages are likely to use more longer utterance, since various intense predictors should be added to each utterance. Accordingly, a summed rather than average intensity score was calculated in this studyz. Two additional coders were employed to code language intensity with the elaborated version of Rogan’s coding scheme. These coders coded the entire data set of 913 strategy units. For increasing the reliability of coding, four training sessions were held. For each training session, both coders practiced coding language intensity with about 50 strategy units of the actual data used in this study. Then, both coders coded the entire data set. For checking reliability, the percentages of two coders’ agreement were calculated on the basis of strategy units for each category. Both coders’ results showed 99.92, 99.8%, 99.5%, and 99.6% of agreement for obscure words, general metaphor, profanity and sex-based statements, and death-based statement, respectively. The high percentages of agreement in these categories were due to low frequencies of these categories. As a matter of fact, any of these categories did 51 not exceed more than 7 frequencies in the entire data set. For qualifiers, the percentages of agreement were 87.3%, 89.8%, 97.6%, 91.4%, 95.4%, and 96.1% for adjective and adverbs, auxiliary verbs, auxiliary verbs with negative contractions, affirmations and negations, pronouns with contractions, and special words, respectively. Average agreement was 95.6%. For the global index, the results of each unit were averaged across every 20 units. ex c ivers't . Generally, lexical diversity is quantified by a type-token ratio, that is, the number of different words in a message divided by the total number of words. Accordingly, here lexical diversity is operationalized by the number of different words divided by the total number of words in a given strategy. This index indicates the extent to which negotiators employ diverse vocabularies in order to transfer a thought. Two coders were asked to measure this index of lexical diversity for reliability checks. Their results showed 88.52 of agreement. For the global index, the results of each unit were averaged across every 20 units. e e t Meas es In the previous section, behavioral measures of face strategies, Pruitt’s task strategies, language intensity, and lexical diversity were discussed. Since this study uses a MTMM approach to test the construct validity of the new coding scheme, perceptual measures of these variables also were necessary. As explicated in the previous section, for perceptual measures, a more global level of analysis was employed. Perceptions of these variables were rated by judges for every global unit, which is composed of 20 strategy units. 52 Before the judges rated their perceptions, they were given the basic information on negotiation such as issues, positions, the results of negotiation, etc. so as to help them understand each global unit. Then, the judges coded one half of one negotiation transcript at each session in order to reduce factors such as fatigue which lower reliability. As a result, they completed their ratings of the whole data across four sessions. Global units were presented to the judges in the order in which they occurred in the negotiation. Each global unit was given separately to the judges so that the judges could not inadvertently code the text before or after the unit. When a global unit began at the middle of a turn (it occurred three times from the whole data), the entire turn was included in both the prior and current global units. In the next section, the perceptual measures will be addressed. Iaagea at fir-naas aag fita;ibility. For perceptual ratings of the images of firmness and flexibility, two novice judges (undergraduate students) were used. Each judge was asked to rate the extent to which negotiators showed firm and flexible images from the negotiators’ own and the opponent’s perspective for every 20 strategies. .This allowed a comparison of perceptions of overall firmness and flexibility with firmness and flexibility as measured by the coding scheme. The examples for this perceptual measure are shown in Appendix D. Each judge responded to three items on seven-point, Likert-type scales for each dimension. Scales were bounded by the endpoints, "strongly disagree" and "strongly agree." Reliabilities for measures were alpha = .86 and .63 for the negotiator’s own and the opponent’s 53 dimension, respectively. Interrater reliabilities, computed by Pearson’s ;, were .55 and .49 for the negotiator’s own and the opponent’s dimension, respectively. Iaak_attatagta_. For the perceptual measure of task outcomes, two different judges rated the degree to which negotiators were concerned about their own and the opponent’s outcomes in every 20 strategies (i.e., global unit). Each judge responded to three items of seven- point, Likert-type scales for each dimension. Scales were bounded by the endpoints, "strongly disagree" and "strongly agree." These items are shown in Appendix D. Reliabilities for measures were alpha = .71 and .83 for the concern about their own and the opponent’s outcomes. Interrater reliabilities, calculated by Pearson’s 2: were .46 and .69 for the concern about their own and the opponent’s outcomes, respectively. Laggaaga intensity. For the perceptual measure of language intensity, two additional judges rated the degree to which negotiators’ expressions were perceived to be intense, or to deviate from neutral expressions, for every 20 strategies. The results were used as the perceptual global index of language intensity. Each judge responded to three item, seven-point, Likert-type scales. Scales also were bounded by the endpoints, "strongly disagree" and "strongly agree." These items are shown in Appendix D. Reliability for these items was alpha = .85. Interrater reliability calculated by Pearson’s x was .62. texigal diversity. A global, perceptual measure of lexical diversity also was devised. For every 20 strategies, two different 54 coders rated the extent to which each negotiator was perceived to use expansive vocabularies. The results of ratings were used as the perceptual global index of lexical diversity. Each judge responded to three item, seven-point, Likert-type scales. Scales were bounded by the endpoints, "strongly disagree" and "strongly agree." These items are shown in Appendix D. Reliability for this measures was alpha = .74. Interrater reliability, as calculated by Pearson’s ;, was .65. Chapter Four RESULTS Methods for testing the construct validity of the new coding scheme were addressed in the previous chapter. In this chapter, statistical methods for testing the hypotheses proposed in the second chapter and the results will be described. e r' 'v a For statistical analyses, scores were assigned to the behavioral and perceptual measures. As indicated in the prior chapter, the behavioral strategies of upgrading, disclosing, and discounting, which lower negotiator’s own face, were coded "1." Respecting, sidestepping, and attacking strategies which maintain current level of negotiator’s own face were coded ”2." Integrating, defending, and distributing strategies which enhance their own facework were coded "3" (see Table 1). For the opponent’s dimension of the behavioral measure of facework, discounting, attacking, and distributing strategies which lower opponent’s face were coded "l." Disclosing, sidestepping, and defending strategies which make the negotiators to maintain the current level of their opponent’s face were coded "2." Finally, upgrading, respecting, and integrating strategies which attempt to enhance the opponent’s face were coded "3." Given this coding, larger scores indicate higher concern by negotiators about maintaining their own autonomy and supporting their opponent’s autonomy. For the behavioral measure of task strategies, scores also were assigned for statistical analyses. As mentioned in the previous «1" chapter, was assigned to yielding strategy for the negotiator’s own 55 56 dimension, while "2" was assigned to problem-solving and contending strategies. For the opponent’s dimension, "1" was assigned to contending strategy, whereas "2" was assigned to problem-solving and yielding strategies. Given this coding, larger scores indicate higher concern about the negotiator’s own and the opponent’s task outcomes. For the behavioral measure of language intensity, the frequency of each predictor within each unit were weighted by Bower’s correlations of intensity and summed across the five predictor categories. For the behavioral measure of lexical diversity, the number of different words was divided by the number of total words within the unit. In addition, for the global indices of the behavioral measures, the coding results of the behavioral measures were averaged across every 20 strategy units. For the perceptual measures of facework, higher facework scores for the negotiator’s own dimension indicate greater attempts to maintain his/her own autonomy, while for the opponent’s dimension, higher facework scores indicate greater attempts by the negotiator to maintain the opponent’s autonomy (i.e., fewer attempts to impose on the opponent). For the negotiator’s own dimension, higher scores of the perceptual measure of task strategies represent higher concern about the negotiator’s own outcomes, whereas for the opponent’s dimension, higher scores of task strategies represent higher concern about the opponent’s outcomes by the negotiator. For the perceptual measures of language intensity and lexical diversity, higher scores indicate more use of intense words and diverse vocabularies. All the perceptual measures employed seven-point scales (i.e., low = 1, high = 7). Prior to testing the hypotheses, descriptive statistics of all the 57 variables were calculated. Means and standard deviations of the behavioral and perceptual measures of facework, task strategy, language intensity, and lexical diversity are shown in Table 2. Overall, the results of descriptive statistics indicated that negotiators balanced managing their own and opponent’s autonomy. Specifically, the behavioral measure of facework showed that negotiators had slightly more interest in maintaining their own autonomy (M = 2.16) than their opponent’s autonomy (M = 1.96; t = 4.27, df = 93, p < .5), while the perceptual measure revealed that negotiators were perceived to maintain their opponent’s autonomy (M = 4.34) more than their own autonomy (M = 3.19; t = 7.79, df = 93, p < .05). The negotiators also showed concern about their own as well as their opponent’s outcomes. Both behavioral and perceptual measures of task strategies indicated that negotiators were highly concerned about their own (Ms = 1.75 and 4.99 for the behavioral and perceptual measures, respectively) and opponent’s outcomes (Ms = 1.45 and 4.11 for the behavioral and perceptual measures, respectively). Although negotiators had much concern about their opponent’s outcomes, however, the results of both measures suggested that negotiators were more concerned about their own than opponent’s outcomes (ts = 6.20, 5.74 for behavioral and perceptual measures, df = 93, pa for both measures < .01). Further, negotiators on average tended to use moderately intense words (Ms = 3.30, 3.54 for the behavioral and perceptual measures, respectively). These results imply that the negotiators attempted to use neutral expressions during negotiation. Similarly, negotiators 58 Table 2 Me_s an ‘tr-‘auaq "' -‘_ 0.18 '01' 1:18 3‘ v'r .40. "reo _- “a: re Wm 9121311 Print: smug Variables Measures Mean SD Mean SD Mean SD NFS behavioral 2.16 .30 2.26 .25 2.08 .32 perceptual 3.19 .99 3.22 1.06 3.17 .94 OFS behavioral 1.96 .37 2.04 .39 1.89 .34 perceptual 4.34 .79 4.20 .66 4.45 .87 NTS behavioral 1.75 .24 1.75 .26 1.76 .23 perceptual 4.99 .66 5.01 .69 4.90 .63 OTS behavioral 1.45 .29 1.50 .29 1.40 .29 perceptual 4.11 1.07 4.20 .96 4.02 1.16 LI behavioral 3.30 1.65 2.77 .89 3.77 2.01 perceptual 3.54 1.27 3.82 1.19 3.29 1.29 LD behavioral .74 .11 .74 .11 .73 .11 perceptual 4.02 1.09 3.45 1.09 4.52 .81 Nata. For all the variables, N’s are 94, 44, and 50 for ’overall,’ ’private,’ and ’public’ negotiations, respectively. ’Overall’ means and standard deviations are means and standard deviations combined for both private (management-labor) and public (school board-teacher) sectors. NFS = Face Strategies for a Negotiator’s Own Dimension OFS = Face Strategies for the Opponent’s Dimension NTS = Task Strategies for a Negotiator’s Own Diaension OTS = Task Strategies for the Opponent’s Dimension LI = Language Intensity LD = Lexical Diversity 59 employed moderately diverse vocabularies (Ms = .74, 4.02 for the behavioral and perceptual measures, respectively). In sum, the findings from the behavioral and perceptual measures of facework, task strategies, and language intensity all suggest that both negotiations contained moderate levels of cooperativeness and concern for joint outcomes. In order to examine possible differences in the use of face strategies, task strategies, language intensity, and lexical diversity across the two negotiations, descriptive statistics were computed separately for each negotiation (i.e., management-labor’s and school board-teacher’s negotiations; see Table 2). The results indicated that negotiators from the private sector negotiation behaviorally had more concern about facework for both self and the opponent’s dimensions than negotiators from the public sector negotiation, while both negotiations were perceived to contain similar levels of facework. For the behavioral measure of face strategies, negotiators from the private sector negotiation attempted to maintain their own face (; = 2.91, df = 92, p < .01; p’ = .08) and to support their opponent’s face (t = 2.01 df = 92, p < .05; g: = .04) more than negotiators from public sector negotiation, although these effect sizes were small. For the perceptual measure, there were no significant differences in face strategies for both dimensions between the two negotiations (ts were p > .05 for both self and the opponent’s dimensions). Similarly, regarding task strategies, there were no significant differences between the two negotiations for the perceptual or behavioral measures (ts were p > .05 for both dimensions in the 60 perceptual and behavioral measures). For language intensity, the results from the perceptual aeasure were contradictory with those of the behavioral measure. The results from the perceptual measure indicated that negotiators from the private sector negotiation were perceived to use more intense language than negotiators from the public sector negotiation (t = 2.06, df = 92, p < .05; I? = .04), while the results from the behavioral measure showed that the former used less intense words than the latter (t = -3.05, df = 92, p < .01; g} = .09). Although language intensity between the two negotiations was statistically significant for the perceptual and behavioral measures, the effect sizes also were small. For lexical diversity, there was no significant difference in the behavioral measure between the two negotiations (t was p > .05), while negotiators from the public sector negotiation were perceived to use more diverse vocabularies than negotiators from the private sector negotiation (t = - 5.42, df = 92, p < .01; ’ = .24). Further, the effect size for the perceptual measure was large. To summarize, the descriptive statistics suggested that negotiators were cooperative with participants who balanced managing their own and opponent’s face. In addition, each negotiation was perceived to contained similar levels of face strategies and task strategies, while the public sector negotiation was perceived to contain more intense message and use more diverse words. Behaviorally, negotiators from the private sector negotiation had more concern about their own and opponent’s face and used less intense language, while there were no differences in using task strategies and diverse vocabularies between 61 the two negotiations. However, since all the significant differences between the two negotiations except for the perceptual measure of lexical diversity were small in effect size, both negotiations were similar in the use of face strategies, task