Development of a Validated Scoring Instrument for Businesses to Help Measure Negotiation Styles in Business Partnerships

Table of Contents

List of Figures

List of Tables

Abbreviations

1. Introduction
1.1. Relevance to Business Practice
1.2. Academic Relevance
1.3. Theoretical Contribution of this Thesis
1.4. Thesis Structure

2. Literature Review - Negotiation Style Scoring Instruments
2.1. Psychometric Properties - An Overview
2.1.1. Reliability Indexes
2.1.2. Validity Indexes
2.2. Historical Development of Approaches to Measure Negotiation Behaviour
2.3. The Managerial Grid
2.4. The Thomas-Kilmann Conflict MODE Instrument
2.4.1. Critical Assessment of the Theoretical Foundation
2.4.2. Critical Analysis of the Psychometric Properties
2.5. The Conflict Management Message Style Instrument
2.5.1. Critical Assessment of Theoretical Foundation
2.5.2. Critical Analysis of the Psychometric Properties
2.6. The Hall Conflict Management Survey
2.6.1. Critical Assessment of the Theoretical Foundation
2.6.2. Critical Analysis of the Psychometric Properties
2.7. The Rahim Organizational Conflict Inventory-II
2.7.1. Critical Assessment of the Theoretical Foundation
2.7.2. Critical Analysis of the Psychometric Properties
2.8. The Organizational Communication Conflict Instrument
2.8.1. Critical Assessment of the Theoretical Foundation
2.8.2. Critical Analysis of the Psychometric Properties
2.9. Conclusion

3. Discussion of an Integrated Theoretical Foundation
3.1. Attitude-Behaviour Theorising
3.1.1. The Theory of Reasoned Action
3.1.2. Critical Assessment of the TRA
3.2. Developing a New Model for the Explanation of an Individual’s Behaviour
3.2.1. Predispositions
3.2.2. Strategy
3.2.3. Tactics
3.2.4. Relational Influencing Factors
3.2.5. Situational Influencing Factors
3.3. A Model for the Classification of Negotiation Styles

4. Development of the Negotiation Styles Scoring Instrument
4.1. Construct Definition
4.2. Object Classification
4.3. Attribute Classification
4.4. Rater Identification
4.5. Scale formation
4.6. Enumeration
4.7. Designing Scenarios to Define Different Negotiation Situations

5. Empirical Validation of the Proposed Negotiation Style Scoring Instrument
5.1. Methodology
5.1.1. Participants
5.1.2. Procedure and Results
5.2. Discussion of the Results

6. Suggestion of a New Negotiation Style Scoring Instrument

7. Summary and Conclusions

8. References
Appendix A: Definitions of Negotiation Styles VIII
Appendix B: Building Blocks to Create One’s Own Scenario X
Appendix C: Questionnaire Items XIII
Appendix D: Questionnaires Used in Our Empirical Study XVI

List of Figures

Figure 1: The Theory of Reasoned Action

Figure 2: A New Theoretical Model for the Explanation of Negotiation Behaviour

Figure 3: A Classification of Four Negotiation Styles

Figure 4: Overview of the Structure Underlying our Questionnaire

Figure 5: Determination of Different Scenarios

List of Tables

Table 1: Results of the Reliability Analysis

Table 2: Results of the Factor Analysis

Abbreviations

Abbildung in dieser Leseprobe nicht enthalten

1. Introduction

1.1. Relevance to Business Practice

Over the last years the number of business partnerships has increased significantly (e.g., Eden and Huxham, 2001; Mohr and Spekman, 1994; Teegen and Doh, 2002). There are various drivers for this kind of strategic re-orientation among companies. Whereas companies have previously relied on profitability of protected home markets, the changed economic climate¹ has fostered collaboration between them (Bleeke and Ernst, 1993). Firms collaborate to mas- ter the challenges caused by these changes. At the same time they benefit from better access to new markets, pooling or swapping technology, sharing of risks, larger economies of scale in joint research and/or production, and economies of scope (e.g., Contractor and Lorange, 1988; Lorange and Roos, 1991).

Business partnerships can be defined as purposive strategic relationships between independent firms who share compatible goals, strive for mutual benefit, and acknowledge a high level of mutual interdependency (Mohr and Spekman, 1994, p. 135). Examples of organizational forms which support this kind of cooperation are strategic alliances, joint ventures, licence agreements, research and development (R&D) partnerships, and franchising (e.g., Borys and Jemison, 1989; Ring and Van de Ven, 1992).

Ring and Van de Ven (1994) explain that business partnerships emerge, evolve and dissolve over time. In spite of the benefits which business partnerships can contribute to a company's success, reported failure rates range between 30 and 70 percent (e.g., Bleeke and Ernst, 1993; Das and Teng, 2000; Visioni, 2002).

Differences between collaborating companies in terms of aims, cultures, structures, proce- dures, languages, power, and accountabilities are said to make the effective management of business partnerships not easy and are admittedly unerringly negative influencing factors on their success (e.g., Eden and Huxham 2001). Lorange and Roos (1991) suggest that matters such as lack of trust and incompatible personal chemistry between partners are responsible for business partnerships failing outright or stopping short of reaching their objectives.

As partners can have inconsistent goals, a certain amount of tension can be expected between them (e.g., Olson and Singsuwan, 1997). Hence, the resolution of conflicting interests and the maintenance of harmony becomes a central role for managers of business partnerships (Borys and Jemison, 1989). Therefore, Mohr and Spekman (1994), and Olson and Singsuwan (1997) emphasise besides important effects of partnership attributes² and communication behaviour³, also the impact of conflict resolution techniques⁴ on the success of partnerships.

The formation of business partnerships creates drawbacks such as increasing complexity, loss of autonomy for the partners, and information asymmetry which result in a multitude of ex- tensive negotiations (e.g., Provan, 1984; Williamson, 1985). Ring and Van de Ven (1994) suggest that the process⁵ of partnerships is influenced by the sequence of events and inter- actions among partners that unfold to shape and modify a business partnership over time (Ring and Van de Ven, 1994). These interactions consist in detail of iterative sequences of negotiation, commitment, and execution stages, which characterise a cyclical process (Ring and Van de Ven, 1994). Given that general conditions and business environment change over time, misunderstandings, conflicts, and changing expectations of the partners may occur (Olson and Singsuwan, 1997). Hence, these changes require rethinking and renegotiations in terms of the relationship between business partners (e.g., Culpan 1993; Reuer and Arino, 2000). As a result managers in commercial relationships find themselves negotiating not only in up-front negotiations over formal contracts and the initial structure of the relationship but also in ongoing (re)negotiations to maintain and manage the partnership that evolves over time (Adler, Brahm, and Graham, 1992; Teegen and Doh, 2002).

