University of Southern California Dissertations and Theses

Coalition Formation In Conjoint Marriage Counseling

(USC Thesis Other)

Coalition Formation In Conjoint Marriage Counseling

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content INFORMATION TO USERS This material was produced from a microfilm copy of the original document. While the most advanced technological means to photograph and reproduce this document have been used, the quality is heavily dependent upon the quality of the original submitted. The following explanation of techniques is provided to help you understand markings or patterns which may appear on this reproduction. 1. The sign or "target" for pages apparently lacking from the document photographed is "Missing Page(s)". If it was possible to obtain the missing page(s) or section, they are spliced into the film along with adjacent pages. This may havp necessitated cutting thru an image and duplicating adjacent pages to insure you complete continuity. 2. When an image on the film is obliterated with a large round black mark, it is an indication that the photographer suspected that the copy may have moved during exposure and thus cause a blurred image. You will find a good image of the page in the adjacent frame. 3. When a map, drawing or chart, etc., was part of the material being photographed the photographer followed a definite method in "sectioning" the material. It is customary to.begin photoing at the upper left hand corner of a large sheet and to continue photoing from left to right in equal sections with a small overlap. If necessary, sectioning is continued again — beginning below the first row and continuing on until complete. 4. The majority of users indicate that the textual content is of greatest value, however, a somewhat higher quality reproduction could be made from "photographs" if essential to the understanding of the dissertation. Silver prints of "photographs" may be ordered at additional charge by writing the Order Department, giving the catalog number, title, author and specific pages you wish reproduced. 5. PLEASE NOTE: Some pages may have indistinct print. Filmed as received. Xerox University Microfilms 300 North Zeeb Road Ann Arbor, Michigan 48106 P 'j Jp-'.f 75-1077 O'CONNOR, Peter Andrew, 1942- f COALITION FORMATION IN CONJOINT MARRIAGE f COUNSELING. j University of Southern California, Ph.D., 1974 j Sociology, family Xerox University Microfilms, Ann Arbor, Michigan 48106 © Copyright by Peter Andrew O'Connor 1 9 1 k THIS DISSERTATION HAS BEEN MICROFILMED EXACTLY AS RECEIVED. .COALITION FORMATION IN CONJOINT MARRIAGE COUNSELING by Peter Andrew O'Connor A Dissertation Presented to the FACULTY OF THE GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY (Sociology) August 1974 UNIVERSITY O F S O U T H E R N C A LIFO R N IA THE GRADUATE SCHOOL UNIVERSITY PARK LOS ANGELES. CALIFORNIA 0 0 0 0 7 This dissertation, written by .............. under the direction of hila ... Dissertation Com mittee, and approved by all its members, has been presented to and accepted by The Graduate School, in partial fulfillm ent of requirements of the degree of D O C T O R O F P H IL O S O P H Y Dtan DISSERTATION COMMITTEE Chairman ACKNOWLEDGMENTS This dissertation is affectionately dedicated to my families, past and present. In particular to my mother who in her gentle attempts to intervene in sibling quarrels un wittingly taught me the rudiments of coalition theory with her frequent articulation of the aphorism "two is company, three is a crowd." To my wife, Margaret and children Felicity and Elizabeth, who have endured and hopefully survived the haz ards of having a graduate student husband and father, I record my deepest love and thanks. I would also like to express my deep gratitude to my major professor and chairman Dr. Carlfred Broderick for his guidance, inspiration and support throughout my gradu ate studies. Sincere thanks are also due to my other com mittee members, Dr. Daniel Glaser, who first introduced me to the works of Simmel and Caplow and Dr. Don Schrader, for their guidance and support. I also owe an immeasurable debt of gratitude to my two sources of financial support. The Winston Churchill Memorial Trust of Australia provided me with the opportu nity of overseas study and the Marriage Guidance Council of Victoria, Australia gave me the necessary leave of absence and also provided the additional financial support that allowed me to complete my studies. In particular I would like to thank the Executive Director Mr. Lloyd Phillips and the Board of Management of the Marriage Guidance Coun cil. Last, but by no means least, I wish to express my gratitude to the counselors and clients who participated in the study. In particular I would like to thank Dr. Clinton Phillips and Mr. Allan Holt of the California Family Study Center and Dr. Jerold Kuhn, Allan Brown and Charles Ryder of the American Institute of Family Relations. TABLE OF CONTENTS Page ACKNOWLEDGMENTS............. ii LIST OF TABLES....................................... vi LIST OF FIGURES......................................vii ABSTRACT..............................................viii Chapter I A REVIEW OF COALITION THEORY..................... 1 Definitions and Assumptions of Coalition Theory Review of Experimental Literature Summary II COALITION THEORY AND MARRIAGE COUNSELING.......................... 35 Decision Resources Payoffs Cheapest Winning Coalition Summary A Model of Coalition Formation in Conjoint Marriage Counseling Type I High Counselor Potency Type II Low Counselor Potency Type III Transitory, Unstable Hypotheses III EXPERIMENTAL DESIGN AND METHODOLOGY ...... 79 Operationalization of the Variables Experimental Procedure Sample Statistical Analysis Chapter Page IV RESULTS AND DISCUSSION........................103 Analysis of Results Discussion Summary V SUMMARY AND CONCLUSIONS ...................... 128 Limitations of the Present Study Suggestions for Future Research REFERENCES................................... 136 APPENDICES........................................... 143 V LIST OF TABLES Table Page I TYPES OF TRIADS AND CAPLOW'S PREDICTED COALITIONS....................................13 II TRIADIC TYPES AND PREDICTED COALITIONS IN THE TRIAD..................................17 III COALITIONS FORMED IN THE SIX TYPES OF POWER PATTERNS IN TRIADS . ................... 19 IV CAPLOW'S REVOLUTIONARY COALITIONS AND GAMSON'S CHEAPEST WINNING PREDICTIONS .... 27 V OPERATIONAL TRIADIC POWER STRUCTURES ........ 87 VI SUMMARY OF MEASUREMENT PROCEDURES . .......... 97 VII TYPE OF POWER STRUCTURE, AS MEASURED BY EGO STRENGTH, AND COALITION PATTERN . . . .105 VIII TYPE OF POWER STRUCTURE AS MEASURED BY LEARY INDEX AND COALITION PATTERN ........ 106 IX COALITION PATTERN AND COUNSELING OUTCOME . . . 109 X POWER STRUCTURE (EGO STRENGTH) AND FREQUENCY OF TYPE OF COALITION..............112 XI COALITION PATTERN AND TYPE OF COUNSELING OUTCOME ......................... 115 vi LIST OF FIGURES Figure Page 1. TRIADIC POWER CONFIGURATIONS IN CONJOINT COUNSELING .......................... 53 2. DYNAMICS OF CASE MANAGEMENT AS A FUNCTION OF COUNSELOR POTENCY .......................... 57 3. DYNAMICS OF REVOLUTIONARY COALITIONS AS A FUNCTION OF LOW COUNSELOR POTENCY ........... 62 vii ABSTRACT COALITION FORMATION IN CONJOINT MARRIAGE COUNSELING Peter Andrew O'Connor University of Southern California, August, 1974 CarIfred Broderick, Ph.D., Chairman Although the use of conjoint interviews in marriage counseling has become a well established clinical proce dure, little attention has been given to developing an ade quate theoretical model of the process of conjoint inter views. Conceptualizing the phenomena as a triad this study utilized the framework of coalition theory, as explicated by Simmel and Caplow to develop a theoretical model and empirically investigate its validity. The major independent variable of the study was power and the major dependent variable coalition forma tion. Power was nominally defined as the distribution of a relevant resource, which for the purposes of this study ■ was conceptualized as dominance and assertiveness. Power i was operationalized by two separate methods that provided ; a measure of power independent of the actual flow of inters i action in the conjoint session. The first procedure uti- viii lized Barron's (1953) ego strength research scale of the Minnesota Multiphasic Personality Inventory. The second method was based on Leary's (1957) index of dominance de rived from four of the clinical M.M.P.I. scales, specific ally Ma, D, Hs, Pt. Individual members of each conjoint triad were classified as either high, medium, or low on interpersonal power. Each triad was then classified as either high or low in terms of counselor potency. The data for the study of the existence of coali tions was obtained by coding audio tape recordings of the first conjoint interview. Coalitions were measured through; the use of Bales' (1970) interaction process analysis meth od. Specifically the categories that refer to positive socio-emotional acts and negative socio-emotional acts were used. A coalition was deemed to exist in any particular dyad, within the triad, when the percentage of positive or negative acts from one member was in excess of 30% of the total positive and negative acts of this person and further there existed a reciprocal number of acts in excess of 30% of the same sign, from the other member of that dyad. In the second phase of the study, concerned with the outcome of counseling, coalition patterns constituted the independent variable and counseling outcome the depen dent variable. Outcome was measured by administering a questionnaire to the counselors six weeks after the initialI conjoint session which obtained information on the current therapeutic status of the case. Twenty counselors who were, with one exception, graduate students doing internships in marriage counseling, participated in the study. Thirty five couples who were drawn from both a university setting and the community com prised the client sample. Four major hypotheses derived from coalition theory were tested in this study. As postulated in hypotheses one and two there was a significant relationship between coun selor power, as measured by ego strength and coalition for mation. Specifically, those conjoint sessions in which the counselor had high potency tended not to form coalitions and those in which the counselor was in a low potency posi tion demonstrated a significant trend to form coalitions. Using the Leary index of dominance the same hypotheses were not confirmed. Hypotheses three and four postulated that there would be a significant relationship between coalition for mation and counseling outcome. These hypotheses were con firmed and this led to the conclusion that if a coalition is present in the first conjoint session then there is a beyond chance probability that counseling will be ineffec- j tive. Conversely, a significant relationship was estab lished between no coalition present and effective counsel ing . Some additional data concerning the particular types of coalitions formed was obtained and this indicated that negative coalitions between the husband and wife were the most prevalent. The study also established that these particular coalitions were the ones most likely to result in clients prematurely terminating from counseling. The same conclusion could not be drawn concerning counselor/ spouse coalitions which appeared to have an ambiguous sta tus with respect to counseling outcome. CHAPTER I A REVIEW OF COALITION THEORY The concept of the triad is both old and new and finds its sociological roots in the work of Georg Simmel. Simmel (Wolff, 1950) was the first Sociologist to differ entiate between the dyad and the triad. He asserts (Wolff, 1950, p. 141) "that the triad is a structure completely different from the dyad but not, on the other hand, spe cifically distinguished from groups of four or more mem bers." Simmel's writings are dialectical in character and he asserts, that "there is no triad in which dissent be tween any two elements does not occur from time to time." (Wolff p. 148-149). Despite Siinmel's valuable contribution to our understanding of the differential structure of a dyad and a triad, it is important to note that he did not discuss the phenomena of coalition formation directly. He was specifically concerned with the question of the addi tion of a third party to an existing dyad. As Caplow (1968, p. 18) points out, Simmel gave little attention to the choice between coalition partners. It is this latter aspect that has dominated the contemporary work in coali tion theory, best exemplified by the writings of Caplow. This contemporary interest in coalition theory has . 1........................................................................... 2 been pursued from three identifiable perspectives. Within the Sociological tradition the triad and the attendant questions of coalition formation have been explored in a series of experimental laboratory type studies and theo retical papers by such scholars as Mills (1954), Strodtbeck (1954), Caplow (1956), Vinacke and Arkoff (1957) and Gamson (1961). The second tradition, which can be labelled the games theory approach, finds its leadership in the work of Von Neuman and Morgenstern (1944) . Their work has subse quently been followed by such scholars as Shapley and Shubick (1954) and Luce and Rogow (1956). The third per spective largely finds its expression in the somewhat mixed field of history, political science, and journalism. In the main this stream is exemplified in the work of Lerner and Aron (1957) and their account of the French National Assembly. This stream or tradition is characterized by descriptive rather than experimental style. Francis' (1962) work on legislative systems and Shubert (1964) study; of judicial behavior are further examples of this descrip- j tive approach to coalition formation. These streams have by and large maintained a rela tively independent existence. Some attempts have been made! to combine them and can be seen in the approach of such scholars as Riker (1962) in his study of the formation of political coalitions. Within the sociological framework attempts at synthesis have primarily focused on the contro versial issue "what determines strategy in a three person game." The game theory approach, primarily based on mathe matics asks the question "which coalition will form" on the! assumption that players in a game will act to maximize their returns and that the same players make a rational analysis of the final outcome, which determines their strab* egy in the game. However, as Luce and Raiffa (1957) point out "it is crucial that social scientists recognize that game theory is not descriptive but rather (conditionally) normative." In essence game theory states how people should behave if they wished to reach certain ends. The alternative consideration is most clearly found in Caplow's work, which focuses on the point at the begin ning of the game rather than the outcome, and is concerned with how a player interprets his position in relation to the other two players with respect to the relative distri- ; bution of resources. This is frequently referred to as perceived power. Vinacke and Arkoff (1957) have experimen-j tally explored this conflict in detail and in essence have ; I tested the question whether persons act according to a ra- ; tional analysis of outcome, or according to their percep tion of the initial power structure. Their conclusions, based on experimental manipulation of power distributions involving ninety students playing an experimental game of 4 parchisi/ lends support to Caplow's theory. They conclude that "rather than employing a strategy based upon recogni tion of the irrelevance of the initial strength of a rela- ; tionship, members of triads act according to their percep tion of the power system." A further objection to the game theory approach is that its most powerful applications have been in the area of two person zero sum games. This is inconsistent with the definition of coalitions offered by Gamson (1964 p. 85) which he defines as "the joint use of resources to deter mine the outcome of a decision in a mixed motive game situ ation involving more than two units." The central notion here is a "mixed motive" situation which can be contrasted with pure co-ordination games and pure conflict games. In the conflict game (following Schelling 1958) two persons are involved in a zero sum situation. Thus any gain ob tained by one participant results in the equivalent loss by the others. According to Gamson (1964 pp.85-86), extended ; to more than two persons, pure conflict implies a situation, in which no person can gain through co-ordinating his ac tions with another. This is either because constraints exist that prohibit co-ordination, or because a person can ! achieve more by acting on his own. Thus, by definition, zero sum games do not fit well with the coalition model. Mixed motive games by contrast, as Stryker (1972) points out, are the ones in which the "maximization of re- ; wards dictates co-ordination with some and conflict with other participants in a system of interaction." Such a setting, in which conflict with some and co-ordination with! others exists, presupposes the existence of more than two social units and it is the three member situation of the triad, that has received most attention. It is this struc ture that is the simplest case in which coalitions are pos sible and consistent with Simmel's propositions it is the structure most clearly differentiated from the dyad. This study is not concerned with exploring further the controversy concerning the conditions that determine players strategies in three person games. Rather it will attempt to combine the experimental with the descriptive tradition and will in the main follow the sociological views of Simmel, Caplow and Gamson. At the same time how ever, the present study differs from the previous experi mental work in so far as it is concerned with a field situ-; j ation, specifically, joint marriage counseling. A review of the literature on coalition theory reveals an almost exclusive concern with experimental groups and/or contrived; tasks in a laboratory setting. Some exceptions to this are; studies conducted, by Turk and Turk (1961) on nursing teams, Torrance (1955) on B-26 combat crews, and Strodtbeck (1954) on families, although in the latter two instances, whilst 6 I the groups were natural, the tasks were contrived. The j relevance of this distinction will be taken up in a later chapter. Suffice it to say for the present that regardless; of the situation, the task, or the nature of the group, certain theoretical tenets and concepts are applicable to the general phenomena of coalition formation. Definitions and Assumptions of Coalition Theory A coalition has already been defined as "the joint use of resources to determine the outcome of a decision in : a mixed motive situation involving more than two units." (Gamson, 1961, p. 373-382). There are several terms within this definition that require further clarification. Gamson (1961) defines a resource as "some weight controlled by the! participants such that some critical quantity of these weights is necessary and sufficient to determine the deci- ; sion." The same author further elaborates that potential resources are almost infinite in their variety, ranging from military force to verbal ability. The critical point is, that resources vary from situation to situation and any quality or object that can exert influence on a decision can be conceptualized as a resource or weight. The "joint use" of such resources re- 7 ! i lates to the mixed motive quality of a triadic structure and specifically refers to two or more persons jointly using the resources relevant to the attendant situation, for the purpose of bearing influence on some decision. This according to Gamson (1961) is simply a "selection among alternatives." Stryker (1972) defines the decision i a little more explicitly as "implying a conscious, deliber ate act of terminating some problematic question or issue." A winning coalition then, is that combination of two or more persons which has sufficient resources to control the decision. Finally each winning decision results in a pay off, a set of rewards, to the participants of the winning coalition. The concept of payoff introduces a further assump tion of coalition theory and is referred to by Gamson (1961) as the "cheapest winning coalition." This states, "that any participant will expect others to demand from a coalition a share of the payoff proportional to the amount j of resources which they contribute to a coalition." (Gam son, 1961). Caplow (1968) calls this assumption the "Equi-i table expectation." Stryker (1972, p. 344) points out that there is a close similarity between this assumption and what Homans (1961) has previously referred to in his ex change theory framework, as distributive justice. In brief, participants bring to the coalition situation a pre- vailing parity norm, which specifies that they ought to receive rewards proportional to their costs. From this assumption we can conclude that in any coalition situation i where total payoff is held constant, the most probable coalition is in fact the cheapest winning coalition, where costs are just sufficient to assure winning. From this discussion of definitions and assumptions we can assert following Gamson (1961), that a full fledged coalition situation will exist when the following condi tions are present: 1. There is a decision to be made and there are more than two social units attempting to maximize their share of payoffs. 