CONFLICT DYNAMICS PROFILE TECHNICAL GUIDE

Transcription

1 CONFLICT DYNAMICS PROFILE TECHNICAL GUIDE Sal Capobianco, Ph.D. Mark H. Davis, Ph.D. Linda A. Kraus, Ph.D Eckerd College

2 I. INTRODUCTION The goal of this report is to provide a general overview of the technical factors associated with the Conflict Dynamics Profile or CDP, particularly its development and its psychometric and normative properties. The content and conventions for the presentation of data have been guided by Standards for Educational and Psychological Testing (American Psychological Association, 1985). Conflict in life is inevitable. Whether in the workplace, home or social gatherings, conflict results from the inescapable fact that people have different (and sometimes opposing) goals, needs, desires, responsibilities, perceptions, and ideas. Despite our best efforts to prevent conflict, we will undoubtedly find ourselves in disagreements with other people. We cannot prevent it entirely, nor would we want to, since some kinds of conflict can be productive. For our purposes, conflict refers to any situation in which people have incompatible interests, goals, principles, or feelings. This is, of course, a broad definition and encompasses many different situations. Conflicts could arise, for instance, over a long-standing set of issues, differences of opinion about strategy or tactics in the accomplishment of some business goal, incompatible beliefs, competition for resources, and so on. Conflicts can also result when one person acts in a way that another individual sees as insensitive, thoughtless, or rude. A conflict, in short, can result from anything that places individuals in opposition to one another. Particularly important to our conceptualization is the idea that conflict is dynamic. That is, conflict is an active process (often) with a beginning, middle, and end; a process with energy and force, capable of movement and change. Conflict may proceed slowly at times, then suddenly and quickly move in a different direction. Because this process revolves around social interaction, how and where this process leads depends on the participants. One's responses to a provocation can determine whether a conflict moves in a beneficial or harmful direction. Although conflict is inevitable, this is not necessarily bad, as some kinds of conflict can be beneficial. Conflict that focuses on ideas, rather than on the personalities and shortcomings of the people involved, can result in creativity and productivity, teamwork and improved group relations. Conflict that focuses on people, on the other hand, can escalate rapidly and unpleasantly, and have quite detrimental and far-reaching effects. The goal then is to try to manage conflict in such a way that its useful functions can develop while minimizing or avoiding entirely the more "toxic" forms. What largely separates useful conflict from destructive conflict is how the involved individuals respond when the conflict occurs. Thus, while conflict itself is inevitable, ineffective and harmful responses to conflict can be avoided, and effective and beneficial responses to conflict can be learned. This proposition is at the heart of the Conflict Dynamics Profile and the accompanying development guide, Managing Conflict Dynamics: A Practical Approach. 1

3 Origins of the CDP. To a considerable degree, the Conflict Dynamics Profile was developed in response to the approach taken by a number of existing measures. Such instruments as the Thomas-Kilmann Instrument (Thomas & Kilmann, 1974), Negotiating Styles Profile (Glaser & Glaser, 1996), and Rahim Organizational Conflict Inventories (Rahim, 1983) are all based on a model similar to or derived from that of Blake and Mouton's (1964, 1970) theoretical model, which argues that conflict style results from one s standing on two underlying dimensions: a desire to satisfy one s own needs and a desire to satisfy the other person s needs. As a result, the constructs measured by those instruments (avoidance, accommodation, compromise, collaboration, and competition) are fundamentally defined in terms of ultimate goals, such as avoiding conflict or winning a negotiation. The Conflict Dynamics Profile, on the other hand, is explicitly based on a behavioral orientation. That is, rather than try to identify conflict styles which represent a combination of behavior, personality, and motivation that can be difficult to change we made the decision to focus exclusively on the behaviors people typically display when faced with conflict. We made this choice for two reasons. First, focusing on specific sets of behaviors would allow detailed examination, and subsequently greater understanding, of how people typically respond to conflict. Second, an explicit behavioral approach, we believed, would provide especially useful information to individuals whose goal is to change. Because we do not focus on personality and motivation, but only on how people act, we take the optimistic position that people can change their behavior for the better. That is, people can change the way they respond to conflict, and the more people know, the better equipped they will be to change. Specifically, the more they know about how they act before, during, and after conflict, what sets them off most easily, and what responses to conflict are especially harmful in their own organization, then the better equipped and (it is hoped) motivated they will be to change. Theoretical Background To begin the process of developing the CDP, we carried out a review of the relevant literature by focusing on four approaches to conflict management: Investigations of conflict in organizations which have gone beyond the style-based approach to examine how conflict naturalistically develops in both positive and negative ways (Amason, 1996; Feeney & Davidson, 1996; Sessa, 1996; van de Vliert Euwema, & Huismans, 1995). Conflict in interpersonal rather than business or work relationships as studied by social psychologists who have examined the constructive and destructive ways of handling interpersonal conflict (Rusbult, Verette, Whitney, Slovik, & Lipkus, 1991; Gottman, 1994; Berry & Willingham, 1997). Popular press accounts such as How to Win Friends & Influence People (Carnegie, revised edition, 1981), Getting to Yes (Fisher & Ury, 1981), and Negotiating for Dummies (Donaldson & 2