Dabholkar, Johnston, and Cathey (1994) highlight explicitly the role of negotiation behaviour in the process of business-to-business partnership formation, the evolution of such relation- ships over time and ultimately their success. They propose specifically that a partner's par- ticular negotiation behaviour influences immediate negotiation outcomes, which have in turn a critical impact on the success of a business partnership⁶. To conclude, negotiations are crucial to every business partnership, and there is a strong and significant link between a manager's specific negotiation behaviour, used in negotiating business partnerships, and the success of these business partnerships (Dabholkar, Johnston, and Cathey, 1994; Mohr and Spekman, 1994; Olson and Singsuwan, 1997; Ring and Van de Ven, 1994). Fisher and Ury (1992) see negotiation in general as a basic means of getting what someone wants from others. It is a back-and-forth communication designed to reach an agreement when you and the other side have some interest that are shared and others that are opposed. People negotiate differently and with varying degrees of success (Scott, 1982; Warschaw, 1980; Williams, 1993). There are several conceptual frameworks to explain an individual's negotiation behaviour. For example Thomas and Kilmann (1974), and Ross and DeWine (1988) presume that the choice of someone's negotiation style is influenced by a general pre- disposition and therefore independent of a particular context, whereas Hall (1969a, b), Putnam and Wilson (1982), and Rahim (1983a) argue that negotiation styles are highly influenced by a particular situation. However, research suggests that understanding of behaviour requires at least as much attention to the situation as it does attention to general predispositions (e.g., Mischel, 1968). In a similar vein, Adler, Schmitz, and Graham (1987) conclude that negotia- tion outcomes are determined by a negotiator's predisposition⁷, strategy⁸ and tactics⁹, and also by situational factors¹⁰.

The way individuals act in a negotiation, is described as their negotiation style (Warschaw, 1980). Negotiation style refers to a characteristic mode or habitual way that a person handles a negotiation (Blake and Mouton, 1964). Negotiation literature describes two main types of negotiation behaviour - integrative and distributive¹¹ (e.g., Isenhart and Spangle, 2000; Neale and Bazerman, 1991; Psenicka and Rahim, 1989; Putnam and Poole, 1987; Walton and McKersie, 1993). Whereas integrative bargaining behaviour pursues a problem-solving ap- proach and usually leads to mutual benefits, the distributive negotiation behaviour emphasises win-lose encounters and is often characterised through an aggressive and selfish behaviour (e.g., Carnevale and Pruitt, 1992; Isenhart and Spangle, 2000; Lewicki et al., 1994).

Different styles produce different sorts of benefits, according to the used negotiation behaviour. Consequently, negotiation outcomes can also be classified as distributive and integrative. Settlements characterised by compromises, trade-offs, and win-lose results fall into the distributive realm, whereas creative proposals are treated as integrative settlements (Putnam and Poole, 1987; Scott, 1982).

In order to reach better negotiation outcomes and arrange the management of business part- nerships more successfully, it is therefore critical to assess a manager's negotiation style used in business partnerships accurately. This is essential, because, as suggested by Robert (1982, p. 14) the more a person learns and knows about how s/he might behave in a negotiation, the greater the chances for selecting an appropriate negotiation style with regard to the intended outcome. Hence, only if one is aware of one's used negotiation style one can make efforts to change it, if necessary, to a more appropriate one, which will eventually result in better individual negotiation outcomes, which will in turn lead to a more successful management of business partnerships (Dabholkar, Johnston, and Cathey, 1994). To our knowledge, there are currently no validated scoring instruments¹² to measure managers’ negotiations styles specifically in business partnerships.

Our aim within this study is to develop and test empirically a scoring instrument which looks separately at a manager’s predisposition as well as strategy, and tactics and measures explicitly a manager's negotiation style used in business partnerships in consideration of influencing variables of this specific situation. Such an instrument should assist managers determine their current negotiation style more accurately.

1.2. Academic Relevance

There is a large range and number of scholarly as well as practitioner publications on negotia- tion behaviour¹³. These publications either describe the characteristics of conflict resolution and negotiation behaviour (e.g., Pruitt and Carnevale, 1993) and/or give specific advice (e.g., Fisher and Ury, 1992) to management practice how to improve negotiation skills and reach better negotiations outcomes. In a similar vein, a multitude of authors deal with the meas- urement of an individual's negotiation styles (e.g., Hall, 1969a, b; Janssen and Van de Vliert, 1996; Putnam and Wilson, 1982; Rahim, 1983b; Ross and DeWine, 1988; Thomas and Kilmann, 1974). The various measurement instruments differ in some attributes. A first attempt to classify strategies for managing conflict is made by Deutsch (1949), who classifies styles along one dimension, defined by competitive and cooperative behaviour as bipolar opposites. Blake and Mouton (1964) enhance this keynote and develop a two-dimensional managerial grid - the dual concern model. They measure an individual’s negotiation style¹⁴ on two independent dimensions - concern for production and concern for people. The authors of the various in- struments¹⁵, which will be discussed and critically analysed within this thesis, have basically adopted the two-dimensional approach by Blake and Mouton (1964) and classify different negotiation styles along two independent dimensions. Whereas the scoring instruments by Hall (1969a), Rahim (1983b), and Thomas and Kilmann (1974) measure five different nego- tiation styles¹⁶, Putnam and Wilson (1982), and Ross and DeWine (1988) find evidence to distinguish only three different negotiation styles¹⁷. Other authors describe four different negotiation styles (e.g., Janssen and Van de Vliert, 1996¹⁸ ; Pruitt, 1983¹⁹ ).

Instruments, which are based on the assumption that an individual’s negotiation style is a re- sult of a general trait and therefore relatively stable throughout different situations, measure negotiation styles independently of a particular context (e.g., Ross and DeWine, 1988; Thomas and Kilmann, 1974). Whereas instruments, which rely on the presumption that an individual’s negotiation behaviour is highly affected by the circumstances of a particular situation, describe specific scenarios or offer various forms of the questionnaire for different situations when measuring negotiation styles (e.g., Hall, 1969a; Putnam and Wilson, 1982; Rahim, 1983b).

There are also a great number of other instruments to measure negotiation styles, primarily used in management practice such as the ‘ Conflict Style Instrument ’ (Warner, 1999a), the ‘ Influence and Negotiation Style Inventory ’ (Gordian Publishing, 2002), the ‘ Negotiating Styles ’ instrument (Warner, 1999b), and the ‘ Negotiating Style Profile ’ (Glaser and Glaser, 1996). As these instruments do not necessarily have a theoretical background and empirical studies on the validity and reliability of these instruments that are, to our knowledge, not available in the academic literature, they are relatively hard to evaluate and are therefore not contemplated in depth within this thesis.

A number of reasons explain why the instruments mentioned above are not entirely suitable when measuring negotiation behaviour within business partnerships. Knapp, Putnam, and Davis (1988) criticise that the existing instruments treat the individual as the sole benchmark to assess negotiation behaviour and they strongly suggest that negotiation behaviour should not be treated as a stable disposition²⁰. As there is high interdependency between the partners in business partnerships (Mohr and Spekman, 1994), Knapp, Putnam, and Davis (1988) stress the importance of incorporating relational²¹ and situational²² factors into the measurement of an individual’s preferences when aiming to predict an individual's negotiation behaviour. Goering, Rudick, and Faulkner (1986) find that the behaviour of one party involved in the negotiation has an important impact on the used negotiation style of the other party. Hence, there are too many factors associated with the other party or the situation than to rely solely on an individual's preference for negotiation strategies (Knapp, Putnam, and Davis, 1988). The only instruments which, at least partly, consider relational variables are Hall's CMS (1969a) and Rahim's ROCI II (1983b, c). The instrument OCCI by Putnam and Wilson (1982) is the only one which considers both relational and situational influencing factors (Wilson and Waltman, 1988). The focus of this work is to measure negotiation styles in busi- ness partnerships and not just an individual's attitudes towards specific negotiation styles. It is therefore critical that a measurement instrument considers both influencing variables - re- lational and situational factors (Knapp, Putnam, and Davis, 1988).