2. No single alternative will maximize the payoff to all participants. 3. No participant has dictatorial powers, i.e. , no one has initial resources sufficient to control the decision by himself. 4. No participant has veto power, i.e., no member MUST; be included in every winning coalition. The first condition clearly establishes that each participant has some stake in the outcome and together with; condition three, it establishes that the true coalition situation is an essential game. Luce and Raiffa (1957, p. 185) describe an inessential game as one in which "no coalition of players is more effective than several players of the coalition operating alone." Conditions three and four are the conditions that clearly establish the differ- i ence between a dyad and a triad, since in the dyad either one member is potentially a dictator or each possesses veto power. Given these conditions the task of coalition theo ry, consistent with the purpose of theory in general, is not only to provide understanding but also to generate pre dictions which can be experimentally verified. In order to meet this requirement Gamson (1961) indicates that the model requires information on the following parameters: (i) Information concerning the initial distribution of resources. Specifically what the relevant re sources are for any given decision and further how much of these resources each participant controls prior to the point when coalitions are formed. (ii) Information concerning the payoff for each coali tion. This review of the major theoretical tenets of coalition theory leads into a review of the experimental literature that has attempted to test hypotheses derived from the theory. 10 ! I Review of Experimental Literature The earliest empirical work was not specifically concerned with testing hypotheses derived from coalition theory per se, but more with the question of testing Sim- mel's original propositions concerning the tendency of the triad to divide into a dyad plus a monad. The two classi cal experiments conducted within this vein are those by Mills (1953) and Strodtbeck (1954). Mills (1953) using student volunteers in 48 three person discussion sessions, had subjects create stories from T.A.T. cards. Using an index of support derived from Bales (1950) categories, he studied Simmel's proposition that the triad segregates into a pair and other. In the main his findings confirmed Sim mel 's general proposition. As a subsidiary finding Mills found that the highest rates of support between any two members occur between the two most active participants. j That is, the two most active members formed a pair and the ; [ i least active member was observed to be the relatively iso lated third party. Strodtbeck (1954) attempted a replication of Mills study on a three person family group where father, mother, | and adolescent son were asked to arrive at an agreed upon decision concerning selected topics relevant to the son's future. Strodtbeck failed to find evidence which supported 11 Mill's findings and his main explanation for this discrep- I ancy is that the family is a totally different group from Mill's ad hoc student groups. However, Strodtbeck, like Mills, did find that decision making power was associated with high participation. The contemporary empirical work on coalition forma tion has been more concerned with testing propositions de rived from coalition theory as espoused by Caplow and Gam son. Caplow (1956 p. 489) argued that the game theoretic approach of Von Neuman and Morgenstern "concealed the assumption of equality of power among the three players, even when they have different possibilities of gain or loss." He further argues that this assumption did not "fit many triads of sociological interest in which the typical game consists of domination over other triad members and not ir an external reward to be obtained by a given coali tion." (1956 p. 489). As an alternative Caplow developed a theory that asserted under certain conditions the formation of particular coalitions depends on the initial distribu tion of power in the triad. He made the further critical assertion that knowledge of the initial power distribution made it possible to predict, to some extent, the coalitions which would be formed. These certain conditions that Cap- low alludes to are specified by him as follows (Caplow 1968 p. 22): (1) Members of a triad may differ in strength. (2) A stronger member can control a weaker member and will seek to do so. (3) Each member of the triad seeks control over the others, preferring control over two others to control over one. (4) Strength is additive. (5) The strength of a coalition is equal to the sum of the initial strengths of its members. Caplow applied these assumptions to six triad types differ ing in the way in which "strength" was distributed and pre dicted the typical coalition(s) that will form. These pre dictions are shown in Table I. The detailed predictions can be adequately under stood by taking each of these triadic types and discussing the dynamic nature of the prediction. In type 1 where all members are of equal strength, Caplow's theory predicts that the coalitions AB, AC, and BC, are all equally likely and each member strives to enter a coalition in which he will be equal to his ally and stronger via the coalition than the isolate. This prediction is consistent with Cap- lows assumptions 3, 4, and 5. In the type 2 structure one member is stronger than the other two, but not stronger than the combined weight of the weaker pair. Again, as in 13 TABLE I TYPES OF TRIADS AND CAPLOW'S PREDICTED COALITIONS TRIAD TYPE PREDICTED COALITION 1. A = B = C 2. A > B, B = C, A < (B+C) 3. A < B, B = C 4. A > (B+C), B = C 5. A > B > C, A < (B+C) 6. A > B > C, A > (B+C) any BC AB or AC none BC or AC none 14 structure 1, all members are assumed to seek control by forming a coalition, since isolation would be a disadvanta geous position relevant to the distribution of rewards. However within this structure, Caplow predicts that all coalitions are not equally likely and that a BC coalition has a higher probability of occurrence. If the position of B is considered, the reasons for the higher probability of a BC coalition emerge. If B formed a coalition with A, he would be stronger than C, (by virtue of conditions 4 and 5), but within that coalition he will be weaker than A. On the other hand, if he forms a coalition with C he will be equal to C within that coalition because of their initial equal weight and at the same time he will achieve control over the stronger member A. Thus a BC coalition is the most advantageous one to B and C and the formation of such a coalition renders the individually strongest member of the triad, the weakest. This is consistent with Caplows (1968 p. 3) assertion that "the most distinctive feature of triadic social systems is the transformation of strength into weakness and weakness into strength." Type 3 triadic structure highlights this feature; it is the situation in which two members are equal in strength, but the third member is weaker. Here A is the member sought after as a coalition partner since both B andj C can gain control over the situation only through a coali 15 ! tion with A. Such a structure in which A is the weakest, and yet the most sought after member, is described by Sim- mel (Wolff 1950, p. 154) as the "tertius gaudens" role. Within the type 4 structure the strength of A exceeds the combined strength of B and C, hence a coalition between these latter two members would have no effect in terms of gaining control. Alternatively, consistent with Caplow1s third condition, A has no incentive to form a coalition, since alone he already has control over the other two mem- : bers. The type 5 triadic structure presents a situation where no two members of the triad are equal in strength, but the combined weight of any two members is greater than the third. Caplow (1968 p. 24) asserts that this is simi lar to the type 3 triad in so far as the weakest member has a decided advantage, and is sure to be included in whatever coalition is formed. Hence the predictions are that either! a BC or AC coalition will form. Gamson (1961), as will be discussed later, produced the theoretical concept of the cheapest winning coalition to predict that in this struc ture it is the BC coalition that has the higher probability: of occurrence. He further reports experimental data to support this prediction. However, it is clear that within this structure an AB coalition is not likely to be formed, since B has nothing to gain from such a coalition and A has: a strong incentive to form one with C in order to gain con-: trol over B. The type 6 triad is similar to type 4, in that there is no incentive for A to form a coalition, since he is already stronger than the combined weight of B and C ! and the latter two members combined would have no effect oh A's control. Vinacke and Arkoff (1957) set out to test these predictions of Caplow1s experimentally. The study, which has become the main paradigm for coalition experiments, also sought to contrast the game theory predictions with Caplow1s (see Table II). The method consisted of 90 male students divided into groups of three playing an experimen tal game of pachisi. Only the outside spaces of the pachi- si board were used giving a total of 67 spaces and the ob ject was to reach "home" first. The different patterns of strength, conceptualized as power, were operationalized by having players draw tokens of varying weights from 1 through 4. A single die was used and each player was en titled to move forward the number of spaces equal to his pre-drawn weight times the number shown on the die. At any time during the game, players could form alliances by of fering some proportion of the hundred point prize, pool their strength, and proceed to a position equal to their combined weight times the die. Each group played a series of three games and each series was comprised of 6 games, 17 TABLE II TRIADIC TYPES AND PREDICTED COALITIONS IN THE TRIAD* TRIAD TYPE WEIGHTS ASSIGNED BY VINACKE & ARKOFF COALITIONS PREDICTED BY CAPLOW COALITIONS PREDICTED BY GAME THEORY 1. A=B=C 1-1-1 any any 2. A>B, B=C, A< (B+C) 3-2-2 any any 3. A>B, B=C 1-2-2 AB or BC any 4. A> (B+C), B=C 3-1-1 none none 5. A>B>C, A<(B+C) 4-3-2 BC or AC any 6. A>B>C, A>(B+C) 4-2-1 none none Source; Stryker and Psathas, 1960 p. 218. one of each type, so that each group played a total of 18 games. Rotation of condition was by type of initial strength and not necessarily by player, which was left to chance in drawing the weights. To control for the poten tial effect of the sequential position of the triad types within the series of games, the order of the games was sys-: tematically arranged according to a Latin square design which varied the sequence of the types in the three series. The results (shown in Table III) of the Vinacke and Arkoff experiment conform with Caplow's predictions and ran count er to the game theory model which led to predictions based on a rational analysis of outcome (see Table II). This model predicted that in types 4 and 6 triadic structures no coalitions will occur, which is consistent with Caplow's original predictions. However, for the remaining triads (i.e., 1, 2, 3 and 5) contrary to Caplow, the game theo retic model predicted that each coalition has an equal probability of occurrence. Vinacke and Arkoff (1957) con- ; elude that "players appear to be influenced primarily by perception of their relative strength at the outset of the game" and that "there was no evidence that players tended to play more in terms of analysis of outcome (rational strategy) in later series of games." This latter assertion is taken as evidence that players did not discover, or at the very least, utilize a strategy based on outcome. 19 TABLE III COALITIONS FORMED IN THE SIX TYPES OF POWER PATTERNS IN TRIADS ALLIES TYPE 1 (1-1-1) TYPE 2 (3-2-2) TYPE 3 (1-2-2) TYPE 4 (3-1-1) TYPE 5 TYPE 6 (4-3-2)(4-2-1) AB 33 13 24 11 9 9 AC 17 12 40 10 20 13 BC 30 64 15 7 59 8 TOTAL 80 89 79 28 88 30 (No Coalition) 10 1 11 62 2 60 p* >.05 <.01 <.01 <.70 <.01 <.50 Source: Vinacke and Arkoff 1957 p. 409. 2 *In computing X , Vinacke and Arkoff assumed that EACH pair would occur an equal number of times by chance. 20 ; Simply put, Vinacke and Arkoff, while finding em pirical support for Caplow's predictions were also able to i empirically refute the rational analysis strategy of the game theoretical model. In other words it was the percep- I tion of initial strength that primarily determined the players strategies. The same study also provided evidence concerning which player initiated the offer to form each winning coalition and how the prize was divided. As ex pected, the weakest players were quickest to propose coali tions. The division of the prize according to the authors (Vinacke and Arkoff 1957) appeared to reflect the influence; of the power structures and where: differences in strength are perceptually pro nounced (types 4 and 6) highly disproportionate divisions of the prize occurred. Where the part ners were equal to begin with, or nearly so (types 1 and 2) equal division was the rule. In situa tions where one of the participants was weaker than the other two, unequal, but not pronouncedly so, deals were made. Although the Vinacke and Arkoff experiment was set ; up explicitly to test the Caplow versus game theory predic- tions, several scholars have questioned the appropriateness; of the design to handle this question. By and large, the details of this debate are beyond the scope of the present ; study, although the main points appear worthy of mention. Stryker and Psathas (1960 pp. 217-230) point out that the bargaining which occurred was face to face and allowed for the intrusion of such variables as bargaining skill and techniques which they claim are extraneous to the immediate concerns of the research. Whilst one can grant that they may be extraneous to the research, it is somewhat dubious to assert that they may be extraneous to the process of coalition formation as originally conceived by Caplow. A more cogent criticism is that the Vinacke and Arkoff paradigm does not allow a valid test of games theory or the rational analysis approach simply because a devia tion from the rational strategy by one player would make the adoption of a game theoretic strategy "irrational" for the others. This criticism has some validity, and again it highlights the point made by Luce and Raiffa (1957) that: game theory is not descriptive but rather (condi tionally) normative. It states neither how people do behave, not how they should behave in an abso lute sense, but how they should behave if they wish to achieve certain ends. In this sense it is reasonable to assert that the game theoretic model is inappropriate and Vinacke and Arkoff1s findings would appear to confirm this assertion. j Thibaut and Kelley (1959) raise the objection that Vinacke and Arkoff"s study did not give game theory a fair ; test because each triad played three games each of 6 dif ferent types of power structures. They claim this neces 22 sitated the employment of different strategies as the triad; type changed and this may have led to confusion. They be lieve that if only one triad type were used and the experi- mental situation were prolonged, the players would eventu ally perform according to game theory. Despite some evi dence in support of this by Kelley and Arrowood (1960), where players were involved in a series of games only with in a type (5) triadic structure, i.e., A>B>C, A<(B+C), Vinacke and Arkoff (1957) produced some contrary evidence. In the types 4 and 6 structures, where one member is strong er than the other two combined and hence a coalition is not necessary, some 31% and 33% of the games with these struc tures did reflect coalitions. Vinacke and Arkoff (1957) state, that although some learning took place it was more a matter of eliminating coalitions between weak players than of avoiding coalitions per se. They conclude "that if any evidence is needed that players are not guided primarily by; 'rational' considerations, it is provided by this result." (Vinacke and Arkoff 1957). Thus Thibaut and Kelley's ob jection that playing in several different types of triadic structures led to confusion and difficulty for the players in implenting a rational strategy would appear to be a du bious one. The types 4 and 6 structures are the ones in which rational play would be the easiest to understand and yet players still showed a moderate trend to form coali 23 tions. In general then, it can be concluded that players act according to their perception of the situation in which they find themselves and not according to what can be re garded as a rational analysis of outcome. This conclusion is the line adopted throughout the present study. It has already been discussed that with reference to type 5 triad A>B>C, A<(B+C) that Caplow predicted either an AC or BC coalition was equally likely to occur and Vinacke and Arkoff broadly interpreted their findings as supportstive of the prediction. However, reference to Table III shows that the BC coalition had a slightly higher frequency of occurrence. Gamson (1961 p. 373-382) proposed several refinements of coalition theory that attempted to account more adequately for the empirical findings, in par ticular the observation that in type 5 triads the BC coali tion occurred more frequently than the prevailing theory predicted. Gamson's main contribution was his introduction of a concept he termed the "cheapest winning coalition." The strategical consequences of this concepts for coalition theory are clearly stated by Gamson (1961 p. 376): When a player must choose among alternative coali tion strategies where the total payoff to a winning coalition is constant, he will maximize his payoff by maximizing his share. The theory states that he will do this by maximizing the ratio of his re sources to the total resources of the coalition. Since his resources will be the same regardless of which coalition he joins, the lower the total 2 4 ! resources, the greater will be his share. Thus where the total payoff is held constant, he will favor the cheapest winning coalition. Using the original Vinacke and Arkoff data, Gamson was able to demonstrate that BC coalitions in type 5 struc tures occurred more frequently than AC, and this can be seen as support for Gamson's view. The cheapest winning coalition concept and the relevant findings have been par alleled in another field by Riker (1962). In his work, which is based on mathematical inference and concerned with data from American political history, he asserts "that participants create coalitions just as large as they be lieve will ensure winning and no larger." This principle he calls the "size principle." (Riker 1962 p. 47). Apart from accounting more adequately for the rele vant empirical findings, Gamson's cheapest winning coali tion concept has a theoretical link with Caplow's (1968) views on organizational triads. These he defines as "a triad whose members belong to an organization and are re quired by its programme to interact with one another." These members, according to Caplow may be either individu als or collectivities. He further states (1968 p. 51) "that an organization is above all a mechanism for achiev ing definite goals, for getting things done that could not j be done as well or at all by unorganized individuals." 