strategies, and language intensity. se H hes 8 Correlation analyses were performed to test most hypotheses of this study. To evaluate convergent and discriainant validity, correlations between the behavioral and perceptual measures of face strategies and the measures of task strategies, language intensity, and lexical diversity were calculated. These statistical analyses, which appear in Table 3, are relevant to testing all the hypotheses except for hypothesis 3. To analyze the relationships between face and task strategies in more detail, crosstabulations and oneway ANOVAs were performed. For oneway ANOVAs, task strategies were treated as the independent variable with three levels (i.e., yielding, problem-solving, a contending) and scores assigned to face strategies from the negotiator’s own and their opponent’s dimensions, respectively (i.e., 1 for flexibility, 2 for maintaining the current level of face, a 3 for firmness) were the dependent variables. The "other" category of task strategies was set aside as a missing value in all the analyses containing the behavioral measures of task strategies. Finally, to answer the research question, confirmatory factor analyses were performed. Hypgthaaia_gaa. The first hypothesis predicted that the global behavioral and perceptual measures of face strategies would share a 62 strong, positive association for both a negotiator’s and the opponent’s dimensions. The results supported the hypothesis. For the negotiator’s own dimension, the perceptual measure of face strategies was strongly and positively associated with the behavioral measure of face strategies (3 = .53, p < .01). Similarly, for the opponent’s dimension, the perceptual and behavioral measures of face strategies also showed a strong and positive correlation (I = .60, p < .01). Consequently, these results support the convergent validity of the behavioral coding system for face strategies. es s w t e . Hypothesis two predicted that the global behavioral and perceptual measures of task strategies would share strong, positive associations for both negotiator’s and opponent’s dimensions. The results also were consistent with this hypothesis. The global behavioral and perceptual measures of task strategies had strong and positive correlations for both dimensions (5 = .77, p < .01 for the negotiator’s own dimension; a = .74, p < .01 for the opponent’s dimension). Hypothesis three predicted that the single-unit behavioral measures of face and task strategies would share a moderate, positive association and that the global perceptual measures of these strategies also would show a moderate, positive association. Further, relationships between specific task and face strategies were predicted for the behavioral measures. The correlations between task and face strategies supported this hypothesis for both global perceptual and single-unit behavioral measures. More specifically, for the negotiator’s own dimension, task and face strategies showed moderate and positive correlations for both 63 global perceptual (g = .35, p < .01) and single-unit behavioral (; = .33, p < .01) measures. For the opponent’s dimension, these strategies relatively shared strong and positive correlations for perceptual (t = .53, p < .01) and behavioral (t = .51, p < .01) measures. For the global behavioral measures, similar patterns of correlations also appeared. The global behavioral measure of face strategies also had moderate correlations with the global perceptual measure of task strategies for the negotiator’s own dimension (2 = .37, p < .01) and opponent’s (E = .44, p < .01) dimensions. Similarly, the global perceptual measure of face strategies showed moderate correlations with the global behavioral measure of task strategies for the negotiator’s (E = .24, p < .01) and opponent’s (t = .53, p < .01) dimensions. In sum, regardless of measurement procedure or unit of analysis, face and task strategies shared moderate, positive relationships. Tables 3 and 4 describe these results. Although face and task strategies for both dimensions had moderate or strong correlations, these correlations were smaller in effect size than the correlations of the convergent validity of face strategies for both the negotiator’s and opponent’s dimensions. Consequently, these results indicate that face and task strategies tap related yet distinct concepts. Oneway ANOVAs were performed in order to clearly interpret the relationships between specific task and face strategies. For a negotiator’s own face dimension, the main effect for task strategies was significant (212,720) = 94.76, p < .01, etnz = .21). Table 5 presents the means for these results. To interpret this main effect, a Student- 64 Table 3 Qazralatign Matrix {9: tha Glabal-Bahavigral and Eergeptual Measures of V ' e h wn 'n e t eses (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (l) NFSb .53 .08 -.14 .37 .37 -.46 -.38 .12 .16 -.05 -.08 (2) NFSp - -.15 -.28 .24 .35 -.57 -.67 .19 .24 -.16 -.22 (3) OFSb - .60 -.66 -.54 .49 .44 -.40 -.36 .15 -.16 (4) OFSp - -.43 -.43 .41 .53 -.31 -.38 .31 .24 (5) NTSb - .77 -.60 -.45 .36 .47 -.15 .19 (6) NTSp - -.49 -.46 .22 .48 -.16 .02 (7) OTSb - .74 -.28 -.31 .12 .02 (8) OTSp - -.36 -.39 .23 .23 (9) LIb - .50 -.58 .04 (10) LIp - -.40 -.15 (11) LDb - .31 (12) LDp - Note. M’s are 94. 3 >= .17 are significant at p < .05 and g >= .24 are significant at p < .01 (one-tailed tests). The abbreviations of variable names used here are the same as Table 2 except that b and p are added to the variable names. ’b’ stands for a behavioral aeasure and ’p’ refers to a perceptual measure. For example, NFSb = Behavioral Measure for Face Strategies for a Negotiator’s own dimension. NFS Face Strategies for a Negotiator’s Own Dimension OFS Face Strategies for the Opponent’s Dimension 65 NTS = Task Strategies for a Negotiator’s Own Dimension OTS = Task Strategies for the Opponent’s Dimension L1 = Language Intensity LD = Lexical Diversity 66 Table 4 19 re a °., eat ' ‘or ‘ 1; e- if_ :‘hiv.0ri, M'a= V ' es (2) (3) (4) (5) (6) (1) NFS .12 .33 -.45 .18 -.01 (2) OFS - -.64 .51 -.31 .23 (3) NTS - -.63 .30 -.22 (4) OTS - -.20 .10 (5) LI - -.61 (6) LD - Mata. M’s are 913 for all the correlations not involving task strategies (i.e., NTS or OTS). M’s are 723 for the correlations involving task strategies, since the ’other’ category from task strategies was excluded from statistical analyses. gs >= .09 for task strategies and .07 for other variables are significant at the level of .05. NFS = Face Strategies for a Negotiator’s Own Dimension OFS = Face Strategies for the Opponent’s Dimension NTS = Task Strategies for a Negotiator’s Own Dimension OTS = Task Strategies for the Opponent’s Dimension L1 = Language Intensity LD = Lexical Diversity 67 Table 5 Means and Standard Deviations of Face Strategies for a Negotiator’s Qua and tha Qppgnent’s Dimension at Eagh Level af Iaak Strategies Yielding Problem-Solving Contending Megatiator’s Own Dimensian Mean 1.73ll 1.86’ 2.49b so .50 .95 .63 Sample Size (N) 167 145 411 Qppgaent’s Qimeasioa Mean 2.92‘ 1.83b 1.56° so .29 .95 .61 Sample Size (N) 167 145 411 Nat . The superscripts with different characters indicate means differ significantly at the level of .05, while the superscripts with same characters indicate means do not differ significantly at the level of .05. 68 Newman-Keuls follow-up test was performed among three task strategies. This test indicated that while there was no difference in face strategies between yielding and problem-solving strategies (p > .05), there were statistically significant differences in face between yielding and contending strategies (p < .05) and problem-solving and contending strategies (p < .05). Put differently, negotiators managed their own face similarly under yielding (M = 1.73) and problem-solving (M = 1.86) strategies. On the other hand, they showed a firmer image under contending (M = 2.49) than yielding or problem-solving strategies. Overall, these results support hypothesis three. Again, for the opponent’s face dimension, the main effect for task strategies was reliable (z[2,7201 = 266.96, p < .01, eta2 = .43). The means for these results appear in Table 5. To interpret this main effect, a Student-Newman-Keuls test also was performed among three task strategies. What this subsidiary analysis revealed was that there were statistically significant differences in face strategies between yielding and problem-solving strategies, yielding and contending strategies, and problem-solving and contending strategies (p < .05 for all three comparisons). Specifically, negotiators attempted to enhance their opponent’s autonomy most under the yielding strategy (M = 2.92), less under the problem-solving strategy (M = 1.88), and the least under the contending strategy (M = 1.56). These results are also consistent with hypothesis three. Finally, crosstabulations were drawn to identify the specific relationships between task and face strategies. These analyses are relevant to hypotheses 3a, 3b, and 3c. These hypotheses predicted that: 69 (1) the yielding strategy would be moderately correlated with the upgrading, respecting, and disclosing strategies, (2) the problem- solving strategy would be moderately correlated with the integrating and discounting strategies, and (3) the contending strategy would be moderately correlated with the defending, attacking, and distributing strategies. The results of crosstabulations supported these hypotheses (see Table 6). Specifically, seventy-nine percent of the total units which were coded as either the upgrading, or respecting, or disclosing face strategies also were coded as the yielding task strategy. Seventy-seven percent of the total units coded as integrating or discounting were included in the problem-solving task strategy. Similarly, ninety-three percent of the total units for defending, attacking, or distributing were matched with the contending task strategy. As a result, eighty- five percent of the total face strategies were matched with task strategies as was proposed in the hypotheses (52: 1759.6, df = 24, p < .01). Several interesting subsidiary patterns are evident in Table 6. First, the distribution of units across strategy categories were much more even for face strategies than for task strategies. This suggests that the coding scheme for face strategies may be more distinctive in identifying the meanings of negotiators’ utterance than is the scheme for task strategies. Second, the most unmatched face strategy was the disclosing strategy which counted the lowest percentage of occurrence (2.3 1 out of the total units). This result implies that negotiators who lower their own firmness do not necessarily yield their own task 70 Yielding Problem-Solving Contending Other Total Upgrading 39 2 2 0 43 (90.7) (4.7) (4.7) (0.0) (4.7) Respecting 112 5 14 8 139 (80.6) (3.6) (10.1) (5.8) (15.2) Disclosing 9 2 6 4 21 (42.9) (9.5) (28.6) (19.0) (2.3) Integrating 4 51 9 2 66 (6.1) (77.3) (13.6) (3.0) (7.2) Discounting 1 73 21 1 96 (1.0) (76.0) (21.9) (1.0) (10.5) Defending 0 5 153 10 168 (0.0) (3.0) (91.1) (6.0) (18.4) Attacking 0 1 115 9 125 (0.0) (0.8) (92.0) (7.2) (13.7) Distributing 1 0 68 0 69 (1.4) (0.0) (98.6) (0.0) (7.6) Sidestepping 2 5 23 156 186 (1.1) (2.7) (12.4) (83.9) (20.4) Total 167 145 411 190 913 (18.3) (15.9) (45.0) (20.8) (100) Nat . g’ = 1759.6, df = 24, p < .01. Numbers outside parenthesis are frequencies; numbers inside parenthesis are the percentage out of total row frequency. 71 goal. Third, most of the strategies coded as "other" category in task strategies were coded as "sidestepping" category in face strategies. This implies that when negotiators’ utterances did not contain information relevant to their task, the negotiators attempted to maintain current level of both their own and opponent’s face. To summarize, the results of the correlations, oneway ANOVAs, and crosstabulations were consistent with hypotheses two and three. As expected, the behavioral and perceptual measures of face strategies shared strong, positive associations for both the negotiator’s and opponent’s dimensions. Face and task strategies also showed substantial correlations for both dimensions, but the correlation between task and face strategies in the opponent’s dimension was stronger than that of the negotiator’s own dimension. 8 s 1 six. Hypothesis four predicted that the measure of language intensity would demonstrate convergent validity. That is, the global behavioral and perceptual measures of language intensity would share a strong, positive association. The results indicated that the hypothesis was supported. The global perceptual measure of language intensity was strongly and positively associated with the global behavioral measure of language intensity (2 = .50, p < .01). This result appears in Table 3. Hypothesis five and six proposed that a negotiator’s own firmness would share a moderate, positive association with the negotiator’s language intensity and that a negotiator’s firmness towards the opponent would have a moderate, negative association with the negotiator’s language intensity. The results supported these hypotheses. The 72 results for these hypotheses are summarized in Tables 3 and 4. For a negotiator’s own dimension, the single-unit behavioral measures of a negotiator’s firmness and language intensity revealed a weak and positive correlation (a = .18, p < .01). The perceptual measures also indicated that the negotiator’s firmness was positively associated with language intensity (3 = .24, p < .01). The global behavioral measure of face strategies had a marginally significant correlation with the global perceptual measure of language intensity (; = .16, p = .07). Similarly, the global perceptual measure of face strategies showed moderate correlations with the global behavioral measure of language intensity (a = .19, p < .05). As a result, these results were consistent in direction with hypothesis five. Further, these results indicated that for the negotiator’s dimension, the correlations of the convergent validity of face strategies were larger than those between face strategies and language intensity. Consequently, these results also supported the discriminant validity for face strategies for the negotiator’s own diaension. For the relationships between a negotiator’s firmness towards the opponent and language intensity, the single-unit behavioral measures of these variables had a moderate and negative correlation (a = -.31, p < .01). Again, the perceptual measures also showed a moderate and negative correlation between a negotiator’s firmness towards the opponent and a negotiator’s language intensity (3 = -.38, p < .01). The correlations of the global behavioral and perceptual aeasures between these variables also had effect sizes similar to those of both the single-unit behavioral measures and perceptual measures. The global 73 behavioral measure of a negotiator’s firmness towards the opponent revealed a moderate and negative correlation with the perceptual measure of language intensity (a = -.36, p < .01). Similarly, the perceptual measure of face showed a similar size of correlation with the global behavioral measure of language intensity (2 = -.31, p < .01). As a result, these analyses confirmed hypothesis six. Further, these results indicated that for the opponent’s dimension, the correlations of the convergent validity of face strategies were larger than those between face strategies and language intensity. Consequently, these results also supported the discriminant validity for face strategies for the opponent’s dimension. In sum, a negotiator’s own firmness consistently was positively correlated with a negotiator’s language intensity with weak but statistically significant relationships. A negotiator’s firmness in upholding the opponent’s autonomy was moderately and negatively associated with a negotiator’s language intensity. Mypathesas aeveg and eight. Hypothesis seven predicted the convergence between the global behavioral and perceptual measures of lexical diversity. Although the correlation between the perceptual and global behavioral measures of lexical diversity was significant (a = .31, p < .01), the size of correlation was not as strong as what was expected. That was partially due to