ABSTRACT AN EXPLORATORY STUDY OF CONFLICT MANAGEMENT AND CONFLICT RESOLUTION STRATEGIES IN PROBLEM SOLVING GROUPS By Barbara Ann Walker This is an experimental field study comparing the effects of con- flict management and conflict resolution strategies in problem solving groups at three points in time. This study finds that conflict resolu- tion groups were significantly more satisfying to members. Over time, groups resolved rather than managed conflicts which occurred. AN EXPLORATORY STUDY OF CONFLICT MANAGEMENT AND CONFLICT RESOLUTION STRATEGIES IN PROBLEM SOLVING GROUPS By Barbara Ann Walker A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Communication 1976 Accepted by the faculty of the Department of Communication, College of Communication Arts, Michigan State University, in partial fulfillment of the requirements for the Doctor of Philosophy degree. C‘ Vs JDJ, Uzi/91A g1 “(In—UL Director of Dissertation Guidance Committee: , Chairman To My Parents 11 [Ii-II 1' ACKNOWLEDGMENTS I am grateful to everyone who has helped me in this process. Professor Erwin Bettinghaus has been an excellent advisor and critic. Professors James Phillips, David Ralph and Joseph Woelfel offered inval- uable advice. Mark Miller and Kook-Ching Huber were exceptionally help- ful and generous in their statistical assistance. Many thanks to Tom Campion, Gary Wilson and Roger Spooner for their help with procedures and equipment. A special thanks goes to my parents for their help and encouragement, which got me this far. TABLE OF CONTENTS CHAPTER ONE: RATIONALE AND HYPOTHESES ........................... —l—J—l—l—l o o o o o m-wa—a Introduction ............................................ Purpose ................................................. Overview of Conflict Resolution Perspective ............. Overview of Conflict Management Perspective ............. Effects of Conflict Management Compared with Conflict Resolution Strategies ................................... l.51 Satisfaction ...................................... l.52 Survival .......................................... l.53 Tension Release ................................... l.54 Problem Solving ................................... CHAPTER TWO: METHODS ............................................ 2.] 2.2 2.3 2.4 2.5 2.6 Experimental Design and Procedures ...................... Pretest ................................................. Experimental Manipulations and Independent Variables.... 2.3l Conflict Rules .................................... 2.32 Time .............................................. Subjects and Group Composition .......................... 2.4l Subjects .......................................... 2.42 Group Composition ................................. Dependent Variables ..................................... 2.51 Description of Variables Measured from Taped Interactions ................................ 2.52 Operationalization of Interaction Variables ....... 2.53 Dependent Measures ................................ 2.54 Dependent Measures of Satisfaction ................ Coder and Observer Training ............................. 2.6l Coder Training .................................... 2.62 Observers and Observer Training ................... iv page n—l-u-l 5 7 ll l2 l5 l5 16 18 18 19 20 20 20 21 21 22 23 24 24 24 26 TABLE OF CONTENTS (cont'd) page 2.7 Experimental Procedures ................................. 26 2.71 Experimental Preparations ......................... 26 2.72 Day One Procedures ................................ 27 2.73 Day Two Procedures ................................ 28 2.74 Day Three Procedures .............................. 28 2.75 Missing Data ...................................... 30 2.8 Ethical Considerations .................................. 30 2.9 Summary ................................................. 31 CHAPTER THREE: RESULTS .......................................... 32 3.1 Overview ................................................ 32 3.2 Inter-coder Reliabilities ............................... 32 3.3 Length of Time of the Interactions ...................... 32 3.31 Time One .......................................... 32 3.32 Time Two .......................................... 34 3.33 Time Three ........................................ 34 3.34 Mean Length of Exercises .......................... 34 3.4 Manipulation Check ...................................... 35 3.5 Self Report Dependent Variables ......................... 38 3.51 Survival .......................................... 38 3.52 Satisfaction Dependent Variables .................. 40 3.53 Accuracy .......................................... 41 3.6 Tension Release ......................................... 42 3.61 Negative Emotions ................................. 42 3.62 Humor ............................................. 43 3.63 Laughter .......................................... 43 3.64 Exploratory Analysis of Tension Release ........... 43 3.7 Conclusions ............................................. 47 CHAPTER FOUR: DISCUSSION AND IMPLICATIONS ....................... 48 4.1 Overview ................................................ 48 4.2 The Experimental Situation and the Manipulation of Conflict ................................ 48 4.3 Survival ................................................ 52 4.4 Accuracy ................................................ 53 TABLE OF CONTENTS (cont'd) page 4.5 Satisfaction ............................................. 54 4.6 Tension Release .......................................... 56 4.7 Research Extension ....................................... 57 BIBLIOGRAPHY ...................................................... 58 APPENDICES Appendix A Group Size Over Time ............................. 60 Appendix B The Search for Leadership Traits ................. 61 Appendix C ................................................. 64 Appendix D ................................................. 67 Appendix E Independent Variables with Negative Emotions ..... 68 Appendix F Independent Variables with Humor ................. 69 Appendix G Independent Variables with Laughter .............. 70 vi LIST OF TABLES Table Missing Data for Groups .................................... Intercoder Reliabilities ................................... Analysis of Variance, Length Day 1 by Conflict ............. Analysis of Variance, Length Day 2 by Conflict ............. Analysis of Variance, Length Day 3 by Conflict ............. Mean Length of Interactions ................................ Analysis of Variance, Time 1 Disagreement by Conflict ...... Analysis of Variance, Time 2 Disagreement by Conflict ...... Analysis of Variance, Time 3 Disagreement by Conflict ...... Condition Means - Number of Disagreements .................. Condition Means - Disagreements Per Minute ................. Repeated Measures Analysis of Variance, Rate of Disagreement, Conflict Condition by Experimental Time ...... Analysis of Variance, Survival 1 by Conflict ............... Analysis of Variance, Survival 2 by Conflict ............... Analysis of Variance, Satisfaction with Solution by Conflict ................................................ Analysis of Variance, Satisfaction with Interaction by Conflict ................................................ 17 Analysis of Variance, Satisfaction with Influence by Conflict ................................................ 18 Analysis of Variance, Accuracy by Conflict ................. 19 Analysis of Variance, Negative Emotion by Conflict - Day 1. 20 Analysis of Variance, Negative Emotion by Conflict - Day 2. 21 Means - Negative Emotions ................................. 22 Analysis of Variance, Humor by Conflict - Day 1 ............ 23 Analysis of Variance, Humor by Conflict - Day 2 ............ 24 Analysis of Variance, Laughter Day 1 by Conflict ........... 25 Analysis of Variance, Laughter Day 2 by Conflict ........... OmNO‘U'l-DwN-fl —l—J—l N-‘O .—a—a._l 01-h“ -—l 0‘ vii LIST OF FIGURES fiflflié page 1 Number of Disagreements by Condition ....................... 