25 j I This organizational focus leads Caplow to distinguish be tween two types of programs, the "official” and the "pri vate." The official program is the organization's program and is inextricably bound up with the status order. Its successful implementation depends on keeping the status order intact. The private program refers to the individual; members' programs, which may or may not, coincide with the official programme. Caplow (1968 p. 51) contends that the most; important transaction between an organization and the members embodies a fundamental conflict of interests, the members wanting to maximize rewards they receive for participating and the organiza tion wanting to minimize the costs of their par ticipation. It is within this context that Caplow introduces his con- , cept of "revolutionary coalitions." These he defines as "a winning coalition that dominates the superior members of an organizational triad." Such coalitions are to be distin guished from two alternative types, respectively, "conser- ; vative" and "improper." A conservative coalition is de fined by Caplow as one which "does not upset the prescribed status order in an organizational triad." An improper coalition is one that is neither conservative nor revolu tionary (Caplow 1968 pp. 50-51). The relevant theoretical link between Caplow and Gamson is that Caplow's list of triadic structures in which! revolutionary coalitions are possible is identical with Gamson's predictions based on the cheapest winning coali tion concept. The relevant predictions are shown in Table . IV, with two additional triadic structures in which the strength of A is equal to the combined strength of B and C. From this agreement between Caplow's and Gamson's views, it can be asserted that in those triadic structures in which revolutionary coalitions are available, they are also the ones most likely to occur. Since by definition they represent the domination of the superior member of the triad and consequently a disturbance of the status order, they can be seen as usually reflecting the imple mentation of the private over the official program (Cap- ; low 1968 p. 52). This particular assertion is basic to the present study and will be expanded on in Chapter 2. Suf fice it to say for the present that with reference to mar- : riage counseling, a revolutionary coalition would indicate that the official goal or programme of counseling had been j subverted. Private programmes would be operating, which would lead to end states inconsistent with established therapeutic goals. The studies reviewed so far have one overriding similarity; that is the utilization of assigned weights or resources. Anderson (1967) points out that a situation may 27 TABLE IV CAPLOW'S REVOLUTIONARY COALITIONS AND GAMSON 'S CHEAPEST WINNING PREDICTIONS TYPE OF TRIAD CHEAPEST WINNING COALITION (GAMSON 1961) AVAILABLE REVOLUTIONARY COALITION (CAPLOW 1968) 1. A = B = C any any 2. A>B, B=C, A<(B+C) BC BC 3. A<B, B=C AB or AC AB or AC 4. A> (B+C), B=C none none 5. A>B>C, A<(B+C) BC BC 6. A>B>C, A>(B+C) none none 7. A>B>C, A=(B+C) inapplicable* none 8. A=(B+C), B=C inapplicable none Source; Caplow (1968, p. 39) ♦Types 7 and 8 are labelled "inapplicable" by Gamson because they do not meet one of his technical conditions (i.e.) no participant must be indispensable to every win ning coalition. exist where all actors have equal resources, in which event! the control of the goal object will be determined by other ! factors in the bargaining situation. Several experimental I investigations have been concerned with some of these "other factors," that may possibly affect the nature of the coalition formed. Vinacke and his associates have been the major contributors to research on other factors in coali tion formation and in the main their studies have concen trated on sex as the variable. Typical of such studies are those conducted by Vinacke (1959) , Bond and Vinacke (1961) , , Uesugi and Vinacke (1963) and Vinacke and Gullickson(1964). ; The general conclusion that emerges from these studies is that the coalitions that formed most frequently in the various triads were similar for males and females. How ever, certain sex differences were observed in relation to ; strategy. For example, in comparison with males, females entered more frequently into triadic alliances where every-; one won something and more often formed coalitions when none were necessary in order to win the game. Females were; also found to divide the prize more equally, and tended to engage in less bargaining. Vinacke and his associates have interpreted these findings to indicate that males utilize a more competitive ! strategy and try to maximize gain, whereas females were more motivated by co-operativeness and maintaining pleasant 29 social relations. Thus, in general the findings have con- tributed very little to the existing theories of coalition formation and provide evidence concerning minor variations of pattern, or playing for different goals, rather than contradictory findings. It can be argued, that in the experimental situations used to test the effects of sex as a significant factor, it was not in point of fact, a rele vant resource. Consistent with Gamson's views, if a re source is not relevant to the goal object, then one would not expect it to have any substantial effect on the coali tion formed. This suggestion finds support in the finding that the coalitions formed in the various triad were simi lar for males and females and conformed to the usual pre dictions made from the theory. The same criticism can be made against a further series of studies concerned with achievement and nurturance in the triadic situation (Chaney and Vinacke [1960], (Amidjaja and Vinacke [1965]). In these experiments, triads were composed of one person high on nurturance and low on achievement, one person low on nurturance and high on achievement and a third intermediate on both. Although some different patterns were observed, for example, in all male triads, high achievers were the ones who usually made the offer and nurturant males tended to behave in a non competitive manner. The major factor determining which coalition was formed, was the distribution of power and not personality characteristics. These findings on the prior- j ity of power do not seem surprising when one takes into consideration Gamson's definition of a resource as "some weight controlled by the participants such that some criti cal quantity of these weights is necessary and sufficient to determine the decision." In the experimental designs used to test the effects of these additional factors, it is a moot point whether some "critical quantity" of them was "necessary and sufficient to determine the decision." It would appear more accurate to assert that they were proba bly irrelevant to the decision to form a coalition in order to win a hundred points in an experimental game of parchisi. If, on the other hand, the experimental situation was an interpersonal one in which the problematic issue to be resolved centered around approval or rejection, then achievement and nurturance may well represent "relevant resources." Other studies, such as one conducted by Buris and Frye (1966) on authoritarianism and need for indepen dence in a five person game can be criticized on the same grounds. They concluded that authoritarianism did not discriminate between winners and losers, or coalition mem bers from non-coalition members. But again, one can raise the question, were such factors "relevant resources?" The overall conclusion that can be drawn from these studies that 31 have attempted to examine the effects of personality fac tors is that the experimental designs did not, in fact, allow an adequate test of the question, in so far that neither theoretical or empirical evidence was put forward concerning their relevance as resources. This issue of relevance will be taken up in the following chapter, where the theoretical model to be tested in the present study is developed. The final area of research to be reviewed covers those studies that have attended to the question of the effect of a series of games on the coalition patterns formed. Although the present study is not specifically concerned with this question, it nevertheless appears use ful to briefly review the research. It is possible that some serial effects could operate within a single period of observation, such as that with which the present study is concerned. These previous studies have focused on the question of sequential changes in coalition formation. Chertkoff (1970) raises this as one of the four major ques-; tions that coalition theorists have to attend to in order to improve the predictive power of the theory. This same author claims that Vinacke and Arkoff ignored the possible influences of differential reinforcement and that no at tempt was made to control or analyze the effects of success: or failure in earlier games on the performance of subjects I in later games. Bond and Vinacke (1961 pp. 61-75) did report that when a cumulative scoring chart was placed in a position where all participants could see it, there was a strong tendency for those who were losing to form a coalition against the leader, regardless of the power structure in the particular games. Hoffman, Festinger and Lawrence (1954 pp. 141-159) report similar phenomena as also does Emerson (1964 pp. 382-398). Such a phenomenon may be high ly relevant to sequential changes in coalition patterns and a satisfactory theoretical explanation for it can be found in exchange theory terms. Within this theoretical frame work, the other player becomes attractive since he offers an opportunity to recoup losses and possibly regain con trol. Emerson (1964) utilizing this framework refers to this phenomena as a type of "balancing operation." There remains a substantial problem in interpreting the effects of cumulative scores and the tendency of two losers to unite against the winner. In using such scores, it is these that become the effective weights and the ini- : tial assigned weights become irrelevant. In fact, as Cap- ; low (1968 p. 31) points out, as soon as any one player de velops a substantial lead, "a coalition between two losing i players becomes almost inevitable." So, although such cumulative scores affect the sequential patterns of coali- ; 33 tion formation, it remains unclear as to what this finding actually tells us, other than that the relevant weights change, and players then conform with the usual predictions! of the theory. The one substantial conclusion we can draw j from these studies on the effects of recording cumulative scores, is that they provide empirical evidence in support ' of Caplow1s (1968) assumption that "each member of the triad seeks control over the other, preferring control over y two others to control over one." In this sense, losers uniting as Emerson (1964) and others have observed, can be seen as behavioral confirmation of this preference and not as refutation of the relevance of assigned weights in any particular game. It is perhaps Caplow's (1968 p. 3) state ment, that the most distinctive feature of a triadic system; is the "transformation of strength into weakness and weak ness into strength" which best accounts for the sequential changes in coalition patterns. Summary From this review of the experimental literature, several conclusions can be drawn which are relevant to the | present study. These are: 1. The perceived distribution of initial resources in a triadic situation, conceptualized as power, en- ables a prediction of the likely coalition patterns' that is superior to predictions based on the game theoretic model which posits a rational outcome strategy. In the main, these patterns reflect a preference for control over other members, where control over two is preferred to control over one, which is preferable to control over none. That resources which are not relevant to the problematic issue to be resolved or the decision, have minimal effect on the overall coalition pattern. Sequential changes in coalitions do occur, and tend to conform to the theoretical statement that in the triadic situation, strength is transformed into weakness and weakness into strength. Members of a triad by and large conform to the distributive justice principle and expect rewards proportional to costs. When it is available, the cheapest winning coali tion is the most probable, and such a coalition is in fact a revolutionary one in Caplow's terminology as it results in a disturbance of the prevailing status hierarchy. CHAPTER II COALITION THEORY AND MARRIAGE COUNSELING The primary purpose of this study is to investi gate empirically the application of coalition theory to conjoint marriage counseling. The distinguishing feature of the studies reviewed in the previous chapter is that they have almost exclusively concerned themselves with ex perimentally contrived tasks in a laboratory setting. Hence, the relevance and validity of the findings for real life situations still largely remains unanswered. It seems both important and useful that some attention be given to the question of what degree these findings can be general ized across various coalition situations. Conjoint mar riage counseling provides a structure which is ideally suited to coalition investigations, since it involves a real life triadic structure. Olson (1970) in his compre hensive review of marital and family therapy suggests that "if there is any one technique which presently best repre- ! sents the field, it is conjoint marital therapy, which in its purest form involves seeing the couple together during all sessions." Empirical support of Olsonrs assertion is 35 36 found in Alexander's (1968) study in which she ascertained,j via mailed questionnaires administered to the 1965 A.A.M.F.C. membership, that among marital therapists some 85% saw marital couples or families conjointly and only 10% saw clients individually without seeing the spouse. How ever, the proliferation in practice has not been accompan ied by a growth in relevant theory and Manus' (1966) state ment that "marriage counseling is a technique in search of a theory" is still an accurate appraisal of the present state of the field. A cursory review of the literature of marriage counseling, and especially that on conjoint counseling re veals a plethora of therapeutic techniques out of which can be abstracted two factors. One, is that the relation ship is the unit of treatment which perhaps explains the extensive use of conjoint interviews. The second, and more fundamental factor, is that the conceptual framework is that of a dyad. The "second patient" appears to have been added to the counseling relationship out of utilitar ian concern for effective treatment and apparently was not accompanied by a shift in the conceptual framework. Clear ly the conjoint marriage counseling situation is a triadic and not dyadic structure and hence the triad is the ap propriate conceptual unit. This establishes, at least at the conceptual level, the relevance of the works of Simmel 37 and Caplow for marriage counseling. The utilizing of a triadic approach is adumbrated in Neubeck's (1973) article entitled "Toward a Theory of Marriage Counseling: A Humanistic Approach." He suggests that the triadic approach will be productive in generating a theory of marriage counseling sui generis. However, his suggestion is a general one and the problem remains actually to transfer the concepts, propositions and empiri cal findings of coalition theory to conjoint marriage counseling. It is this problem that the remainder of this chapter will attend to. The previous chapter established that the major concepts of coalition theory are resources, decision point, cheapest winning coalition and payoffs. The major proposi tions derived from these concepts were concerned with the relationships among them and with the determinative role they play in coalition formation. As previously stated, Gamson (1961) indicates that in order to have a full fledged coalition situation there must be a "decision to be made and there are more than two social units attempt ing to maximize their share of payoffs." It, therefore, seems necessary that as a first step attention should be focussed on what is the decision, or in Stryker's (1971) terms the "problematic issue to be resolved." If this question can be answered satisfactorily, one can deductive- 38 ly arrive at what are the relevant resources, what consti tutes the cheapest winning coalition and what is the pay off. Proceeding in this manner will serve to circumvent the problem associated with previous studies that have attempted to examine the effects of personality factors on coalition formation, namely the ambiguity of their status as a resource. Decision In the field of family therapy, Jay Haley (1962) , from a systems theory perspective views the family as a homeostatic system which he considers has two types of governing processes. On the one hand, family members be have in what he terms an "error activated way" if one member exceeds a certain established limit. On the other hand, Haley sees individual family members as attempting to be the ones who set or establish this limit. He consid ers that the central struggle in disturbed families is concerned with this question of who is going to tell whom what to do and so govern the limits of behaviour. It is Haley's view of the relevance of this phenomenon for family therapy that Is particularly pertinent to the present dis cussion. Succinctly put, it is, that members of disturbed families Cand here for the purposes of the present dis cussion one can substitute disturbed marriages) respond to the therapist "as if he has been included in the family (marriage) system and so he is subject to the same strug gles that they have with each other over the problem of who is to set or circumscribe their range of behaviour." The goal of the therapist within this context indicated by Haley, is to gain control or become the "governor" of the system. This then represents the proble matic issue to be resolved or the decision to be made and the formation of a coalition can be viewed as the major strategy by which this is achieved. Within Caplow's (1968) theory the successful implementation of the official pro gramme would occur when the therapist achieved the goal of becoming the "governor" of the system and the one who set the limits on the range of behaviour, at least within the confines of the conjoint situation. The formation of a revolutionary coalition would indicate the operation of the unofficial programme resulting in the subversion of therapeutic progress since it would deny the therapist con trol of the situation. In political jargon, he, at best, could govern only via a coalition and not with a majority and hence the necessity of compromise. The logical conclusion that can be derived from the foregoing argument is that the therapist can only achieve the status of "governor" of the system when he is in what Caplow terms a type 4 or type 6 triadic structure, 40 [ A>(B+C), B=C and A>B>C, A>(B+C)], where he is invulner able to revolutionary coalitions. This position will be referred as "high counselor potency" in this study. Those triadic structures in which he (she) is vulnerable to revolutionary coalitions will be referred to as "low counselor potency." These structures are specifically Caplow's types 1 [A=B=C], 2 [A>B, B=C, A<(B+C)], 3 [A<B, B=C] and 5 [A>B>C, A<(B+C)]. The two remaining triads are those in which the counselor would be equal to the combined weight of the husband and wife, i.e., Caplow's types 7 