4 Donaldson, 1996) which offer highly specific and useful advice with regard to effectively resolving conflict situations. Popular press writings and books on managerial advice which offer suggestions regarding situations and circumstances likely to evoke conflict in the first place (Bramson, 1981, 1992; Glass, 1995; Davis, Skube, Hellervik, Gebelein, & Sheard, 1992). Two key findings emerged from our examination of these sources. First, it is clear that distinctions can be made among types of conflict and it is possible to have conflict that is useful rather than harmful. Cognitive conflict is the type that focuses on ideas, rather than personalities. It is the kind of conflict that can at times result in creativity and productivity. While cognitive conflict tends to increase arousal, the affect associated with it is positive, and it should be positively related to good group functioning, or at worst, should be unrelated to functioning. Affective conflict, on the other hand, is the type that focuses on people, not on ideas. This is the kind of conflict that can escalate rapidly and unpleasantly. It increases arousal, but the affect associated with it is negative. Affective conflict is negatively related to good group functioning. Thus, conflict is generally most beneficial when the emphasis is on issues and problem-solving, and it is most detrimental when focused on personalities and competition. The goal of conflict management then is to minimize the occurrence and escalation of harmful (affective) conflict while allowing the useful forms of (cognitive) conflict to unfold. The second finding to emerge from our examination of the literature is that behavioral responses to conflict can be important shapers of the course that conflicts may take. Specifically, responses made in the early stages of conflict can determine whether a conflict becomes cognitive or affective. Constructive responses have the effect of not escalating the conflict further. They tend to reduce the tension and keep the conflict focused on ideas rather than personalities. Destructive responses, on the other hand, tend to make things worse; they do little to reduce the conflict and allow it to focus on personalities. Responses to conflict also differ in terms of how active or passive they are. Active responses are those in which the individual takes some overt action in response to the conflict or provocation. Such responses can be either constructive or destructive; what makes them active is that they require some overt effort on the part of the individual. Passive responses, in contrast, do not require much in the way of effort from the person. In fact, they typically involve the person deciding to refrain from some kind of action. Again, passive responses can be either constructive or destructive; that is, they can make things better or they can make things worse. Given then that responses can be either constructive or destructive, and either active or passive, we view responses to conflict as falling into one of four categories: Active-Constructive, Passive-Constructive, Active-Destructive, and Passive-Destructive. The distinctions between types of conflict and types of responses to conflict led us to develop a model of conflict. 3

5 The Dynamic Conflict Model The starting point for the model (see Figure 1) is a precipitating event something that sets the stage for a conflict to develop. This event could be anything: a single behavior by another person which is upsetting or frustrating, a long-standing set of issues between people, a difference of opinion about strategy or tactics in the accomplishment of some business goal, and so on. The precipitating event can be anything that places the interests of individuals in opposition to one another. PRECIPITATING EVENT/HOT BUTTONS INITIATE CONFLICT CONSTRUCTIVE RESPONSES ACTIVE Perspective Taking (PT) Creating Solutions (CS) Expressing Emotions (EE) Reaching Out (RO) PASSIVE Reflective Thinking (RT) Delay Responding (DR) Adapting (AD TASK-FOCUSED CONFLICT Focus on task and problem solving Positive affect Tension decreases Group functioning improves DESTRUCTIVE RESPONSES ACTIVE Winning at all Costs (WI) Displaying Anger (DA) Demeaning Others (DO) Retaliating (RE) PASSIVE Avoiding (AV) Yielding (YL) Hiding Emotions (HE) Self-Criticizing (SC) PERSON-FOCUSED CONFLICT Focus on personalities Negative affect Tension increases Group functioning derailed CONFLICT DEESCALATES CONFLICT ESCALATES Figure 1. The Dynamic Conflict Model The presence of a precipitating event sets into motion the dynamics of conflict, but the end result of that process is still to be determined. One of the biggest influences on how things unfold, we argue, will be the behavioral responses (Active-Constructive, Passive-Constructive, Active- Destructive, and Passive-Destructive) of the people in the conflict. Constructive responses, both active and passive, will on average tend to prevent precipitating events from developing into emotional, person-oriented conflicts. In contrast, destructive responses especially active ones will on average tend to exacerbate the situation and thus make it more likely that an emotional, person-oriented conflict will result. 4

6 One's behavioral responses to provocation, which can determine whether the potential conflict evolves in either the cognitive or affective direction, can also play a role later in the conflict sequence. That is, after a conflict develops in one or the other general direction, it can also "change course" depending on the behaviors which occur later. For example, a situation can begin as a cognitive conflict with controllable levels of arousal and centered on some non-personal issue, but destructive responses during this phase could change the direction of this sequence and lead to affective conflict instead. A misinterpreted statement or a hasty response that should have been inhibited can engage personal feelings in a way that is not helpful to the resolution of the matter at hand. Alternatively, it is possible that a dispute that started out as an affective conflict could be "reined in" by careful behavioral work and transformed into a less destructive cognitive conflict. For example, if at least one party to the affective conflict can begin inhibiting his or her destructive responses (an admittedly difficult task), and begin substituting constructive ones, then the affective conflict is denied the fuel it needs to perpetuate itself. In this sense, affective conflict can be thought of as a kind of "fire" destructive acts are the fuel needed to burn and constructive acts are like the wet blankets that dampen the conflagration. The goal of successful conflict management is to remove fuel and add water to a volatile situation. Another important feature of our behavioral approach to conflict is the concept of Hot Buttons those situations and individuals that are especially annoying, frustrating or upsetting. An individual s Hot Buttons can be thought of as the kinds of people or behaviors that are especially likely to serve as precipitating events for that person. When pushed, Hot Buttons can provoke one into starting or escalating a conflict. The "hottest" Hot Buttons (that is, those that are most upsetting) will be the ones most likely to evoke a quick and automatic set of destructive responses, while the "cooler" Buttons are more likely to evoke a mixture of responses that include some constructive behaviors. By understanding and examining the links between provocation and response, it becomes easier to control one's behavior. 5