Moreover, Knapp, Putnam, and Davis (1988) criticise that the underlying conceptual frame- work - the dual concern model - is oversimplified to measure negotiation behaviour in or- ganizational settings appropriately. Specifically they find fault that the static, two-dimen- sional model fails to cope with the fact that individuals alter their strategy and tactics used in negotiations. As some of the existing instruments only focus on pre-planned strategies, they overlook the fact, that negotiation behaviour is sometimes automatic, mindless, impulsive and hence can change during a particular negotiation situation (Knapp, Putnam, and Davis, 1988). Therefore, Womack (1988a) concludes that it is critical that measurement instruments explic- itly differentiate between negotiation strategies and tactics, and combine these two approaches to measure negotiation behaviour accurately. We will take up Womack’s (1988a) suggestion and discuss a theoretical foundation for the measurement of negotiation behaviour which con- siders a manager's negotiation predisposition, strategy, and tactics separately.

Despite the fact that many of the instruments, which are mentioned in the previous para- graphs, are widely used, Womack (1988a) criticises that they mostly exhibit only weak to moderate reliabilities. Likewise the instruments show partially dissatisfying results in terms of content, construct, and predictive validity (Womack, 1988a)²³. In order to gain useful re- sults when measuring negotiation behaviour, such instruments should be validated. To summarise, the specified instruments measuring negotiation behaviour focus too narrowly on the individual, disregard organizational settings and other important influencing factors, and do not differentiate between an individual's negotiation predisposition, strategy, and tac- tics. However, when measuring negotiation behaviour in business partnerships, an appropri- ate instrument requires a separate measurement of someone's predisposition, negotiation strat- egy and tactics, in consideration of relational and situational factors as influencing variables (Adler, Brahm, and Graham, 1992; Adler, Graham, and Schwarz Gehrke 1987; Womack, 1988a). Despite the fact, that the mentioned instruments have some unique strengths, none of them fulfil all of the requirements and therefore do not measure negotiation behaviour in business partnerships accurately.

Hence, our aim within this thesis is to develop and empirically test a new scoring instrument to measure negotiation styles that focuses explicitly on negotiations in business partnerships and takes into account all of the requirements. By doing this we contribute to the existing academic literature and fill an existing gap.

1.3. Theoretical Contribution of this Thesis

The adequate measurement of negotiation styles requires an understanding of the causes for human behaviour. There exist various influential approaches for the explanation of an indi- vidual's behaviour, like for example the ‘Attitude Behaviour Theorising’ (e.g., Ajzen, 1988), the ‘Situational Approach’ (e.g., Davis-Blake and Pfeffer, 1989), or the ‘Interactionism Ap- proach’ (e.g., Endler and Magnusson, 1976). To integrate these approaches for the explana- tion and measurement of negotiation behaviour, it is essential to measure an individual’s pre- disposition, strategy and tactics separately in consideration of important influencing variables. We discuss an integrated conceptual framework that satisfies these requirements and therefore can be used as an underlying theoretical model for instruments that measure negotiation styles. The suggested model consists of five linked elements: on the one hand predisposition, strategy, and tactic are used to measure negotiation behaviour and on the other hand relational and situational influencing factors reflect the impact a certain situation might have on an indi- vidual’s negotiation behaviour.

Many theorists agree that predispositions have an important impact on an individual’s behav- iour in negotiations (e.g., Knapp, Putnam, and Davis, 1988; Ross and DeWine, 1988; Thomas and Kilmann, 1974). However, empirical research fails to provide evidence that predisposi- tions alone cause behaviour (e.g., Mischel, 1968; Thompson, 1990; Wicker, 1969). Follow- ing the prominent ‘Theory of Reasoned Action’ (Fishbein and Ajzen, 1975; Ajzen, 1988) and heeding the assumptions of the ‘Interactionism Approach’ it is concluded that predispositions influence an individual’s negotiation strategy, which in turn, has an impact on specific tactics, which serve the implementation of a chosen strategy. In contrast to predispositions, which are stable enduring characteristics, an individual’s strategy and tactics are susceptible to situ- ational and relational factors.

Our suggested instrument measuring negotiation styles in business partnerships considers the underlying model inasmuch that it consists of three main parts, each measuring either a manager’s negotiation predisposition, strategy, or tactics. In addition, different scenarios are described in the last two parts of the questionnaire to take into account that situational and relational factors may have an impact on both negotiation strategy and tactics.

1.4. Thesis Structure

The remaining thesis begins with an extensive literature review. Various well-known scoring instruments that measure negotiation styles will be described and their strengths and weak- nesses with regard to the discussed requirements will be elaborated on in Chapter 2. In Chapter 3, we discuss a theoretical framework, which will serve as the foundation for the de- velopment of a new scoring instrument that measures negotiation styles in business partner- ships. In Chapter 4, we will outline the proposed procedure for scale development and apply each of the recommended steps to develop a new negotiation style scoring instrument. In Chapter 5, we will describe the empirical survey which was conducted to test and refine our proposed scale and represent the results of this survey. Building on these findings, in Chapter 6 a new validated scoring instrument to measure negotiations styles in business partnerships will be suggested. Finally we will present a summary and conclusion. Herein, we emphasise the contribution of our instrument to management practice and theory. We also call attention to apparent limitations and offer possible avenues for future research.

2. Literature Review - Negotiation Style Scoring Instruments

Initially this chapter presents a brief description of the basics of psychometric properties, needed within this work. Following the introduction an overview of the historical development of approaches to determine negotiation behaviour is presented. The majority of the presented instruments does not explicitly talk of negotiation behaviour and most of them are designed to assess conflict behaviour. However, since both, a conflict as well as a negotiation situation, can be generally described as situations in which the wishes of at least two parties differ from each other (Pruitt, 1981; Pruitt and Carnevale, 1993; Thomas and Kilmann, 1974), we use the terms negotiation behaviour and conflict behaviour synonymously. Additionally, other researchers have already used instruments originally developed to assess conflict behaviour to measure negotiation behaviour (e.g., Perdue, Day, and Michaels, 1986). Hence, the most important and widely used instruments to measure conflict behaviour as well as instruments to measure negotiation behaviour will be presented successively and described in greater detail in this chapter. Firstly, instruments which measure an individual’s negotiation behaviour independent of a particular context²⁴ are investigated, which is followed by the description of instruments which incorporate situational and/or relational factors²⁵.

Here, their theoretical foundation and the structure of each individual questionnaire will be described and an analysis will be carried out afterwards to assess whether the instruments meet the requirements that are elaborated in the introduction. However, the articles on which the scales are based vary greatly in depth, length, and detail. Consequently, the summaries themselves are dependent upon characteristics of the original sources. In order to find out if a manager’s negotiation behaviour in business partnerships is measured accurately by the analysed instruments, each will be critically examined according to:

- its underlying theoretical framework with specific attention to its ability to distin- guish between an individual’s negotiation predisposition, strategy, and tactics and to incorporate situational and relational factors as influencing variables;
- its psychometric properties, i.e., the fact whether the instrument meets the necessary standards for reliability and validity.

Concerning the latter point, the evaluation of instruments on the basis of their psychometric properties is complicated and ambiguous for the two following reasons. First, according to Churchill (1979) there are too many different reliability and validity indexes and often their terms are used interchangeably by researchers²⁶. Second, and more important, there exists no convergence regarding which of them are actually convincing indexes to prove and consequently ensure the quality of a measurement instrument²⁷. For this reason, in the next paragraph a short overview is given of the psychometric properties relevant within this work. This chapter will be concluded with a short summary of the advantages and drawbacks of the presented instruments with specific regard to the measurement of managers’ negotiation styles used in managing business partnerships. These findings will serve as a basis for the discus- sion of an integrated theoretical foundation of a new scoring instrument in the third chapter of this thesis.