A>B>C, A= (B+C) and 8 [A= (B+C] B=C] and these are referred to as "transitory or unstable" since it is hypo thesized that they will drift into one of the other two categories. Resources In a sense the above discussion has begged the question by elucidating the types of structures in which the therapist will or will not have control, without answering what are the relevant resources or weight that determine his likelihood of having control. It is here that Erving Goffman's views, within the symbolic interac- tionist perspective, have considerable relevance. Goff- man (1959 p. 31) states that "regardless of the particu lar objective which the individual has in mind and of his 41 motive for having this objective, it will be in his inter ests to control the conduct of others, especially their re sponsive treatment of him." This "control", Goffman asserts, is achieved largely by influencing the definition of the situation which the others come to formulate. It is information about an individual that is critical to the definition of a social situation and en ables others to anticipate what to expect and thus "in forms" others in the group how best to act in order to call forth a desired response. Goffman implies that the underlying theme of all interaction is the desire of each participant to direct and control the responses made to him by others. There is an obvious consistency among this view, Caplow's assumptions within the coalition framework concerning control, and Haley's views regarding the battle for control in family therapy. Critical to the present discussion is Goffman's assertion that the information a person initially displays to others is the crucial infor mation in defining the situation. Later information is seen as merely modifying or adding to the initial informa tion. Goffman (1959 p. 11) , utilizing the work situation of service type occupations states that work adjustment will "often hinge upon the capacity to seize and hold the initiative in the service relation, a capacity that will 42 require subtle aggressiveness of the server when he is of lower socio-economic status than his clients." Although this last condition may not be generally applicable to marriage counselors, it nevertheless appears valid to ab stract the general principle that taking the initiative is the critical factor in determining the definition of the situation. In relation to psychotherapy, Haley (1973) says that "therapy can be called strategic if the clinician initiates what happens in therapy. That is when the therapist identifies the problem, sets goals, designs in tervention to achieve those goals, examines the responses he receives to correct his approach and ultimately exam ines the outcome of therapy to see if it has been effec tive." In short, the therapist implements the official programme. Haley (1973) further claims that under the in fluence of psychoanalysis, Rogerian therapy and psychody namic theory in general the idea developed that the person seeking help should determine what happens in the thera peutic session. He sees this historically determined pas sive approach as resulting in a loss for the clinical pro fession of many effective therapeutic strategies. What emerges from this discussion of Goffman and Haley is the critical role that taking the initiative and being assertive plays in the establishment of the defini 43 tion and ultimately of what Goffman would term the "work ing consensus." Thus, it can be seen that the degree of assertiveness is central to who establishes control and thus by definition it becomes a central ingredient in what constitutes a relevant resource in counseling. Some em pirical evidence for this assertion exists. For example, a study by Shapiro and Budman (1973) focusing on defection, continuation, and termination in both individual and family therapy, found that 70% of families who continued in treat ment i.e., for more than three sessions "lauded what they perceived as therapists high level of activity." Converse ly 13 out of 21 respondents remarked on the therapists lack of activity as a major reason for discontinuing therapy. Follow up comments obtained by Budman and Shapiro such as "they should have spoken more" or "he just listened, he didn’t say much" provide apposite examples of clients' objections to this passive attitude on the therapist's part. A further finding from the Shapiro and Budman (1973) study which is relevant to the present one, con cerns the difference between terminators from individual therapy and those from family therapy. The former group of terminators were more concerned with what they inferred as therapists lack of empathy or concern, while the latter group were more concerned with how active the therapist 44 was. This would appear to lend considerable support to Haley's assertion concerning the need for the therapist to gain control and establish the limits of behaviour in the session. Shapiro and Budman (1973, p. 65) conclude from their study that the "majority of respondents seemed to expect the interviewer to initiate activity and direct the meeting." They further state that "a practical conclusion from our findings is that the family therapist has to fore go or abandon the silent listening disposition of the typical individual therapist if he is to have success with families in treatment." It seems legitimate to state that the foregoing conclusion is equally applicable to conjoint marriage counseling. It also seems reasonable to consider that what terminating clients in this study were complaining about was that they, and not the therapist, dominated the session. If this is a valid point, then one can conclude that dominance as a personality trait is a relevant re source in the therapeutic session. In this sense talking a lot is seen as the behavioral manifestation of dominance and it was the lack of therapist talking to which the ter minators were objecting. Although the whole area of personality traits of effective therapists is beyond the scope of this study, a brief discussion is warranted, in so far as it is relevant 45 to establishing what constitutes a resource. A comprehen sive review of the experimental literature on therapists interpersonal skills as they relate to process and outcome has been provided by Truax and Mitchell (1971). A detail ed review of these findings is not relevant to the present discussion but the general content of the studies is per tinent here. Essentially the research literature exhibits an almost exclusive concern with four basic variables: therapists non-possessive warmth, congruence, genuineness and empathy. These in the main reflect a legacy to the pioneering psychotherapy research of Carl Rogers. That such variables are important appears to have been reason ably well established particularly by the work of Truax and his colleagues. However, what is noticeable in all this research is the almost total absence of focus on action- oriented variables such as dominance and assertiveness. It can be argued that the research into psycho therapy to date reflects the operation of what Gouldner (1970) would term "domain assumptions" concerning the necessary and sufficient ingredients for personal growth. Such, "domain assumptions" appear to be implicit throughout many prevailing theories of counseling, epitomized in the nondirective approach. Further, they appear to have played a systematic and determinative role in what variables have been chosen for experimental investigation. A few re- 46 searchers such as Plummer (1966), Swensen (1967) and Bergin (1966) have investigated the question of dominance. Bergin (1966) for example, using E.P.P.S. profiles found a posi tive relationship beyond the .05 level between dominance and change and effective therapists. Phillips (1970) bas ing his argument on Leary's (1957) reduction of the person ality to two primary continua, dominance and love, suggests that in the "interpersonal interacts of the typical coun seling situation, the counselor is expected to be a re source person, a leader." In the same article Phillips (1970) using M.M.P.I. profiles of 149 marriage counselors, 6 provides some evidence for the necessity of effective counselors to be above average on both dominance and warm th. However, compared to the plethora of research on the warmth type of variables these few studies on dominance and assertiveness are relatively insignificant and there appears to have been a systematic exclusion of such vari ables from the research field. At the theoretical level, Talcott Parsons (1955) has developed a paradigm of psychotherapy which is related to his broader theoretical views of the processes of socialization and social control. A brief discussion of this paradigm will serve to highlight how the existing re search has concentrated on only one half of the psycho therapeutic process. Using the basic A.G.I.L. paradigm 47 developed by Bales and his associates (1953) from small group studies, Parsons (1955) argues that psychotherapy is the A.G.I.L. process in reverse. The basic A.G.I.L. paradigm refers to the phases of development which a task oriented group goes through. These are, adaptive-instrumental, goal gratification or consummatory phase, intergrative and latency (Parsons and Bales 1955). Parsons (1955 p. 38) argues that both therapy and socialization involve the same basic phases as task performance, but in the reverse order. The overall pro cess can be seen as one in which solidarity is built up during the permissive (L) and supportive (I) phases which facilitates the successful implementation of the phases concerned with the actual modification of behaviour, that is, phases G and A. What is particularly relevant to the present discussion is that these phases can be conceived of as incorporating two basic therapeutic orientations, passi vity and activity. The first two phases in which permis siveness and support are central, can be seen as being systematically related to such therapist traits as empathy, congruence, non-possessive warmth and genuineness. In short, the variables which have received the most research attention. The third and fourth phases specifically demand action-orientated behaviour, directed toward establishing control. The relevant variables here would appear to be 48 dominance, initiative and assertiveness. Incomplete as this discussion is, it nevertheless highlights the convergence between Parsons' paradigm of psychotherapy, particularly phases G and A, and the theor etical views of Haley concerning the importance of thera pist control in psychotherapy. As has already been indi cated the present study is specifically concerned with this question of control. From a theoretical perspective and from the limited empirical evidence, it seems reason able to conlcude that dominance and assertiveness con stitute a relevant resource within this context. This is not meant to imply that such variables as warmth and con gruence are irrelevant; but merely to assert that they are not central when the focus is on control in the therapeu tic session. The present study, however, is not solely concern ed with the therapist in the conjoint setting. Rather is it more concerned with the broader question of coalition formation involving the three persons in conjoint coun seling. So although the question of dominance and as sertiveness as a relevant resource for effective thera pists has been established, this now has to be placed within the conceptual framework of the triad. If these traits are in fact resources, then this is because they facilitate the counselors ability to govern the session. 49 But the critical point here is "the ability to govern or control" and by extension of the evidence one can conclude that clients possessing above average degrees of these traits would also have relevant resources to effect control over the situation and thus impose their private or unoffi cial program. What is crucial if one is to predict the likelihood of coalitions or not in conjoint counseling, is the relative distribution of the resources possessed by each member of the triad. The pragmatic problem of what constitutes more than and less than will be discussed in the next chapter on experimental design and measurement. For the present purposes, the discussion will confine it self to the conceptual level, from which a theoretical model of coalition formation will be developed. It is this model which forms the basis of the hypotheses to be tested in this study. Before the model can be developed it is necessary to attend to the final two concepts in coalition theory, that is payoffs and the cheapest winning coalition. The relevance of these concepts falls into the realm of specu lative theorizing and will not be tested empirically in this study. They are rather introduced as explanatory aids in the development of the conceptual model. 50 Payoffs Stryker (1972) from a symbolic interactionist per spective conceptualizes coalitions as representing shared definitions of the situation. In his words (1971 p. 375) "bargaining ends when members of a projected winning coali tion come to share definitions." At a strictly speculative level, it can be suggested that couples seeking marriage counseling come with predetermined definitions of what is the cause of the current marital conflict. The fact that conflict exists provides sufficient grounds for one to as sume that they do not share a common definition of causa tion. Conjoint counseling can then be seen by each spouse as the opportunity to gain credence for their own causal assumptions. The payoff then becomes the absolving of guilt and responsibility for the problem. A coalition based on a shared definition of causa tion provides the situation in which blame can be placed upon the person not in the coalition. In pragmatic terms, in a coalition between the counselor and the wife, it is being hypothesized that the payoff is that the wife is relieved of the responsibility for the problem and hence the necessity for change. In a coalition between the hus band and wife against the counselor, the couple can place responsibility for change upon the counselor and thus ab solve and protect themselves from actually having to cope 51 with change. Hence within this study, payoffs will be con ceptualized in this speculative manner as the removal of responsibility and guilt for the marital problem(s). Haley (1962) provides some support for this view when he states "the strain on the therapist as he is pulled in several directions at once in family therapy is compounded by the determination of family members to prove their inno cence to the therapist." Cheapest Winning Coalition As discussed in Chapter I, this concept can be adequately handled by the social exchange theoretical framework of Homans (1961) and Blau (1967). In this sense, the cheapest winning coalition would be that coalition in which costs are kept to a minimum and just sufficient to ensure winning. Consistent with the discussion concerning payoffs it can be speculated that the cheapest winning coalition is that one which requires the least change in the relevant members causal assumptions concerning the marital problem. Bales (1970) has developed a three di mensional spatial model of group structure which includes a tentative theory of coalition formation. Although he is mainly concerned with group roles and value directions, his views are relevant to the present discussion. Coali tions, for Bales, form between people having similar value 52 directions. He asserts (1970 p. 37) "that individual mem bers prefer to minimize the amount of degree of change they need to make in their particular value direction." It ap pears reasonable to suggest that this view is applicable to the above discussion concerning unwillingness to change causal assumptions. Thus the cheapest winning coalition is that one which requires the least change, as change is con ceived of as costly since it involves loss of face. Summary This chapter began with the goal of establishing the applicability of concepts, propositions and empirical evidence of coalition therapy for conjoint marriage coun seling. It was established that the decision or problem atic issue to be resolved was the gaining of control of the therapeutic session. In Haley's terms, the establish ment of who sets the limits of behaviour in the conjoint session. Using the Parsonian paradigm it was argued that the existing research had concentrated on only one half of the therapeutic process, that of establishing solidarity and that such variables as therapists dominance had been ignored. From this it was argued that dominance and as sertiveness were the relevant resources when the focus is on the question of therapist's control and modification of clients' behaviour. Some limited empirical evidence that 53 FIGURE 1 TRIADIC POWER CONFIGURATIONS IN CONJOINT COUNSELING TYPE I HIGH COUNSELOR POTENCY C>(H+W) ("invulnerable to Revolutionary Coalitions) H C>(H+W)* H = W (4) H C>(H+W) W H C>(H+W) W H<W (6a) H>W (6b) TYPE II LOW COUNSELOR POTENCY C<(H+W) (Vulnerable to Revolutionary Coalitions) H C=H=W W H C<H (1) H=W (3) W H O H W H C>H>W H=W C<(H+W) (2) C<(H+W) (5a) W H W C>W>H C<(H+W) (5b) TYPE III TRANSITORY, UNSTABLE C=(H+W) H C>H>W C= (H+W) (7a) W H C>W>H C=(H+W) (7b) W H C>H H=W C= (H+W) (8) W * (Adapted from Caplow 1968— Numbers in parentheses refer to original typology). 