7 Conflict Dynamics Profile Drawing upon theoretical work, academic research, and popular press writings as well as the Dynamic Conflict Model, we developed the Conflict Dynamics Profile, a multi-rater instrument designed to help individuals better understand the way they typically respond to conflict and to help them improve those areas that are most problematic. The way in which the CDP does this is by measuring an individual's behavior from several different vantage points. First, the CDP asks the individual (self) to describe how s/he thinks s/he responds before, during, and after conflict. Second, the CDP asks other people specifically, one's boss, peers, and direct reports how they see the individual responding before, during, and after conflict. The purpose of asking the same questions of many different people is to highlight any differences that may exist between the way an individual sees his/her behavior and the way one's friends and colleagues see it. Because this design describes the full panorama of one's behavior, it is referred to as a 360 or multi-rater instrument. The Conflict Dynamic Profile provides a complete "conflict profile" by providing feedback on: What provokes an individual (Hot Buttons). How that individual perceives the way s/he typically responds to conflict. How others view that individual as responding to conflict. How the individual responds before, during, and after conflict (Dynamic Conflict Sequence). Which responses to conflict have the potential to harm one's position in his/her particular organization (Organizational Perspective on Conflict). Table 1 outlines these and other features of the CDP Feedback Report. 6

8 Table 1. Guide to the CDP Feedback Report. Active-Constructive Response Profile Passive-Constructive Response Profile Active-Destructive Response Profile Passive-Destructive Response Profile Scale Profile Discrepancy Profile Dynamic Conflict Sequence Organizational Perspective on Conflict Hot Buttons Profile Developmental Feedback Developmental Worksheets Four ways of responding to conflict which require some effort on the part of the individual and which have the effect of reducing conflict: Perspective Taking, Creating Solutions, Expressing Emotions and Reaching Out. Three ways of responding to conflict which have the effect of dampening the conflict, or preventing escalation, but which do not require any active response from the individual: Reflective Thinking, Delay Responding, and Adapting. Four ways of responding to conflict which through some effort on the part of the individual have the effect of escalating the conflict: Winning at All Costs, Displaying Anger, Demeaning Others, and Retaliating. Four ways of responding to conflict which due to lack of effort or action by the individual cause the conflict to either continue or to be resolved in an unsatisfactory manner: Avoiding, Yielding, Hiding Emotions, and Self- Criticizing. How one s typical responses during conflict are viewed by his/her boss, peers, and direct reports. The particular responses to conflict on which one s self-perceptions and those of others differ most markedly. How constructively and destructively one responds to conflict before it begins, after it is underway, and after it is over. The particular responses to conflict which are especially discouraged in one s organization; regularly engaging in these responses can have severe negative effects on one s career. The types of people and situations most likely to irritate the individual and provoke conflict. Direct comments about one s responses to conflict from his/her boss, peers, and direct reports. Two worksheets to aid the individual in identifying his/her clearest opportunities for development. 7

9 II. ITEM SELECTION AND SCALE CONSTRUCTION The Conflict Dynamics Profile is the result of two years of development and revision. Three preliminary versions were developed, administered, analyzed, and revised in order to produce the current CDP. Item analyses and factor analyses were carried out on each version of the instrument, and based upon these analyses, items and item sets were deleted, added, or in some cases combined in order to produce a cleaner and more psychometrically sound instrument. The rating scales used in the CDP are straightforward. Respondents indicate the frequency of occurrence of a particular Response to Conflict along a five-point rating scale continuum (1=Never, 2=Rarely, 3=Sometimes, 4=Often, 5=Almost Always). For Hot Buttons, respondents indicate, again along a five-point continuum, the degree to which they are upset by a particular individual or situation (1=Not At All, 2=A Little, 3=Moderately, 4=Considerably, 5=Extremely). A three-point rating scale is used for the Organizational Perspective on Conflict; respondents indicate the kind of effect a particular response to conflict would have on a person's career in their organization (No Negative Effect, Moderately Negative Effect, and Severely Negative Effect). Version 1. Drawing upon the theoretical work, research, and popular press writings described earlier, we created the first version of the Conflict Dynamics Profile in the spring of We wrote 97 items designed to tap 20 different possible Responses to Conflict as well as 116 items tapping 17 Hot Buttons. Table 2 charts the scale construction process of the Responses to Conflict while Table 3 does the same for the Hot Buttons. This first version of the instrument was then administered to a convenience sample of 160 undergraduates, continuing education students, and college staffers. 8