49 2 Mean Length of Interactions Across All Groups .............. 50 viii CHAPTER ONE RATIONALE AND HYPOTHESES 1.1 Introduction This paper studies conflict in small groups. Two contrasting views of small group processes are considered. Both points of view deal with the strategies groups use to deal with conflict. One perspective empha- sizes the resolution of conflict. An alternative perspective holds that conflict ought to be managed and maintained for the benefit of the group and individual group members. 1.2 Purpose This is an exploratory study of the effects of conflict management and conflict resolution strategies on small group processes. Specific- ally, the purposes of this study are to: (1) Compare the effect of conflict management and conflict resolution strategies on individual member satisfaction in small groups. (2) Analyze the communication patterns in groups which either manage or resolve conflicts, focusing on tension release communication. (3) Compare groups which utilize conflict management and conflict reso- lution strategies in terms of the accuracy of solutions reached. (4) Compare conflict management and conflict resolution groups in terms of their survival. In this chapter, the conflict management and conflict resolution perspectives will be described. Hypotheses will be developed related to l four variables defined as potential effects of managing or resolving conflict. These variables are group survival, group satisfaction, ten- sion release, and accuracy of group solutions. 1.3 Overview of Conflict Resolution Perspective The conflict resolution perspective is a set of assumptions about the nature of conflict, commonly held by scholars in speech communica- tion. This perspective emphasizes communication strategies which can be used to resolve conflict once it occurs. Hawes and Smith (1973) define this perspective as follows: Normative theories of conflict assume that the only good conflict is a resolved conflict. The rationale is that people are happier, healthier, and more pro- ductive when they are 'cooperating' and 'getting along' with one another. . . . A central thesis throughout these discussions is that the effect of conflict is destructive and the best means for bring- ing it to a speedy end should be sought. (1973: 425-425) According to Simons (1972), Rogers and Roethlisberger popularized this perspective when they presented a paper to speech communication scholars and psychotherapists at the Centennial Conference on Communica- tions at Northwestern University in 1951. Comparing conflict to a breakdown in a mechanical system, they argue that the barrier to effec- tive communication is evaluation of someone else's perspective from our own point of view. Communication breakdowns can be prevented by non- evaluation. The conflict resolution perspective is utilized by scholars who ad- here to a systems perspective. From this point of view, social systems are likened to organisms and conflict to a disease. This analysis is characteristic of the work of Talcott Parsons (1951) who treats conflict as a form of sickness in the social body. Systems theorists in communication typically view conflict as a disruption of homeostasis. According to Hawes and Smith (1973) the con- flict resolution approach assumes that cooperation is the normal state of a system: Equilibrium and stability of systems are thought of as conflict-free states in which all components are coordinated and acting in harmony. Conflict occurs as a temporary disruption of the system. The disrup- tion has a beginning, in all likelihood a cause, and is terminated allowing a return to a state which, although possibly altered by the conflict, will remain stable until the next episode of disruption. (1973: 425) Thus, underlying the conflict resolution perspective is the assump- tion that the effect of conflict is destructive. The quickest means to bring conflict to an end is studied. Frequently analyzed according to systems theory, this perspective views conflict as a temporary disrup- tion of the normal state of c00peration. There is an alternative set of assumptions underlying the study of conflict in small groups. The conflict management approach rejects the idea that conflict is a destructive deviation from normality. This approach holds that conflict should be managed and maintained for the benefit of the parties in conflict. 1.4 OverView of Conflict Management Perspective Kenneth Boulding, editor of the Journal of Conflict Resolution, states that the title of his journal may not be appropriate for the na- ture of the issues covered in his journal: Perhaps 'management' would have been better, for the distinction between constructive and destructive conflicts is not necessarily the distinction between those which are resolved and those which are not. Conflicts are sometimes resolved in ways which are highly undesirable for one party if not for both. Sometimes there is a need for protracting conflict and for keeping it unresolved, perhaps by diminish— ing its intensity and increasing its duration. (1968: 410) Hawes and Smith (1973) analyze the assumptions underlying the study of conflict in the field of speech communication. They find that the conflict management perspective has received little attention. They suggest that the most fruitful way of analyzing conflict assumes that "conflict is a perpetual condition to be managed and maintained" (1973: 425). The conflict management perspective has been discussed by scholars in other fields. Sociologist Georg Simmel analyzes conflict as a posi- tive, constructive force. Simmel (1955) argues that if we look at con- flict in isolation, it may appear to be destructive. However, if we look at the total effect of conflict in a system, it may be positive force: If every interaction among men is association, conflict - after all one of the most vivid interac- tions, which, furthermore, cannot possibly be carried on by one individual alone ~ must certainly be considered as sociation. And in fact, gjssocia- tion factors - hate, envy, need, desire - are the causes of conflict; it breaks out because of them. Conflict is thus designed to resolve divergent dualisms; it is a way of achieving some kind of unity, even if it be through the annihilation of one of the conflicting parties. (1955: 13) In summary, the conflict management perspective looks at the poten- tially constructive effects of conflict. If viewed from this perspec- tive, conflict contributes to group unification, and is a continuous process to be managed and maintained. Four major effects of managing conflict in small groups have been theoretically isolated, and research hypotheses can be developed related to each of these four major effects, which are group member satisfaction, survival of the group, tension releasing and group problem solving abil— ity. 1.5 Effects of Conflict Management Compared With Conflict Resolution Strategies Several authors discuss potentially beneficial effects of conflict management strategies in small groups. Although their attention is fo- cused on conflict management, theorists contrast the effects of conflict management with the effects of conflict resolution. Now, group member satisfaction will be described as a dependent variable effected by the management or resolution of conflict. 