54 related to dominance was presented and this evidence was then extended as applying to all members of the triad. Further, it was argued that the payoff for a winning coali tion is the sharing of a definition of cause and this ab solved the members of the winning coalition from the re sponsibility for the problem and the necessity to change. Finally the dynamics of the cheapest winning coalition were discussed in exchange theory terms as that coalition in which costs are kept to a minimum and are just sufficient to win. A Model of Coalition Formation in Conjoint Marriage Counseling The final section of this chapter is concerned with developing a conceptual model along systems theory lines, that adequately incorporates these concepts. Figure 1 de picts the eight basic triadic structures divided into the three major classifications discussed previously. It can be seen that the key classificatory variable is the coun selors possession of resources, which can be conceptualized as power. While discrepancies in the distribution of re sources between the husband and wife are of some interest, it is the power relationship between the counselor and the client couple as a unit that is the area of major focus. Type I High Counselor Potency In this set of triadic structures the counselor is invulnerable to revolutionary coalitions. This set in cludes Caplowk (1968) types 4 and 6 and as already estab lished, a coalition between the two less powerful members within this structure will not result in control of the situation. It is within this type I structure of high po tency, that the counselor can exert considerable influence and thus establish control of the therapeutic session. It is his definition of the situation which will prevail, since he establishes the limits of behaviour within the triad. Within Caplow's framework, when a counselor is in a type I set, of high potency, he is able to implement the official programme and thus circumvent attempts by the client to activate the unofficial programme, resulting in therapeutic resistance. Consistent with the previous dis cussion, it is within this structure that we would expect the counselor to be substantially more dominant and warm than either of his clients. Type II Low Counselor Potency This set includes all those triadic structures in which revolutionary coalitions are available. It is with in these structures, as Gamson (1961) has elucidated, that 56 revolutionary coalitions represent the cheapest winning coalitions and are therefore the most probable coalitions. Hence any coalition formed within this low counselor poten cy structure, is likely to result in the implementation of the unofficial program and consequently the subversion of therapeutic progress. It has already been argued, in ac cordance with both Caplow (1968) and Haley (1962), that members of a triad seek control and within this low counse lor potency set, such control can only be achieved via a coalition. Since these coalitions, by definition, will be revolutionary ones, then it can be concluded that this structure is an unsatisfactory one within which to attempt conjoint marriage counseling. This is paramount to saying that regardless of what therapeutic technique is used, whether it be behaviour modification, psychoanalysis, transactional analysis, or whatever, the counselor will be unable to implement an effective therapeutic relationship within the confines of these type II triadic structures. Type III Transitory, Unstable In this set, the counselor is equally matched by the pair. It is assumed that this will result in a stand off, which is regarded as temporary until one or other party can successfully wrest control of the situation. This statement regarding the temporary nature of the stand- FIGURE 2 DYNAMICS OF CASE MANAGEMENT AS A FUNCTION OF COUNSELOR POTENCY High Counselor Potency (Type I) ------ Transitory, Unstable Status (Type III) i Effective Counseling Single Sessions to Try to Eliminate Revolutionary Coalitions Low Counselor Potency (Type II) Revolutionary Coalition to Subvert Counseling Termination/Referral Appropriate Termination By Counselor Revolutionary Coalition (engi neered or spon taneous) , to Terminate by Couple i Interminable, Ineffective Counseling VI 58 off finds justification in Caplow's view, concerning the desire of each member to seek control in the situation. Thus it is further assumed that there will be a shift in this structure into either a type I or type II. The impli cations of this shift have already been discussed under the previous types. Figure 2 explicates in model form, the dynamics of case management as a function of counselor potency. The far left of the model shows the types of counselor power structures. It can be seen that high counselor potency leads to effective counseling and ultimately to appropriate termination. Appropriate here is taken to mean that previ ously set therapeutic goals have been achieved. The termi nation can be initiated by either the counselor, the couple or both. Within coalition theory terms a revolution ary coalition either engineered or spontaneously arrived at, would not reflect subversion of the official programme at this stage of contact. This is so, since it simply reflects that counseling has been effective in so far as the couple have re-established their relationship and no longer need the counselor to set the limits on their be haviour. It is both possible and probable, that within this type I set, that the counselor will form alternating coalitions with the spouses as part of his effective thera py. Zuk (1971) refers to this technique as the "go be- 59 tween process," aimed at breaking up pathogenic relation ships . The critical point of the above analysis, however, is that his counseling technique will only be effective within this particular set, for two reasons. Firstly, the counselor initiates and controls the coalition formation and does not need an alliance in order to control the situ ation. Secondly, within this set, he is invulnerable to revolutionary coalitions and thus any coalition will not result in subversion of the official program. Thus one would expect that if coalitions are formed within this type I set, no one type of coalition will predominate, since the counselor will be concerned with balancing the triadic structure. The flow from the type II set, of low counselor potency, consistent with the previous discussion, indicates the formation of a revolutionary coalition, resulting in subversion of therapy or in more traditional terms, thera peutic resistance. The end result of this is shown as interminable or ineffective counseling. However, the same figure indicates that two case management strategies are available to the counselor to avoid this end state. The first strategy, which can be seen as the preferable one within the terms of effective counseling, is for the coun selor to utilize single sessions to try to eliminate the 60 revolutionary coalitions. By temporarily declining to work in a conjoint structure, it is assumed that the counselor can work toward re-establishing his control of the situa tion, redefining it, and examining the basis of the revo lutionary coalition. If this move is successful, then as the model indicates, the counselor will establish himself in a type I structure of high potency. In short, he as serts his position as governor of the system. However the model also indicates that this strategy may fail, pre sumably for a variety of reasons such as poor timing and ineffective technique. If this should occur then a return to the type II triadic structure is anticipated. It is possible that a different revolutionary coalition may be in operation at this stage of return to type II structure, but the crucial point is that it is still a revolutionary coalition. For example, a counselor, faced with a coalition between himself and the wife, may choose the strategy of single sessions to re-establish his concern for the husband and in the process find himself forming a more positive relationship with this spouse. A return to the triadic structure of a conjoint interview will then result in a husband/counselor coalition, rather than the counselor/wife. This cycle could repeat itself intermin ably resulting in ineffective counseling. The second strategy is for the counselor to termin- 61 ate counseling prematurely or to make a referral to another! counselor or mental health agency. This strategy is seen as less satisfactory, since although it indubitably re solves the coalition, it does so at the expense of thera peutic improvement and as such represents ineffective therapy. It seems feasible to suggest that such a strategy will be motivated by the counselor's own sense of frustra tion or inadequacy and not based on well articulated diag nosis and treatment plan. Essentially, it represents in adequate case management and possibly accounts for those counseling clients who have a long history of seeking therapeutic help. Such premature termination could also be implemented by the clients and these then become what are colloquially referred to as therapeutic "drop outs." This particular phenomena will form a central part of the following discussion. Because the existence of revolutionary coalitions is seen as critical to the outcome of counseling, it ap pears warranted that additional attention be focussed on this type II set of low counselor potency. Figure 3 is a detailed systems model of the flow resulting from those triads in which revolutionary coalitions are available. It is in discussing these that exchange theory becomes par ticularly useful in so far as it provides the theoretical mechanism of the model. It further provides the rationale FIGURE 3 DYNAMICS OF REVOLUTIONARY COALITIONS AS A FUNCTION OF LOW COUNSELOR POTENCY Type II Power Configurations (3) (1) (2) (5) C=H=W C< (H+W) C>H>W or C>W>H C< (H+W) CH Coalition Exit to Single Interviews CW Coalition i HW Coalition ♦Figures in parentheses refer to Caplow's (1968) original Typology. One Spouse Feeling Rejected * Counselor Ineffective, Feeling Inadequate Rejected Spouse Withdraws Rejected Spouse Forms Coalition Counselor Forms Co alition or Existing one Continues Both spouses or counselor withdraw. r Exit to Single Interviews H J a o h i - a > a h g t o w H a for the cheapest winning coalition concept. The far left of the model depicts the four relevant triads, with Caplow's original figures in parentheses. The first triad (Caplow's type 3) is the situation in which the counselor is the least powerful person in terms of the relative distribution of resources. The hus band and wife are shown as being equal to each other in power. Consistent with previous theory, the model indi cates that this structure will lead to either a coalition between the counselor and wife, or the counselor and the husband. The latter, by forming a coalition with the counselor is able to assert control over the situation and consequently have his definition of cause, at least par tially accepted as the working definition. In short, he is able to dominate his wife, with whom he is, without the coalition, equal in power. The reverse situation obviously holds true for the wife. Thus it is the counselor in this situation who will be sought after as a potential coalition partner, since he represents the cheapest winning coali tion. Simmel (Wolff, 1958, p. 154) would consider the counselor in this structure to be in the "tertius gaudens" or "enjoying third" role, since he holds the critical bal ance of power. Consistent with exchange theory it can be suggested that both spouses would attempt to make themselves "at- 64 tractive" to the counselor as a potential coalition part ner. "Attractive" in the sense that a coalition would be perceived by the counselor as one in which rewards were high and costs low. Thus before this second stage of the model is reached there will be both verbal and non verbal maneuvers to seduce the counselor into a coalition. These may take the form of readily agreeing with the counselor, supporting him, being exceedingly co-operative, in essence building up the counselors self esteem. In terms of the operational measures to be used in this study, which will be discussed in detail in the next chapter, this behaviour would show up as a high frequency of positive-emotional acts directed toward the counselor. Alternative strate gies on the client's part, such as appeals to the counsel or's own values in an attempt to establish solidarity with the counselor may also be used. Undoubtedly there are innumerable ploys that could be utilized by clients at this stage, but the present study is more concerned with the end result of such ploys rather than with their variety. The essential point is that they are all seen as being specifically directed to ward enhancing the client's attractiveness as a coalition partner for the counselor. The second stage of the model depicts the point at which a coalition is formed. This is when the problematic issue of who is going to assert control is established and the payoff for the partner in the winning coalition is the absolving of responsibility for the marital problem, and justification for their own behaviour. Over a period of a therapeutic session this would manifest itself in this particular triad as a re ciprocal positive emotionally supportative relationship between the coalition partners. To reiterate, the model indicates that at the theoretical level this coalition would be either a counselor/wife or counselor/husband coalition. Reference to Figure 3 indicates that the formation of a coalition between either spouse and the counselor will lead to feelings of rejection for the isolated spouse. It is anticipated that this will be reflected behaviorally by the isolated spouse becoming increasingly negativistic toward the counseling process. One would expect such be haviour as overt agression toward the coalition members, or perhaps more covert expressions such as frequent postpone ments, late for appointments and generally resistant and non cooperative behaviour. The model, as discussed previously, indicates that the counselor can circumvent this increasing negativism by exiting to single sessions and attempting to re-establish his control of the situation and to reinforce the official program. By adopting this strategy a positive feedback 66 loop can be avoided, since the single sessions provide a structural solution to this anticipated spiraling effect. It also temporarily removes the counselor from the clients struggles with each other and consequently their struggles with the counselor over who is to set the limits on be haviour. This is seen as a temporary strategy to resolve the negative therapeutic effects of the coalition. If successfully handled the counselor would return to the conjoint session with himself in type I structure of high potency and invulnerable to revolutionary coalitions. Failure to adopt this strategy as the model sug gests, will result in the isolated spouse doing one of three things. He (or she) could sabotage the marriage counseling process by physically withdrawing and refusing to attend (exit 2). If he (or she) perceives the cost of such a strategy as being too high, for example, likely to result in a complete breakdown of the marriage, then he (or she) could choose the second alternative. This is to attempt to form a coalition with one of the two members in the existing coalition. Consistent with the previous discussion on exchange theory, this would require the iso lated spouse to make himself (or herself) "attractive" as a potential coalition partner. Several strategies aimed at achieving this coali tion objective can be speculated. The rejected spouse may 67 increase his (or her) frequency of positive-emotional acts towards the person he (or she) selects as a possible partner. Within this structure of equal weight between the husband and wife the perceived possible partner is likely to be the counselor. Therefore he (or she) may be come increasingly cooperative toward the counselor by such maneuvers as psuedo-insights, thereby increasing his positive reinforcement potential for the counselor. Al ternatively, or perhaps simultaneously, he (or she) could devalue the spouses attractiveness, by introducing negative information about him or her. Further, he (or she) could increase the cost of maintianing the coalition for the counselor by accusing him of taking sides and thus gener ating discomfort in the counselor to the extent that con tinuing the coalition is perceived as too costly. If the isolated spouse is successful in any of these attempts, then the model, as shown in Figure 3, indicates a return to the second stage and the formation of the alternative spouse/counselor coalition. This, as shown in Figure 2, will lead to ineffective and often interminable counseling.; It is possible that the isolated spouse could decide that a coalition with the marital partner was the least costly and work at establishing this. One possible strategy is to increase the costs for the spouse of main taining the coalition, by threatening not only to withdraw 68 from counseling, but also to withdraw from the marriage. This would present the partner in the coalition with the choice of continuing counseling . < or the marriage and hence the former alternative may be perceived as too costly. Theoretical validation for this possible outcome can be found in the previously quoted statement of Bales (1970, p. 37) where he asserts that individuals prefer to minimize the amount of change they need to make in their particular value direction. One can utilize this Bales principle to suggest that in those situations in which maintaining the marriage may be a stronger value than continuing counseling, the latter alternative would be too costly. Conversely, it could be argued that this strategy would not be successful in those marital situations where the spouse in the coali tion did not value the continuance of the marriage. Howe- ever, if these attempts to establish a husband/wife coali tion are successful then the resulting phenomena is what has traditionally been referred to as resistance. The successful implementation of a husband/wife coalition is depicted in the model as leading to a return to stage two, but this time with a different but still revolutionary coalition, i.e., Caplow's type 5. Finally stagnation in the counseling process would result if the rejected spouse perceived all of the 69 alternatives spelled out in the far right of the model as non-viable. This would result if (s)he felt that with drawal from counseling was too costly and further that (s)he had insufficient resources to make himself or herself attractive as a potential coalition partner. Mills (1954) in a study of coalition patterns in experimentally contrived three person groups of students, provides some insightful data and conclusions concerning the likelihood of coalitions changing. His experimental model was designed to test the persistence of coalition structures and to specify the conditions under which the structure is more persistent. Specifically he used a switching of allegiance technique where a pre-arranged coalition was experimentally disolved to test the effects of this defection on continuing allegiance. Although his findings tended not to support his underlying assumptions, his general conclusion is relevant to the present discus sion. Contrary to Mills' expectations he found that per sons experiencing the greatest loss from a change of alli ance resulting in their becoming isolates, i.e., the anx ious, insecure person, defended themselves, rather than capitulating to group demands and set themselves off from the group. Conversely, the less threatened persons ap parently accepted the reality of change produced by the switched coalition and adjusted to it by altering opinions 70 and feelings to fit in with the rest of the group members. Mills (1954 p. 666) concludes by tentatively sug gesting that "one important personality variable affecting the persistence of the coalition structure is the level of anxiety that isolation generates; the higher it is, the more persistant the structure, the lower, the more likely is the coalition structure to dissolve." From this study of Mills we can reasonably conclude that if rejection is accompanied by high anxiety then the isolated partner will not be inclined to exercise any of the alternatives shown on the far left of the model. In fact, counseling will simply stagnate as the existing coalition will continue with no attempt by the isolate to engineer an alternative coalition or to dissolve the existing one by threats of withdrawal. Within the type II set of low counselor potency, the two triadic types most likely to result in a revolu tionary coalition of the husband and wife against the coun selor are Caplow's original types 2 and 5. In these structures the counselor is less powerful than the com bined weight of the spouses, but more powerful than either spouse individually, (see Figure 1). Figure 3 depicts the detailed flow resulting from these structures. It is high ly probable that before a coalition is consolidated between the husband and wife that there will be psuedo attempts by 71 both spouses to form a coalition with the counselor. How ever, consistent with coalition theory, it is anticipated that these will be sporadic and that the steady state of the system will be a husband/wife coalition. Mills (1953) provides a typology for classifying individual matrixes of emotional support which is useful to introduce at this point since it gives a method of classifying the type of revolutionary coalition which is not maintained by reciprocal positive emotional support. Using Bales (1950) categories, Mills in a study of stu dents in 48 three person discussion sessions, compared in dividual matrixes of emotional support indexes with the median index for the entire smaple. Individual matrixes were then classified depending on whether they were above or below the median, into a four fold typology of solidary, contending, dominant and conflicting. The fourth type, the conflicting support type in which two individuals are in a negative reciprocal emotional relationship, is indi cated by their positive emotional support indexes both being below the median. In pragmatic terms this indicates that the coalition is being maintained by negative emotion al acts rather than positive. Mills (1953) makes the fur ther suggestion, supported by empirical evidence, that next to the solidary pattern the conflicting pattern is the most stable coalition structure. 