2.1. Psychometric Properties - An Overview

The quality of an instrument can be assessed by its reliability and validity (Bearden, Netemeyer, and Mobley, 1993). The measurement instruments, discussed in the next paragraphs, are critically analysed with respect to the following reliability and validity indexes²⁸.

2.1.1. Reliability Indexes

In general, reliability denotes the precision of a measurement (Zimbardo and Gerrig, 1996). There are two broad types of reliability referred to in the psychometric literature: internal reli- ability (or internal consistency) and test-retest reliability. Internal reliability can be under- stood as the correlation among items or sets of items in the scale for all who answer the items (Bearden, Netemeyer, and Mobley, 1993, p. 4). For negotiation style scoring instruments, this indicates how consistently individuals prefer one of the styles measured by a given instrument (Womack, 1988a). Internal reliability is expressed by a so-called alpha coefficient represent- ing the correlation between different items. The correlations among items measuring the same construct should, in general, be much higher than those of items measuring different constructs of interest. The major theory of reliability is based on the domain sampling theory (Nunnally, 1967). Each construct that should be calculated is assumed to have a domain, which is specified by the definition of the construct. An exact evaluation of the construct would include all possible items contained in the domain. Because all possible items can never be thoroughly detailed, measures of a construct will necessarily be imperfect representations of the construct. Measures will be acceptable representations if an adequate sample of items is drawn from the construct domain. Coefficient alpha is an indicator of how well a sample of items depicts the construct domain. Coefficient alpha augments as the correlation between the items increases (John and Roedder, 1981).

Therefore, a typical test of reliability is the coefficient alpha computation (Rossiter, 2001). The most widely used coefficient for assessing the internal reliability is the Cronbach’s (1951) Alpha. Nunnally (1967, p.226) determines a value of the Cronbach’s Alpha, as 0.8 for adequate reliability for basic research and 0.9 for applied settings, which Konovsky, Jaster, and McDonald (1989, p. 271) call ‘generally accepted’.

Another test of the reliability of an instrument is the test-retest reliability. Test-retest reli- ability denotes the correlation between the results of two administrations of an instrument under similar circumstances at different times for the same population (Womack, 1988a).

2.1.2. Validity Indexes

Validity, on the surface, determines the overall usefulness of an instrument. If a scale achieves to measure what it is intended to measure, it is valid and thus useful (Nunnally, 1967). In detail, according to Churchill (1979, p. 65), a ‘measure is valid when the differences in observed scores reflect true differences on the characteristic one is attempting to measure and nothing else’. Accordingly, validity is the most important property to ensure an instrument’s quality (Cronbach, 1984). Content validity, construct validity, and predictive validity are three types of validity which are frequently considered.

Content validity denotes the adequacy with which a specified domain of content is sampled: the scale items should be consistent with the theoretical domain of the construct which is measured (Nunnally, 1967). Content validity can be used to test the theoretical framework of an instrument. With respect to most negotiation style scoring instruments, content validity refers to the ability of measuring the two independent underlying dimensions and the different styles classified along these two dimensions (Womack, 1988b). An approach often used to prove content validity is the factor analysis, because it confirms whether the number of the conceptualised dimensions can be verified empirically (Churchill, 1979).

Construct validity deals with the extent to which a specific instrument correlates with other measures that are theoretically alike (Carmines and Zeller, 1979). Hence, this validity index rests on the assumption that a particular construct can be measured by different methods (Churchill, 1979). Construct validity²⁹ is usually assessed through the Multitrait Multimethod Matrix³⁰ (MTMM).

Predictive validity ³¹, on the other hand, refers to the ability of an instrument to predict behav- iour (Womack, 1988b). Nunnally (1967) uses the term predictive, when describing functional relations between an instrument and events that occur before, during, and after the use of the instrument. Predictive validity is determined by the degree to which the instrument, that is to be tested, correlates with a criterion or an outcome measure (Nunnally, 1967). However, it has to be acknowledged that there are researchers who doubt the appropriateness of this methodology. Rossiter (2001, p. 27) claims that, even though widely-used, ‘construct validity and predictive validity are not conclusive’. Regarding construct and predictive valid- ity he criticises that a high correlation between two constructs does not show which one of the two is valid and therefore he concludes that validity should be established for a scale inde- pendently. Rossiter (2001, p. 4) emphasises that the only type of validity which is essential to verify a scale’s validity is ‘content validity’. But in his view, content validity cannot be shown by statistical testings after the scale is developed³². Content validity cannot be proved by a high alpha and factor analysis, it is rather ‘an appeal to reason, conducted before the scale is developed, that the items will properly represent the construct’ (Rossiter, 2001, p. 4). To ensure content validity, the developers should follow a specific procedure³³ when develop- ing a scale. With respect to reliability, Rossiter (2001, p. 17 ff.) claims that it should rather be seen as a ‘precision of score for a particular application’ and that a high coefficient alpha can- not be provided as a sole justification for employing a scale. Furthermore, in Rossiter’s (2001, p. 17) words ‘test-retest reliability can be dismissed immediately because it provides no information about the precision of scores obtained from the test’.

With this in mind, we will have to treat the results reported for the psychometric properties of the various instruments presented in this chapter with caution. Accordingly, the results should not be regarded as a definite evidence that the instruments are indeed valid even though some results may indicate so.

2.2. Historical Development of Approaches to Measure Negotiation Behaviour

One of the first approaches to explain negotiation behaviour is the so-called Game Theory (e.g., Nash, 1950; Rapoport 1973; Von Neumann and Morgenstern, 1947). The Game Theory explains behaviour with an underlying assumption that the interacting individuals always act rationally and try to maximise self-interest throughout.

From the Game Theory evolved a one-dimensional approach that describes two fundamental ways of managing conflicts. Deutsch (1949) classifies behaviour along this one dimension, defined by cooperation and competition as bipolar opposites. The main differences between these two distinct ways of conflict management lies in the nature of the relationship of the participants’ goals in a conflict situation: in a cooperative situation the goals are linked in a way where everybody “sinks or swims” together, while in a competitive situation where someone aims to increase his or her chance of goal attainment, the other party’s chance automatically decreases (Deutsch, 1973).

In 1964, Blake and Mouton propose for the first time a two-dimensional model to classify managerial behaviour, while calling it the ‘Managerial Grid’. Later research reveals that the cooperative-competitive dichotomy as introduced by Deutsch (1949) does not capture the wide range of alternative behaviours available to individuals in a conflict situation (e.g., Ruble and Thomas, 1976; Van de Vliert and Prein, 1989). According to Thomas (1976), the cooperative-uncooperative dichotomy appears to oversimplify the more complex range of options available to the conflict parties to a great extent. Hence, for instance, Cosier and Ruble (1981) suggest that the one-dimensional conceptualisation of styles of conflict man- agement must be rejected in favour of the two-dimensional grid. The resulting conflict styles, specified in the grid, can be regarded as a function of the two underlying independent dimen- sions (Pruitt and Carnevale, 1993). As Blake and Mouton (1964) label the two underlying dimensions ‘concern for people’ and ‘concern for production’, the two-dimensional model is also well-known as the ‘Dual Concern Model’. Since Blake and Mouton’s dual concern model (1964) inspires many authors (e.g., Hall, 1969a; Rahim, 1983a; Ross and DeWine, 1988; Thomas and Kilmann, 1974) to develop variations of it, and thus remaining the theo- retical basis for most instruments that measure negotiation behaviour, it is a very important contribution in this field and will be described in greater detail in the next paragraph. Ac- cording to Lewicki et al. (1994, p. 11) ‘the dual concern model and its variation have become some of the most ubiquitous and durable vehicles for approaching and understanding conflict management strategies’. Carnevale and Pruitt (1992, p. 539) refer to the dual concern model as ‘the chief theory of negotiator motivation’.