1.51 Satisfaction. According to Simmel (1955), in—group conflict has positive functions for the individual as well as for relationships. By expressing opposition, an individual is making a potentially unbear- able situation at least tolerable. Simmel argues that expressing con- flict gives the individual psychological relief and satisfaction. Simmel asserts that just as a fighter must "pull himself together," individuals in conflict must similarly focus their energies. The inner changes which occur when an individual is in opposition to another indi- vidual strengthen each party. Daily experience shows us how easily a quarrel be- tween two individuals changes each of them not only in his relationship to the other but also in himself. There are first of all the distorting and purifying, weakening or strengthening consequences of the conflict for the individual. In addition, there are the conditions of it, the inner changes and adaptations which it breeds because of their usefulness in carrying it out. (1955: 88) Simmel (1955) argues that the changes which an individual goes through in order to engage in conflict are positive and satisfying. The individual must take on a concentration of his or her energies when engaged in conflict. In a conflict-free state, however, these energies can be unfocused. Thus, one of the satisfying consequences of conflict for the individual is in the concentration of his or her efforts. Weick (1969) explains that group members need to express opposing points of view in order to be satisfied with a group product. Conflict results from the expression of what he calls "self-centered" responses. Weick argues that the expression of self-centered responses, or individ- uated action, provides satisfaction with group interaction. In addi- tion, it produces a greater degree of satisfaction with the group prod— uct, since members feel they have an input in the group. Torrance (1957) reports the results of a number of field studies of small groups under conditions of survival. Through intensive inter- views, he determines the extent to which conflict was expressed in these groups and analyzes the effects of expressing conflict on group pro- cesses. One field study he reports deals with the effects of expressing conflict on group member satisfaction. He interviews groups of equip- ment technicians who were caught in a blizzard in the high Sierras. The groups adapted poorly to adverse conditions, as a majority of the group members became severely frostbitten as a result of continuing their ex- cursion after marching through an unfrozen stream. Each member said that he wanted to stop and dry his feet before continuing, but was afraid of expressing dissent from the group. The instructors claimed that they wanted to have the trainees stop and do the same, but thought the trainees were too apathetic to take care of themselves. Thus, the instructors interpreted fear of disagreement as apathy. Torrance concludes that none of the group members were satisfied with the solution reached by the group to continue marching. He attri- butes their lack of satisfaction to their unwillingness to express dis- agreement. The work of Simmel (1955), Weick (1969) and Torrance (1957) lead to the following hypothesis: H]: In groups where conflict management techniques are utilized, there is greater individual member satis- faction than in groups where conflict resolution techniques are utilized. 1.52 Survival. A number of authors have looked at the effects of expressing conflict on small group survival. Each individual operation- ally defines the variable, but the conceptual definition of group survi- val is not discussed. A definition of group survival can be developed by looking at definitions of a group. Sociological definitions of group emphasize structural character- istics and role relationships. Sherif and Sherif (1956) provide the following definition: A group is a social unit which consists of a number of individuals who stand in (more or less) definite status and role relationships to one another and which possesses a set of values or norms of its own regulat- ing behavior of individual members, at least in matters of consequence to the group. (1956: 144) McDavid and Harari (1968) state a similar definition: A social-psychological group is an organized system of two or more individuals who are interrelated so that the system performs some function, has a standard set of role relationships among its members, and has a set of norms that regulate the function of the group and each of its members. (1968: 237) Using this definition, group survival can be defined as the exis- tence over time of a collective of individuals with a common structure and function. Group members can be added or dropped, and the gpggp_ would survive. Individuals' affective relationships to other group members can change without affecting survival of the group. A psychological definition of group leads to a different concept of group survival. This type of definition stresses the fulfillment of in- dividual member needs, rather than structure and function. Bass (1960) provides a psychological definition of group: We define group as a collection of individuals whose existence as a collective is rewarding to the indi- viduals. (1960: 39) Cattell (1951) also provides a psychological definition of group: A group is a collection of organisms in which the existence of all (in their given relationships) is necessary to the satisfaction of certain individual needs in each. (1951: 167) Using a psychological definition of group, group survival is the fulfillment of the needs of a collective of individuals over time. According to this definition, groups which fail to meet the needs of individual members fail to survive. A group could survive in the socio- logical sense by remaining intact as a structural unit, but fail to sur— vive in the psychological sense because member needs were not fulfilled. All of the authors surveyed who discuss the effects of managing conflict on group survival consider survival in the sociological sense. Sociologists Simmel (1955) and Coser (1956) equate survival with a group or dyad remaining as a social-structural unit. Campbell (1965), Weick (1969) and Torrance (1957) look at survival according to the group's ability to adapt to changes in the environment. These authors take a sociological approach, implicitly defining survival as the ability of the group to persist as a structural unit. Now, the relationship be— tween conflict management and survival will be discussed. Simmel describes the results of ppt_expressing conflict as leading to the destruction of relationships. The feeling of oppression increases in a relationship if conflict is not expressed. Finally, silent opposi- tion causes such relationships to split apart. Coser (1956) argues that the indirect expression of conflict has a negative effect on survival of the group. If group members choose to act out conflict toward alternative objects, the direct expression of conflict is inhibited. This leads to rigidity in a group structure, which would otherwise be modified by the direct expression of conflict. The expression of conflict in indirect ways may function like lightening rods, to clear the air, but it cannot prevent conflict from recurring. Campbell (1965) discusses the survival value of altruism and of self-centered behavior. He argues that there are natural selection pro- cesses operating which cause one individual to survive at the expense of the other individual. Also, there are selection processes operating that cause one gppgp_to survive at the expense of another group. Campbell holds that there is survival value for the individual to behave cooperatively in order for his or her group to survive. There is also survival value for the individual to behave in a self-centered way for ego-gratification. Since individual member satisfaction is ul- timately beneficial to the group, it is desirable for the survival of the group to maintain selfish, as well as altruistic, responses. Weick (1969) relying on Campbell's analysis, argues that in a small group, whenever compromise responses are emitted, the survival of pppp_ altruistic and self-centered motives is destroyed. Although compromise responses appear to be satisfying both individual and group interests, Weick maintains that these responses satisfy neither interest. 