the differences in the unit of analysis between the behavioral and perceptual measures. Regarding the perceptual ratings of lexical diversity, they rated lexical diversity on the basis of the degree to which negotiators used diverse vocabularies in a given global unit. On the other hand, the global index of the 74 behavioral measure was calculated by averaging scores of lexical diversity across twenty single-units. It is likely to have high levels of lexical diversity for each single-unit behavioral measure but to have low levels of lexical diversity between single-units. In other words, words of single-unit 1 might be similar to words of single-unit 2. For example, A1: I just don’t want to do it. B: Do the rest of you feel this way as well? A2: We don’t want to do it. B: No, I just wonder if you felt the same way? (A1 and A2 are negotiators of the same side) This tendency can raise perceptions of high levels of lexical diversity at a specific single-unit and weak or moderate levels of lexical diversity at a global unit. The global unit of lexical diversity is different from other variables in the degree to which perceptions at a global unit and single-units in a given global are consistent. Accordingly, given the difference in the unit of analysis between the global behavioral and perceptual measures, the moderate correlation between these two measures seems to be plausible. Hypothesis eight proposed that face strategies and lexical diversity would be unrelated. The results appear in Tables 3 and 4. According to the results, these hypotheses were partially supported. For a negotiator’s own dimension, while the perceptual measure of lexical diversity was negatively correlated with that of face strategies (I = -.22, p < .05), the single-unit behavioral measures were not correlated (; = -.01, p > .05). The global behavioral measure of face 75 strategies also was unrelated with the perceptual measure of lexical diversity (a = -.08, p > .05), whereas the correlation between the perceptual measure of face strategies and the global behavioral measure of lexical diversity was marginally significant (3 = -.16, p =.07). Thus, the results were contradictory according to the types of measures. These contradictory results may be partially due to the low convergent validity for the measures of lexical diversity. However, for the negotiator’s own dimension, these results supported the discriminant validity of face strategies. Any of the correlations between face strategies and lexical diversity was smaller than the correlation of the convergent validity for face strategies. For the opponent’s dimension, lexical diversity was positively correlated with face strategies in both the perceptual (; = .24, p < .05) and single-unit behavioral (a = .23, p < .01) measures. In other words, as negotiators enhance their opponent’s face, they tend to employ diverse vocabularies. The correlations of the different measures between these two variables, however, were also inconsistent with those of the same measures. While the correlation between the global behavioral measure of face strategies and the perceptual measure of lexical diversity was marginally significant in a negative direction (a = -.16, p = .07), the perceptual measure of face strategies showed a positive and moderate correlation with the behavioral measure of lexical diversity (I = .31, p < .01). These inconsistent correlations also may be due to the low convergent validity of lexical diversity. However, for the opponent’s dimension, the results also supported the discriminant validity for face strategies by indicating that the 76 correlation of the convergent validity for face strategies was larger in effect size than any of the correlations between face strategies and lexical diversity. To summarize, the results suggest that face strategies for a negotiator’s own dimension and lexical diversity were unrelated or had a weak and negative correlation. These variables had weak to moderate correlations in conflicting directions for the opponent’s dimension. However, these results indicated that the coding scheme for face strategies had discriminant validity for both the negotiator’s own and the opponent’s dimensions. Mypgtheses aine and ten. Hypotheses nine and ten proposed general predictions from the MTMM approach. Hypothesis nine predicted that monoconstruct-heteromethod correlations (i.e., validity diagonal) would be larger than heteroconstruct-heteromethod correlations. One of the conventional evaluation criteria which test these general hypotheses drawn from MTMM approach is to compare all the correlations in ranking order. As is indicated in Table 7, monoconstruct-heteromethod (MTHM) correlations ranged from .50 to .77 except for lexical diversity. On the other hand, all the heteroconstruct-heteromethod (HTHM) correlations except for the correlation between face strategies for the opponent’s dimension and task strategies for a negotiator’s dimension were below .47, which implies that monoconstruct-heteromethad correlations were larger than heteroconstruct-heteromethod correlations (average as after Fisher’s ; to a transformation = .69 and .29 for MTHM & HTHM, respectively; Fisher’s Z = 3.82 for the difference in the average as, df = 91, p < .01). Further, the monoconstruct-heteromethod correlations of 77 Table 7 e r e t r QIchalzflshsxioral_hsasure Perceptualiueaaure (1) (2) (3) (4) (5) (6) (1) (2) (3) (4) (5) (6) - v1 M r (1) NFS .83 (2) OFS -.08 .83 (3) NTS .37 -.66 .87 (4) OTS -.46 .49 -.60 .87 (5) L1 .12 -.40 .36 -.28 .96 (6) [AD -005 015 -015 012 -058 089 Eezcentual_hea§urs (1) NFS .53 .55 (2) OFS -.14 .60 -.28 .49 (3) NTS .37 -.54 .77 .35 -.43 .46 (4) OTS -.38 .44 -.45 .74 -.67 .53 -.46 .69 (5) L1 .16 -.36 .47 -.31 .50 .24 -.38 .48 -.39 .62 (6) LD -.08 -.16 .19 .02 .04 .31 -.22 .24 .02 .23 -.15 .65 Nata. The diagonals of the same measures (e.g., behavioral measures) are reliability, while validity diagonals (i.e., diagonals of the different measures) are correlations of convergent validity. Upper, left off-diagonals and bottom, right off-diagonals are correlations of heteroconstruct-monomethod (HTMM), while bottom, left off-diagonals are correlations of heteroconstruct-heteromethod (HTHM). 78 face strategies for the opponent’s dimension (2 = .60) and task strategies for a negotiator’s own dimension (2 = .77) also were larger than the heteroconstruct-heteromethod correlation of face strategy for the opponent’s dimension and task strategies for a negotiator’s own dimension. Again, the monoconstruct-heteromethod correlation of lexical diversity was stronger than any heteroconstruct-heteromethod correlations involving lexical diversity. Thus, hypothesis nine was supported. These results implicate that face strategies had the discriminant validity. Hypothesis ten predicted that monoconstruct-heteromethod correlations would be larger than heteroconstruct-monomethod (HTMM) correlations. As expected, monoconstruct-heteromethod correlations were larger than most heteroconstruct-monomethod correlations except for three correlations (average a for HTMM = .36; Fisher’s z = 3.15 for the difference in average as, df = 91, p < .01). Face strategies for the opponent’s dimension had a strong correlation with task strategies for a negotiator’s own dimension in the behavioral measure (a = -.66). Face strategies for a negotiator’s own dimension also shared a strong correlation with task strategies for the opponent’s dimension in the perceptual measure (a = -.67). One plausible explanation for these relationships might be that system for coding task strategies caused an artificially high correlation between some face and task strategies. As a matter of fact, the weighing score systems assigned to both dimensions in task strategies may have caused the high correlation between these two dimensions. Following the weighing score system adopted in this study, 79 the contending and yielding task strategies in a negotiator’s dimension always had perfect negative correlations with the contending and yielding task strategies in the opponent’s dimension, respectively. This might have naturally led to the high correlation between these two dimensions. Accordingly, since face and task correlations for the same dimensions also were moderately correlated, face strategies for a negotiator’s own dimension and task strategies for the opponent dimension, or vice versa might have ended up being highly correlated, via the high correlation between these two dimensions in task strategies. l To test this possibility, partial correlations were calculated. The partial correlations indicated that these correlations were substantial. When controlling task strategies for a negotiator’s own dimension, partial correlation between face strategy for the negotiator’s own dimension and task strategies for the opponent’s dimension was strong (pa = -.61). Similarly, when controlling task strategies for a negotiator’s own dimension, partial correlation between face strategies for the opponent’s dimension and task strategies for the negotiator’s own dimension also was strong (pa = -.52). As a result, these results suggested that negotiators strengthened (conceded) their own autonomy when they had a low (high) concern about their opponent’s outcomes and that negotiators also supported (unsupport) their opponent’s autonomy when they had a low (high) concern about their own outcomes. The heteroconstruct-monomethod correlation of language intensity and lexical diversity in the behavioral measure also showed a negative 80 and strong correlation (a = -.58), implying that a negotiator adopted intense message by repeating many words. In sum, although a few heteroconstruct-monomethod correlations were larger than monoconstruct- heteromethod correlations, on average most heteroconstruct-monomethod correlations were significantly lower than heteroconstruct-monomethod correlations. Accordingly, overall the findings supported hypothesis ten which implicates that face strategies had the discriminant validity. Reseaach Qaestion In this study, one research question was raised about whether a negotiator’s own and the opponent’s dimensions of face strategies would be correlated. For perceptual measure, these two dimensions of face strategies turned out to be negatively and moderately correlated (a = - .28, p < .01). On the other hand, these two dimensions had a weak and positive correlation for the single-unit behavioral measure (a = .12, p < .01). Although the correlation from the behavioral measure was significant, it is not substantial. The significance test for correlations is a function of sample size. Since the sample size of the single-unit behavioral measure in this study is large (a = 913), all the correlations with small effect size can be statistically significant. In this sense, the correlation of .12 is not substantial. To summarize, the perceptual measures of a negotiator’s own and their opponent’s dimensions in face strategies shared a negative, moderate correlation, while the behavioral measures had a weak correlation. Further confirmatory analyses were performed in order to identify whether these two dimensions were different. LISREL analysis was performed under the assumption that these two dimensions were 81 independent (see Table 8). The results of this two factor model showed that the factor loadings of the behavioral and perceptual measures for the negotiator’s own dimension were significantly different from 0 (ts = 3.53 and 4.00 for the behavioral and perceptual measures, respectively). Similarly, for the opponent’s dimension, the factor loadings of the behavioral and perceptual measures also were significantly different from 0 (ts = 3.50 and 4.03, respectively). Further, the chi-square for the goodness of fit was not significant (32 = 4.94, df = 2, p > .05). This indicates goodness of fit for the measurement model proposed above, i.e., the two factor model in which the negotiator’s own and opponent’s dimensions are assumed to be different. This model also had better goodness of fit than a single factor model in which these two dimensions are assumed to be unidimensional (52:: 44.33, df = 2, p < .01)3. Similarly, the goodness of fit index confirmed this fact by indicating that the index of two factor model (=.98) was higher than that of the single factor model (=.84) (see Table 9). As a result, the results of LISREL indicated that these two dimensions were different constructs. To summarize, the results of the correlations revealed that the perceptual measures of the negotiator’s own and their opponent’s dimensions in face strategies shared a negative and moderate correlation, while the behavioral measures had a weak correlation. In addition, the results of LISREL indicated that these two dimensions were different constructs. Thus, we can conclude that the negotiator’s own and opponent’s dimensions are empirically as well as conceptually distinct. 82 Table 8 i m 'k i st' te 4 tems d ors Variable Factor One (K811) Factor Two (K812) T-Value NFSb o 17 " 30 53 NFSp a 92 - 4000 OFSb " o 19 30 50 OFSp - e 90 4003 Chi-Square DF Chi-Square/DF Two Factor Model 4’94: 2 2.47 Single Factor Model 44.33 2 22.17 Nate. * = significant at p = .01. Chapter Five DISCUSSION This study elaborated the conceptual definition of face in negotiation by integrating the ideas of mutual interdependency from the negotiation literature and the concept of face from the politeness literature. Here, face as the images of firmness and flexibility was defined as autonomy from both negotiators’ perspectives. Further, the coding scheme for facework in negotiation was developed on the basis of this new conceptualization. Then, this research empirically investigated the construct validity of this coding scheme. In the following section, the results and implications of this investigation will be summarized. Specifically, findings regarding research hypotheses will be described. Further, important findings unrelated to research hypotheses will also be addressed. Then, the implications of the new coding scheme and findings in this study will be addressed. Fi r ' ese t eses This study examined the validity of the coding scheme by adopting both nomological and multitrait-multimethod approaches. Based on the prior findings about face strategies, task strategies, language intensity, and lexical diversity, face strategies were hypothesized to share moderate, positive associations with task strategies and language intensity, while face strategies were expected to have no correlation with lexical diversity. Coders who were blind to the purpose of this study coded face strategies, task strategies, language intensity, and lexical diversity from two naturalistic negotiations; one from the private sector and the 83 84 other from the public sector. Judges also rated the degree to which these variables were perceived in the data. All the codings and ratings were reliable. The findings from the MTMM approach showed that the coding scheme for face strategies was valid and reliable. Convergent validity was established. The perceptual measure of face strategies was strongly correlated with the behavioral measure of face strategies for both a negotiator’s own and the opponent’s dimension. These results were expected, since these two measures tapped the same construct. These results also imply that the theoretical conception of face underlying the new coding scheme has representative validity (Folger, Hewes, & Poole, 1984). In other words, naive observers similarly interpret the theoretical meaning of face embedded in the negotiation coding scheme. The coding scheme for face strategies also had discriminant validity. The results revealed that the correlations of face strategies between heteromethods (i.e., perceptual and behavioral measures) were larger than most correlations of heteroconstruct-monomethods (HTMM) and heteroconstruct-heteromethods (HTHM). These findings were expected, since correlations between variables were lowered by trait variance in HTMM or trait and method variances in HTHM. The results from the nomological approach also indicated that the coding scheme for face strategies was valid. The correlations between face strategies and similar constructs (i.e., task strategies and language