10 9 Table 2. Constructing the CDP Responses to Conflict scales. VERSION 1 VERSION 2 VERSION 3 VERSION 4 Active-Constructive Active-Constructive Active-Constructive Active-Constructive Gather Information 4 Perspective Taking 4 Perspective Taking 4 Perspective Taking 4 Creating Solutions 6 Creating Solutions 4 Creating Solutions 4 Creating Solutions 4 Open Communication 5 Open Communication 4 Open Communication 5 Expressing Emotions 5 Reaching Out 5 Reaching Out 5 Reaching Out 4 Reaching Out 4 Accept Responsibility 3 Repairing Emotions 5 Repairing Emotions 5 Respecting the Other 4 Respecting the Other 5 Passive-Constructive Passive-Constructive Passive-Constructive Passive-Constructive Reflective Thinking 6 Reflective Thinking 4 Reflective Thinking 5 Reflective Thinking 4 Tactical Avoidance 6 Delay Responding 4 Delay Responding 4 Delay Responding 5 Acceptance 6 Acceptance 4 Acceptance 4 Adapting 5 Controlling Emotions 5 Establish Distance 5 Adapting 5 Active-Destructive Active-Destructive Active-Destructive Active-Destructive Winning at all Costs 6 Winning at all Costs 5 Winning at all Costs 4 Winning at all Costs 4 Displaying Anger 6 Displaying Anger 5 Displaying Anger 4 Displaying Anger 4 Demeaning Other 7 Demeaning Other 5 Demeaning Other 5 Demeaning Other 4 Retaliating 3 Retaliating 5 Retaliating 4 Retaliating 4 Criticize Other 3 Express Opposition 4 Passive-Destructive Passive-Destructive Passive-Destructive Passive-Destructive Total Avoidance 5 Total Avoidance 4 Avoiding 4 Avoiding 4 Total Accommodation 3 Yielding 5 Yielding 4 Yielding 4 Withdrawal 4 Hiding Emotions 4 Hiding Emotions 4 Hiding Emotions 4 Internalizing 6 Self-Criticizing 4 Self-Criticizing 4 Self-Criticizing 4 Stonewalling 4 Stonewalling 5 Stonewalling 4 Total Number of Items 97 Total Number of Items 90 Total Number of Items 77 Total Number of Items 63

11 Table 3. Constructing the CDP Hot Buttons scales. VERSION 1 VERSION 2 VERSION 3 VERSION 4 Unreliable 6 Unreliable 4 Unreliable 4 Unreliable 4 Expertise 7 Overly-Analytical 4 Overly-Analytical 4 Overly-Analytical 4 Unappreciative 7 Unappreciative 4 Unappreciative 5 Unappreciative 4 Aloof 5 Aloof 4 Aloof 5 Aloof 4 Intrusive 7 Untrusting 5 Micro-Managing 4 Micro-Managing 4 Self-Centered 7 Self-Centered 5 Self-Centered 5 Self-Centered 4 Abrasive 8 Abrasive 4 Abrasive 4 Abrasive 4 Dishonest 7 Untrustworthy 5 Untrustworthy 4 Untrustworthy 4 Hostile 6 Hostile 4 Hostile 4 Hostile 4 Bad Attitude 5 Arrogant 5 Passive 8 Passive 4 Competitive 6 Autocratic 7 Communication 8 Critical 8 Inflexible 7 Unfair 7 Total Number of Items 116 Total Number of Items 48 Total Number of Items 39 Total Number of Items 36 10

12 To evaluate the adequacy of our efforts to create items, we carried out a series of factor analyses. Our strategy in regard to the Responses to Conflict was to separately examine the four domains of Active-Constructive, Passive-Constructive, Active-Destructive, and Passive-Destructive responses. Therefore, separately for each domain, we carried out a principal components analysis (oblique rotation) on the items intended to reflect the behavior set within that domain. For example, the Active-Constructive domain was represented by 27 items intended to tap six different behavioral responses. The number of items in the other three domains ranged from 22 to 25, and the number of intended behavioral responses ranged from four to five. A similar series of factor analyses (principal components analysis, oblique rotation) was conducted on the 17 Hot Buttons and the five to eight items intended to reflect each one. Not unexpectedly, the factor analyses revealed that some of the intended sets fared better than others. In some cases, dimensions expected to emerge from the factor analyses performed as anticipated; in other instances, the expected dimensions failed to appear at all. Sometimes items from two different anticipated dimensions loaded together to form a single construct; other times, items from a single anticipated dimension split apart to form separate dimensions. Subsequently, based on the results of these analyses, as well as an examination of individual item characteristics, we then prepared the second version of the CDP. Version 2. Based on the insights provided by the factor analyses of the CDP Version 1, we combined some of the original intended dimensions, split apart others, and generally attempted to produce dimensions which more closely resembled those suggested by the analyses of the first sample. Revising or dropping some items as well as writing several new ones subsequently led to the second version of the CDP. This version, containing 90 items designed to tap 20 Responses to Conflict (see Table 2) and 48 items aimed at tapping 11 Hot Buttons (Table 3), was then administered in the fall of 1998 to 263 adult workers from 13 different organizations and companies. To evaluate the adequacy of the second version of the CDP, we carried out a series of factor analyses. Our strategy in these analyses paralleled that employed in Version 1: separate principal components analyses (oblique rotation) on the Hot Buttons as well as on the items reflecting the Responses to Conflict sets within the four domains of Active-Constructive, Passive- Constructive, Active-Destructive, and Passive-Destructive responses. The number of items falling into the four response domains ranged from 17 to 27, and the number of intended behavioral responses within each domain ranged from four to six. As in the analyses of Version 1, some dimensions that we had expected to emerge from the factor analyses performed as anticipated, while others did not. In general, however, the results of these analyses began to suggest the Responses to Conflict and Hot Buttons that were most strongly and consistently reflected in the workplace. Based on these results then, we created a third version of the CDP that focused more narrowly on a smaller number of dimensions. 11