10 According to Weick, compromise responses fail to satisfy group in- terests because they level out responses from individual members which may have great adaptive value for the group. The expression of conflict may trigger the use of conflict resolution techniques in a group, which can be destructive for the group when conditions change: No one questions that if left on its own, the group could destroy itself with conflict and ambivalence. Our point is that the presence of conflict does not necessarily indicate that a group is dissolving; it merely signifies that the group retains heterogeneous responses and preferences, all of which may be adap— tive under some circumstances. (1969: 104) By interviewing United States Air Force jet pilots in combat in Korea, Torrance (1957) compares pilots who have exceptional fighting records with average pilots. Compared to other pilots, "aces" were more unwilling to accept "no" as an answer. They were also more willing to oppose accepted procedures than less successful pilots. Torrance con- cludes: If willingness to disagree is related to individual ability to adapt, it is only reasonable to expect that group processes are affected accordingly. Willingness to disagree has meant the difference be- tween survival and failure to survive in group situa- tions. (1957: 314) Torrance (1955) describes an unpublished study by Howard which sup- ports his conclusions about the effect of disagreement on group pro- cesses. Howard analyzes survival experiences of groups which were stranded in the Southwest Pacific in World War II. He finds that group prejudices against eating strange foods cause groups of individuals to starve to death, rather than overcoming these prejudices. Groups which survived under these adverse conditions contained individual members who dissented against the group opinion. 11 The work of Simmel (1955), Coser (1956), Campbell (1965), Weick (1969) and Torrance (1957) leads to the following prediction: H2: Groups which utilize conflict management techniques will be more likely to survive than groups which utilize conflict resolution techniques. 1.53 Tension Release. Theory relating to the tension—relieving characteristics of conflict is characteristic of Lewis Coser (1956). He criticizes the work of Georg Simmel for over—simplification of the tension-relieving characteristics of conflict. Coser characterizes Simmel's perspective as a "safety—valve" theory of conflict. That is, conflict releases hostile feelings that would otherwise build up and destroy relationships. Coser's criticism of the "safety-valve“ characterization of con- flict is based on Freudian analysis, which became popular subsequent to the original publication of Simmel's work. Coser argues that the hos- tilities and tensions generated in a conflict situation can be acted out in other ways besides the overt expression of conflict. Speaking about the feelings of hostility generated by conflict, Coser (1956) states: The relevant expression of these feelings in behavior are of at least three possible kinds: (1) direct ex- pression of hostility against the person or group which is the source of frustration, (2) displacement . of such hostile behavior onto substitute objects, and j (3) tension-release activity which provides satisfac- tion in itself without need for object or object sub- stitute. (1956: 41) The discussion of Coser leads to the following prediction about tension release in small groups. Groups in which conflict is resolved rather than managed would not be able to release tension through direct expression of conflict. Therefore: AI ' I I “ II-(ll' [[1 Elli-l l lint“ Ci.." I ll ll Ill I'll All II. II | f- 12 H o 3. Groups in which conflict is resolved display more tension release behavior than groups in which conflict is managed. 1.54 Problem Solving. Irving Janis (1972) studies the relation- ship between the suppression of conflict and the problem solving ability of policy making groups. From historical records, recollections of par- ticipants and descriptions of observers, Janis reconstructs the condi- tions under which a number of major national policy decisions were made. He finds that groups which made policy decisions based on incomplete and distorted information, or on false premises, show the following in- teraction patterns: Direct pressure on any member who expresses strong arguments against any of the group's stereotypes, illusions, or commitments, making clear this type of dissent is contrary to what is expected of all loyal members; Self-censorship of deviations from the apparent group consensus, reflecting each member's inclination to minimize to himself the importance of his doubts and counterarguments; A shared illusion of unanimity concerning judgments conforming to the majority view (partly resulting from self-censorship of deviations, augmented by the false assumption that silence means consent); The emergence of self-appointed mindguards-members who protect the group from adverse information that might shatter their shared complacency about the effec- tiveness and morality of their decisions. (1972: 198) These symptoms are among eight characteristics of groups which em- ploy what Janis labels "groupthink." An example of a "groupthink" de- cision was made by John Kennedy and his staff of advisors, prior to the Bay of Pigs invasion. Individually, each man was brilliant and capable of shrewd analysis. Collectively, they did not detect serious errors in the invasion plan: The President and his key advisors approved the Bay of Pigs invasion plan on the basis of six assumptions, each of which was wrong. In retrospect, the President's advisors could see that even when they first began to discuss the plan, sufficient information was available 13 to indicate that their assumptions were much too shakey. They could have obtained and used the crucial informa- tion beforehand to correct their false assumptions if at the group meetings they had been more critical and probing in fulfilling their advisory roles. (1972: 19) Weick (1969) argues in favor of retaining heterogeneous responses in a problem solving group. A group will be more flexible in dealing with environmental changes if competing reSponses are maintained. Re- sponses which were appropriate at one time may not be appropriate when circumstances change. Groups which preserve a larger repertoire of re- sponses will have more resources available in solving problems, espe- cially problems which cannot be foreseen in advance. Weick believes that if conflict is resolved, the solution should not constitute a compromise. That is, such a solution must not destroy the original, adaptive responses present in the conflict. The only conflict resolution strategy acceptable to Weick is one which allows both points of view of uncompromised expression. Bell (1974) argues that an emphasis on the goal of consensus, and the corresponding assumption that conflict should be resolved, destroys the group's primary reason for being together. One of the basic rea- sons for having groups, rather than individuals, solve problems is that inaccuracy of perception and judgment is modified in a group. This takes place when group members critically compare one idea against an- other. When conflict is resolved because of an emphasis on the goal of consensus, inaccurate solutions are likely to result. Based on the discussions of Weick (1969) and Bell (1974) and the observational research of Janis (1972), we would expect that groups which utilize conflict resolution techniques would produce more accurate solutions to problems because they are drawing from a wider range of {I'll‘l‘lll‘aolllll‘l llllvl I Illlllla' .Illllll. 