72 This study of Mills1 provides a basis from which one can argue for the development and maintenance of a negative emotional coalition as opposed to the more tradi tional view of a coalition as necessarily being a positive ly supportative relationship. It is this type of conflic ting coalition that it is anticipated will be characteris tic of a husband/wife coalition in the types 2 and 5 tri adic structures. The behavioral manifestations in the counseling sessions will be typically resistant behaviour on the client's part towards the counselor and the domina tion of the session by the clients. Operationally, we would expect that the highest rates of participation will be between the couple, and the counselor and will be ren dered relatively impotent. It is this result that is de picted in the model as that stage when the counselor be comes increasingly ineffective, loses control of the ses sion and presumably experiences increasing feelings of inadequacy. As the model indicates the counselor can choose to exit into single sessions at this stage in order to re establish control of the counseling process, and further attempt to alter the power distribution to a type I set of high counselor potency. If the counselor does not choose this strategy then the model depicts four possible out comes. Firstly, counseling will be prematurely terminated 73 by the couple. This, as has already been mentioned, pro vides a possible explanation for client defection from counseling. Secondly, counseling may merely continue in terminably and ineffectively. Such an outcome presumably accounts for those cases that go on for months and even years without any appreciable change. Factors external to the system may play a determining role in whether the clients choose to defect or continue. For example, if the spouses are separated or seriously contemplating it at the time of seeking counseling, then it can be suggested that premature termination will result. In this instance coun seling can be seen as being subverted by the unofficial programme resulting from the revolutionary coalition be tween the spouses. The agenda of this program could be seen as centering on seeking approval and removal of guilt for a prior decision to terminate the marriage. Alterna tively, a marriage characterized by what has traditionally been referred to as neurotic collusion would be more likely to result in continuance without change. The unofficial program in this instance would be maintenance of the status quo. Regardless of the actual outcome chosen from these two possible outcomes, the payoff for the clients is seen asi the removal of pressure to change. The payoff in the second outcome is specifically the opportunity to maintain the existing symptomatic relationship. 74 ; The third possible outcome is that the counselor, under increasing pressure of feelings of inadequacy may decide to terminate the case. This is shown on the model as leading to premature termination on the counselor's part. Such a termination is seen as being motivated by the counselor's own sense of frustration when confronted with a revolutionary coalition and not based on the achievement of prior set of treatment goals. A slight variation on this strategy is for the counselor to refer his clients elsewhere and as discussed previously this possibly accounts for those clients who have a long and varied history of con^ tact with therapeutic agencies. Finally, like the rejected spouse, the counselor can attempt to break the existing coalition by increasing his attractiveness as a potential coalition partner to either of the spouses. Such a maneuver as increasing his frequency of positive emotional acts toward one spouse, represents one such strategy. However, even if the counse lor should be successful in this move, as the model indi cates, its value in terms of therapeutic effectiveness is questionable, since he is still within the confines of a type II set of low potency. This then leaves the counselor open to further revolutionary coalitions and the probabil ity of therapeutic subversion. Because of this, it is possible that the counselor will perceive termination or 75 referral as the least costly maneuver. The final type of triadic structure (Caplow's type 1) within this type II set of low counselor potency, is that structure in which all members are equal in power. Thus by definition, any coalition is a revolutionary one since it results in the domination and control of the third member. All three coalitions, as the previous theory and evidence indicates, are equally likely, since all three are theoretically equal in cost. Factors extraneous to the power distribution may serve as constraints on the sys tem within this triadic type and play an important role in determining which coalition will be formed. For example, the sex of the counselor may be one such factor that would increase the probability of a particular coalition, e.g., male counselor with female client. Other factors such as the counselor's own values, needs and the compatibility of these with either one of the clients may serve to facili tate a particular coalition being formed. As the model stands at present it predicts any coalition as being equal ly probable. The formation of one predominant over the others would require a posteriori reasoning along the lines indicated above with respeet to specific constraints. The central point is that this type of structure, iUe., C=H=W, is an unsatisfactory one in which to attempt conjoint counseling, since by definition it is a revolu 76 tionary one. Such a structure highlights the point made by Haley that the counselor must be assertive, take the ini tiative and establish control of the therapeutic situation. If he fails to do this then the outcome of counseling with in this triadic type is likely to follow any of the pre viously indicated ones for revolutionary coalitions. It may well be that when confronted with this structure, where all members are equal, that the counselor could avoid revolutionary coalitions and consequently subversion of therapy, by initially working in single sessions with the aim if establishing himself in a type I structure of high potency. This chapter has now established the relevance of coalition theory for marriage counseling and on the basis of this a conceptual model, along systems lines was de veloped. The next logical step in the process of theory construction is to deduce from the conceptual model em pirically testable propositions in the form of hypotheses. Such hypotheses have been implicit throughout the discus sion in this chapter, the task now remains to state them explicitly. Due to the relatively small number of subjects only the major hypotheses 1, 2, 3 and 4 will be investiga ted. The sub-hypotheses will be excluded from the present study. Hypotheses Conjoint sessions in which the counselor is in a type I set of high potency, will not exhibit any particular coalition pattern. Conjoint sessions in which the counselor is in a type II set of low potency will exhibit a particu lar coalition pattern. (a) Those conjoint sessions in which the counselor is in a type 1 triadic structure, will exhibit as the predominant coalition pattern, either a husband/wife, husband/counselor or a coun selor/wife coalition. (b) Those conjoint sessions in which the counselor is in a type 2 or type 5 triadic structure, will exhibit as the predominant coalition pattern, a husband/wife coalition. (c) Those conjoint sessions in which the counselor is in a type 3 triadic structure will exhibit as the predominant pattern a counselor/husband or counselor/wife coalition. Conjoint sessions in which there is no predominant coalition pattern will tend to result in effective counseling and appropriate termination. Conjoint sessions in which there is a predominant coalition pattern will tend to result in ineffective 78 counseling and inappropriate termination. (a) Those conjoint sessions in which the predomi nant coalition pattern is either counselor/ husband or counselor/wife will tend to result in ineffective counseling and/or premature termination hy the isolated spouse. (b) Those conjoint sessions in which the predomi nant coalition pattern is a husband/wife, will tend to result in continuing ineffective coun seling or premature termination by the counse lor, or by both spouses. CHAPTER III EXPERIMENTAL DESIGN AND METHODOLOGY The previous chapter established the theoretical model of the present study and from this the relevant hy potheses to be tested were derived. The next step in the process of theory building is the operationalization of the concepts to be investigated. It is this task, along with the overall experimental design, that is the focus of the present chapter. Chapter I established that the major concepts of coalition theory are resources, decision point, cheapest winning coalition and payoffs. Chapter II concentrated on establishing the relevance of these concepts for conjoint marriage counseling. It was argued that the decision, or problematic issue to be resolved in the conjoint situation,; is who is to be in control of the interaction. Dominance and assertiveness were seen as the relevant resources with in this context. A coalition was viewed as consisting of aj sharing of the definition of the situation, specifically a i consensus between two members of the triad regarding the cause of the marital difficulty which results in domination; by the pair of the session. The payoff for a winning co- 79 80 | alition was the removal of responsibility for the marital problem(s), the absolving of guilt for the problem, and the need for change. From the discussion in Chapter II and specifically the hypotheses, it can be seen that the major ! concepts to be investigated are resource as the predictor variable and coalition patterns as the dependent variable. In the second stage of the study coalition patterns become : the predictor variable and counseling outcome the dependent variable. It is the operationalization of these variables that now requires discussion. Operationalization of the Variables Power Configurations It has already been argued that dominance and as sertiveness represent the critical resources in conjoint counseling. The relative power of the counselor and spouses is viewed as the direct expression of the differ- [ ential dominance and assertiveness of the three parties. One obvious measure of relative assertiveness is the number! of assertive acts that each party initiates in the conjoint session. In the present study, however, the style of in teraction in the conjoint session is precisely what is being predicted from the various power configurations. If , ] 81 ! the independent and dependent variables are not to be oper ationally confounded then it is necessary to find a measure of the relative dominance of the three members which is independent of the interaction among them. Such a measure was obtained by administering a personality test to the individual members of the triad which purported to measure dominance and assertiveness. The specific test used was the Minnesota Multi- phasic Personality Inventory developed by Hathaway and McKinley (1951), hereafter referred to as the M.M.P.I. Although this test was designed as a measure of psychiatric symptoms it has nevertheless proven useful in delineating personality types among normals as well. The original test contained three validity scales and nine clinical scales of which the present study will utilize four plus an addition al research scale developed by Barron (1953) and referred to as ego strength. In order to enhance the predictive power of the independent variable two separate methods of measuring it were utilized one based on the work of Leary i (1957) and the other based primarily on Phillips' (1970) study of marriage counselor profiles. Leary (1957) as previously mentioned, developed a two dimensional model of personality in which he reduced the traits to two primary continua of dominance and love. In addition# he developed several levels of measuring in- terpersonal behavior, since he conceived of personality as "a multilevel pattern of interpersonal responses (overt, conscious or private)." (Leary 1957, p. 15). Specifically the levels are level 1, which is objective or public, level 2 which refers to conscious self description or self per ception and level 3 which consists of private symbolism and; specifically refers to the unconscious level of personali ty. Level 1 behavior according to Leary (1957), can be satisfactorily measured by the M.M.P.I., since it measures the symptomatic impact of the patient(s) upon the clini cian. Although no doubt a substantial argument could be put forward contesting the so called objectivity of the M.M.P.I., the present study accepts Leary's position that the test is a valid measure of public behavior and endorses his view that "every psychological symptom seems to have an; interpersonal meaning, i.e., implications as to what the patient is communicating through the symptom, and what the patient expects to be done about it etc." (Leary 1957 p. 107). Using this orientation Leary converted the M.M.P.I. into an instrument for predicting interpersonal behavior and specifically developed a set of indices for predicting dominant or submissive interpersonal behavior. The in tended use of this measure in the present study is exactly in line with Leary's views of its appropriate use. He asserts (1956 p. 443) that at the time of "intake evalua 83 tion the symptomatic pressure exerted by the patient is of crucial importance in planning a therapeutic program." If I one accepts Haley's (1963 p. 19) view that "the primary gain of symptomatic behavior in a relationship could be said to be the advantage of setting rules for that rela tionship," then utilizing a symptomatic measure is consis tent with the central question of this study concerning the; issue of control in conjoint counseling i.e., who sets the rules. The Leary index of dominance then is the measure which will be used to provide an independent predictive measure of the ability of each member of the triad to achieve control of the conjoint session. The specifics of how Leary arrived at this formula are adequately discussed in his major work (Leary 1957) and the present discussion will confine itself to the final formulation. The method uses four of the M.M.P.I. scales, Ma, D, Hs and Pt. The Mai or hypomania scale, purports to measure the personality factor characteristic of persons with marked overproductiv ity in thought and behavior. The D, or depression scale, measures the symptom complex, depression. Hs, or hypochon driasis scale, is a measure of abnormal concern about bodi-; ly function and the Pt, or psychasthenia scale, measures the similarity of the subject to psychiatric patients who are troubled by phobias or compulsive behavior. Leary's system involves subtracting the M.M.P.I. standard scores of 84 Ma from D and Hs from Pt and summing the result (all M.M.P.I. scores are K corrected for social desirability of responses). The rationale, is as follows (Leary 1957 p. 440) : If Ma>D, a+ score results. This indicates that interpersonal strength, assertion and confidence are emphasized. If Ma<D, the opposite is indicated. Weakness, immobilization and lack of con fidence are suggested. If Hs>Pt, a+ score results. The subject seems to be indicating that his physical health concerns him more than emotional worries. This is the wounded warrior theme often expressed by psychosomatic patients. The subject admits to some bodily weakness, but emotional strength is by comparison stronger. Hs<Pt, the opposite is true. The subject is more concerned with his emotional problems and is emphasiz ing fears, worries or immobiliza tion. The two scores provide an arithmetical summary point, which is then converted to a standard score by referring to a table developed by Leary based oh a sample of 787 psychi atric patients. This then provides a standard dominance score within the context of Leary's scheme. The problem for the present study was to decide on what constituted high and what constituted low. Basing the! argument on standard scores it was decided that dominance Hypomania— Depression: Hypochondriasis— Psychasthenia: scores above one standard deviation i.e., standard scores of 60+, would be considered high. Ideally then, a standard score below 40 i.e., below one standard deviation, ought to be classified low. However, as Phillips (1970) points out,! "typically there are few low scores on the M.M.P.I. i.e., below a T score of 45." Because of this it was decided that a standard score of 49 or less would be considered low. Hence, a standard score within the range of 50- 59 was classified as medium. By using this method each indi vidual member of a triad could be classified in terms of their degree of dominance. As previously mentioned, Barron's 1953 ego strength research scale was utilized in this study as an additional predictor variable of power. The inclusion of this scale is based more on intuitive grounds than either theoretical or empirical. However, because it purports to measure feelings of self confidence, alertness, adventurousness, determination, etc., it is reasonable to suggest that it would measure behavior which is consistent with dominance and assertiveness. Phillips (1970) provides some empirical! justification for its inclusion. He states on the basis of! his study, that "ego strength, along with warmth are the irreducible minima for personality traits of successful marriage counselors." In his conclusion, based on M.M.P.I. profiles of successful counselors, he states that "it is 86 ! hypothesized that ego strength should be the highest or second highest scale on the profile of effective marriage counselors." Because effective marriage counseling within ! the confines of the conceptual model of this study is seen as occurring when the counselor is invulnerable to revolu- : tionary coalitions i.e., in a type I structure of high po- ; tency, it seems logical to include the ego strength measure as a predictor variable. The question of what constitutes high or low was solved in the same manner as for the domi nance scores. Specifically, a standard ego strength score of 60 plus was classified as high, 50 - 59 as medium, and 49 or less as low. Utilizing the preceeding discussion, Table V depicts in tabular form the basis of the operational defi nitions of type I and type II counselor potency structures.