13 Version 3. For the third version of the CDP, we combined some of the intended dimensions from the second version, dropped others, revised and re-wrote items as we continued in our goal to produce scales that would accurately reflect the life experiences of our respondents. Version 3, containing 77 items designed to tap 18 Responses to Conflict (Table 2) and nine Hot Buttons assessed by 39 items (Table 3), was then administered throughout the winter of to 354 adult workers. To evaluate the adequacy of the third version of the CDP, we again carried out a series of factor analyses paralleling those previously described. A majority of the intended dimensions clearly emerged from these analyses, strengthening our conclusions as to the most stable and important Hot Buttons and Responses to Conflict. However, as in earlier factor analyses, a few inconsistencies also emerged. Two scales intended to be separate from one another instead merged into a single factor, while a handful of other dimensions emerged only partially or not at all. In general, however, the results of these analyses reinforced our conclusions as to which dimensions were most strongly and consistently reflected in the lives of working adults. Based upon these results, we created the final version of the Conflict Dynamics Profile. Final version. In most cases, decisions about the particular items and scales making up this final version were based solely on the factor analyses conducted on the Version 3 data. However, in a few instances, items or scales were included in the final version despite their failure to emerge clearly from the Version 3 analyses. First, in the Version 3 factor analysis of the Active-Destructive domain and in contrast to all earlier analyses the items tapping the Displaying Anger and Demeaning Other dimensions loaded together on a single factor rather than splitting into two factors. Given their separation in earlier analyses, and the substantive distinction that we believe exists between them, we retained both scales in the final version of the CDP. Second, for three of the Responses to Conflict scales Expressing Emotions, Delaying Responding, and Adapting we were not entirely satisfied with the items making up these measures. Therefore, in addition to the three items for each scale that were acceptable, we wrote two new items for each construct; consequently, these three scales consist of five items rather than four. Finally, in the Hot Buttons analyses of the Version 3 data, two scales that had emerged in previous analyses Aloof and Abrasive did not so clearly emerge for this version. However, given the earlier findings, and our belief that there was value in including these dimensions, these two scales were retained for the final CDP version. To reiterate, however, in the vast majority of cases, decisions about final item inclusion were based solely on the result of the Version 3 analyses. At the present time, the final version of the CDP has been completed by over 2000 individuals (and their bosses, peers, and/or direct reports). Factor analyses based on the responses of this group strongly support the psychometric structure of the CDP. Separately for each of the four Responses to Conflict domains, principal components analysis using oblique rotation, and specifying the intended number of factors (either three or four),were carried out; in each case the results almost perfectly replicated the intended pattern of item loadings. In the case 12

14 of the Hot Buttons items, the factor analyses also largely supported the factor structure. Five of the nine Hot Button scales were perfectly or substantially reproduced by the factor analyses; in addition, items from two somewhat similar scales (Untrustworthy and Unreliable) loaded on a single factor, as did the items from two other similar scales (Self-Centered and Abrasive). See Tables 4 and 5 for scale definitions and factor loadings for items making up the final version of the CDP. 13

15 Table 4. CDP Responses to Conflict scale definitions and mean factor loadings based on data from 9318 working adults. Scale Name Perspective Taking (PT) Creating Solutions (CS) Expressing Emotions (EE) Reaching Out (RO) Reflective Thinking (RT) Delay Responding (DR) Adapting (AD) Winning at All Costs (WI) Displaying Anger (DA) Demeaning Other (DO) Retaliating (RE) Avoiding (AV) Yielding (YL) Defined as Responding to Conflict By Putting one's self in the other person's position and trying to understand his/her point of view. Brainstorming with the other person, asking questions, and trying to create solutions. Talking honestly with the other person and expressing one's thoughts and feelings. Reaching out to the other person, making the first move, and trying to make amends. Analyzing the situation, weighing the pros and cons, and thinking about the best response. Waiting things out, letting matters settle down, or taking a "time out" when emotions are running high. Staying flexible and trying to make the best out of the situation. Arguing vigorously for one's own position and trying to win at all costs. Expressing anger, raising one's voice, or using harsh, angry words. Laughing at the other person, ridiculing his/her ideas, and using sarcasm. Obstructing or retaliating against the other person, and trying to get revenge later. Avoiding or ignoring the other person, and acting distant and aloof. Giving in to the other person in order to avoid further conflict. N Items Mean Factor Loading Hiding Emotions (HE) Concealing one's true emotions even though feeling upset Self-Criticizing (SC) Replaying the incident over in one's mind later, and criticizing one's self for not handling it better