14 possible responses. This leads to the following prediction about the solutions of prob- lem solving groups: H4: Groups which utilize conflict management techniques produce more accurate solutions to problems than groups which utilize conflict resolution techniques. We have described four theoretic hypotheses which were investigated in an experimental field study. In the next chapter, we will describe a study which was conducted to test these hypotheses. CHAPTER TWO METHODS 2.1 Experimental Design and Procedures This is an experimental study of small group behavior at three points in time. Students in Communication 210, Leadership were randomly assigned to twenty groups, which worked on problem solving tasks related to topics covered in class. Individuals received instructions on how to deal with conflict which occurred in their groups. Individuals in half of the groups received instructions to manage and maintain any con- flicts which occurred, and subjects in the other half of the groups were told to quickly resolve any conflicts which occurred. Students remained in their experimental groups for three sessions, held one week apart. After the third group meeting, students were asked to choose people whom they wanted to work with on their group pro- jects. They were also asked how much they wanted to work with each group member on their project. The responses were used as a measure of survival of the groups. Group interactions were tape recorded. Sets of coders were trained to measure dependent variables from the tapes. The variables coded were humor, laughter, negative emotions, disagreement, statements of agree- ment, and use of reasoning. Also, singular and plural pronouns were measured and were used to create an index of group cohesiveness. The following discussion describes the experiment in detail. First, 15 16 the pretest will be described. Then, the independent variables will be discussed. Next, subject and group composition will be described. After discussing the subjects and group composition, we will dis- cuss coder and observer training. Then, the dependent variables will be described. Next, the experimental procedures will be discussed. Finally, ethical considerations will be described. 2.2 Pretest A pretest was conducted as a trial run for the experimental proce— dures, and to test the effectiveness of the experimental manipulations. Students in Communication 205, Persuasion, received extra credit for participation in the pretest. Before the pretest experiment, six Communication 100 students were trained to code the sequence of inter- action in the groups. Before the students were assigned to experimental groups, they were given a list of the following ten health professions to rank from one to ten according to their prestige: plastic surgeon, neurosurgeon, pedia- trician, dentist, orth0pedic surgeon, registered nurse, opthamologist, dermatologist, psychiatrist, and cardiologist. After completing the rankings, the students were placed into six three or four person groups. Three of these groups were then physically separated from the other groups. Three of the groups were given instruc- tions labelled "Conflict Management,“ which contained instructions to manage conflicts which occurred. The other half of the groups received instructions labelled "Conflict Resolution," instructing them to quickly resolve all conflict which occurred. Each set of instructions contained a description of theory supporting either conflict resolution or con— flict management assumptions. r1 1. 17 The groups were told that it was their task to come up with one group ranking. Each group member had a card with "CONFLICT" printed on it in large letters. Each time a conflict occurred, group members were to hold up this conflict card. In the conflict resolution groups, mem« bers were told that they were to have their cards gpgg most of the time. In the conflict management groups, students were told that they should have their cards gp_most of the time. This was to reinforce the set of procedural rules. An observer went to each of the groups, started the tape recorder which was in the center of the groups, and asked each member to state his or her name. Every five minutes group members were reminded of the procedural rules they were to use to deal with conflict. After completing the task, observers and participants discussed the experimental procedures with the author. Most students agreed that the theoretical introduction was unnecessary. Several observers stated that the conflict cards were distracting, and students agreed. Students said that being interrupted every five minutes was also quite distracting. From listening to tapes of the interactions, it was confirmed that both the cards and the interruptions distracted groups from their task. A majority of the students indicated that there was at least one profession they did not know. During the exercise, a number of students asked the instructor what "opthamologist" meant. After this discussion, a number of changes were made from the pree test situation to the experimental situation: 1. Conflict cards were not used. 2. The groups were not reminded of the rules. 3. The groups were not given an extensive theoretical 18 explanation of conflict management or conflict resolution. 4. A different set of instructions was used for the ranking task. The author listened to the tapes to determine if the manipulation was effective. In the conflict management condition, there were an average of sixteen disagreements expressed per group. In the conflict resolution conditions, there was an average of one disagreement expressed per group. 2.3 Experimental Manipulations and Independent Variables One variable was manipulated in this experiment - the procedural rules the groups used in their problem solving tasks. One set of rules was labelled conflict resolution and the other set was labelled conflict management, creating two levels of this variable. Time was a measured independent variable, at three different points. This created a two-by- three design. 2.3l Conflict Rules. One set of procedural rules was designed to bring about the resolution of conflict in the groups. Based on the pre- test manipulation, the following set of rules, labelled "Conflict Reso- lution," was used: In any group, people have different values, attitudes, and beliefs that relate to the on-going activities of the group. When you consider the different back- grounds and experiences of each person in a group, the emergence of conflict is not surprising. Since conflict can occur frequently, most groups develop ways of resolving it. One method of resolving con- flict is through the specification of communication rules. Below are some communication rules to follow in your small group. If you carefully follow these rules, conflict should be effectively resolved. 1. YOUR GROUP SHOULD FUNCTION TO AVOID CONFLICTS. IF THEY OCCUR, TRY TO QUICKLY RESOLVE DISAGREEMENTS. Il‘l‘||lll IIIU‘I“-IIITIII1 ..... 19 2. WHEN YOU AGREE WITH SOMEONE, OPENLY EXPRESS THAT DISAGREEMENT. 3. DON'T CRITICIZE THE IDEAS OF OTHER PEOPLE IN YOUR GROUP. 4. IT IS NOT NECESSARY THAT YOUR OWN POINT OF VIEW IS EXPRESSED. Another set of rules was designed to bring about the management of conflict, labelled "Conflict Management." In any group, people have different values, atti- tudes, and beliefs that relate to the onegoing activities of the group. When you consider the different backgrounds and experiences of each per- son in a group, the emergence of conflict is not surprising. Since conflict can occur frequently, most groups develop ways of regulating it. One method of regulating conflict is through the specification of communication rules. Below are some communication rules to use in your small group. If you carefully follow these rules, conflict should be effectively maintained. 1. YOUR GROUP SHOULD FUNCTION TO ACCEPT AND MAIN. TAIN CONFLICT. WHEN CONFLICT AND DISAGREEMENT OCCUR, TRY TO MAINTAIN OPPOSING POINTS OF VIEW. 2. WHEN YOU DISAGREE WITH SOMEONE, FREELY EXPRESS YOUR OWN POINT OF VIEW. 3. FREELY CRITICIZE THE IDEAS OF OTHER PEOPLE IN YOUR GROUP. 4. MAKE SURE YOUR OWN POINT OF VIEW IS HEARD. 