intensity) supported the validity of the coding scheme for face strategies. Table 10 summarizes the relationships between the predictions in the hypothesis and the results of the statistical 85 Table 10 r e o s t rti th es s r Va 'dit est Variable Measure Prediction Result with Eaaa Staatagies tar a Negatiator’s an Dimensign Task Strategies behavioral + + perceptual + + Language Intensity behavioral + + perceptual + + Lexical Diversity behavioral O - perceptual 0 0 t F e r te 'es 0 the e t’s im 8' n Task Strategies behavioral + + perceptual + + Language Intensity behavioral - - perceptual - - Lexical Diversity behavioral 0 + perceptual 0 + 86 analyses. As shown in Table 10, nine out of twelve predictions from the nomological networks were confirmed. These results clearly evidence the validity of the coding scheme for face strategies. As expected, face strategies were moderately associated with task strategies for a negotiator’s own and their opponent’s dimensions in both perceptual and behavioral measures. Put differently, the more firm (or the less flexible) image a negotiator reflected in his/her interaction, the more concern about his/her own outcomes the negotiator had. Similarly, the more firm a negotiator was about upholding the opponent’s image, the more concern about his/her opponent’s outcomes the negotiator had. These results supported the hypotheses proposed in this study and were consistent with Hiltrop and Rubin’s (1981) and Pruitt and Johnson’s (1970) findings that task goals were positively related to face goals. One interesting trend among the findings was the relatively stronger correlation between single-unit behavioral face and task strategies for the opponent dimension (2 = .51) than for a negotiator’s own dimension (3 = .33). This implies that negotiators include the information about face strategies towards their opponent clearly in their task strategies. On the other hand, when they have concern about their own outcomes, they tend to present information about their own face in their task strategies only moderately. One plausible explanation for these results is that the socio- cultural rules of social face or politeness which people learn are usually directed towards others’ face (Brown, 1977), which might lead people to be more sensitive to others’ face than their own face. 87 Perhaps this implies that negotiators are more sensitive to the management of their opponent’s face when they perform their task in negotiation. Another possible explanation is that task and face strategies are less intertwined for the self than the opponent’s dimension. That is, it may be more difficult for negotiators to show varying levels of firmness in upholding their opponent’s autonomy when showing concern for their opponent’s task outcomes than to present various levels of firmness about their own face when showing concern for their own task outcomes. It seems to be natural that rejecting or criticizing an opponent’s position may be strongly associated with the opponent’s loss of face, since he/she might feel affronted. As a matter of fact, Tjosvold and Huston (1978) found that criticizing a negotiator’s position was highly correlated with the negotiator’s loss of face. On the other hand, when a negotiator supports his/her own position, he/she might lower his/her own face easily since the negotiator might feel that his/her own positional commitment compensates his/her own loss of face. Accordingly, task and face strategies may be more intertwined for the opponent than self dimension. As was predicted, face strategies also were moderately correlated with language intensity. A negotiator’s own firmness had a moderate and positive correlation with the negotiator’s language intensity, while the negotiator’s firmness towards the opponent shared a moderate and negative association with the negotiator’s language intensity. Put differently, when negotiators fostered their own firm image and attempted to limit their opponent’s flexibility, they employed more intense messages. The correlation between language intensity and a 88 negotiator’s firmness towards the opponent (i.e., face strategies for the opponent’s dimension), however, was larger in effect size than that of the negotiator’s own firmness (i.e., face strategies for a negotiator’s own dimension). This pattern is consistent with those of the relationships between face and task strategies in which the opponent’s dimension was stronger than the negotiator’s own dimension. These results imply that the use of language intensity might be strongly related to the social and cultural rules of politeness directed towards other’s face. Accordingly, people may manage another’s face effectively through limiting their biased attitude towards a matter or concept. As indicated above, another possible explanation is that face and language intensity might be less intertwined for the self than the opponent dimension. In sum, these findings supported the proposed hypotheses, which increases the validity of the coding scheme for face strategies. The hypotheses on the relationships between face strategies and lexical diversity were partially confirmed. Although face strategies were expected not to be correlated with lexical diversity, there were some significant correlations between these two variables. For a negotiator’s own dimension, the perceptual measures of face strategies and lexical diversity were negatively correlated, whereas the behavioral measures were not associated with each other. For the opponent’s dimension, both perceptual and behavioral measures of face strategies and lexical diversity were positively correlated. Put differently, when negotiators foster their own flexible image and the firm image towards their opponent (i.e., negotiators are polite to their opponent), they tend to command different numbers of words. These results are 89 understandable, since people tend to use diverse words with longer sentences for politeness expressions (Brown & Gilman, 1960). In conclusion, the convergent and discriminant validity of the coding scheme of face strategies was established frbm the MTMM approach. Most hypotheses from the nomological approach also were confirmed. In this study, one research question was raised: that is, are a negotiator’s and the opponent’s face dimensions related? The results for this research question were interesting. These two dimensions were not correlated for the behavioral measure, while they were negatively and moderately associated for the perceptual measure. Since the behavioral measures were theoretically drawn under the assumption that these two dimensions are distinct, naturally their correlation was nonsignificant. However, it is possible that increasing a negotiator’s own autonomy will be perceived to reduce the opponent’s autonomy, since negotiators are highly interdependent on each other to achieve their goals. Given no correlation in the behavioral measures and only the moderate correlation in the perceptual measures between these two dimensions, the findings indicate that these theoretical dimensions embedded in the coding scheme also are empirically independent, and support Brown and Levinson’s argument that a speaker’s and hearer’s face are conceptually distinct. Further, these results imply that negotiators should also manage their own autonomy effectively for successful negotiations. To summarize, the perceptual measures of a negotiator’s own and their opponent’s dimensions for face strategies shared a negative, moderate correlation, while the behavioral measures had a weak 90 correlation. We can conclude that the negotiator’s own and opponent’s dimensions are empirically as well as conceptually distinct when considering that the correlation of the perceptual measures was only moderate. In the following section, some interesting findings unrelated to the hypotheses will be discussed. her ' di 8 There were some interesting findings which were not directly related to the hypotheses. First, for task strategies, a negotiator’s own dimension shared a strong and negative correlations with the opponent’s dimension. This inverse correlation occurred for both behavioral and perceptual measures with similar effect size. These results of the perceptual measures were not surprising. When a negotiator is highly concerned about his/her own outcomes, the negotiator can be naturally perceived to have low concern about the opponent’s outcomes. As argued above, for the behavioral measure, a negotiator’s own and the opponent’s dimensions for task strategies are theoretically oblique. The results from the perceptual and behavioral measures of task strategies supported this relationship between these two dsmensions. Further, the correlations between the self and opponent dimensions for task strategies were larger than those of face strategies. These results imply that the self and opponent dimensions for task strategies may be both empirically and conceptually less distinct than these two dimensions for face strategies. Another interesting finding is that a negotiator’s own and the opponent’s dimensions for face strategies had strong and negative correlations with the opponent’s and a negotiator’s own dimensions for 91 task strategies, respectively. Put differently, when negotiators have high concern about their own outcomes, they attempt to lower their opponent’s face. Similarly, when negotiators have low concern about their opponent’s outcomes, they try to enhance their own face. These substantial correlations seem to make conceptual sense. For example, when negotiators are highly concerned about their own face, they might criticize or attack their opponent’s task positions. In sum, high and negative correlations between the different dimensions of face and task strategies not only empirically make sense but also are reasonable, given the relationships between the different two dimensions of task strategies and between the same dimensions of face and task strategies. Third, the findings revealed that the perceptual and behavioral measures of lexical diversity had only a relatively moderate correlation. In other words, lexical diversity showed weak convergent validity. One plausible explanation is that the global measure calculated from the behavioral measure of lexical diversity had weak validity. As argued above, this global index might be different from an actual index of lexical diversity calculated on the global unit as a whole. For example, a negotiator might employ diverse vocabularies within strategy units but use the same words across strategy units. This may explain why the perceptual measure of lexical diversity had only a moderate correlation with the global, behavioral measure calculated on the basis of an average of behavioral units within the global unit. Finally, the behavioral measures of language intensity and lexical diversity had a strong and negative correlation. Put differently, 92 negotiators frequently repeated the same words or phrases when they expressed their biased or aggressive attitude toward various issues in negotiation. This finding also seems to make empirical sense. For example, negotiators can repeat the same phrases in which they express their own position in order to show an aggressive and strong attitude toward their positional commitment. To summarize, there were some interesting findings in this study. A negotiator’s own and the opponent’s dimensions for task strategies were negatively correlated. The different dimensions for face and task strategies also were negatively correlated. Besides, language intensity shared a negative correlation with lexical diversity. Further research which will explain why these relationships exist seems to be necessary. In the following section, the implications of the new coding scheme and findings in this study will be addressed. I i ti s The findings of this study have implications for both negotiation and politeness research. Further, since the coding scheme for face strategies turned out to be valid and reliable, it is useful to speculate on strengths and implications of the coding scheme. In the following section, the implications related to the findings in this study will be addressed. Then, the general implications of the new coding scheme will be discussed for future research. iaplicatiaas gaiatad ta tha {inning . There are some interesting implications related to the findings in this research. First, it may be more appropriate to examine negotiation process from the self and other dimensions for face strategies than for task strategies. According to 93 the results in this study, when negotiators are concerned about their own outcomes, they tend to have low concern about their opponent’s outcomes. In contrast, they present varying levels of their own firmness (or flexibility) when they uphold their opponent’s autonomy. In other words, the self and other dimensions for face strategies are empirically and conceptually more distinct than these two dimensions for task strategies. This first implication suggests one direction for future research. Given that the self and other dimensions for face strategies are conceptually and empirically distinct, future research on situational conditions under which face strategies for the self dimension are strongly, moderately, or weakly associated with face strategies for the other dimension seems to be necessary. For example, if a negotiator has close relationship with the opponent, he/she might use face strategies between the self and opponent dimensions with positive and strong association. Thus, in future research, various situational factors which might influence the relationship between the self and other dimensions in the use of face strategies can be addressed. Second, the management of face in negotiation may be more other- oriented than self-oriented. The results indicated that face strategies for the other dimension had stronger associations with task strategies and language intensity than face strategies for the self dimension. One interpretation of these results is that negotiators are more sensitive in presenting the information about their opponent’s face than about conveying information about their own face. This implication also suggest directions for future research. For 8X81 |0l" in st) wh sh' UH st CO 'A 94 example, in the future research, the issue of why negotiators may be more other-oriented in using face strategies should be explicated. Further, the issues of when negotiators are more self-oriented in using face strategies and whether the self-oriented or other-oriented face strategies differ in predicting the outcomes of negotiation (that is, which face strategies are more efficient in maximizing task outcomes) should be examined. Finally, face strategies may be more useful than task strategies in understanding negotiation process. According to the results of this study, the nine face strategies were evenly distributed, whereas a contending task strategy was dominant in both negotiations. These results imply that task strategies may be less effective than face strategies at identifying the dynamics on various use of message strategies. This implication also suggests directions for future research on negotiation process. As a matter of fact, in the past, negotiation research focused mainly on the relationship between the frequency of use a: of specific strategies and other negotiation-related variables, such as outcomes (e.g., impasse vs settle-ent, distributive or integrative settlements; for a review, see Putnam & Jones, 1982). As a result, past research has tended to neglect investigation of the negotiation process due to the paucity of appropriate coding schemes and/or the limitation of methods for interaction analysis (Bender & Curington, 1983). Scholars currently are calling for more research which views negotiation as a process (Donahue, Diez, & Stahle, 1983; Putnam, 1985; Rubin, 1983). As argued above, the new coding scheme could be very useful for 95 understanding negotiation process. Accordingly, the new coding system can solve research questions concerning negotiation process such as the differences in the use of face strategies across phases of negotiation, types of outcomes (distributive vs integrating settlements), and