16 Table 5. CDP Hot Buttons scale definitions and mean factor loadings based data from 9318 working adults. Scale Name Defined as Becoming Especially Irritated with People Who: N Items Mean Factor Loading Unreliable Overly-Analytical Unappreciative Aloof Micro-Managing Self-Centered Are unreliable, miss deadlines, and cannot be counted on. Are perfectionists, overanalyze things, and focus too much on minor issues. Fail to give credit to others or seldom praise good performance. Isolate themselves, do not seek input from others, or are hard to approach. Constantly monitor or check up on the work of others. Are self-centered or believe they are always correct ** * Abrasive Are arrogant, sarcastic, and abrasive * Untrustworthy Hostile Exploit others, take undeserved credit, or cannot be trusted. Lose their tempers, become angry, or yell at others ** * Items from these two scales loaded on the same factor. ** Items from these two scales loaded on the same factor. 15

17 The final version of the Conflict Dynamics Profile is divided into three sections which vary depending on which questionnaire one is completing: Self comprises 114 items while Respondents (i.e., boss, peers, and direct reports) contains 80 items. The first section (63 items) taps fifteen Responses to Conflict by having respondents answer ten or so questions at a time in response to six different stems denoting circumstances that exist early, during, or late in a conflict episode. Thus, in addition to providing a more detailed context within which people can answer the questions, the CDP provides the basis for specific feedback about the person s typical behavior before, during, and after conflict (Dynamic Conflict Sequence). Another section assesses the Organizational Perspective on Conflict, that is, the kinds of conflict-related behaviors within one's particular organization that have the most negative effect on an individual's career. Respondents indicate the kinds of conflict-related behaviors that are viewed most negatively by their particular organization. The fifteen items correspond to each of the fifteen Responses to Conflict scales of the CDP. Unlike the two aforementioned sections, the 36 items assessing the nine Hot Buttons are completed only by the target individual and not by his/her boss, peers, and direct reports. This is because it seemed like a very difficult task for other people to answer with any authority about the kinds of people and situations which most irritate and upset the target person. Finally, the form completed by one's boss, peers, and direct reports asks two openended questions regarding the individual's approach to conflict (Developmental Feedback). These items are included to capture any additional insights that the 360 evaluation process can provide about the target individual. 16

18 III. RELIABILITY Internal Reliability The internal reliability (alpha) coefficients for the fifteen Responses to Conflict scales appear in Table 6. The first column depicts the reliabilities (for self-ratings) for 9318 working adults who completed the final version of the CDP. The other three columns contain reliability data for the bosses, peers and direct reports of these individuals. Table 6. Internal reliability of CDP Responses to Conflict scales based on data from 2374 working adults and their raters. Self (N=9318) Boss (N=8780) Peers (N=33,275) Direct Reports (N=24,474) PT CS EE RO RT DR AD WI DA DO RE AV YL HE SC

19 As Table 6 indicates, internal reliability estimates are for the most part quite acceptable, with alpha coefficients exceeding.70 over 80% of the time, and exceeding.80 over 50% of the time. Moreover, the fact that most of the scales consist of only four items makes the size of these coefficients somewhat more impressive. Table 7 depicts the internal reliability for the nine Hot Buttons scales, based on the responses of the same 2374 working adults mentioned previously. As with the Responses to Conflict scales, the reliabilities range from acceptable to very good. Table 7. Internal reliability of CDP Hot Buttons scales based on data from 9318 working adults. Unreliable 0.86 Overly-Analytical 0.74 Unappreciative 0.86 Aloof 0.70 Micro-Managing 0.86 Self-Centered 0.75 Abrasive 0.71 Untrustworthy 0.76 Test-Retest Reliability To determine the stability of the 15 Responses to Conflict scales, 83 undergraduate students completed that portion of the CDP at two time points; the interval between the two administrations ranged from 77 to 91 days. As displayed in Table 8, scores at Time 1 and Time 2 were significantly and positively correlated for each scale; the lowest test-retest value was.43, and the highest was.73, with a mean value of.64. This pattern suggests that the tendency to display specific behavioral responses to conflict is at least somewhat stable over a period of weeks. However, the fact that the two lowest associations were found for scales in the Active-Destructive domain (Demeaning Other and Retaliating) is worth noting. Relative to most of the other CDP dimensions, both of these are low-frequency behaviors. One possibility is that this infrequency may contribute to the somewhat lower temporal stability of these dimensions. Simply put, the evidential basis on which respondents make their self-judgments at any given time may be smaller and less stable for those dimensions. Alternatively, of course, it may be that these behaviors simply are less stable over time, perhaps more under the influence of features of the particular conflict episode (e.g., the kind of provocation; prior relationship with the other party). 18

20 IV. VALIDITY Social Desirability One way in which to evaluate the validity of a new instrument is to examine its relations with measures of social desirability. Table 8 displays the relationship between the fifteen Responses to Conflict and two measures of social desirability subscales of the Balanced Inventory of Desirable Responding (BIDR; Paulhus, 1988). The Self-Deception scale of this instrument is designed to tap overly-positive self-perceptions, and is made up of items assessing a person s inflated views of his or her judgment and rationality. The Impression Management scale, in contrast, is designed to tap a concern for how one comes across to others, and is made up of items that measure overly-positive claims about overt behaviors of which others would be aware. Table 8. Correlations of CDP Responses to Conflict scales with measures of social desirability (N = 137 undergraduates). BIDR Self- Deception BIDR Impression Management PT 0.18* 0.03 CS 0.21* 0.07 EE 0.23** 0.04 RO RT 0.23** 0.14 DR * AD 0.23** 0.21* WI DA * DO RE -0.22** -0.25** AV -0.21* YL HE SC *p <.05 **p <.01 19