2.32 limp, Time was a second independent variable. The experi- mental sessions took place at three different points in time. These were spaced one week apart, as class met once per week. Groups re- ceived the same procedural rules all three times. With the exception of two groups, which will be discussed later, group composition remained the same at all three points in time. 20 2.4 Subjects and Group Compositipp‘ 2.4l Subjects. Several requirements were necessary for subjects. First, a collective of approximately eighty people was needed. Second- ly, these people needed to meet over time, since time was one of the independent variables. Thirdly, these people needed to be in a situa. tion where they could continue working in a small group, in order to measure survival of the group. Fourth, we needed a situation where working on a problem solving task was natural. All these requirements were met in the author's Communication 210, Leadership, class. This class had a projected enrollment of about eighty peOple, and met once a week for nine weeks. Leadership classes are frequently taught with problem solving exercises. From their ex- perimental groups, students could choose other students to work with for the rest of the term on their group projects, making survival of the small groups a relevant variable. Seventy—seven students participated in the experiment; one fresh- man, seventeen sophomores, thirtyuone juniors and twenty-eight seniors. Thirty-two students were female and forty-five students were male. There were twenty different academic majors represented by students in this class. 2.42 Group_Composition. Since we were measuring survival of the groups, it was important that group membership remained constant across time. Therefore, several incentives were given to the students to attend all three sessions. They were offered an automatic 4.0 (A) as a participation grade in class, constituting ten percent of their final grade. Also, they were urged to attend to understand the class material, since class met once a week. 21 Students were randomly assigned to twenty groups of four or five students each. Females were assigned to groups first, since there were fewer females in the class, to ensure an equivalent distribution of fe- male students across groups. Two students who anticipated being absent were not assigned to groups. On day one of the experiment, five persons were absent who had been assigned to groups. On day two of the experiment, two people present on day one dropped the class. Any new people were not assigned to groups. On day three, group membership remained the same as on day two. The stable size of the groups, after day one, was as follows. In the conflict resolution condition, there were three three~person groups, four four-person groups and three five-person groups. In the conflict management condition, there were two groups which had one member less than the conflict resolution groups. See Appendix A, a summary of the size of the experimental groups over time. 2.5 Dependent Variables A number of dependent variables were measured, both from the taped interactions and from a questionnaire given to each participant. In this section, we will describe the variables measured from the tapes of group interactions. Then, we will describe the variables measured from a questionnaire. 2.51 Description of Variables Measured from Taped Interactiogs. The number of times overt disagreement occurred was measured as a mani- pulation check. The number of times statements were made expressing agreement or positive reinforcement was also measured. 22 Three variables were measured which were indicators of tension re- lease. The number of times laughter occurred was measured. The number of humorous statements was measured. Another measure of tension release was the number of statements which expressed negative emotion. The number of statements using reasoning or evidence was measured. In addition, two variables were measured to create an interactional mea- sure of cohesion. These were the number of times first person singular pronouns were used and the number of times first person plural pronouns were used. The more plural, as Opposed to singular, pronouns which are used, the more cohesive the group (see Pacanowsky, 1975). 2.52 Operationalization of Interaction Variables. The eight inter- action variables were operationalized as follows: (1) Humorous statements - any statement that was intended to be funny by the source was coded as humorous. Humor could be directed at self, at others in the group, at the situation, or at the class and the instructors. (2) Laughter - laughter was divided into two categories, individual and group laughter. If only one person laughed, this was considered individual laughter. If more than one person laughed simultaneously, this was considered group laughter. (3) Negative emotions - any negative feelings that were expressed through tone of voice and were indirect expressions of hostility were coded as negative emotions. If a negative vocal tone accompanied an overt disagreement, coders were told not to code it. (4) Disagreement — any overt, direct expression that one person did not agree with another person was coded. This had to be directly and intentionally related to another person's comment. 23 (5) Agreement - any statement directly expressing that one person was in agreement with another was coded. (6) Reasoning — any statement indicating reasons for a person's point of view was coded. Included were statements of evidence or fact. (7) Singular pronouns - any first person singular pronoun or con- traction was coded. (8) Plural pronouns — any first person plural pronoun or contrac- tion was coded. 2.53 Dependent Measures of Survival, After the third group meet- ing, a measure was taken to determine whether the small groups would survive for the rest of the term. In the time four questionnaire, each group member was asked whom he or she wanted to work with on the group project. They were also asked how much they wanted to continue working with each person in their group. See Appendix B for the questionnaire used. The group project was a separate assignment from the small group interactions. Students were told in the beginning of the class that they would interact the first three weeks in problem solving groups, and then choose people to work with on a class project. The first three in- teractions, they were told, would give them experience working in groups before they started the class project. The class project was for the students, in groups with other stu- dents of their choice, to study leadership patterns in another group, such as a fraternity, social club, or task—oriented group. This project was to terminate in a class presentation, counting twenty percent of the student's grade. 24 2.54 Dependent Measures of Satisfaction. In addition to questions related to survival, students were asked the following questions related to their satisfaction with the group. If 100 units is the average amount of satisfaction and zero is the complete absence of satisfaction, how satisfied are you with the solutions reached by your group? If 100 units is an average amount of satisfaction and zero units is the complete absence of satisfac- tion, how satisfied were you with the interaction in your group? If 100 units is an average amount of satisfaction and zero units is the complete absence of satis- faction, how satisfied were you with the amount of influence you had in your group? 