2.3. The Managerial Grid

The Managerial Grid by Blake and Mouton (1964) is designed to classify managerial behav- iour. It enables to measure more than just the two possible opposing management styles as suggested by Deutsch (1949). Their approach therefore turns away from binary oppositions and acknowledges the possibility of showing characteristics of a variety of leadership styles. According to Blake and Mouton (1964) three key variables are used to describe an organiza- tion. Firstly, the main purpose of an organization is to produce and make profits; secondly, an organization could not exist without people; finally, every organization consists of a hierarchy - the process of achieving organization purpose therefore results in some people are bosses while others are being bossed (Blake and Mouton, 1964).

The fundamental theoretical idea of the Managerial Grid derives from these characteristics. According to Blake and Mouton (1964) managerial behaviour can be classified according to the amount of emphasis which is placed on achieving production, in other words - the concern for production and equally the concern for people, i.e. the productive unit of every organization. The manner in which these two concerns are combined by a manager shows how s/he uses the third key variable - hierarchy.

Concern for people on the vertical axis and concern for production on the horizontal axis form a two-dimensional coordinate system, which is called the ‘Managerial Grid’. The axes range from 1, indicating a low concern, to 9, indicating a high concern for production or peo- ple respectively. The term ‘concern for’ does not indicate how much is actually produced, nor does it reflect the degree of whether the needs of people are actually met. It simply shows in how far a manager is concerned with production or people respectively and how these con- cerns intertwine and influence each other when following the aim of making profits for the organization.

Concern for production can be expressed in a variety of ways. It is not only limited to produced and countable goods, but production also covers whatever it is that organizations engage people to accomplish, for instance, the quality of policy decisions, creative ideas, procedures or processes (Blake and Mouton, 1964). In a similar vein, a concern for people covers a range of considerations. In the centre of attention are for example the aim to establish and maintain good working conditions and an equitable salary structure, the desire for security at work, and social relations between associates.

Blake and Mouton (1964) identify five different theories of managerial behaviour, evolving from high or low concern for people and production, which are located in the corners and in the middle of the Managerial Grid³⁴. Each of these five broad types of managerial style de- fines a definite set of assumptions regarding how managers use hierarchy in situations of pro- duction where people are involved. They describe different alternatives of handling these situations - all of them are realistic and can be found in management practice (Blake and Mouton, 1964).

Blake and Mouton (1964) indicate that these managerial styles are not fixed and might change over time or in different situations. They are influenced by a range of situational and person- ality factors. For example, in some organizations, the managerial style of an individual might, due to strict internal policies, only reflect little of a man’s personal thinking. The situation itself can also be an influencing factor; since in an exceptional situation people might behave differently from how they would under ordinary circumstances. Individuals hold values and beliefs regarding on how to behave in the “right” way, which have an important impact on their behaviour (Blake and Mouton, 1964). Deep-rooted character traits may predispose peo- ple to prefer one style over another. The following five managerial styles are distinguished by Blake and Mouton (1964):

9,1 Management (high concern for production / low concern for people):

Efficiency in operations results from arranging conditions of work in such a way that human elements interfere to a minimum degree.

9,9 Management (high concern for production / high concern for people):

Work accomplishment is from committed people; interdependency through a ‘common stake’ in organization purpose leads to relationships of trust and respect.

1,9 Management (low concern for production / high concern for people):

Thoughtful attention to needs of people for satisfying relationships leads to a comfortable friendly organization atmosphere and work tempo.

1,1 Management (low concern for production / low concern for people):

Exertion of minimum effort to get required work done is appropriate to sustain organization membership.

5,5 Management (intermediate in both concern for production and people):

Adequate organization performance is possible through balancing the necessity to get out work while maintaining the morale of people at a satisfactory level.

Blake and Mouton (1964) offer a questionnaire to help individuals to determine their man- agement style. In order to determine one’s managerial style, the respondent is asked to rank- order five paragraphs from most to least typical. Each of these paragraphs consists of six dif- ferent statements reflecting either decisions, convictions, conflict, emotions, humour, or ef- fort.

However, as the structure of the questionnaire does not allow a comprehensive statistical as- sessment of the validity and reliability of the Managerial Grid (Rahim and Magner, 1995; Thomas and Kilmann, 1978), within this work it will not be presented as an instrument to concentrating on the measurement of negotiation behaviour. At this point it is however men- tioned, since it serves as a role model for the other instruments, which will be described and analysed in the following paragraphs and which, to a great extent, evolve from the theoretical framework of this ‘first’ dual concern model.

2.4. The Thomas-Kilmann Conflict MODE Instrument

The Thomas-Kilmann Conflict MODE³⁵ Instrument (TKI) is one of the world’s most widely used questionnaires of its type in both training and research (Putnam, 1988; Womack, 1988b). The TKI is designed to measure an individual’s conflict-handling modes in conflict situations, i.e. situations in which the concerns of two individuals appear to be incompatible (Thomas, 1988). In other words, the wishes of one person differ from those of another person (Thomas and Kilmann, 1974).

In a two-dimensional model, this approach classifies five distinct conflict-handling modes along two more basic dimensions of strategic intent. Firstly, cooperativeness, i.e., the extent to which the individual attempts to satisfy the other’s party concerns and secondly, assertiveness, i.e., the extent to which the individual attempts to satisfy his/her own concerns (Thomas, 1988). The five conflict-handling modes, which are classified along these two dimensions are the following (Thomas and Kilmann, 1974 p.10):

Competing (high assertiveness / low cooperativeness):

An individual pursues his/her own concerns at the other person’s expense. Competing means using all kinds of power, standing up for one’s own rights, defending a position which one believes is correct, or simply trying to ‘win’.

Collaborating (high assertiveness / high cooperativeness):

Collaborating involves an attempt to work with the other person to find some solution which fully satisfies the concern of both parties. It means identifying the underlying concerns and finding an alternative which meets both sets of concerns.

Compromising (intermediate assertiveness / intermediate cooperativeness):

The objective is to find some expedient, mutually acceptable solution which partially satisfies both parties. Compromising means splitting the difference, exchanging concessions to reach a quick middle-ground position.

Accommodating (low assertiveness / high cooperativeness):

When accommodating, an individual neglects his/her own concerns to satisfy the concerns of the other person. It means obeying to another person’s order or yielding to another’s points of view.

Avoiding (low assertiveness / low cooperativeness):

An avoiding person does not address the conflict, s/he pursues neither his/her own concerns nor the other person’s concerns. Avoiding means either sidestepping or postponing an issue, or withdrawing from a threatening situation.

In order to determine an individual’s conflict-handling modes, the respondent is asked to choose statements which reflect his/her most likely behaviour when dealing with conflict situations. The respondent is advised that there are no right or wrong answers. The question- naire consists of 34 distinct statements describing the five different modes³⁶. Statements representing one mode are compared with the items of every other mode three times, making a total number of 30 different pairs of items (Kilmann and Thomas, 1977). Hence, some of the items are used more than once. The respondent is ‘forced’ to select the statement in each pair which would best describe their own response in a conflict situation. Even if neither the first nor the second statement seems typical for a respondent’s own behaviour, s/he should select the more likely one to use. The administration time, including self-scoring, takes ap- proximately 12 minutes (Kilmann and Thomas, 1977).