21 Although there were some significant correlations between the CDP scales and the BIDR scales, they were for the most part not large, hovering around.20 or so. Thus, it does not appear that the CDP has serious problems with social desirability. It also appears that scores on the CDP scales were somewhat more associated with an inflated view of the self (Self-Deception) than with a conscious desire to appear admirable in the eyes of others (Impression Management). It is important to note that although the data here are quite acceptable, social desirability concerns are of somewhat less importance with a multi-rater instrument such as the CDP. After all, the point of asking peers and bosses and direct reports to complete the measure is to overcome inaccuracies in self-perceptions such as a desire to appear in a desirable light. Scale Inter-correlations A second means by which to evaluate the validity of an instrument such as the CDP is by examining the relations of its subscales to one another. Table 9 displays the inter-relationships among self-ratings on the fifteen Responses to Conflict for 2398 working adults; several features of the table are worth noting. First, inter-correlations among scales within the same general domain (e.g., Active-Constructive) were always positive in sign, statistically significant, and substantial in size. Inter-correlations within the two active domains (mean r for the four Active-Constructive scales was.45, and for the four Active-Destructive scales was.43) were slightly higher than for the two Passive domains (mean r for both the Passive-Constructive and Passive-Destructive scales was.34). 20

22 21 Table 9. Inter-correlations of CDP Responses to Conflict scales (N = 9318 working adults). PT CS EE RO RT DR AD WI DA DO RE AV YL HE Perspective Taking (PT) Creating Solutions (CS).56** Expressing Emotions (EE).34**.52** Reaching Out (RO).46**.51**.45** Reflective Thinking (RT).57**.57**.27**.38** Delay Responding (DR).21**.04* -.13**.14**.30** Adapting (AD).44**.43**.18**.43**.48**.27** Winning at All Costs (WI) -.21** -.10**.06** -.16** -.18** -.20** -.19** Displaying Anger (DA) -.31** -.24**.07** -.17** -.40** -.20** -.30**.42** Demeaning Others (DO) -.30** -.32** -.13** -.29** -.37** -.07** -.27**.36**.53** Retaliating (RE) -.25** -.31** -.18** -.28** -.27** -.05* -.23**.37**.42**.55** Avoiding (AV) -.28** -.46** -.35** -.35** -.27**.22** -.21**.17**.28**.41**.44** Yielding (YL) -.09** -.26** -.31** -.08** -.18**.27**.08* -.03**.11**.20**.18**.34** - - Hiding Emotions (HE) -.09** -.25** -.54** -.17** **.08** -.09** -.10**.06*.14**.35**.38** - Self-Criticizing (SC) -.12** -.24** -.24** -.08** -.18**.10** -.07**.13**.20**.20**.19**.33**.31**.34** **p <.001

23 With regard to correlations across domains, the scales in the Active-Constructive and Passive-Constructive domains were almost all positively correlated with one another, and in most cases these associations were substantial (mean r =.28). In contrast, there was considerably more variety in the associations found between scales in the Active-Destructive and Passive-Destructive domains. Some of these associations were close to zero, and the mean r was only.14. Thus, it appears that constructive responses of any kind tend to be associated with constructive responses of every kind, while the relationships between differing forms of destructive responding are somewhat more varied. Relations between Self-Ratings and Ratings by Others Another method by which to evaluate the validity of the CDP scales is to compare the way that test-takers rate themselves with the ratings made by bosses, peers, and direct reports. If the scales are valid indicators of the intended constructs, then self- and other-ratings should tend to be positively and significantly correlated. Table 10 displays these correlations for over 2000 working adults who were also rated by bosses, peers, and/or direct reports. Ratings by multiple peers of a single test-taker were averaged together to produce a mean peer rating for each scale; a similar process was carried out for ratings by multiple direct reports of a single test-taker. In almost all cases, only a single boss rated the test-taker; when multiple boss ratings were collected, however, these were also averaged for each scale. 22

24 Table 10. Correlations between self-ratings working adults and others ratings for CDP Responses to Conflict scales. Self-Boss (N=7773) Self-Peers (N=8709) Self-Direct Reports (N=7256) PT 0.18** 0.21** 0.17** CS 0.14** 0.16** 0.14** EE 0.18** 0.21** 0.18** RO 0.17** 0.19** 0.16** RT 0.16** 0.20** 0.17** DR 0.11** 0.17** 0.15** AD 0.14** 0.16** 0.14** WI 0.19** 0.24** 0.19** DA 0.27** 0.33** 0.32** DO 0.19** 0.22** 0.20** RE 0.11** 0.15** 0.11** AV 0.12** 0.16** 0.13** YL 0.16** 0.20** 0.15** HE 0.17** 0.20** 0.16** SC 0.14** 0.17** 0.13** **p <