2.6 Coder and Observer Trainipg: 2.61 Coder Training. Forty—eight coders were chosen from Communi- cation 350R, as an optional class assignment. Coders worked in pairs. Each pair coded one of the eight variables at one point in time. Pairs of coders were trained individually, and received the same introduction to the research project. They were told the general nature of the research, but were not told the hypotheses of the study. Coders were instructed to code the frequency with which each of the eight in- teraction variables occurred in the tape they were listening to. Speci- fic training for each variable took place as follows: (1) Humorous statements — a pretest tape was played, and coders marked down the number of humorous comments which were made. After com- paring their answers, the author described various forms of humor which they might hear, such as banter, irony, teasing, joking, overstatement, understatement, sarcasm, etc. (2) Laughter - coders listened to a small segment of the pretest 25 tape at a time, marking down the number of times both group and individ- ual laughter occurred. They compared the number of times they coded laughter in each segment. (3) Negative emotions - several training sessions were used for this variable. After the nature of the variable was explained, a pretest tape was played. Each time a coder heard a negative emotion being ex- pressed he or she signalled and the tape was stopped. Coders listened to that segment of the tape again, and discussed the nature of the nega- tive emotion. This procedure continued until the coders could easily discriminate the variable. Coders made a brief list of negative emotions which might occur, which they read before each coding session to refresh their memories. Such emotions included anger, resentment, frustration, bitterness, disgust, and other emotions expressing hostility. (4) Disagreement - the same procedure, stopping the tape each time the variable occurred, was used. (5) Agreement — coders listened to a pretest tape and coded the num- ber of statements of agreement which they heard. They were specifically instructed not to code statements such as "Yes" or "0.K." as agreements if they were used simply as transition statements. (6) Reasoning - coders listened to a pretest tape and were told to listen for statements starting with because as expressions of reasoning. Coders were told to code reasoning from the source's perspective. That is, if an individual was attempting to use reasoning, this was to be coded as a statement of reason. The coders were told not to evaluate the responses according to what they felt sppp1g_be reasoning. (7) Singular pronouns - listening to a pretest tape, students coded every time one of the following words was used: I, me, my, mine, I'll, l__' «I I III“! I III ...I l '11 26 I've, I'd, I'm. (8) Plural pronouns - using the same procedure as with singular pro- nouns, coders listened for: we, our, ours, we've, we'll, we'd. 2.62 Observers and Observer Training. Observers were used to code the sequence of interactions in the tapes, to identify a statement with a particular person, in case such data was needed for secondary analysis. Observers were Communication 100 students. Each observer was given a coding sheet, which contained columns representing people in the group and rows representing time sequence of interaction. Coders wrote group members' names across the top of the page, and marked down an "X" in the appropriate column every time a person spoke. When the following person spoke, an "X" was marked in the next row in that person's column, and so forth. 2.7 Experimental Proceduresp 2.7l Experimental Preparations. On the first day of class, and on their syllabus, students were told that they would be participating in a series of three exercises, which would give them knowledge of group pro- cesses before they worked on their group project. They were also told they could choose people they knew from these groups to work with on their class project at the end of the three weeks. Cassette tape recorders were used to record the interactions of group members. Recorders were numbered from one to twenty corresponding to the groups, and tapes were placed in them with the same numbers. Ten tape recorders were set up in an adjacent room while students were in class. These tape recorders were set up in the middle of a group of chairs. Groups were placed as far away as possible. The remaining 27 ten groups were set up in the classroom. While class was still in session, twenty observers arrived in an ad- jacent room. They were given a brief training session on how to code the sequence of interaction. Due to a snow storm on the third day of the experiment, only fourteen observers arrived. To maintain consisten- cy across conditions, they were not assigned to observe groups. 2.72 Day One Procedure. The author handed out group assignments. After students found their groups, the author handed out a sheet en— titled "The Search for Leadership Traits," as in Appendix C. This con- tained a number of conflicting findings about the relationship between leadership and personality traits, and was abstracted from an article entitled "Leadership" by Cecil Gibb in the 1968 Handbook of Social Psy- chology. Groups were told that this was actual data, and that it was their task to derive three conclusions from this data that they felt were the most important conclusions that could be reached. At the end of the class, they were to hand in the conclusions. The author stressed that next week's lecture was going to be based on the conclusions they reached. Groups one through ten went to the adjacent room, and were told to look for a tape recorder with their group number on it. Students in groups one through twenty were directed to different parts of the class- room, and an assistant followed with the appropriate tape recorder. The experimenter handed out a set of procedural rules to each per— son in groups eleven through twenty. These rules instructed the groups to manage conflict, and are described in Section 2.61. She explained that the use of these rules in their groups was very important to help the group run smoothly. She then read the four rules listed on their 28 handout sheets. An assistant followed the same procedure in directing groups one through ten in the adjacent room. He handed out the procedural rules designed to help the groups quickly resolve conflict. He explained the importance of these rules and read them to the groups. Observers then went to each group, turned on the tape recorders, and asked people to introduce themselves. When the groups finished the exercise they handed in a group solution and left class. 2.73 Day Two Procedures. Each class member was given the follow- ing assignment before they broke into their groups. They were assigned to rank fifteen professions according to their prestige. This exercise is “Consensus Seeking: A Group Ranking Task" as found in Pfeiffer and Jones' A Handbook of Structured Experiences in Human Relations Training, Volume II. After they finished their individual rankings, students were told to develop one group ranking. When groups were assembled, the experimenter or an assistant passed out copies of the procedural rules and stressed the importance of follow- ing these rules. The appropriate procedural rules were read to each set of groups. The groups were given a form to record a group decision, and handed in this form when they were finished. See Appendix D for the group task. 2.74 Day Three Procedures. As an introduction to the exercise, the experimenter told students that it was extremely difficult for an instructor to assign individual grades to students for a class project. She described the following problems: It was difficult to determine how much each individual contributed to the group's effort. If the same grade was assigned to everyone in the group, this was unfair to people 29 who carried all the work of people who did nothing. When group members graded each other, they usually gave everyone a high grade, or gave their friends high grades and people they didn't like low grades. Students were told that it was their task to come up with a method to use in grading their class project. For the class project students were to be reassigned to groups which would study leadership patterns in a group of their choice. As a group, they were to develop exact criteria the instructor should use in determining each individual's grade on this project. The best solution was going to be adopted by the instructor. As the experimenter was setting up a tape recorder at group thir- teen, a member of the group told her that they did not need a tape re- corder because they were finished. She responded that they were not fin- ished until the group deliberated the problem according to the procedur- al rules given to them. As she was talking someone wrote down a solu- tion. Group members insisted that this was a group product. When the tape recorder was being set up at group fourteen, a member of the group said that