An individual’s score on a given conflict-handling mode is the number of times which of the statements representing a particular mode are selected in favour of other statements. Hence, each score ranges from 0 (very low use) to 12 (very high use). In order to have a comparison, the profile scores of the respondent can be graphed in relation to the scores of 339 middleand upper-level manager in business and government who have already completed the MODE Instrument (Womack, 1988b). The respondents are told that high or low use of a particular mode is not necessarily bad, since specific situations may require extreme use of a given conflict-handling mode (Thomas and Kilmann, 1974). Thomas and Kilmann (1974) furthermore stress that each mode may be useful in certain situations.

In the interpretation section of the MODE Instrument, Thomas and Kilmann (1974) offer a list of situations in which a particular mode might be most appropriate. Additionally, specific questions are asked to respondents who score extremely high or low on a given mode. These questions are supposed to help and give advice to the respondents on how to change their con- flict-handling modes if they feel that their actual mode is inappropriate in a given situation.

2.4.1. Critical Assessment of the Theoretical Foundation

The theoretical foundation of the Thomas and Kilmann MODE Instrument is basically adopted of the Managerial Grid introduced by Blake and Mouton (1964). Even though it was reinterpreted by Thomas (1976), it is in its essentials still comparable to the first Dual Concern Model. The main changes made by Thomas (1976) refer to the underlying two dimensions describing the modes. While Blake and Mouton consider two concerns of indi- viduals as sufficient to classify managerial behaviour, Thomas (1976) introduces the dimen- sions of assertiveness and cooperativeness, to specify behaviour more precisely, for the first time.

However, the theoretical foundation of the TKI seems to be sketchy and ambiguous, as will be discussed in the following paragraphs. The Thomas-Kilmann Conflict MODE Instruments is originally designed to assess an individual’s behaviour in conflict situations (Thomas and Kilmann, 1974) and this behaviour is classified into five conflict-handling modes. But what do these modes actually measure? Do they measure the individual’s actual behaviour as often contended?

Kilmann and Thomas (1977) claim that the five specific conflict-handling modes reflect inde- pendent dimensions of interpersonal conflict behaviour. According to Thomas (1976), be- haviour consists of the three interrelated components: orientation, strategic objectives, and tactics (actual behaviour). All these components and their influencing variables are important for the understanding of the resulting behaviour. Hence, when assessing an individual’s be- haviour in conflict situations, it would be appropriate to measure all of these components and consider their influencing variables at the same time. The MODE instrument on the other hand does not make the necessary distinction between these three components.

Thomas (1988) admits that the exact nature of the modes has not been clear for a long time. In his article on conflict management, Thomas (1976) uses the two-dimensional model to classify individuals’ conflict-handling orientation towards conflict behaviour. However, in a later publication, Thomas (1988) views the conflict handling modes as intentions. The con- flict-handling modes do not describe specific behaviour. According to Thomas (1988) they can rather be seen as strategic intentions to perform a specific behaviour. In other words a specific mode does not appear to be a behaviour, but rather an intended outcome. Of course, intentions are acted out through observable behaviour, but sometimes there are different ways of behaviour that reach a specific intended outcome and, vice versa, many specific behaviours are ambiguous (Thomas, 1988). Even earlier, in a comparison of four measurement instru- ments, Thomas and Kilmann (1978) conclude that the TKI is rather designed to assess specific intentions for handling conflict. Yet, in this article they do not define the term intention unambiguously. It remains unclear what is meant: an intended outcome or the intention to perform a specific behaviour? This shows the importance of defining styles accurately and of measuring individuals’ predisposition, strategy and tactics (actual behaviour) separately. The TKI does not meet these requirements.

The two-dimensional grid, which is used to classify the five distinct conflict-handling modes along two more basic dimensions of intent is purely a taxonomy (Thomas, 1988). It is not causal, i.e. the modes cannot be understood as a result of neither a high nor a low level of the assertiveness and the cooperativeness dimension. According to the interpretation section of the MODE Instrument (Thomas and Kilmann, 1974), the conflict-handling modes are a result of both, personal predispositions as well as of the requirements of the situation in which an individual finds him-/herself. In contrast to this explanation, Thomas (1988) refers to an arti- cle he wrote in 1976 on “Conflict and Conflict Management” (Thomas, 1976) to explain the cause of the modes. In this article Thomas (1976) describes two distinct models that represent conflict - the Process Model and the Structural Model - which focus on separate aspects of conflict. The Process Model concentrates on the sequence of events which transpire in a con- flict episode, while the Structural Model focuses upon the underlying variables which shape conflict behaviour (Thomas, 1976). Conflict behaviour therefore plays a central role in both models. When referring to the Process Model, the conflict-handling modes of an individual arise from their cognitions made in a conflict episode. Not only the desirability of different outcomes but also the behaviour of the other party have a great impact on the conflict-han- dling modes. Hence, conflict modes can change or evolve during an interaction (Thomas, 1988). In structural terms , on the other hand, the conflict-handling modes are a result of a variety of relatively stable characteristics of the parties (e.g., behavioural predispositions) as well as of the relational context (e.g., social pressures on the individual, incentive structure of the conflict situation, rules and procedures, and power structures) (Thomas, 1976; 1988). However, it remains unclear whether both models or which one of them is used to form the theoretical foundation and explain the conflict-handling modes measured by the MODE Instrument.

Depending on the causes of the modes, the items of the questionnaire should be fitting. To put it differently, do the items reflect an individual’s predispositions, cognitions, and/or in- corporate relational factors? A closer look at the items shows evidence that they are designed to measure individuals’ predisposition however incorporate relational and situational variables in no way. Hence, the TKI measures an individual’s modes independent of a particular con- text. This thus reflects the assumption that the modes can be assessed across different situa- tions (Womack, 1988b). In the instruction of the instrument the conflict situation is described as generally as possible. Hence, it seems impossible to find if it is someone’s intention, predisposition, or the situation that actually coins someone’s score on the modes. But in order to measure managers’ negotiation styles in business partnerships accurately this specific fac- tor is especially crucial.

Even though Kilmann and Thomas (1974) view the ipsative nature of the TKI as strength, the forced-choice format leads to additional drawbacks. For example, since respondents have to select between two modes to lower the social desirability bias of the instrument, the scores on the modes are unambiguously interrelated. As scoring on one style automatically consequents a reduction of the possible score on an other style, the total score on each mode may not reflect the respondent’s true preference (Putnam and Wilson, 1982; Womack, 1988a). As a conclusion, the underlying theoretical foundation of the TKI MODE is insufficient to measure managers’ negotiation style used in business partnerships accurately.

2.4.2. Critical Analysis of the Psychometric Properties

There is an impressive amount of empirical studies testing the reliability and validity of the MODE Instrument. In the following paragraphs a short overview of the most prominent ones is presented. However, due to its forced-choice format, in contrast to the other instruments presented in this chapter, the MODE Instrument produces ipsative data, which is not suitable for tests with inferential statistics (e.g., factor analysis) and constitutes a weakness in the researchers’ use and interpretation of the results of empirical surveys (Putnam and Wilson, 1982; Putnam and Poole, 1987; Womack, 1988a).

Reliability: Testing the internal reliability of the TKI, the Cronbach’s Alpha varies from study to study³⁷. The smaller the size of the used sample, the higher the internal reliabilities. However, in no study the MODE Instrument meets the standards determined by Nunnally (1967). In a study made by Kilmann and Thomas (1977) the mean coefficient alpha of the five modes is .60. Even though this is somewhat higher than the mean coefficient alphas of the other instruments³⁸ tested in this study (Thomas and Kilmann, 1978), the internal coefficients could still not meet Nunnally’s criteria (1967).