25 As the table reveals, each of the 45 correlations were positive in sign and statistically significant, thus supporting the validity of the 15 Responses to Conflict scales. For the most part, associations between self-ratings and ratings by others were roughly equivalent across rater category (boss, peers, and direct reports) and behavior category (Active-Constructive, Passive-Constructive, Active- Destructive, and Passive-Destructive). One exception to this general pattern was the finding that agreement regarding Active-Destructive behavior was slightly stronger than for the other three categories. Relations with Other Conflict Measures A third means by which to evaluate the validity of a new instrument is to compare it to existing measures of the same construct. Accordingly, correlations between the fifteen Responses to Conflict scales and scores on two existing conflict measures, the Thomas-Kilmann Instrument (TKI) and the Negotiating Styles Profile (NSP), appear in Table 11. These values come from two different sets of respondents: 122 undergraduate students who completed both the TKI and NSP, and 146 working adults who completed the TKI. 24

26 25 Table 11. Correlations between CDP Responses to Conflict scales and the TKI and NSP subscales (N = 122 undergraduate students and 146 working adults). Collaborate Accommodate Compromise Defeat Withdraw Students Adults Students Adults Students Adults Students Adults Students Adults NSP TKI TKI NSP TKI TKI NSP TKI TKI NSP TKI TKI NSP TKI TKI PT 0.21* * 0.22* * CS 0.20* 0.22* 0.17* ** -0.20* EE ** * ** RO 0.23** ** 0.27** 0.18* 0.25** ** -0.33** -0.23** RT * * -0.20* -0.23** DR * ** 0.32** ** ** -0.31** 0.30** 0.22* 0.22** AD 0.21* ** 0.24** ** -0.32** -0.20* * 0.08 WI ** -0.21* * 0.37** 0.42** 0.48** * DA ** * 0.33** 0.25** 0.18* DO ** ** 0.36** 0.25** 0.16* RE * ** * 0.40** 0.39** AV * ** 0.27** -0.20* * ** YL * * 0.24** 0.22** ** 0.34** 0.27** 0.27** HE ** * * 0.24** 0.24** SC 0.22* ** *p <.05 **p <.01

27 Predictably, the Active-Constructive responses (PT, CS, EE, RO) displayed their most consistent associations (which are relatively modest) with the dimensions of collaboration and accommodation both of which reflect a concern for meeting the needs and interests of the other party. There were also some negative associations between Active-Constructive responses and the conflict style emphasizing the defeat of the other. The Passive-Constructive responses (RT, DR, and AD) displayed some of their strongest associations (which are negative in sign) with the defeat conflict style; those who act in Passive- Constructive ways were less likely to try and prevail over the other person. In addition, the Passive- Constructive behaviors were associated with a greater tendency to accommodate the other person, and also to withdraw from conflict. Thus, one way to think about these Passive-Constructive responses is that while they can be associated with helpful responding (inhibit defeat ), they may also be associated with a tendency to avoid engagement with the other person. As expected, the Active-Destructive responses (WI, DA, DO, and RE) displayed their strongest association with the defeat style; those who possess that style are more likely to display anger, demean the other, and so forth. Active-Destructive responses were more sporadically associated with lower accommodation and compromise scores. Interestingly, these Active-Destructive responses were almost completely unrelated to collaboration and withdrawal. One implication of this is that people can possess an unpleasant set of conflict behaviors (angry, demeaning) yet still have a desire to collaborate, or to withdraw from conflict. This may reinforce the idea that Active-Destructive responses are simply a set of behaviors that can be changed; they do not necessarily imply that the person is opposed to a collaborative approach to conflict resolution. In contrast to the Active-Destructive scales, the Passive-Destructive scales displayed relatively little association with the defeat style, and stronger associations with the withdrawal style. Both the Yielding and Hiding Emotions scales displayed consistent positive associations with the withdrawal style, and to some degree with accommodation; the Avoiding scale displayed similar but weaker associations. What all of these associations suggest is that Passive-Destructive responses especially Yielding and Hiding Emotions are associated with a lower willingness to engage the other person. In contrast, the Self-Criticizing scale is, for the most part, not associated with any of the five conflict styles; this scale seems to tap something largely unrelated to the dimensions captured by the TKI and NSP. Relations with Broad Personality Measures An additional kind of validity comes from comparing the CDP to a broad measure of personality to see how strongly related particular responses to conflict are with general personality domains. This was tested in two ways. First, a sample of 266 undergraduate students completed the 15 Responses to Conflict scales from the CDP and a widely-used measure of personality, the NEO-PI (Costa & McCrae, 1985) which taps 5 broad dimensions of personality. Table 12 displays these associations. 26

28 Table 12. Correlations between CDP Responses to Conflict scales and the NEO-PI scales (N = 266 undergraduates). Extraversion Agreeableness Openness Neuroticism Conscientiousness PT 0.24** 0.28** 0.27** CS 0.43** 0.25** 0.25** -0.21** 0.12 EE 0.42** RO 0.26** 0.40** * RT 0.19* 0.29** 0.23** ** DR -0.19* 0.41** 0.22** * AD 0.26** 0.37** 0.24** -0.18* 0.09 WI ** DA ** -0.21** DO ** -0.26** RE ** -0.28** AV 0.39** ** YL ** ** HE 0.39** 0.28** ** 0.06 SC ** ** 0.08 *p <.05 **p <.01 27