the impact of the sequential patterns of the use of face strategies on outcomes. In the following section, the general implication of the new coding scheme will be discussed. Gener im lic ti 8 f r the ew o ' 3 he e. In the above section, the implications related to the general findings in this study were addressed. The new coding scheme also has general implications for negotiation and politeness research. First, the coding scheme could be used to examine how negotiators’ discourse contributes to their interpersonal relationship. Negotiators are concerned with relational development in order to increase common interests and expand cooperation, as well as to maximize their own interest. In particular, relational development between negotiators is achieved through interactions which promote cooperative or competitive moves (Donahue et al., 1983), since negotiators’ talk involves not only the content of messages themselves but also relational information (Bavelas, Rogers, & Millar, 1985; Owen, 1987). Most prior research on negotiation interaction, however, has focused on negotiation strategies with regards to the content of talk. For example, Marley and Stephenson’s (1977) Conference Process Analysis (CPA), developed from Longabaugh’s (1963) categories, classifies negotiators’ communication into four modes of interaction (accept, reject, seek, offer), nine categories of type of information, and seven categories of referent (the 96 subject of the message - self, opponent, party, etc.). These coding systems overlook relational dimensions that accompany the content of negotiators’ talk. Recently, however, students of negotiation have highlighted that relational communication should be investigated, since relational dimensions of negotiators’ talk plays an important role in negotiation (Donahue et al., 1983; Rubin, 1983). In fact, there is an evidence that the false management of relational messages can spoil a whole negotiation game plan. For example, Brown (1968, 1977) found that if negotiators feel loss of face from their opponent’s talk, they retaliate regardless of the cost. Cycles of retaliation might lead to escalating conflict between negotiators (Wilson & Putnam, 1990). In contrast with the prior coding systems, the new coding system was constructed on the ground of relational dimensions, that is, the images of firmness and flexibility which are manifest in negotiators’ interactions. Thus, the new coding scheme might generate findings concerning how negotiators’ talk contributes to relational development, the effect of dimensions such as intimacy and power on sequential patterns of relational messages, and the effect of relational messages or the patterns of relational messages on outcomes. Second, from the opponent’s perspective, this coding scheme seems to be hierarchical in terms of the extent to which a negotiator is perceived to be polite. While politeness is a multidimensional construct, the underlying concept of politeness is regarded as face (Brown a Levinson, 1978; Goffman, 1967; Lim, 1990). In other words, politeness can be achieved by managing another’s positive or negative 97 face. The underlying concept of the new coding scheme is negative face, one politeness dimension. Thus, when it is assumed that other factors related to politeness hold constant, then enhancing the opponent’s autonomy is more polite than giving no information on his/her autonomy, i.e., face-neutral by definition. Further, giving no information is more polite than lowering the opponent’s autonomy. On the other hand, since negotiators are mutually interdependent, a negotiator who constrained or lowered his/her own autonomy might be perceived as enhancing the opponent’s autonomy. Accordingly, it is tentatively assumed that lowering a negotiator’s own autonoay is more polite than being face neutral, and that being face neutral is more polite than enhancing one’s own autonomy. Following this logic, we can divide nine categories of the new coding system into five clusters with regards to the degree of politeness. The first cluster is ’upgrading’ which is the most polite, since it employs the most politeness from the ’I’ and ’You’ dimensions (i.e., it constraints the negotiator’s own autonomy and directly enhances the opponent’s autonomy). The second cluster is ’respecting’ and ’disclosing’ which are relatively polite. The third cluster is ’integrating’ and ’discounting’ which are neither polite nor impolite. The fourth cluster is ’depending’ and ’attacking’ which are relatively impolite. The final cluster is ’distributing’ which is the most impolite. Although it seems to be very difficult to determine the hierarchical order of politeness within the clusters, it can be resolved by Lakoff’s (1973, 1975, 1977) politeness rules. According to Lakoff, indirect expressions are more polite than t t-f-u-é-u-l 98 direct expressions when threatening the other’s autonomy. As argued above, due to mutual interdependence, there are two ways that a negotiator might be perceived as lowering his/her opponent’s autonomy: constraining an opponent’s autonomy and enhancing the negotiator’s own autonomy. While constraining an opponent’s autonomy is a face strategy that directly threatens autonomy, enhancing a negotiator’s autonomy is an indirect strategy. Thus, ’depending’, the indirect strategy to attack other’s autonomy may be more polite than ’attacking’, the strategy to directly attack his/her autonomy. In contrast, when a negotiator saves an opponent’s autonomy, the direct way of enhancing the opponent’s autonomy seems to be more polite than the indirect way of lowering his/her own face, since the former makes the opponent feel better than the latter (Lakoff, 1975, 1977). Based on this rule, ’respecting’ is more polite than ’disclosing.’ When also adopting both of Lakoff’s rules, ’integrating’ is more polite than ’discounting,’ since the former directly saves and indirectly attacks the opponent’s autonomy while the latter directly attacks and indirectly saves the opponent’s autonomy. Thus, we can logically arrange the hierarchical order of nine face strategies in terms of the extent to which an opponent perceives politeness from a negotiator’s discourse’. The hierarchical order of politeness corresponds to the number assigned to each face strategy in Appendix A. That is, face strategies with lower numbers are more polite , while face strategies with higher numbers are less polite. Although there were no data to test the degree to which the nine face strategies are perceived or related to politeness in this study, 99 the behavioral measures for face strategies, language intensity, and lexical diversity indirectly provide the ground for testing the relationship between face strategies and politeness. As argued above, politeness might be inversely associated with language intensity and positively with lexical diversity. To indirectly test the relationship between nine face strategies and politeness, subsidiary correlation analyses were performed. In these analyses, mean scores for language intensity and lexical diversity for the nine face strategies were calculated. Then, correlations were computed between the theoretical politeness of nine face strategies (scored 1-9) and the mean levels of language intensity and lexical diversity for those strategies. The results appear in Table 11. The results indicated that the nine face strategies were closely related with politeness. As expected, the nine face strategies were inversely related to language intensity (; = -.74, p < .05) and positively associated with lexical diversity (3 = .43, p ) .05). Although the correlation between the nine face strategies and lexical diversity was not significant at the level of .05, the effect size was substantial when considering the small sample size (a = 9) and low power. To summarize, these subsidiary analyses suggest that the nine face strategies might be positively associated with politeness. However, future research which accurately investigates the relationship between the nine face strategies and perceived politeness should be conducted. This hierarchical feature might generate further interesting future research on negotiation. For example, the hierarchical coding scheme enables students of negotiation to investigate the impact of situational 100 Table 11 s'-'_-18 a.“ 10. e», 01: 0‘ 4-11-8.5 teni ' a-“ ‘ ' -;, _iV ' ' (or Wages Face Strategies Language Intensity Lexical Diversity Upgrading 1.86 .79 Respecting 1.53 .92 Disclosing 1.99 .82 Integrating 3.88 .81 Discounting 4.34 .75 Sidestepping 2.57 .86 Depending 3.40 .83 Attacking 3.60 .82 Distributing 8.52 .70 Mata. as = .74 (p < .05) between face strategies and language intensity and -.43 (p > .05) between face strategies and lexical diversity. For these correlations, Ms are nine (face strategies) and distributing was coded as "1," whereas upgrading was coded as "9." 101 variables such as relational history, intimacy or power differences, and individual difference variables such as the need of aspiration or social desirability on the extent to which negotiators are polite to their opponents. Finally, since the new coding scheme reflects the mutual interdependence of negotiators, it is likely to stimulate research on how negotiators manage tensions between competing goals through the manipulation of their images. Negotiation can be described as a process in which incompatible and competing goals are transformed into compatible goals (Wilson & Putnam, 1990). Negotiation necessitates communication and coordination in order to solve the problems caused by perceived incompatible goals. In other words, participants are mutually interdependent for problem solving. The new coding scheme reflects this characteristic. Thus, the new coding system seems to be an effective tool for investigating how negotiators manage images in order to reduce tensions between competing goals. For example, the coding scheme makes it possible to investigate what patterns of image management lead to changing and reestablishing negotiators’ goals. In sum, the coding scheme for face strategies developed here seems to have several implications for future research by resolving the conceptual problems present in prior coding systems. Specifically, the coding scheme was developed an the basis of negotiation as a process involving mutual interdependence and relational messages. As indicated above, these features show strengths with regards to generating future research. Thus, the new coding scheme seems to serve as a useful tool which leads to new trends for research on negotiation. 102 A few limitations of the current study, such as the use of global measures, the use of a task coding scheme modified from Pruitt’s task strategies, and the lack of applying the face coding scheme to various types of negotiation, should be acknowledged. First, the use of global perceptual measures might have influenced this study’s findings. The perceptual, global measures might be different from perceptual measures for each unit. It is possible that such differences work differently for variables investigated in this study. Fortunately, for most variables, behavioral measures produced findings comparable to those generated by the perceptual, global measures. Studies which employ perceptual-micro measures, however, should be conducted in the future in order to accurately test convergent validity. Second, the use of Pruitt’s task coding scheme might also be reflected in this study’s findings. Pruitt’s coding scheme originally focused on a macro-level strategies, not micro-level tactics. Accordingly, in this study, the coding scheme for task strategies modified from Pruitt’s coding scheme was inevitably employed. Prior to this study, that coding scheme should have been validated. However, without validating the revised task coding scheme, this study employed it to validate the coding scheme for face strategies. Although overall results of this study indicated that the coding scheme was valid, a measure for task strategies which focuses on a micro-level analysis for interactions of negotiations should be validated and elaborated. Then, future studies should be conducted to investigate the relationships between task and face strategies with more elaborated and valid measures for task strategies. A third limitation in this study was that the images of firmness 103 and flexibility were considered only in terms with autonomy, i.e., negative face. It is likely to manage the images of firmness and flexibility by using strategies related to positive face. Put differently, negotiators might manage their images by varying their own or an opponent’s positive face within the same strategy for negative face. Accordingly, further studies should be conducted to develop a coding scheme for positive face and to examine the relationships between negative and positive face. A fourth limitation in this study was the use of naive rather than expert judges for perceptual data. As Donahue, Diez, and Hamilton’s (1984) study indicated, naive judges might differ in perceptual ratings from expert judges. These differences might be reflected in the results of this study. Thus, future research should be done with expert judges in order to further validation for the new coding scheme. Finally, the data of this study were limited to two types of naturalistic negotiations. Future studies should apply the coding scheme for face strategies to a variety of negotiation situations in order to enhance external validity of this coding scheme. Further, studies on whether the coding scheme for face strategies can be applied to other situations such as conflict should be conducted. If not, it seems to be desirable to develop general coding scheme for face strategies which can be applied across situations. 104 NOTE 1 Brown and Levinson acknowledge that the speaker’s and hearer’s face are interdependent. However, their politeness strategies don’t reflect this features only by considering a hearer’s face. 2 Actually, summed language intensity without being divided by total words showed high convergent validity with the perceptual measure of language intensity (a = .50), whereas averaged language intensity divided by the total words indicated low convergent validity (a = -.27). Thus, summed language intensity seems to be more valid measure. 