they had finished, too. Several members of the group said that they had discussed the problem and that they already had a solution. When the author returned to these groups while the other groups were discussing the assignment, she found that they were studying to- gether for the midterm exam, which was to be given after everyone com- pleted the exercise. When they were questioned, both groups said that they agreed to hand in a solution quickly so they could study for the test together. They indicated that studying was more important than do- ing the exercise. After all of the groups finished, they were separated from each 30 other so that they could fill out a questionnaire in confidence. The author reminded them that now they could choose people to work with on their class project, which was to start the following week. They were given the questionnaire in Appendix C. 2.75 MissinggData. Tape recordings of some of the group inter- actions were missing. For group fourteen on day one, the background noise was too great to hear voices of group members. The same was true for group thirteen at time two and group seventeen at time three. When the first set of coders was listening to tape nineteen, the tape got caught in the machine and a large portion of it was "chewed." When the tape was fixed, the first eight minutes of day two's exercise were spliced off. This tape was not used in analysis of day two's in- teraction. There were also missing data on day three for groups thirteen and fourteen, who finished their exercise before the recorder was turned on, as described. Thus, the missing data for the twenty groups at the three points in time were as follows: Table 1 Missing Data for Groups Time one Time two Time three Groups: 14 l3 l3 19 14 17 2.8 Ethical Considerations Students were offered a reward for participating in all three exer- cises, which was 4.0 (A) as their participation grade in the class. This 31 grade was independent of their class project grade, since the class pro- ject started after these three exercises were completed. Students were told that they did not have to participate, although no student declined. Exercises were designed to integrate with classroom topics, so that the experiment could be a learning situation. After the three experimental sessions, students were debriefed about the hypotheses. Students were asked about any complaints they had about participating in the experiment. One student told the experimenter after class that she felt that her privacy has been invaded by the use of the tape recorders. The experimenter explained that her individual responses were not being analyzed, since we were looking for responses on the group level. The student indicated that she felt more comfortable after this explanation. 2.9 Summary Subjects were assigned to small groups which worked on problem solv- ing tasks at three points in time, while their interactions were tape recorded. Half of the groups received instructions to manage conflict which occurred and half of the groups received instructions to resolve conflicts. At a fourth point in time, subjects were given a question- naire measuring group satisfaction and survival. Coders listened to tapes of the group meetings to measure interaction variables. In the next chapter, analysis of the interaction and questionnaire data will be discussed. CHAPTER THREE RESULTS 3.1 Overview In this chapter, we will present results relevant to the hypotheses. First, intercoder reliabilities for the interaction measures are pre- sented. Second, length of time of the interactions are discussed, which is relevant to the analysis of other measures. Next, the manipulation check is presented. Then, the analysis of the self report questionnaire is presented. Finally, the analysis of interaction data is discussed. 3.2 Inter-coder Reliabilities The following table (Table 2) represents the inter-coder reliabili- ties for the interaction measures. As discussed earlier, a different set of coders measured each variable at each point in time. This table rep- resents zero-orcer correlations between coders' estimates of the freu- quency with which each variable occurred at each point in time for each of the twenty groups. As shown, correlations are .92 or higher. (Table 2, next page) 3.3 Length of Time of Interactions 3.31 Time One. The length of time each group worked on the problem solving task was measured. Analysis indicates that there is no signifi- cant difference between experimental conditions in the length of time on day one (Table 3). 32 33 Table 2 Intercoder Reliabilities Experimental time Variable Time one Time two Time three Disagreement .99 .97 .95 Agreement .98 .90 .99 Negative emotions .96 .97 .99 Reasons .99 .97 .99 Humor .99 .99 .99 Laughter .94 .99 .97 Group laughter .92 .99 .99 Singular pronouns .99 .99 .96 Plural pronouns .94 .99 .99 Table 3 Analysis of Variance Length Day 1 by Conflict Source D.F S.S. M.S. E_ 3_ Between groups 1 .61 .6101 .029 .867 Within groups 17 357.45 21.0206 Total 18 357.96 34 3.32 Time Two. At time two, groups in the conflict management condi- tion interacted significantly longer than groups in the conflict resolu- tion condition. Table 4 Analysis of Variance Length Day 2 by Conflict Source D.F. S.S. M.S. f_ P_ Between groups 1 108.46 108.46 8.16 .011 Within groups 15 212.61 13.29 Total 17 321.07 3.33 Time Three. There was no significant difference between exper- imental conditions in the length of time of the interactions at time three. Table 5 Analysis of Variance Length Day 3 by Conflict Source D.F. S.S. M.S. f_ ‘3 Between groups 1 17.26 17.26 .80 .386 Within groups 15 324.09 21.61 Total 16 341.34 3.34 Mean Length of Exercises. The mean length of time of each exercise, across groups, was as follows: 35 Table 6 Mean Length of Interactions Time 1 Time 2 Time 3 28.4 Min. 21.4 Min. 11.6 Min. 3.4 Manipulation Check The number of times disagreement occurred in each group was used as a manipulation check of the conflict variable. One-way analysis Of vari- ance at time one indicates that the manipulation was effective at this point in time. Table'7 Analysis of Variance Time 1 Disagreement by Conflict Source - D.F. S.S. M.S. E_ .3 Between groups 1 618.84 618.84 22.81 .000 Within groups 16 434.10 27.13 Total: 17 1052.94 At time two, the conflict manipulation was also effective. 36 Table 8 Analysis of Variance Time 2 Disagreement by Conflict Source D.F. S.S. M.S. E_ ‘3 Between groups 1 160.00 160.00 8.58 .010 Within groups 16 298.50 18.66 Total 17 485.50 At time three, the manipulation of conflict was not effective. Table 9 Analysis of Variance Time 3 Disagreement by Conflict Source ' D.F. s.s. M.S. 5 3 Between groups 1 .53 .53 .200 .661 Within groups 15 39.36 2.62 Total 16 39.88 These results can be understood by considering the mean number of times disagreement was expressed at each point in time. In both groups, the expression of conflict decreased over time. 37 Table 10 Condition Means - Number of Disagreements Time 1 Time 2 Time 3 Conflict Resolution 3.6 2.3 1.4 Conflict Management 13.7 8.5 1.8 Overall Means 8.7 5.4 1.6 Thus, the experiment at time three generated few expressions of dis- agreement, although the groups were exposed to the same experimental in- structions regarding conflict as at the first two points in time. Since the length of the interactions decreases over time, as illus- trated in Table 6, the rate of disagreement was computed by dividing the number of disagreements in each interaction by the length of that inter- action. The followint table presents the data which was used in a re- peated measures analysis of variance, expressing disagreement as a rate. Table 11 Condition Means - Disagreements Per Minute Time 1 Time 2 Time 3 Overall Conflict Resolution .13 .13 .12 .13 Conflict Management .88 .48 .43 .60 Overall .51 .31 .28 .37 38 Repeated measures analysis of variance was conducted using rate of disagreement as the dependent variable. This indicates that the main effect for conflict condition was significant, p