Similar results can be found in a study for the test-retest reliability. Again, the mean coeffi- cient was somewhat higher than those of other instruments, however, revealing a value of 0.64, the MODE Instrument does not meet the criteria of Nunnally either (Thomas and Kilmann, 1978). In conclusion, even though the tested statistical values of the TKI are better than those of other instruments, they are far away to meet Nunnally’s criteria for acceptable use in both research and applied settings.

Validity: In two empirical studies Ruble and Thomas (1976) investigate the content validity of the MODE, i.e., they test whether individuals use the two dimensions, assertiveness and cooperativeness, to classify conflict-handling behaviour. A factor analysis generally confirms the expected two-dimensional structure underlying the five conflict-handling modes. How- ever, possibly due to the inappropriate ipsative data produced by the MODE Instrument, only very few studies have been carried out to analyse the item structure in order to confirm the five distinct modes (Womack, 1988b). Konovsky, Jaster, and McDonald (1989) adapt the MODE by converting the forced-choice format to a 5-point Likert (1932) scale to assess the content validity. In contrast to Ruble and Thomas (1976) they find no support for the under- lying two dimensional structure in their study. In addition, due to the revised scale structure, Konovsky, Jaster, and McDonald (1989) are able to conduct a factor analysis to prove whether the items measure adequately the proposed five conflict-handling modes. The factor analysis reveals that only the items measuring three of the five modes loaded as predicted.

Another way to test the validity is to test the construct validity of an instrument, i.e., the inter- correlations among different instruments, which measure the same modes, are investigated. Depending on a given conflict-handling mode Thomas and Kilmann (1978) find only low to moderate correlations between Blake and Mouton’s (1964) Managerial Grid, Hall’s (1969a) CMS, and the Lawrence-Lorsch (1967) Instrument. In another study made by Putnam and Wilson (1982), the authors report moderately high correlation between the five styles MODE Instrument and three styles that were depicted by the OCCI by Putnam and Wilson (1982).

There are quite a few studies which test the predictive validity of the MODE Instrument (e.g., Duane, Azevedo, and Anderson, 1985; Goering, Rudick, and Faulkner, 1986; Kabanoff, 1987). However, as the MODE is designed to assess conflict-handling modes independent of a particular context, the testing of predictive validity poses a serious problem in so far as a particular behaviour cannot be judged as representative of any mode measured independent of a specific context (Kabanoff, 1987).

2.5. The Conflict Management Message Style Instrument

The Conflict Management Message Style (CMMS) Instrument developed by Ross and DeWine is a self-report instrument to assess an individual’s conflict styles through the analysis of messages used in interpersonal conflicts (Ross and DeWine, 1988). Ross and DeWine use the following three message types to classify styles for handling interpersonal conflicts in organizations:

Self-oriented:

An individual with a high score in this subscale emphasises his/her personal interests in an interpersonal conflict.

Issue-oriented:

An individual with a high score in this subscale argues that a problem can be solved without jeopardising the relationship and that both parties must deal with the problem.

Other-oriented:

An individual with a high score in this subscale wants to overlook the problem and keep the other party happy.

In contrast to the other instruments presented in this thesis, the CMMS is unique with regards to its explicit focus on communication. The questionnaire , which measures the styles of the CMMS, consists of two main parts. In the first part, the respondents are initially asked to describe their usual behavioural ways in a general conflict situation with another person. After that, the respondent is asked to recall a particular conflict situation in detail and to give information on what exactly was said, and which responses were made. According to Ross and DeWine (1988) the proceeding on this cover page should help the respondents to become more aware of their actual behaviour and hence, encourage congruity between the individuals’ assessment of the message styles in the questionnaire and their actual behaviour in a conflict situation.

The second part is the main part of the questionnaire. It comprises 18 different messages, six dealing with each style. In order to determine an individual’s conflict management style, the respondent is asked to rate these messages separately on a 5-point Likert scale ranging from remark such as ‘I never say things like this’ to ‘I usually say things like this’. The term ‘things like this’ indicates that the available messages may not exactly reflect a specific response of the subject. Hence, the respondent is advised to choose an answer that describes his/her behaviour in that particular situation best. Furthermore it is indicated that there are no right or wrong answers. After completing the questionnaire, the respondent receives three subscales, one on each style. Each subscale can range from 6, indicating a very low use of the behaviour described, to 30, reflecting a very high use of messages similar to the ones of a given message style. According to Ross and DeWine (1988) the CMMS is easy and possible to administer within a short time.

The CMMS doesn’t offer a special interpretation section. However, as the CMMS is designed for training purposes, the respondents are directed to interpret their scores by comparing them to an average use of a sample group and to discuss their results within a trainings group (Ross and DeWine, 1988).

2.5.1. Critical Assessment of the Theoretical Foundation

The CMMS instrument was initially developed to assess the five styles distinguished by Blake and Mouton (1964) to classify managerial behaviour (Ross and DeWine, 1988). Hence, it can be assumed that the original theoretical foundation of the CMMS was adopted and therefore is similar to the one of the Managerial Grid or the MODE Instrument. However, as a factor analysis in a pre-test of the CMMS could not maintain the assumption of the existence of five distinct styles, Ross and DeWine (1988) create a three factor solution: Factor 1 (now called ‘concern for self’) contains the statements which were originally in- tended to measure the competing and partly the avoiding style; the second factor (now called ‘concern for issue’) is created from messages reflecting the collaborating and compromising styles; factor 3 (now called ‘concern for other’) emerged from messages initially meant to measure the accommodating and partly the avoiding styles (Ross and DeWine, 1988). The theoretical foundation does therefore seem to be somewhat vague. The unusual devel- opment of the CMMS Instrument implies that the theoretical foundation of the CMMS is methodological driven rather than theory driven. Additionally, there are no studies testing a correlation between the two dimension that underlie the styles, e.g., concern for peo- ple/production as introduced by Blake and Mouton (1964) or cooperativeness/assertiveness as denoted by Thomas and Kilmann (1974) and the three styles distinguished by Ross and DeWine (1988). The two dimensions originally used simply to classify (e.g., MODE Instrument) or explain the causes of the distinct styles (e.g., Managerial Grid) are never men- tioned in the description of the CMMS.

However, the theoretical framework seems unambiguous with regard to what the CMMS Instrument is supposed to measure. According to Ross and DeWine (1988) the CMMS is designed to measure styles for handling interpersonal conflicts in organizations. Yet, what exactly do these styles measure?

As a conflict situation is manifested and managed through communicative behaviour, the CMMS focuses on the wording of messages to assess these communicative interchanges in an interpersonal conflict (Ross and DeWine, 1988). By concentrating on the formulation of the utterances, which reveal a given tactic when following a particular strategy in a conflict situation, the CMMS assesses the actual behaviour of an individual. A closer look at the items gives evidence that they are actually designed to measure communicative behaviour accurately. However, although the items of the CMMS Instrument are aimed to measure actual behaviour, they do not focus on nonverbal behaviour. An individual’s predisposition and strategy are definitely not considered by the CMMS.

According to Ross and DeWine (1988) the CMMS Instrument is designed to measure conflict styles in organizations. However, neither the cover page nor the questionnaire itself incorpo- rate relational or situational factors which take into account organizational settings or other influencing variables. Hence, the CMMS Instrument does not meet all the ideal requirements elaborated in the introduction to measure managers’ negotiation styles accurately.