^ It is important to point out that the classification is not; dealing with precise quantitative dimensions and it is not i being implied that a "high" equals in any precise sense two "lows" etc. An attempt merely is being made to opera- j tionally establish the relationship between the predictor ! variables of the present study and Caplow's (1968) original! triadic power configurations. Because of the inexact nature of the classification; scheme, certain assumptions had to be made regarding the type III or transitory set. Consistent with the theoreti- 87 TABLE V OPERATIONAL TRIADIC POWER STRUCTURES TYPE I HIGH COUNSELOR POTENCY C>(H+W) Counselor Husband Wife Triadic Structure (4) High Low Low High Med Med Triadic Structure (6) High Med Low High Low Med TYPE II LOW COUNSELOR POTENCY C<(H+W) Counselor Husband Wife Triadic Structure (1) High High High Med Med Med Low Low Low Triadic Structure (3) Low High High Low Med Med Low Med High Low High Med Med High High Med High Med Med Med High Triadic Structure (2)&(5) High Med High High High Med High High Low High Low High Med Med Low Med Low Med Med Low Low cal point put forward in Chapter II, it was assumed that this type III structure was basically unstable and would drift into either a type I or type II structure. For the purpose of empirical convenience it was assumed that in the counselor high and both clients medium situation, which intuitively appears to be close to type III structure, that this would drift into a type I of high counselor potency. Alternatively with the other triadic situation which ap pears to fit the type III set, i.e., counselor medium and both clients low, it was assumed that this would drift into a type II structure of low potency. The rationale behind these assumptions is that the counselor being high, consis tent with the previous discussion, would be able to assert sufficient control to avoid a revolutionary coalition. Conversely, with a medium level of dominance or ego strength it was assumed that this was less likely. A further modification contained in Table V of the ; original classification shown in Figure 1, is that, where the predicted coalition is the same these triadic struc tures were bracketed together. This specifically applies to triadic types 4 and 6 in the high counselor potency set ; and types 2 and 5 in the low counselor potency set. The final modification of the original classifica- I tion scheme concerns the type 3 triadic structure. The original indicated that a type 3 structure existed when the 89 counselor was less than the clients who were equal to each other. Because of the inexact nature of the classification method it was decided that in those situations where the counselor was either medium or low and one or both of the clients were higher than him that this would constitute a type 3 triad. Consistent with the original theory the counselor in this modified type 3 would still represent the cheapest winning coalition and hence be the sought after coalition partner. It is indisputable that this entire classification scheme is arbitrary, perhaps with the excep tion of the type 1 triad. However, because the present study is exploratory there appeared to be no alternative. The empirical data are seen as the means by which the va lidity of the arbitrariness can be assessed, and hence as indicating where necessary changes need to be made. Coalition Patterns The operationalization of this variable, in the main, follows the previous non-laboratory studies of coali tion formation, in particular, those of Mills (1953, 1954) and Strodtbeck (1954). These studies used Bales (1950) method of interaction process analysis which records the initiator, the recipient and the type of each. act. Bales (1950) scheme consists of 12 categories for coding interaction. Categories 1, 2, 3, refer to positive socio-emotional acts and responses. Categories 4, 5, and 6 refer to the task dimension of group interaction and in clude such behavior as giving information, opinions and suggestions. Categories 7, 8 and 9 also relate to instru mental or task orientated behavior and include asking for opinions, suggestions and information. The final three categories 10, 11, and 12 are concerned with negative socio-emotional acts. In the original studies conducted by Mills (1953, 1954) and replicated by Strodtbeck (1954) on families, an index of support was derived from the Bales categories and this was used as the measure of coalition formation. The index purportedly "reflects the tendency of a particular actor, number 1, to give positive responses to the attempts: at problem solutions by another actor, number 2." (Strodt-> beck 1954 p. 103). It is derived by summing the positive emotional acts originated by number 1 and directed toward number 2, and subtracting from this sum the number of nega-| tive emotional acts number 1 directs towards number 2. This forms the numerator of the index and the denominator is obtained by summing the total number of task orientated ; acts (i.e., categories 4, 5, 6, 7, 8, and 9) originated by number 2 and directed toward the group as a whole. Finally, the result is multiplied by 100 to give the index of sup 91 port for person 2 by person 1. The use of this index involves the assumption that i it measures person number 1's support of person number 21s problem solving attempts. In analyzing the data of the present study it became apparent that this assumption was at times a difficult one to make. Many of the emotional responses that were made by participants in conjoint coun seling did not necessarily reflect support for another per sons problem solving attempts. In fact, in some instances the number of socio-emotional responses were in excess of the problem solving acts. While this may reflect an arti fact of the coding, it appears equally plausible to suggest that in early conjoint sessions, particularly where a coun selor is theoretically orientated toward encouraging ex pressions of emotion, that one would expect an excess of socio-emotional responses. In an attempt to solve this measurement problem an i alternative method of measuring coalitions was devised. Essentially it consists of summing the positive emotional acts (categories 1, 2, and 3) that person 1 directs towards person 2, subtracting from this sum the total number of negative emotional acts (categories 10, 11, and 12) origi nated by person 1 and directed toward person 2, to form the numerator. The denominator consists of person number I's total number of emotional acts, ignoring whether they 9 2 ! I were positive or negative and to whom they were directed. This ratio was then multiplied by 100. The schematic form of the procedure appears below: Number of Pos. Number of Neg. emotional acts emotional acts person 1 directs MINUS person 1 directs toward person 2 toward person 2 --------------------------------------------- X 100; Total number of Emotional Acts Person 1 Initiates The procedure was then repeated for each of the dyads in a particular triad. Using this method a random check of 10 cases re vealed that in nine of the ten, the exact same coalition was obtained as that ascertained by the original method devised by Mills (1953). Since the alternative method had the two fold advantage of not having to make the assumption concerning the support for problem solving attempts and that it seemed to fit the data better, it was the alterna- j tive method that was used in this study. A further modifi-i cation that was made was that the present study used Bales' (1970) revised categories (see Appendix 1). Bales (1970, p. 477) claims that the changes to his original categories i were made "in order to increase the sensitivity of the in teraction categories to underlying emotional currents in the group process." It may well be that this provides a partial explanation as to why in some instances an excess of socio-emotional responses was obtained. This formula was applied to each possible dyad within each conjoint situation. Specifically, the counselor/wife, counselor/ husband and husband/wife. The problem that still remained was how to decide when a coalition existed or not. Again this decision had to be made in a somewhat arbitrary manner. It was assumed that if no coalition existed, using the alternative scoring method that was developed, then a support index of zero would be obtained since the number of positive acts would equal the number of negative acts. Using this ideal point as a bench mark, it was decided that a cutting point of 30% or more in either direction would indicate a dispropor-: tionate flow of emotional acts toward any one person. How ever to qualify as a coalition, as much of the previous discussion has indicated, some degree of reciprocity must exist. Hence in those particular dyads where an index of 30% either positive or negative, was reciprocated by a 30% i or more response in the same value direction, then a coali tion would be deemed to exist. This method, allows for the existence of either a positive or negative coalition and as discussed in Chapter II, it was assumed that negative or conflicting coalitions, would be characteristic of husband/l wife coalitions. When more than one coalition was found to 94 exist, then this was categorized as a "no predominant co alition pattern." In order to meet the criteria of a coalition a particular dyad had to be the sole one within the triad that manifested a reciprocal emotional interaction of 30% or more. If one member of a dyad directed in excess of 30% positive emotional acts towards the other member and this was reciprocated by an excess of 30% negative then this was not considered a coalition. In other words the sign had to be the same. Justification for the above procedure can be found at both the theoretical and empirical level. Theoreti cally, a relationship characterized by opposite signs is inconsistent with the definition of a coalition as devel oped in Chapter II of this study. Essentially, this stated that a coalition consisted of a shared definition of cau sality. At the empirical level Mills (1953) developed a typology for classifying individual matrixes of support. Two of these have already been mentioned, the solidary pat-j tern or ++ structure and the conflicting or — pattern, the other two are dominant, +-, and contending -+. Mills (1953, p. 355) concludes from his study, that the single stable pattern is the solidary one; next in stability is the conflict pattern. The more transient patterns, the dominant and the contend ing tend to shift to the conflicting. Temporally, 9 5 j i the solidary and the conflicting patterns seem to be terminal tendencies. The data of the present study confirmed this finding and not a single instance of a dominant or contending struc tures, as measured by the method used in this study, oc curred. This alone provides sufficient justification for rejecting the structure of opposite signs as indicative of a coalition. Counseling Outcome This area is perhaps one of the most difficult that has permeated the field of counseling research. The pres ent study and the attendant measurement procedures are not ; in any way an attempt to answer the broad question of ef fective therapy per se. The measurement technique utilized in the present study was specifically designed to test the outcome predicted by the conceptual model shown in Figure 3; and as stated explicitly in hypotheses 3, 4 and 5. To re iterate, these were, that when no predominant coalition was; present then effective counseling would result. Converse ly, when a counselor/husband or counselor/wife coalition was present, ineffective counseling would result. The same prediction was made for those conjoint sessions in which a husband/wife coalition was present. One modification that 96 is not stated in the relevant hypotheses, but is shown in the conceptual model, is that if a counselor implements the strategy of single interviews when confronted with a revo- ! lutionary coalition, then effective counseling will result,! if such a move is successful in overcoming the revolution ary coalition. Operationalization of this phase of the study was simply achieved by the administration of a short question- : naire approximately six to eight weeks after the initial conjoint session. The relevant questionnaire is included in Appendix 2. Because this method relies exclusively on the counselor's assessment of the effectiveness of coun seling, it is obviously biased. However the question of bias only refers to the assessment of the progress of con tinuing cases and the bulk of the data obtained by this method can be considered objective. For example, that which refers to clients who have defected from counseling does not suffer from counselor bias. The questionnaire also obtained objective information on the patterns of interviews and this data provided the opportunity to exam ine the effects of the use of single interviews to circum vent revolutionary coalitions. Table VI displays in sum mary from the measurement procedures utilized in this study. TABLE VI SUMMARY OF MEASUREMENT PROCEDURES Variable Method Operational Definition (1) Individual Dominance and Assertiveness Minnesota Multiphasic Personality Inventory (Hathaway & McKinley 1951) (a) Ego Strength (Barron, 1953): Standard Score of 60+ = high Standard Score of 50-59 = med. Standard Score of 0-49 = low (b) Dominance: Leary's (1956) formula Ma - D Hs - Pt Sum converted to standard scores where 60+ = high 50-59 = med. 0-49 = low (2) Power Structure See Table V (3) Coalition Interaction Process Analysis, utilizing Bales (1970) categories in the following manner: 1+2+3 - 10+ll+12v 1+2+3+10+1T+12 x u A reciprocal emotional interaction consisting of 30% or more emotional acts of the same sign between two persons, where interactions between all other dyads do not meet this cri teria . (4) Counseling Outcome QUESTIONNAIRE (a) Ineffective counseling: where one or both clients have de fected from counseling or counselor assesses that no progress has taken place. (b) Effective Counseling: where the counselor assesses that progress was made resulting in termi nation or progress is currently being made. 98 | Experimental Procedure This consisted of obtaining thirty five audio-tape recordings of initial or first conjoint counseling ses sions. The initial conjoint was chosen because it rendered the data more directly comparable, particularly with ref erence to predicted outcome. In some instances the initial conjoint had been preceded by a single interview with the wife and/or husband, however this was not considered to be a serious contamination of the data. If several single interviews had preceded the initial conjoint then it was assumed that a relationship had already developed between one spouse and the counselor and such initial conjoint interviews that conformed to this patterns were not in cluded in the present study. The M.M.P.I.'s usually were administered between the first and the second interviews, however in some instances, due to administrative difficul- ; ties, this procedure was not followed and the tests were administered at a later stage in counseling. Nevertheless it is reasonable to assume that the M.M.P.I. profiles are sufficiently stable that this time lapse would not invali date them. As an added precaution against coder subjective bias, all tapes were coded prior to the M.M.P.I.'s being scored. Three objections can be raised concerning the use of audio tapes. The first, and most cogent one, is that it results in a serious loss of information. That this is true, is indisputable, and the conclusions of the present study are drawn with full awareness of this limitation. Bales (1950, 1970) whilst recognizing the inadequacies of sound recordings alone, nevertheless indicates that they are a legitimate means of collecting data for his type of analysis. The second objection centers around the question of the effect of tape recorders on the counseling session it self. Lamb and Mahl (1956) studied the reactions to infor mation that the interview was being recorded of 35 patients, none of whom had had interviews recorded before. They found that 80% of these patients showed no overt disturbance concerning recording after the first few minutes of the interview. The same study (Lamb and Mahl 1956) also re ported findings on questionnaires administered to thera pists concerning their attitude toward recording. The findings indicated a positive correlation between the de gree of disturbance felt by therapists while recording and : the dual belief that they conducted therapy differently when recording, and the feeling that the patient was ef fected adversely. The authors' interpretation of this find ing was that it may reflect the therapists projection of their own concern about recording. The present study ac- 1 ioo ! cepts the position that recording does not have a signifi cant effect on the counseling process beyond some initial and temporary anxiety. In fact, for a large percentage of counselors participating in this study recording of coun seling sessions was routine. The third objection that can be raised against using this type of data is the question of the reliability of the coding. This issue has been traditionally handled by having one or more additional persons as coders and assessing their agreement. The present study adopted this approach and a random sample of eight interviews were coded by an independent person. The issue on which the study sought agreement was specifically whether the two coders arrived at the same decision concerning the presence or absence of a particular coalition. Although it could be argued that this ignores the question of possible disagree ment within the Bales categories, it was agreement on the existence of an actual coalition that was the central mea- ; surement issue in the study. This was simply assessed by expressing as a percentage the number of agreements over the total of agreements plus disagreements. Using this method an inter-coder reliability rate of 75% was obtained. j There is a further point concerning the reliability of coding and this centers around difficulty of coding such complex data as human interaction. While this is partly 101 coped with by having two coders, an additional precaution was adopted in the present study. Essentially this was to use the data from the entire counseling session and not to j do as some previous researchers have, to select at random various parts of the interview for analysis. By using the entire session it was anticipated that some of the errors arising from the inferential nature of the data would be reduced. The sessions averaged 60 minutes and this pro vided an average of 400 - 450 separate coded acts. Sample The sample consisted of 20 graduate students en rolled either in a Masters or Ph.D. level programme in marriage and family counseling at 3 accredited training centers. One of the counselors was in fact a professional counselor who at the time of the study was also enrolled aS a graduate student. At the time of the study, the students were doing their internship in marriage counseling. In order to increase the number of tapes available for analy- ; I sis, some students gave up to, and not exceeding, three separate tapes of initial conjoint interviews. An objec tion could be raised that this builds into the data a cer tain redundancy. It can be equally well argued, however, that this procedure is not inconsistent with the laboratory 102 studies on coalition formation where the same members were involved in different triadic power structures. In this sense, although the same counselor was involved in three sessions, each conjoint interview represents a unique triadic structure. It is pertinent to mention that all students who participated in the study were unaware of its purpose and were merely asked to volunteer tapes for an investigation of the process of conjoint interviews. Statistical Analysis Because of the unequivocally nominal nature of the data and the small sample, it was decided to use non- parametric statistics. As the purpose of the statistical analysis was to assess the degree of association between certain types of power structures, coalition patterns and counseling outcomes, the non-parametric statistics of chi- square and lambda were chosen for use throughout the analy sis. CHAPTER IV RESULTS AND DISCUSSION The four major hypotheses tested in this study fall; into two groups. The first is concerned with the relation ship between the types of power structures and coalition formation. Within this group the independent variable, types of triadic power structure, were measured by two dif ferent procedures. The first was an index of dominance based on Leary's (1957) theory of interpersonal behavior and the second used Barron's (1953) ego strength scale. These hypotheses were aimed at investigating whether cer tain types of power configurations, based on Caplow's (1956, 1968) theory resulted in coalition formation in the conjoint marriage counseling situation. The second group of hypotheses, specifically three and four, were concerned with the consequences of coalition formation on the outcome of marriage counseling. As men tioned previously, due to the relatively small sample it was not possible to investigate empirically the various sub hypotheses that referred to the relationship between specific types of type II, or low counselor potency struc- ■ tures, and the outcome of counseling. However the latter 103 I 104 j part of this section includes a frequency table which pro- | vides some indication of the outcome pattern associated with the specific types of low potency structures. The underlying rationale for the various hypotheses has been adequately explicated in the previous chapters and the present chapter will be restricted to an analysis and dis- ; cussion of results. It will be noticed that a discrepancy : exists in the sample size for the two groups of hypotheses. This is due to the fact that M.M.P.I.'s could not be ob tained from four couples, for a variety of reasons. Hence the analysis of the relationship between power and coali tion patterns is restricted to 31 cases instead of 35. Analysis of Results The first hypothesis postulated that conjoint ses sions in which the counselor is in a type I structure of high potency, will not exhibit a particular coalition pat- tern. Conversely hypothesis two postulated that conjoint sessions in which the counselor is in a type II structure of low potency will exhibit a particular coalition pattern. Tables VII and VIII display the relevant data for both the ; ego strength scale and the Leary dominance index and the relationship between these independent variables and the coalition pattern. 