3 For LISREL analyses, the model adopted to test construct validity in this study, i.e., the measurement model of four underlying constructs measured by two different measures was locally underidentified. Accordingly, it is unlikely to test the construct validity for face strategies with LISREL. However, comparison between single-factor and two-factor models employed to test the relationship between self and other dimensions for face strategies was likely, since both measurement models were averidentified. ’ In order to logically determine the hierarchial order of nine face strategies in terms of the degree of politeness, two assumptions were established. First, it was assumed that the strategies could be rank- ordered in the degree to which a negotiator is polite as judged by an opponent rather than by the negotiator him/herself. Since politeness is a social value (Lim, 1990), it seems to be reasonable that a hearer or a third party judges a speaker’s politeness. Second, it was assumed that Lakoff’s (1975, 1977) rules of politeness in directives could be applied to the new coding scheme, since the underlying concept of the rules is relevant to that of the new coding scheme. Directives are a speech act threatening a hearer’s autonomy (Brown & Levinson, 1978) and the new coding scheme is also focusing on an autonomy. By using Lakoff’s two assumptions, the hierarchical order of nine face strategies was determined not empirically but logically. Thus, further study for empirical evidence seems to be necessary. APPENDI CES Appendix A CODING MANUAL: NEGOTIATOR’S IMAGES OF FIRMNESS & FLEXIBILITY This manual contains instructions on coding negotiators’ images of firmness and flexibility which appear in their verbal interactions. It outlines nine general strategies in which negotiators manage their images of firmness and flexibility. This manual is divided into three sections. Section one defines the concept of the images of firmness and flexibility and differentiates it from related concepts. Section two describes rules for unitizing. Section three presents rules for categorizing image-management strategies present in negotiators’ verbal interactions. Section One: Key Concepts (A) Defining the Images of Firmness and Flexibility. The images of firmness and flexibility are conceptualized in terms of autonomy. The image of firmness in negotiation results from attempts to maintain or protect the rights to behave independently and be unimpeded in pursuing a position, proposal, or course of action. The image of flexibility is just the opposite. That is, by giving up autonomy, negotiators cultivate the image of flexibility. Since negotiators are highly interdependent, they can do this in two ways. Specifically, the image of flexibility can be defined as the degree to which negotiators concede their own autonomy and protect their opponent’s autonomy. The image of firmness reflects the degree to which negotiators maintain their own autonomy and constrain their opponent’s autonomy. Section Two: Rules for Unitizing (A) Definition of strategy A strategy is an abstract behavioral unit and provides general guidelines for action (Berger, 1985). A strategy refers to a sequence of actions or to a family of related actions (Wheeless, Barraclough, 5 Stewart, 1983). The coding scheme intends to code the extent to which negotiator’s message content as well as expressions contain firmness and flexibility. Since negotiators manage their images of firmness and flexibility through strategies, a strategy seems to be appropriate for coding the extent to which negotiators intend to show these images. Most often a single argument consists of one strategy, since it usually expresses one intention. When negotiators attempt to transfer their intentions, their face strategy contains explicitly and implicitly a claim (i.e., what negotiators believe) intended for their opponent to undertake and evidences which support the claim. Accordingly, one negotiator’s argument (i.e.,a claim and its supportive evidences) is regarded as one strategy. Negotiators, however, sometimes can manage 105 106 their face strategies without providing clear and supportive evidences for their arguments. For example, negotiators can make assertion ( e.g., a negotiator simply says, "I don’t want it.") or simply express motivation for supporting their assertion (e.g., threat for strike). (B) Conceptual Differences between Strategy and Tactic Militarily, a tactic is a method of deploying forces in combat, whereas a strategy is the employment of political, psychological, and military forces so as to afford maximum support to policies. In other words, a tactic is a more specific act carried out in support of an overarching strategy. In conflict theory, a tactic refers to a single action, or a single message in the case of communication, while a strategy refers to a sequence of actions. (C) General Rule for Unitizing Generally, cutting points for face strategy as a unit exist, where - negotiators’ claim is altered - main current issue (e.g., grievance, salary, insurance, etc.) is shifted (D) How does Strategy Compare to Other Units? 1. Turn - turn is each speaker’s utterance. every turn has more than one strategy. - most of one turn are often one strategy, but not always. (e.g.) M: David Krause didn’t even speak to the person, director. - if turn is larger than 5 sentences, it is coded into two or more strategies. (e.g.) M: We, we have discussed that we do not doubt that Mr. Krause was sick. He was sick. However, um I cannot reiterate enough that Article 8-2c clearly state that in order to be eligible for holiday pay an employee must work his last scheduled work day prior to and his last scheduled work day after a holiday. He knew this and Conig one of these people who testified for the company’s consistent requirement that employee work at least part of the day before and after a holiday in order to be eligible. He reiterated that sick leave required company approval. A doctor’s certificate is insufficient and one day absence had never been construed to be a sick leave. - look at within turn cues to determine the number of strategies (see below). 107 2. Sentences - strategies 49080) M: (ease) L: - strategies can be one or more sentences. Unacceptable. I’m not asking you if you’re a doctor. You just told me if a doctor agrees that if David Krause is sick that you being management would agree to it. Is that not what you just said? typically start at beginnings of new sentence. 3. Arguments (Toulmin’s sense) - claim + supportive evidence = one strategy. (e.g.) M: We say there are arguments that we don’t feel Mr. Krause is eligible for sick pay. I mean holiday pay, because he did not show up the day after the holiday he was informed by our assistant personal director. His wife spoke with her and told her that he needed to be there on the day of work. He was made aware of that by his wife and yet he made no further attempts to call us, talk with us and so at this time that is our position. (E) What are Cues for Beginning/Ending of Strategies? 1. Cues for start of strategy - beginning of sentence (capitalization) (e.g.) M: We, we have discussed that we do not doubt that Mr. Krause was sick. He was sick. However, umm I cannot reiterate enough that an Article 8-2c clearly state that in order to be eligible for holiday pay an employee must work his last scheduled work day prior to and his last scheduled work day after a holiday. // He knew this and Conig one of these people who testified for the company’s consistent requirement that employee work at least part of the day before and after a holiday in order to be eligible. He reiterated that sick leave required company approval. A doctor’s certificate is insufficient and one day absence had never been construed to be a sick leave. - markers for different arguments (not different supportive evidence for same argument) (e.g.) "first, second, ..." (e.g.) L: Article 2G employee who is granted a sick leave shall be paid for holidays following within the first 30 calendar days of his approved sick leave. ... 26 was read in such a way that its intent was to be interpreted according to each individual case. // Article 20 is a second one. ... 108 - long silent pauses (e.g.) M: Oh. O.K. Well, I know that many people have and I find it very hard to believe that we all have once in our life. Like I said I’m not discrediting the doctor (pause) // but he did not go about the right way of getting an approved sick leave. If that doctors say he was sick I have to take that doctors worked for it because that doctor is laying his medical eligibility on the line. 2. Cues that 2nd sentence in part of same strategy of prior sentence - cues that 2nd sentence elaborates 1st sentence (e.g.) "I mean," "for example" (e.g.) M: We say there are arguments that we don’t feel Mr. Krause is eligible for sick pay. I mean holiday pay, because he did not show up the day after the holiday he was informed by our assistant personal director. His wife spoke with her and told her that he needed to be there on the day of work. He was made aware of that by his wife and yet he made no further attempts to call us, talk with us and so at this time that is our position. - cues that the 2nd sentence is claim which is supported by lst sentence (e.g.) "therefore, thus, accordingly, consequently..." 3. Cues for end of strategy - end of sentence (punctuation) or turn (F) When an Overlap (Simultaneous Talks) Occurs - In the next turn from the same person, if he/she continues or repeats the same argument, the previous + next turns are regarded as one unit. (e.g.) L: Oh, so this is not uh, a max, this a, we’re not M: We’re not talking about... L: (can’t) talking about maximums ... - If negotiators make new argument or shift topics, the previous and next turns are coded as separate unit. (e.g.) L: Would you give us an indication first of all... M: Yes, may I speak with my business manager here. We have some costs we have already taken into consideration. L: Well, I recognize that. - Meaningless utterance caused by an overlap is coded into the same person’s next turn. 109 (e.g.) L: If it’s going to be a three hour caucus we just soon not sit here and wait for you. M: Uh... As ... L: I’m assuming that when you return that you will respond to our proposals in the area of benefits. M: I’m involved in caucus. Section Three: Rules for Categorizing (A) Summary of Nine Face Strategies in Negotiation 1. Upgrading: The strategy enhances an opponent’s autonomy, while lowering a negotiator’s own autonomy. (a) Agreeing to an opponent’s proposal/opinion through a concession * L: We want protection under 21.12 because of use of Part -timers. It’s obvious you want Part-time in some areas, and we’re concerned. M: Q,§. Wg’ii check. Wa gag pa zaaaaa ghx they agen’t Regaiap sum * M: Our proposal is simple. We have a requirement for Temporary Supervisory replacements. We don’t overuse them. We don’t need more management. They should have the same treatment in Network as other departments. L: W s v ’ M: A good move. A significant move. * M: I’ll check that. Now the phone centers are different along with the operators. Seniority should be adjusted two times a year. L: It’s not just the phone centers. Like Term employees. The lack of following, or reporting, we don’t know whether they do it or they don’t do it. Mzwmmwmmm (b) Combinations of disclosing + respecting strategies given unit of analysis * L: We want the money. M: O.K. Then he would have contacted, he should have come in on Monday. 110 2. Respecting: The strategy is to enhance an opponent’s autonomy without affecting the negotiator’s own autonomy. (a) Supporting/understanding/accepting an opponent’s arguments (agreement with ideas, not proposals) * L: We can use your argument that money is one thing but how many people left the small town to come to these type of jobs. Most of the people hire into this job. The dollar is sure a part of the consideration, but we realize there are other problems. mishanuhmmgleuauliammmm W (b) Emphasizing the opponent’s positive value/attitude/behaviors * M: Our proposal is simple. We have a requirement for Temporary Supervisory replacements. We don’t overuse them. We don’t need more management. They should have the same treatment in Network as other departments. L: We made a move. The monkey’s on your back. WW * L: The job functions and monthly quotas are different. They have more time to increase the evaluations made. Any calls of a failure must be noted and they must have knowledge of all company policies. It’s as increase in the technical nature. With competition, it’s a more important tool for customer contact in the Bell system. =Iha_t_’_s_an_lnteneatins_9hseuatlons (c) Linguistic expressions minimizing the opponent’s impositions (As far as I’m concerned,) 3. Disclosing: The strategy lowers a negotiator own autonomy without affecting the opponent’s autonomy. (a) Disclosing a negotiator’s own weakness (but not making concession) 4 M: For certain specific functions. And from within ourselves we try to cover all the needs that we’ve got. I’ll still have the bloody need, whatever you do. U: Yeah. WWW (b) Apologizing to his/her opponent * M: And er, I would suggest that if this arises again on a question of interpretation of the thing, this is all we can do. You know, I mean there’s no other solution for this but in fact to raise it and discuss it. 111 03W (C) Accepting negotiator’s own fault/responsibility, etc. * L: M: L: Seaton is supposed to be working on it to reclassify all the employees in Racine stated as term. We have to take a look at that. No one told them of Term employees in Racine. M: Ma’re guilty, Raging is baiag fixad. * L: The work has value to you. You were willing to pay money before. M: That waa a aistata. (d) Offering disclaimers expressed in the current tense *M=WWJMW 4. Integrating: The strategy integrates both negotiators’ face by enhancing both negotiators’ autonomy. (a) Making new proposal with which both negotiators’ demands can be satisfied. : With 18 months, are you telling me you would accept the rest of the provisions? 2 and 1/2 years was a big concern with our people. We were saying that we would look at the rest. But the 2 and 1/2 years was much too long. People designated as Term employees, we can’t bargain away what they already have. . - 9 = Wanna immune . We tell people to make the schedule suitable to the needs. We tell them no to develop a schedule with no relationship to requirement. Be a se a ' h s s d it d w d disc 8 s e es w t e ocked. O ce t e b i ess h s 't res ric ' WW restrictiaas, : Sure. If we consider expanding restrictions, some areas are good for us and bad for you and visa versa. : We have no problem living with present restrictions if you give us more open mouth. Sounds fair to me. Everything is fair. (b) Fostering both negotiators’ positive feelings by emphasizing mutual interdependency * L: Redefine qualifications for Call-Out Pay Plant. We want to see the Plant application built in to all appendixes. 112 . With a two hour minimum? . It’s difference of when you get paid and the two hour minimum. We get calls at home in the middle of the night. W You had same requirements before and they don’t make sense. You violate contract as far as I’m concerned. Don’t you think we should have some control on movement? . I don’t have a problem on lateral movement. It’s the way the policy is presented. = When It’s been a problem with us too. (c) Combination of defending + respecting strategies * understanding the opponents + repeating their prior arguments L: There are currently three people that we are looking at. They have technical responsibilities along with other specialist work. We feel that they should have a title higher than a Group 5. This is not the year for heroic job upgrades. We do not want to upgrade in ’80 and have to downgrade in ’83. e ers ° w e t w ' r e ‘ s. * Avoiding criticisms made by the opponent + Supporting the opponent’s argument M: L: it's-=5 What other problems do you have with supervisors? If employee talks to supervisor and tells them the problem, they say "no." We have an operator in Racine that was threatened with discipline if she kept a doctor’s appointment in the years of her 1977 absence record to use as a discipline. Well, I don’t dictate to my boss when I want to go. Ann, what did we do for you when you went on leave? I got what I deserved, but only through grievance. I had to fight for it. We have made attempts to talk to manager and grievances have been dropped. Ann made statement about the blind operator getting caught in the elevator, took it up to third step before it was settled. When they loss control at the first level in network, that is where we lose communication. 113 5. Discounting: In this strategy, a negotiator tries to lower his/her own as well as his/her opponent’s autonomy. (a) Compromising (the outcome forces both parties to make sose concessions) * U: This is a must item. Equalize the salary for coordinator and chair -- they perform the same function. :1’ m ngt sure 1’s P11 lag ta g9v agaai bagka (a at aaiazy) r ECA b t i th (b) Suggesting a willingness to compromise * L: We’ve had some problems with bulletin boards. M: What type of problems? L: We’ve had problems. Three of them. Example: District Supervisor and Network removing information from the bulletin board. M: We’ ' ' s ' i t'v e to s. (c) Using displays of weakness to pressure the opponent to make concessions * M: That is, the four of us, or five of us in the talk. So you’ve now come back now the full circle back to where you begin. U: Yes. M: Now, could I ask one or two things? U: Yes. M: _hi: - - z a, - .a 5;: t at ve vered. is is m ‘ e t bec se mewhere there as t to be ver e. w est' 8 k ' st ' t 1 me e attem our t t ri e r t is e te e .? 