2.5.2. Critical Analysis of the Psychometric Properties

There are no studies testing the reliability and the validity of the CMMS Instrument except for the ones made by the developers Ross and DeWine themselves. According to Womack (1988a) the results of this kind of studies have to be interpreted with caution, as she supposes that for some reason the developers of instruments find higher reliabilities than other research- ers testing the instrument.

Reliability: Testing the internal reliability of the CMMS, the instrument was given to a sam- ple of 123 students. The mean coefficient alpha for the three styles is 0.76 (Ross and DeWine, 1988). Even though this mean score is higher than those of most of the other in- struments presented in this chapter, it still does not meet Nunnally’s criteria. In order to examine the test-retest reliability, in Ross and DeWine’s (1988) study the CMMS was administered to a sample of 118 students in a one-week interval.

[...]

---

¹ e.g., emergence of intense global competition, economic integration among countries, formation of regional markets, technological innovation, and shortening of product life cycles (cf. Olson and Singsuwan, 1997).

² Partnerships attributes are, for example, trust, interdependency, commitment, culture, cooperation, and coordination.

³ Communication behaviour: quality, participation, and information sharing.

⁴ Conflict Resolution techniques: joint problem solving, persuasion, domination, and harsh words.

⁵ According to Ring and Van de Ven (1994) business partnerships emerge, evolve, and dissolve over time.

⁶ Dabholkar, Johnston, and Cathey (1994) assume that the use of an integrative negotiation style increases the likelihood that partnerships do succeed. Similar assumptions are made by Mohr and Spekman (1994), Olson and Singsuwan (1997), and Pruitt and Carnevale (1993).

⁷ We define negotiation predisposition as a relatively stable and enduring characteristic that exerts pervasive influence on an individual to negotiate in a specific way (see also p. 50).

⁸ We define negotiation strategy as a planned guideline on how to come to an agreement (see also p. 52).

⁹ We define negotiation tactics as the set of specific activities, which serve the implementation of a chosen negotiation strategy (see also p. 53).

¹⁰ The same assumptions are made by Adler, Brahm, and Graham (1992).

¹¹ Some authors like Deutsch (1973), Perdue et al. (1986), Pruitt (1981), and Williams (1993) use different terms such as competitive/aggressive and coordinative/cooperative describing the two main types of negotiation styles.

¹² Within this work we use the terms ’scoring instrument’, ‘measurement instrument’, ‘measure’, and ‘scale’ as synonyms.

¹³ e.g., Fisher and Ury (1992); Hall (1993); Lewicki et al. (1994); Mnookin, Peppet, and Tulumello (2000); Neale and Bazerman (1991); Pruitt (1981); Pruitt and Carnevale (1993); Putnam and Roloff (1992); Rahim (2001); Rubin, Pruitt, and Kim (1994).

¹⁴ Blake and Mouton (1964) originally talk of managerial style but within this thesis we use managerial and negotiation style as synonyms for each other.

¹⁵ Authors: Hall (1969a), Putnam and Wilson (1982), Rahim (1983b), Ross and DeWine (1988), Thomas and Kilmann (1974).

¹⁶ CMS (Hall, 1969a): Win-Lose Style, Yield-Lose Style, Lose-Leave Style, Compromise Style, Synergistic Style; ROCI II (Rahim, 1983b): Integrating Style, Compromising Style, Dominating Style, Obliging Style, Avoiding Style; TKI MODE-Instrument (Thomas and Kilmann 1974): Competing, Collaborating, Compromising, Avoiding, Accommodating.

¹⁷ OCCI (Putnam and Wilson, 1982): Solution-Orientation, Control, Nonconfrontation; CMMS (Ross and DeWine, 1988): Concern for Self, Concern for Issue, and Concern for Other.

¹⁸ Four negotiation styles are classified by Janssen and Van de Vliert (1996): Contending, Problem Solving, Yielding, Avoiding.

¹⁹ Four negotiation styles are classified by Pruitt (1983): Concession Making, Problem Solving, Inaction, Contenting.

²⁰ The CMMS (Ross and DeWine, 1988), and the MODE-Instrument (Thomas and Kilmann, 1974) reflect the assumption that an individual's disposition can be assessed across different situations.

²¹ Relational factors are able to capture the effects the other party has on an individual’s behaviour, e.g., sort of interdependency between the negotiation parties or the negotiation behaviour of the other party (please see Chapter 3.2.4. for a discussion in greater detail).

²² Situational factors describe the impact the particular context has on negotiation behaviour, e.g., the negotiation issue (please see Chapter 3.2.5. for a discussion in greater detail).

²³ For a closer look in terms of validity and reliability for each instrument, please see Chapter 2.

²⁴ The Thomas-Kilmann MODE Instrument and the Ross-DeWine CMMS.

²⁵ Hall’s CMS, Putnam and Wilson’s OCCI, and Rahim’s ROCI II.

²⁶ Some researchers (e.g., Bearden, Netemeyer, and Mobley , 1993; Churchill, 1979) use content validity and face validity as synonyms whereas others (e.g., Nunnally, 1967; Rossiter, 2001) clearly emphasise an important difference in the meaning of these two indexes.

²⁷ For example, according to Thomas and Kilmann (1978) predictive validity is the most rigorous and demanding test of the usefulness of an instrument, whereas Nunnally (1967, p. 77) claims that ‘other validity checks are much more important’. Furthermore, Bearden, Netemeyer, and Mobley (1993, p. 4) stress that developers of scales ‘should give a priori consideration to assessing test-retest reliability’ which totally contradicts Churchill’s (1979) view that test-retest reliability should not be used.

²⁸ However, due to insufficient data regarding the psychometric properties provided by some of the developers, some instruments cannot be analysed completely with respect to all mentioned validity and reliability indexes.

²⁹ Some authors (e.g., Bearden, Netemeyer, and Mobley, 1993) claim that construct validity is a superordinate concept composing other validity indexes such as ‘ convergent ’, ‘ discriminant ’, ‘ nomological ’, and ‘ known group validity ’.

³⁰ For a discussion in greater detail on this issue please see Campbell and Fiske (1959).

³¹ Predictive validity is sometimes referred to as concurrent validity or criterion-oriented validity (see Nunnally, 1967 and Cronbach, 1984).

³² Rossiter (2001) stresses that content validity should not be confused with ‘face validity’, which is a ‘post hoc claim that the items in the scale measure the construct’. Hence, ‘face validity’ is measured after the scale is developed. Consequently, there might be a reason for using the term ‘face validity’ instead of ‘content validity’ for the psychometric properties presented in the next paragraphs. However, since the developers of the instruments use the term ‘content validity’ and since it is them who we refer to, we will follow their example.

³³ We will discuss this procedure, called C-OAR-SE, in greater detail in Chapter 4.

³⁴ In a 9-point system it is theoretically possible to distinguish 81 different managerial styles. But as it is not the aim of the Managerial Grid to pigeon-hole managers according to the exact proportion of their two concerns, only five very characteristic styles will be contemplated in depth.

³⁵ MODE is an acronym for Management-of-Difference-Exercise.

³⁶ There are 7 statements reflecting the competitive mode, 7 the collaborating mode, 7 items describe the compromise mode, 7 the accommodating mode, and 6 the avoiding style.

³⁷ A clear overview of the results of different studies on the reliability of the MODE Instrument is given by Womack (1988b).

³⁸ In the same study the internal consistencies of the Lawrence-Lorsch (1967) instrument and the Hall (1969a) CMS are tested as well (Kilmann and Thomas, 1977).