3 M: What you’re saying is, ’This is your problem’. U: Well we felt it was impossible at the beginning. M: All right. U: But we still agreed, we done it for the last five years. M: r' ’ ' ' es mm (d) Threats which indicate that one can take any loss if the opponent does not accept one’s proposal r"——fi 114 (e) Combination of disclosing + attacking strategies * accepting one’s own fault + disagreeing with the opponent’s argument L: O.K. Do you think David Krause was taking advantage of that type of behavior as you speak up? It’s very possible that he has taken advantage of that because. Ums. Because. L: Do you think David Krause would to the doctors and get a doctor’s notice and take advantage of this. M: Did you ever get a doctor’s note for missing a day of school, because you wanted to stay home, I know I have. L: I haven’t. M: W t v ' i v h d t be ieve t t we ve ce in r i e ike s i ’m not iscred t n t e ct r b t e i about the ri t w ett'n an rove e v . If that doctor says he was sick 1 have to take that gggtgzg rke r i s t ' eligibility on the line. 6. Sidestepping: This strategy maintains both negotiators’ current levels of autonomy. (a) Procedural statement re v r w ich we 0 r e (b) Sumsarizing/clarifying others without an evaluation * U: If I was on bank holiday I would have said, ’Nell look, I did the last bank holiday’. 9 ) (c) Neutral expressions for information-seeking: asking questions which do not imply negative qualities about the other, or do not ask for suggestions for creative proposals * : We don’t want to make a scene. MW 7. Defending: Through this strategy, a negotiator tries to enhance his/her own autonomy without affecting the opponent’s autonomy. U M (a) Emphasizing prior concessions as a means of refusing more 115 * M: What you’re saying is, ’This is your problem’. U: Well we felt it was ispossible at the beginning. M: All right. U: But w w e ' e t 'v (b) Rebutting/avoiding criticisms made by the opponents without demanding/constraining * U: The fitters was there, and we were there. Then came out the original callout procedure, which has never been altered on paper. And it said 365 days per year, and we kicked up then that we did not agree to this, ’cos it said there ’as agreed by us’. We kicked up then, us fitters, that this had never been agreed on. M: When was this now? This was way back, was it. U: Yes. This was the very original... U: (continue) callout procedure. M: s 9 ~ ' ' t e U: Yes. 14:me ( 4 intervening turns ) U: but we’ve never wanted to do it. uzmmmmmmmmm e t ' ' ’ ev i w e v t (c) Persisting in a Positional comaitment * M: There you go again using snow days. What’s a snow day? L. City declared snow emergency. (d) Supporting/repeating one’s own prior argusents/thoughts/proposals * M: In 1974, company people got better treatment. We leveled up to the best in the system. Four companies are less liberal. None are more liberal. We will testify to that. Cannot do any better. Will not do any better. L: I didn’t say this would come cheap. (e) Making one’s own proposal only satisfying one’s own desire t (starting discussion about new item) L: www.mnmmnmm 8. Attacking: The strategy is to lower an opponent’s face by attacking his/her autonomy without affecting the negotiator’s own autonomy. (a) Questioning the accuracy/veracity of the opponent’s prior comment * L: So you’re saying that David Krause are flat on his back and be asked to come into work? M: And, what’s his name? David Eddie did when he was sick. (b) Rejecting/disagreeing with an opponent’s arguments * L: Item 17: Standby Pay Provisions - M: That’s new. Give example. L: There is so much pressure fron management. We’re willing to discuss it. WWW * U: We’ve always had bank holidays different to the other people. 11anth (c) Expressions for interrupting the opponent’s talk * M: The fact was though after he was told you must show up to work neither did he call us back to try to talk that out with our assistant personal director because our director would have said O.K. mm (d) Criticizing/faulting the opponent * L: This illustrates the inportance of safety and health language in the contract. We have to get in on the engineering and the design of equipment. That’s important. Some avenue other than just looking at the equipment after we get it. M: Read section 5 of the Exhibit 13 aloud. You want this in our contract? What are you going to do with this money? L: Too bad several telephone company employees did not enjoy this language before. In some cases, it nay have saved their lives. 117 8 L: More people than this came to us to get job upgraded. We asked then to write it out. M: You made independent decisions? L: Everybody has an opportunity to put in for a job upgrade. M: ’v w t i 19.4mm (e) Hostile questions (Direct/leading questions) * M: Changes are normal mistakes. L: Injury to one is an injury to all. We don’t pull on intensity of problem. M: e 9 s t ‘ 9 ’ r e ? * L: How do you feel about 2c? M: To be eligible for holiday pay an employee must work his last scheduled work day prior to and his repeat or either be there on the 21st to work and the 26th in order to get his holiday pay. And if he was not going to be there the let or the 26th and or he would have got an approved sick leave through our company. 9. Distributing: In this strategy, a negotiator enhances his/her own autonony and simultaneously lowering the opponent’s autonomy. (a) Positional commitment accompanied by disagreement * U: I just wondered if you could have a certain person, say, on these holidays that you could say if there is something happening one Saturday or on a bank holiday you get in touch with him. And try and get one of the electricians. Instead of one electrician being on call, perhaps, if this is agreeable with the other two. mmmmuhmmmum W (b) Suggesting win/lose proposal - beneficial to self and detrinental to other : You post 14 days in advance for most part. Where’s the late posting? Some post 11 days before and there is no penalty. How about posting 7 days in advance. It’s hard to make plans. Well, it’s hard to make the schedule. Schedule is usually It‘ll-‘3!" 118 Iade out three days before it’s posted. ( 4 Intervening turns )8 We (c) Attributing all the responsibility (or faults) to the opponent, not to self. By the sale token, negating one’s responsibility WWO * M: Now suppose we recruit an electrician, and say, ’well now, part of your job will be in fact to cover these days.’ U: What, bank holidays? Oh, the guys wouldn’t think that’s fair. M: We’ll ask the recruit. U: No, no, no, no. For us to sit at home, go, to go out there five days and leave Joe Soap in for work. M: W w we ' ? t ’ e n 1 gan’t,.. (d) Combinations of defending + attacking strategies * supporting one’s own position + criticizing the opponent L: What are the restrictions? Must it be taken in a calendar month or say it be taken week to week? M: We will check it out and get back to you. e e (e) Threat/command emphasizing that the opponent should take a loss if he/she does not concede. (B) Rules for Categorizing 1. If two or more than different strategies can be applied in a given unit, code them as higher number of strategy. * code the following example as respecting (although it can be coded into disclosing) 1’ 11 1993 at you; prongsgl. 2. Several sane strategies found in a given unit are coded as one strategy. * an example of defending (three same positional commitments) 9 holiday.- e o w t e . w t b k 3. 119 If contradictory strategies which do not appear in the coding scheme (e.g., defending & disclosing; respecting & attacking) can be applied in a given unit, code then as combination strategies (upgrading, integrating, discounting, & distributing) according to the following next rules. (a) Code disclosing to respecting. * an example of integrating L: When other people call out sick in other department, or call out because of a snow emergency, are they marked absent? M: Yes, they are. L: Are you sure of that? M : Yes, we are. We don’t like the work "Snow dgys" and we don’t like this. We’ll get bagk to you on it. (b) Code attacking to defending. Distinguish overlaps which are aggressive, nonsupportive (interruptions) from those that are supportive (confirmations). * interruption M: We’re not arguing that fact that ... L: But you’re trying to ask is sound like he was not so sick. * confirmation L: Yes. I mean that David Krause ... M: Go ahead. * simple overlapping (1) L: Oh, so this is not uh, a max, this a, we’re not M: We’re not talking about... L: (con’t) talking about laximuns ... (2) L: Would you give us an indication first of all... M: Yes, say I speak with my business aanager here. We have sone costs we have already taken into consideration. L: Well, I recognize that. (3) L: If it’s going to be a three hour caucus we just soon not sit here and wait for you. M: Uh... As ... L: I’m assuming that when you return that you will respond to our proposals in the area of benefits. M: I’m involved in caucus. Appendix B PRUITT’S TASK STRATEGY CODING SCHEME 1. Problem-solving strategy (a) conceding with the expectation of receiving a return concession * M: Let’s look at item 2-G. inistr t' t i t e N tw r rtle t w ’ ve Other e rt ents re 9.!E. (b) mentioning possible compromises as talking points * L: Let’s love eh. gee g: 9 he 2 a ll: yeeze en yen; effer, Like to see it cht in hell, With l8 mohths eh the Iere ~ employees. Nine lenths ls whet they eye leylhg in hhege he! w itin or u rade M: Glad to see you are looking at this one. 18 months is a little tight for our taste. With 18 months, you are telling me you would accept the rest of the provisions? L: Yes. M: I would agree with that. As I said, I’m glad to see you are looking favorably. * L: Item 16 - Eliminate entirely. M: You don’t want to permit us to have Temporary Supervisors? It’s going to be a long summer. Network is the problen. It won’t sell. We’re not hurting people. We don’t need it. We need it. We don’t need a lot of the contract. Wings. You mean to promote? Yes. F'ZP‘ZI" (c) revealing one’s interest (i.e., one’s goals and values); * M: We know we talked about Holiday business. That’s a tough one. . We’re looking for more than one week, day at a time. What do you have in mind? Limb... I want to modify modify relief provisions. What are you looking for on this. MW Lengthen them? We’ ' ' . 1*") N) FSFSF 5.3.5? .0 O. O. O. O. 120 121 (d) retracting a proposal in the face of resistance * L: No other out than have a 30 day leave of absence. M: You can’t go on leave for less than 30 days. (4 intervening turns) L: What are the restrictions? Must it be taken in a calendar month or may it be taken week to week? M: w' ec 't t t t (e) expresses a willingness to compromise * U: This is a lust item. Equalize the salary for coordinator and chair -- they perform the same function. M: 1’! net sure 1’; wllllhg te ge egnel bneke (z 91 eelegy) fer C t ’ ' i ve ' * L: We’re going to catch a lot of flak on this. We have lots of probless on this one. M: Well, I feel maybe this has become a way of life and easier for the people to accept. (f) promises reward or withdrawal of sanction if the opponent behaves in a stated manner * M: Why all of a sudden has this come up? L: We’re the only ones with differentials in different job titles in any appendix. M: If we gave you this, you’d really get on our case for using S.A.’s as Operators. o y o ‘ L: This item doesn’t cost you a dime. We will have a contract a with the ability to arbitrate a discipline. M: r t i i we’re 1]", A .J 0’ e we sheuld sit doyhuend-discues'sehedhles end yhy they gee" ke e t b si ess h s i at ' ' ifti eo e i ch e w' :eetrietions. M: Sure. If we consider expanding restrictions, some areas are good for us and bad for you and visa versa. 122 * M: Don’t you think we should have some control on movement? L: I don’t have a problem on lateral movement. It’s the way the policy is presented. M: s we t r s t er L: It’s been a problel with us too. (h) sending disavowable intermediaries (i) talking in back channels (j) communicating through a mediator 2. Contending strategy (a) making desands that far exceed what is actually acceptable (i.e., one’s resistant point) (b) making comaitments (to unalterable positions) * L: Itel 16 - Eliminate entirely. M: You don’t want to permit us to have Temporary Supervisors? It’s going to be a long sum-er. Network is the problem. It won’t sell. We’re not hurting people. W ’ e t We e 't. We n’t ee t t n r t * How can we be a part of something we don’t believe in? Like I said, you put the Union above the law. No, we challenged it. We are willing to talk about certain items such as tine in title in certain areas. MMAWMLWLLM =r=¢ as (0) making persuasive arguments aimed at convincing the other that (concessions are in his/her best interests) * M: Union item 12 - Reduce nileage requirements for moving. Reduce 35? L: Reduce to 30. M: They’re not aware of anyone physically moved over 30 miles. L: ’ e ri r e fut re it e st . e ’ w bo iv'n w o 7 mi e hehz. § hile geductieg wen’t he eheh, (d) using threats * M: Can’t agree to any change in arbitration language. We have no intention of changing this at this time. L: This is going to give you trouble. M: I’m sorry, that’s the way it is. I wish you could help me on 123 this one. I have a boss that will kill me if we love on this issue. L: 's v r ' e t e e w 1 be a settles t ' t is iss e w’th. * M: ' we w 1 ive in his 0 d i u t e t t n ess we’re i write a s. (e) rejecting (the opponent’s proposal/position) * L: No other out than have a 30 day leave of absence. Yo can’t e ve r ess t (f) demanding that the opponent makes accommodations, or concessions * L: Some do get out. The value of the work is compared to others. They always end up on the short end of the stick. M: Autonobile industry is right down the tubes. Industry has gotten greedy Unions have gotten greedy. We’re pricing ourselves right out of the market. L: We e ki a f 'r w e r e f b in Qenetreller, There is get e hell ef e let e1 thee, (g) Faulting/criticizing the opponent’s thought or position * L: All this stuff is hocus-pocus. Breaking codes is just not so much a title but what people are doing. Where week is the sale, titles are the same. One on one is easy. Lumping creates confusion. M: ’ e s ' ' ' * M: Housekeeping proposals - Titles #1 - Delete Supply Expeditor. Add Supplies Coordinator. #2 - Delete Cook, Cafeteria Attendant. (reading six more items as their proposals) L: Throw the hell te yen. Den’t yoe knew ohe eeeh thieg eheet ' ? W ’ e 9 e n es b t w e e ' on o 's in uish t ? (h) Avoiding or minimizing one’s own responsibility * L: There is so such pressure from management. We’re willing to discuss it. M: There can’t be too such talk. L: You haven’ t talked to the right people. 124 (i) Requesting a specified change in the opponent’s position or thought * L: ltes A 24,9 . We’d like e eerled net elte; the word "leyeff" in the thled line eee seretch the beleeee, (j) Asserting rights/needs * L: M: L: We wonder why we have Terms. There is a definite need for Ters esployees. Milwaukee is big enough to absorb the people as jobs and offices change. There is no problem with need. The problem is with use. It allows the company to escape any obligation to the employee. We’re looking to have employees reclassified after one year of service. : Twelve sonths is a short time. We have interest in Term employees. They should have no rights in the Upgrade and Transfer plan. T e union ri t w t e e 's i u 1 w t ' t e e is t? We e re ei e it. The Uhion is entitled to know semethigg thet effeets its members. : This one will go down hard. . We are suggesting 6.01 be eliminated and 6.02 be modified. Are you saying everything should be arbitrable? W ' r t ri t t st ' e. (k) demonstrating that there is more (tise) pressure on the other than on oneself (l) retracting a previously sade concession so as to sake the position clearly less agreeable to the opponent 3. Yielding strategy (a) making concessions by giving up one’s position * L: How can we be a part of sosething we don’t believe in? M: Like I said, you put the Union above the law. L: No, we challenged it. M: We are willing to talk about certain items such as time in title in certain areas. L: We are putting a blessing on your plan. No way will this put a difference on challenging your plan. We will see what you come up with. 1‘1qu (b) agreeing to the opponent’s proposal/position without modification * L: Item 7-A. We would like some inforsation on the "bench- 125 marks" for wage credits. M: H ve e w'th t . (c) Accepting one’s own responsibility * L: Seaton is supposed to be working on it to reclassify all the employees in Racine stated as Term. We have to take a look at that. : No one told thes of Ters employees in Racine. We’re geilty, Reelee ie helng Iiyee. (d) Expressing understanding, acceptance, or positive regard for the opponent .Zt": * M: We are not sure what you want. L: We do not need separate units. We don’t see any large obstacles and we bargain and ratify as total units. M: We will leeh et it. (e) Soliciting complaints about one’s own self 4. Other: all the strategies which cannot be coded into the above strategies (a) Procedural statement * M: W ’ t is ime L: w t ' i e e s? M: We will be back with proposals. (b) Sumsarizing/clarifying others without an evaluation * M: Xee’he eeyleg thet theee eereeetege tignhe le 19. 19 is right? (c) Neutral expressions unrelated to tasks (i.e., position or outcomes) * M: I’m t i i wit ew r r s t t ve. u MW APPENDIX C LANGUAGE INTENSITY CODING SCHEME Obscure words: Instances in speech which involve the use of "uncommon" words or phrases. This includes foreign words, polysyablic words, and rarely used words. For example, "lascivious" rather than lustful; "coup d’etat" rather than a rebellion; "aggrandizing" rather than enriching; "despotic" rather than severe. General metaphor: This category involves words or phrases in which the denotative meaning of a term is used in a sanner other than it is conventionally associated. For example, "I feel like I’s sinking.’ coaster." and "My life is a roller Profanity and sex-based statesents: This category includes the use of words to convey profanity and terms associated with the practice of and traffic in the sex act and related events. For example, "bullshit," "piss-off," "asshole,"; references to fashion publicists as "pisps" for fashion industry, progressiviss as "prostitution” of education, students in progressivism as being "raped by" the system, college admission standards allowing a "perversion of" the aims of education, and so forth. Death-based statements: This category includes words or phrases which involve the use of death-based statements. For exasple, "I’m going to waste him," "I’m have to end it, and "I’ll kill her,"; references to women’s fashions as "ghastly," the attribution rate asong college students as the "fatality" rate, the "decay" of education, the "death" of conservatism, the "murder" of traditional education, and others. 9' Qualifiers This includes adjectives and adverbs, auxiliary verbs, affirsations, negations, contractions, pronouns with contractions, and certain "special" words. Adjective and adverbs Code all adjectives and adverbs. For example, code ”beautiful" as one qualifier, not just "incredibly" in the statement "incredibly beautiful." Auxiliary verbs Code such words as: will should done do might could can 126 C. F. 127 have as had does would are is were was did say ought to Auxiliary verbs with negative contractions Code such things as: won’t shouldn’t can’t wouldn’t couldn’t don’t aren’t isn’t haven’t hasn’t hadn’t weren’t wasn’t doesn’t didn’t. Affirsations and negations Code the following: yes no yah yeah not nah Pronouns with contractions Code the following: I’m They’ll He’ll I’ve You’ve That’s He’s It’s There’s I’ll They’ll She’s We’ve etc. Special words Code the following special words which describe the quality of something or someone: all anything nothing too also whatever nobody everybody somebody now right anybody right now Appendix D QUESTIONNAIRE FOR PERCEPTUAL MEASURES s 'r as 'bi 't (1) Negotiator’s (teacher) own disension a. The teachers’ negotiators try to show an ispression that they are flexible in dealing with disagreement. b. The teachers’ negotiators try to show an impression that they are willing to make significant concessions in order to reach an agreesent. c. The teachers’ negotiators try to show an ispression that they are willing to sacrifice sose of their desires in order to reach an agreesent. (2) Opponent’s (school board) disension a. The teachers’ negotiators attespt to sake their opponents look like weak representatives. b. The teachers’ negotiators attempt to pressure their opponents to sake the largest concessions. c. The teachers’ negotiators try to reach agreement without forcing their opponents to sacrifice their desires. 128 129 Ereitt’e Ieek Steetegy (2) a. Negotiator’s (teacher) own dimension The teachers’ negotiators are eager to obtain what they want without concession. The teachers’ negotiators highly concerned about their own outcoses. The teachers’ negotiators try to maximize their own outcoses. Opponent’s dimension The teachers’ negotiators sees to be concerned about their opponent’s as well as their own outcoses. The teachers’ negotiators sees to have little concern about their opponent’s outcomes. The teachers’ negotiators try to sinisize their opponent’s outcomes. 130 Leegeege Intensity a. The teachers’ negotiators tend to select unbiased words. b. The teachers’ negotiators tend to choose intense words which show biased or aggressive attitude. 0. The teachers’ negotiators try to use neutral expressions when they talk to their opponents. Lexi Dive s't a. The teachers’ negotiators use the same words and phrases repeatedly. b. The teachers’ negotiators seem to have limited vocabularies. c. The teachers’ negotiators seem to command their speech with sophisticated vocabularies. BIBLIOGRAPHY BIBLIOGRAPHY Bavelas, Rogers, & Millar. (1985). Interpersonal conflict. In T. A. Van Duk (Ed.) Heedheoh et diseeerse eeelyele, A, (pp. 9-26). London: Academic Press Benton, A. A., 1 Druckman, D. (1974). Constituent’ s bargaining orientation and intergroup negotiations. leeteel_e1_eeellee 629191_2_12h91_sx A. 141- 50. Bender, D., & Curington, W. (1983). Interaction analysis: a tool for understanding negotiations. t ' bo e i £21121. §§. 389-401. Berger, C. (1985). Social power and interpersonal cossunication. In M- Knapp & G- Miller (Edso). Handb99k_ef_in1ernersensl eeeeeeleetleh. (pp. 439-499). Beverly Hills, CA: Sage. Bowers, J. (1963). Language intensity, social introversion, and attitude change. §eeeeh_hehegheehe, 99, 345-352. Bowers, J. (1964). Some correlates of language intensity. Qu1112111.1922nsl_21_§nss9h. 59. 415-420. Bradac, J., Bowers, J., & Courtwright, J. (1979). Three language variables in communication research: Intensity, immediacy, and diversity. flu_1n_§2ssun191tign_flsseszsh i. 259-69- Breaugh, J., & Klisoski, R. (1981). Social forces in negotiation sinulntions- .2_rs2nali1x_1nd_a99111.2szghglgsx_8ulle_in 1 290-95. Brown, 8. (1968). The effects of need to saintain face on interpersonal bargainins- inurnal_gf_Exnszissanl_§gsial Earshglgsx. i. 107-122- Brown, B. (1977). Face-saving and face-restoration in negotiation. In D. Druckman (Ed.), Negetletlenee_eeelel nsxshglgsissl_nez§nestixes- (pp- 275-299). Beverly Hills: Sage. Brown, R., & Gilman, A. (1960). The pronouns of power and solidarity. In T. Sebeok (Ed.), 9tyle_le_leeghege. (pp. 253- 276). Casbridge, Mass: MIT Press. Brown, P., & Levinson, S. (1978). Universals in language usage: politeness phenomena. In E. Goody (Ed.), Qeeetlehe_ene eelltehese; stteteglee in soclel lhteeeetlen. (pp. 56-289). Cambridge: Casbridge University Press. Brown, P., & Levinson, S. (1987). ' : iv 'n Lengeege_9eege. Cambridge: Cambridge University Press. 131 132 Carnevale, P., Pruitt, D., & Britton, S. (1979). Looking tough: the negotiator under constituent surveillance. Pehsoeellty 1nd.§2sisl.£sxsh21231.flulletin. i. 118-21- Caspbell, D., & Fiske, D. (1959). Convergent and discrisinant validation by the sultitrait-multimethod satrix. Psychological Bulletin, 99, 81-105. Cook, T., & Caspbell, D. (1979). '- rim ntat o : ' & eealysis issues for flele settieg . Boston: Houghton Mifflin Cospany. Cronbach, L., & Meehl, P. (1955). Construct validity in psychological tests. £eyehelegleel_9elletlh, §_, 281- 301. Deutsch, M. (1969). Conflicts: productive and destructive. goereal of aeclel leeeee, 99, 7-41. Deutsch, M., & Krauss, R. (1962). Studies of interpersonal bargainins- louznsl_of_Qonfli£1_Eesolntion. 6. 52-76. Diez, M. (1983). Ne ti i n s ete e: t r eti it t o ifferen e 'n e h ° e. Unpublished Dissertation. Michigan State University. Donohue, W. (in press). 1 ' m r' i 'v seeiatien. New Brunswick, NJ: Laurence Earlbaum. Donohue, W., Allen, M., 1 Burrell, N. (1988L Mediator comsunication cospetence. geehehieetleh_heeegheehe, 59, 104 -119. Donohue, W., & Diez, M. (1985). Directive use in negotiation interaction- gossun1£1112n_!bnosrsnhs. 52. 305-318. Donohue, W., Diez, M., & Hasilton, M. (1984). Coding naturalistic negotiation interaction. Hus1n_Qo__un121_ion_Resesrgh. 19. 403- 425. Donohue, W., Diez, M., & Stahle, R. (1983). New directions in negotiation research. In R. N. Bostrom, & B. H. Westley (Eds.), Coeeueleetien yeeehee , vel , (pp. 249-279). Beverly Hills: Sage. Druckman, D., & Bonoma, T. (1976). Determinants of bargaining behavior in a bilateral monopoly situation 11: Opponent’s concession rate and similarity. Beheyleyel_§eleet1§t. 21. Folger, J., Hewes, D., 1 Poole, M. (1984). Coding social interaction. In B. Dervin 1 M. Voigt (Eds.), Eregress ie ti cien s, yel 4, (pp. 115-161). Norwood, NJ: 133 Ablex Publishing Co. Folger, J., & Poole, M. (1984). Working through conflict; A eomsunieetive nersnective. Glenview, IL: Scott Foressan. Goffsan, E. (1955). On face-work. Psychiatry, l9, 213-231. Goffman, E. (1967). Interection rituel: essays on fece-to-feee behavlo . NY: Doubleday Anchor Books. Greenberg, B. (1976). The effect of language intensity sodification on perceived verbal aggressiveness. Consunicetien Menegrenhs, 19, 130-39. Guetzkow, H. (1950). Unitizing and categorizing probless in coding qualitative data. Journel o: Clinicel Bsycholegy, 9, 47-58. Hiltrop, J., & Rubin, J. (1981). Position loss and isage loss in bargaining. Journal 0! Confliet Eesolntien, 99, 521-34. Hornstein, H. (1965). The effects of different sagnitudes of threat upon interpersonal bargaining. J a E erime t 1 Social Esyeholegy, l, 282-293. Kelley, H. (1966). A classroom study of the dilessas in interpersonal negotiations. In K. Archibald (Ed.), fitretegie inter tio an co ' t: r ers 'sc ss ons. (pp. 49-73). Berkeley, CA: Institute of International Studies. Jackson, S., & Jacobs, 8. (1983). Generalizing about message: suggestions for design and analysis of experiments. finnen Cosnnnieetlen Eeseene , 9, 169-191. Jesperson, O. (1964). Eeeentiele el English grensa . Alabama: Univ. of Alabasa Press. Lakoff, R. (1973). Logic of politeness: or minding your p’ and q’s. In C. Colus et al. (Eds.), ers fros e Ni h e iona Meeti Chi o i ' ti iet . (pp. 292-305). Chicago: Chicago Linguistic Society. Lakoff, R. (1975). Lengnege and wosen’s nlae . New York: Harper 1 Row, Publishers. Lakoff, R. (1977). What you can do with words: politeness, pragmatics and performatives. In A. Rogers, B. Wall, 1 J. Murphy (Eds.), Eneceeelngs ef Ieeee Conference en er rsat’ves Pres s’ti s m ' t . (pp. 79- 106). Arlington, VA: Center for Applied Linguistics. 134 Lewicki, R., & Litterer, J. (1985). Negotietion. Hosewood, IL: Richard D. Irwin, Inc. Longabaugh, S. (1963). A category systes for coding interpersonal behavior as social exchange. fieelenetny, 99, 319-344 0 Lim, T. (1990). Politeness behavior in social influence situations. In J. Dillard (Ed.), 9eehlng_eennlieneee_the t’ n 0 er e so inf e ess e. Gorsuch-Scarisbrick Publishers. McEwen, W., & Greenberg, B. (1970). The effects of sessage intensity on receiver evaluations of source, sessage and topic. Journel of Cossnnieetien, 99, 340-50. Morley. 1., & Stephenson. G. (1977). 122.222121_2sx22212sx_21 bargeining. London: George Allen & Unwin Ltd. Nunally, J. (1967). i t r . New York: McGraw-Hill. Owen, W. (1987). Marital interaction of discourse structures and relational pragmatics in conversation influence attespts. The Sonthern Seeeeh Qennnnleatlen Jounnel, 99, 103-127. Podell, J., & Knapp, W. (1969). The effect of mediation on the perceived firsness of the opponent. gennnel_et_9entllet Resolution. 2. 511-20- Pruitt, D. G. (1981). N i t' v' . New York: Academic Press. Pruitt, D. G. (1983). Strategic choice in negotiation. Amerieen 9222112221_§21221121..21. 167-94- Pruitt, D. G., Carnevale, P. J. D., Ben-Yoav, 0., Nochajski, T. H., 1 Van Slyck, M. R. (1983). Incentives for cooperation in integrative bargaining. In R. Tietz (Ed.), Aenlretien v ' b r 'ni e i ecisi . (pp. 22-34). Berlin: Springer. Pruitt, D., & Johnson, D. (1970). Mediation as an aid to face- aavins in negotiation. 1222n21_21.22222221121_2nd_§22121 2212221231. 11. 239-246- Pruitt, D., & Ssith, D. (1981). Impression managesent in bargaining: isages of firsness and trustworthiness. In J. Tedeschi (Ed.), e ' n nt t eo 1sn2_2212__ssn2ses2____h__rx_222_s22121 nsyehelegicel reseereh. (pp. 247-267). NY: Acadesic Press. Putnam, L. (1985). Bargaining as task and process: multiple functions of interaction sequences. In R. L. Street, Jr., 1 135 J- N. Cnppella (Ede-) MW heheyley. (pp. 225-242). MA: Edward Arnold. Putnas, L. (1989). Bargaining. er s n e i sm 1 ti , vel. 1, (pp. 176-178). Amnerst, PA: Oxford University Press. Putnam, L., 1 Geist, P. (1985L Argument in bargaining: An analysis of the reasoning process. Ihe_§entheyn_§neeeh Wm. 29. 225- 45. Putnas, L., 1 Jones, T. (1982). The role of communication in bargaining. WW. 2. 262- 280- Putnam, L., 1 Wilson, 8. (1988). Argusentation and bargaining strategies as discriminators of integrative outcomes. In M. A. Rabin (Ed ). MW enn_eeeh. (pp. 121-141). New York: Praeger Publishers. Putnas, L., Wilson, 8., Waltsan, & Turner, D. (1986). The evolution of case argusents in teachers’ bargaining. genteel efi the Anerieen Eerenele esseeletlen, 99, 63-81. Quirk. Ro. 8: Greenbnul. 8- (1973). MW eenteneeneny_§nglleh. NY: Harcourt, Brace, & Jovanovich. Roloff, M., 1 Campion, D. (1987). On alleviating the debilitating effects of accountability on bargaining: authority and self Ionitoring- 92W. 5.1. 145-64- Rosan. R. (1990) WWW tensit messa e be v ’ with' t - ’7 ' t" -°H- scr tive i v ti at t e e ' ' . Unpublished Dissertation. Michigan State University. Rubin, J. (1983L Negotiation: an introduction to some issues and the-em MW. 21. 135-147 Rubin, J., 1 Brown, B. (1975). s b i end_negetletlen. New York: Academic Press. Shisanoff, S. (1987L Types of emotional disclosures and request conpliance between spouse. 92222212121211.11222312222.21. 85- 100. Sillars, A. L. (1980). The sequential and distributional structure of conflict interactions as a function of attributions concerning the locus of responsibility and stability of conflicts. In 0. Nimso (Ed. ), Qennnnleetlen yeenheeh, vel, 5, (pp. 217- -235L New Brunswick: Transaction Books. 136 Sillars, A., Coletti, S., Parry, D., 1 Rogers, M. (1982). Coding verbal conflict tactics: nonverbal and perceptual correlates of the avoidance-distributive-integrative distinction. finnen Cossnnieetion Eeseere , 9, 83-95. Spradlin, A., & Bhargava, C. (1989). e-w k i ' t: eview e9 the literetnre. Paper presented to International Association for Conflict Management second bi-annual convention. Stevens, C. (1963). t te c e tive bar i i e o ’ t . New York: McGraw-Hill. Tjosvold, D. (1977). Low power person’s strategies in bargaining: negotiability of demand maintaining face, and race. lntennetienel Jennnel ef green Tensiens, 9, 494-510. Tjosvold, D. (1983). Social face in conflict: a critique. lnternetienel genrnel e9 Grenn Tensiens, 19, 49-64. Tjosvold, D. & Huston, T. (1978). Social face and resistance to cosprOIise in bargaining. W. 19.4. 57-680 Tracy, K. (1990). The sany faces of facework. In H. Giles & P- Robinson (Eds-L WW neyehelegy. (pp. 209-226). Chichester, England: John Wiley and Sons. Walton, R., & McKersie, R. (1965). e v eor r e;0ti:t'09:° ' ' ', ° : :'= Academic Press. Wheeless, L., Barraclough, R., & Stewart, R. (1983). Compliance- gaining and power in persuasion. In R. Bostrom 1 B. Westley (Eds.), gemsnnicatlen yeenheeh, vel, 7, (pp. 105-145). Beverly Hills, CA: Sage. Wilson, 8. R., & Putnas, L. L. (1990). Interactional goals in negotiation. In J. Anderson (Ed.), Qennnnleetlen yeenboeh, vel, 19, (pp. 374-406). Beverly Hills, CA: Sage. Wilson, 3., Meischke, H., 1 Kim, M. (1990, June). A neylsee englnegetletlen. Paper presented at the 2nd ISSA conference on argumentation, Amsterdas.