105 TABLE VII TYPE OF POWER STRUCTURE, AS MEASURED BY EGO STRENGTH AND COALITION PATTERN Power Type No Coalition I High Counselor Potency II Low Counselor Potency Coalition 8* 4 3 16 12 19 11 20 31 X =6.24 P<.05 with 1 df. (X cor rected for continuity) A = .57 c ♦Figures refer to actual observed frequencies. 106 TABLE VIII TYPE OF POWER STRUCTURE AND COALITION PATTERN AS MEASURED BY LEARY (1957) INDEX Power Type No Coalition Coalition I High Counselor Potency 5 3 8 II Low Counselor Potency 6 17 23 11 20 31 = 2.03 P>.05 (Not significant) Xc = .18 107 j Table VII indicates that there is a significant association beyond the .05 level between types of power structure as measured by ego strength and the coalition pattern. Specifically, hypothesis one and two are con firmed with a chi square (corrected for continuity) of 6.24. Lambda of .57 provides further confirmation of the hypotheses and indicates that knowledge of the power struc ture results in a 57% proportional reduction of error in predicting the type of coalition pattern that will form. Table VIII refers to the same hypotheses with the Leary index as the independent variable. A chi square of 2.03 with one degree of freedom, fails to reach the .05 level of probability and leads to a rejection of hypotheses one and two and the acceptance of the null hypotheses that there is no relationship between type of power configura tion as measured by the Leary index of dominance and coali tion pattern. Lambda of .18 indicates that knowledge of the power structure as measured by this procedure only re- | suits in an 18% reduction in error over using the marginal totals. Some possible reasons for this lack of signifi cance will be elaborated in the discussion section of the present chapter. Tables VII and VIII provide support for the general proposition of coalition theory, that a triad tends to dis-j solve into a dyad plus a monad. Reference to the Tables reveals that in 20 of the 31 cases, or 64.5%, a predominant coalition pattern was observed. Table VII further indi cates that it is unlikely that coalitions will not form under the conditions of low counselor potency. Of the 20 cases in which a coalition was observed 16 or 80%, of these! were associated with low counselor potency and only 3 or 15%, resulted in no coalition. Again this would appear to be strong confirmation of the theory developed in Chapter II concerning the vulnerability of the counselor to revolu tionary coalitions when he is within a type II structure of low potency. Hypothesis three postulated that conjoint sessions in which no coalition was observed would tend to result in effective counseling. This, as previously discussed, was measured by obtaining the counselor's assessment of thera peutic progress via the administration of a questionnaire six weeks after the initial conjoint interview. Conversely,; hypothesis four postulated that those conjoint sessions in which a coalition was observed would tend to result in in effective counseling. In assessing ineffective counseling, the counselors rating of the current therapeutic progress of the marital relationship was used and not the counselors! assessment of individual progress. The reason for this is I that consistent with the theory developed in Chapter II, under certain conditions one spouse might defect from coun- 109 TABLE IX COALITION PATTERN AND COUNSELING OUTCOME Coalition Pattern No Coalition Observed Coalition Observed 16 19 35 X2 = 6.08 P<.05 with 1 df (X2 corrected for continu ity) X = .52 c ♦Refers to cases "dropped out" and those which are continuing with no change. **Of these 6 cases, four did not follow the pattern of exclusive use of conjoint interviews. In two cases the counselor used single interviews and in the other two an additional counselor was brought in and hence co-therapy was conducted. When these specific cases are removed from the analysis X2 increases to 11.31 P<.001 and A to .58. Effective Ineffective* Counseling Counseling 10 3 6** 16 110 seling and this according to the model represents ineffec tive counseling. Table IX provides the data relevant to those hy potheses. A chi square of 6.08 with one degree of freedom and corrected for continuity, indicates that hypotheses three and four are supported beyond the .05 level of prob ability. Lambda of .52 suggests that having knowledge of the existence of a coalition leads to a 52% proportional reduction of error in predicting the outcome of counseling. Detailed examination of Table IX reveals that of the 19 cases resulting in effective counseling, 16 or 84% occurred when a coalition was observed in the first conjoint ses sion. The data referring to effective outcome at first appears less pronounced, with only 10 or 62.5% of these occurring under the condition of no coalition and 6 or 37.% when a coalition had been ascertained. However, of these six cases in which an effective outcome was the re sult, four of these deviated from the exclusive use of conjoint interviews. In two of these four cases the coun selor conducted a minimum of two single sessions with each spouse and in the other two a co-therapist was introduced for the entire length of the counseling. These facts pro vide support for the model developed in Chapter II, where it was argued that a counselor, when confronted with a co alition, could subvert therapeutic failure by utilizing the strategy of single interviews. What was not anticipated by the model, but which nevertheless makes theoretical sense, is that ineffective counseling could be avoided by the in- ! i troduction of an additional therapist. The reason for this! will be discussed in the latter section of this chapter. When these four cases are removed from the data on coun seling outcome, a chi square of 11.31 is obtained and with one degree of freedom this is significant beyond the .001 level. Lambda under these conditions increases to .58, which means that knowledge of the coalition patterns leads to a 58% reduction in error when predicting the outcome of ! counseling. These figures provide relatively strong support for hypotheses three and four and further provide strong empir ical validation of the theoretical model developed in Chap-; ter II, particularly that part of the model that deals with* the strategy of single interviews to subvert therapeutic failure. In summary then, it can be concluded that both hypotheses three and four are confirmed. Specifically this; leads to the conclusion that when a coalition is present there is a beyond chance level that counseling will be in effective. Tables X and XI provide descriptive data which is relevant to the various sub hypotheses derived in Chapter II. Due to the relatively small number of subjects and the; 112 TABLE X POWER STRUCTURE (EGO STRENGTH) AND FREQUENCY OF TYPE OF COALITION Counselor/ ounselor/ Husband/ No Coalition husband wife wife Type I (8) ♦ (1) (1) (2) High Counselor 66.7^ 8.3 8.3 16.6 Potency 12.7*** 25.0 28.6 11.1 Type II (3) (3) (5) (8) Low Counselor 15.8 15.8 26.3 42.1 Potency 27.3 75.0 71.4 88.9 11 4 6 10 ♦Numbers in parentheses refer to observed frequencies. ♦♦Refers to Row percentages. ♦♦♦Refers to Column percentages. comparatively high number of categories, it was decided not to subject this data to statistical analysis. However, the| descriptive data portrayed in the tables in terms of ob served frequencies, row and column percentages, does pro- j vide some further interesting trends relevant to the pres ent study. Due to the lack of statistical significance of I the Leary index, Table X refers only to ego strength as the independent variable. This table displays the frequency of the various types of coalition patterns that are associated with the two different power configurations. Hypotheses 2 (a), (b) and (c) are the relevant hypotheses for this table. Due to an insufficient number of cases in the various types of triadic power structures associated with a low counselor potency set, all such triads were collapsed into the one category of type II or low potency. Examination of Table x! reveals that 66.7% of those cases in which the counselor was in a high potency situation no coalition was found and ■ this of course is consistent with the data relevant to hy potheses one and two. There appears to be no pronounced pattern of a particular type of coalition being formed when it did occur under the conditions of high counselor potency. The table reveals one case in each of the possible coun selor/spouse categories and two cases in the husband/wife categories. However the same lack of pattern does not ap 114 pear to hold for the type II set of low counselor potency. The table reveals that under this condition, 8 or 42% of the cases displayed a husband/wife coalition, with the next most prevalent pattern being counselor/wife (26%). It could be tentatively suggested from this that if a coali tion does form under the type II condition of low counselor potency then it is more likely to be a husband/wife nega tive coalition than any other type. Further, it can be speculated that this particular pattern is unlikely to oc cur in any systematic way under the condition of high po tency. The implications of this trend will be discussed in the latter section of this chapter. At this descriptive level it can be tentatively concluded that this data offers some support for the sub hypotheses 2 (a), (b) and (c). Table XI is concerned with the sub hypotheses 4 (a) and 4 (b). These postulated that under the condition of different types of coalitions, different outcomes would be observed. Specifically, when a counselor/spouse coalition ; exists premature termination by one spouse will result or counseling will continue ineffectively. Alternatively hy pothesis 4 (b) states that under the conditions of a hus band/wife coalition both spouses will terminate prematurely or counseling will continue ineffectively. Although these : hypotheses were not tested in any statistical sense, Table XI does provide some interesting data which tends to run, 115 TABLE XI COALITION PATTERN AND TYPE OF COUNSELING OUTCOME Coalition Pattern Effective Continuing but not effective Clients dropped out (10) ♦ - (3) No Coalition 76.9** - 23.1 62.5*** - 18.8 Counselor/ (4) (2) (4) Spouse 40.0 20.0 40.0 Coalition 25.0 66.7 25.0 Husband/ (2) (1) (9) Wife 16.6 8.4 75.0 Coalition 12.5 33.3 56.2 16 3 16 35 ♦Numbers in parentheses refer to observed frequencies ♦♦Refers to Row percentages. ♦♦♦Refers to Column percentages.. 116 at least in part, contrary to the theory. With reference to the "no coalition" category, again the same finding as previously discussed is observed, i.e., effective counseling occurs 77% of the time, ineffec tive only 23% and no cases that are continuing with inef fective treatment. The counselor/spouse coalition pattern is suggestive of an interesting trend. The table indicates that while 40% of these cases resulted in the clients ter minating prematurely i.e., dropping out, an equal percent age resulted in effective treatment. By comparison in the husband/wife coalition category, 75% resulted in the cli ents dropping out and only 16.6% were found to have made any therapeutic progress. From these figures it can be speculated that a husband/wife coalition is substantially more likely to result in ineffective treatment than a coun-i selor/spouse coalition. Under the latter condition there appears to be an equal probability of effective counseling. Again some possible reasons for this will be given in the discussion section of this chapter. The column percentages of Table XI also provide some interesting speculative data. Of those cases in which counseling was continuing ineffectively, 66.7% of them were in the counselor/spouse coalition category and only 33.3% were in the husband/wife coalition group. The reverse trend can be observed with reference to the premature ter 117 mination outcome. Here, 25% of the cases were those in which a counselor/spouse coalition was present and 56% where a husband/wife coalition was observed. Thus again it can be tentatively concluded that the latter coalition pat-: tern is more likely to result in defection from counseling.| Overall these two tables of descriptive data indi cate that the husband/wife coalition pattern is probably the one that is most likely to occur and the one that has the highest probability of resulting in the clients drop ping out of counseling. The counselor/spouse coalition pattern appears to have an ambiguous status with respect to the outcome of counseling. Discussion The major theme of this study has been to investi gate the relationship between power and coalition formation; in conjoint marriage counseling. The results indicated that there is a significant relationship between these variables and further, that coalition patterns are system atically related to counseling outcome. Of particular interest in the results is the phe nomenon of husband/wife coalition. Not only does this pat tern appear to be the most prevalent one, but it is also the one most likely to lead to premature termination. It was discovered that under the conditions of low counselor potency, as measured by ego strength, this particular type of coalition occurred with relatively high frequency. It had been theoretically anticipated that husband/wife coali tions would be predominantly negative and further that next to the ++ coalition structure they had the highest solidar ity. The data revealed that all 16 of the husband/wife co alitions were in fact, negative. The high percentage of these that resulted in premature termination by the clients can also be seen as indicative of the strength of this co alition pattern. When these two observations are combined with the knowledge that the occurrence of a husband/wife coalition was most prevalent under the conditions of low counselor potency, some plausible theoretical arguments can be put forward to explain this phenomenon. First, it ap pears reasonable to assume that a sizeable percentage of clients seeking marriage counseling would exhibit a nega tive coalition pattern, or in Haley's terms, (1962) "an error activated system." This is consistent with the dis cussion in Chapter II, concerning the issue of a struggle for control in the marriage over who is going to set the limits on behavior. Under the condition of low counselor potency it would appear from the data that the counselor in; this situation is unable to assert sufficient influence to become what Haley (1962) refers to as the "governor of the 119 i system." In pragmatic terms this means that the counselor is rendered impotent and the clients continue, if not strengthen, their already existing negative coalition. It is in connection with this that the strategy of single interviews, or co-therapy, becomes important. It is through the use of these strategies that the counselor can, at least temporarily, circumvent the problem of an error activated system characteristic of a negative coalition be tween the husband and wife. The removal of one person from the triad, or the addition of another through the introduc tion of a co-therapist, alters the basic triadic configura tion and opens up the possibility of alternative ways of behaving that are not error activated. In the single ses sion the counselor could concentrate on building an indi vidual relationship with each spouse and in the process enhance his credibility and power as a mediator. With the introduction of a co-therapist, the combined power of the two therapists may well be the critical factor that pre vents the marital system from maintaining its characteris tic negative pattern and thus enhance the probability of effective treatment. Within the theoretical context dis cussed in Chapter II a husband/wife coalition under the conditions of low counselor potency could be seen as the cheapest winning coalition and hence the one that is most likely to form and maintain itself. 120 A somewhat more abstract explanation within the framework of systems theory can also be advanced to explain this phenomenon of the husband/wife coalition. Lynn Hoff- ; man (1971) in an article on deviation— amplifying processes in natural groups, discusses also the question of devia tion— counteracting processes or negative feedback. She discusses the interplay between the two mechanisms and their role in maintaining homeostasis and preventing run aways in the form of positive feedback loops. It is possi ble that it is the actual threat of exceeding the homeo static plateau that motivates couples who exhibit negative coalitions to seek marriage counseling. Given the validity of this assumption, counseling can then be perceived as providing a temporary deviation counteracting force that prevents the feared runaway of the system. Thus it is not change that clients who exhibit a negative coalition are seeking but, system maintenance. Hoffman makes an inter esting point in connection with this when she states "it is clear that any intervention that tries to reverse the devi ation— amplifying sequences in this set without figuring out how to deal with the deviation— counteracting sequences; will fare badly." (Hoffman 1971, p. 304). It seems plausi ble to suggest that this provides a reason why negative co alitions tend to result in ineffective counseling. Essen tially it can be postulated that the counselor failed to 121 grasp the couples motivation for seeking counseling i.e., maintaining the status quo, and thus were unable to arrest i control of the counseling situation or redefine their role in terms other than deviation counteracting. In this sense they had accepted the clients definition of counseling. Hoffman (1971) proposes a solution to this problem which emanates from her theoretical view that "the problem is that the patient is caught in an increasingly inflexible set of patterns and the task is to loosen them up." She therefore suggests that therapy "may require the introduc tion of complexity rather than restoration of order." (Hoffman 1971, p. 306). The introduction of "complexity" could be seen as being produced by the strategies of single interviews or co-therapy, which may have the effect of unstabling the system and thus opening up the possibility of change rather than maintenance of the status quo. The exclusive use of conjoint interviews could be seen as pro- i viding a relatively safe environment for clients caught in I a negative coalition to test the upper limits of the posi tive feedback loop without the risk of destroying the sys- ; tern. Thus, when this has occurred and homeostatis has been restored, the clients drop out of counseling, with a "no change" assessment by the counselor. Regardless of the explanation for this phenomena, it is clear that if coun selors do not utilize the previously mentioned strategies, when confronted with a husband/wife coalition, then the probability of therapeutic failure is high. The somewhat ambiguous findings concerning the ef- 1 feet of counselor/spouse coalitions on outcome, appears worthy of further discussion. Table XII shows that 40% of these resulted in ineffective treatment and 40% in effec tive treatment. If for the purposes of the present dis cussion the question of the small number of cases is tem porarily ignored, some interesting theoretical possibili ties emerge, which may serve to explain this ambiguity. Of particular interest is the 40% of cases in which a coun selor/spouse coalition existed which also resulted in ef fective treatment. The present study, as it was designed, did not allow for an investigation of sequential effects of coalition formation, however, it is possible that this is an important factor in explaining the cases within this category that resulted in effective treatment. In these instances the counselor may have become aware of the exist ing coalition and would have been able to correct it in the following interviews, by forming a coalition with the pre viously rejected spouse. This appears to be consistent with Zuk's (1971) technique of the "go-between process" to : break up pathogenic relationships. It is possible that the entire counseling process consisted of these sequential changes in counselor/spouse 123 ; I coalitions which prevented the outcome of ineffective ther-| apy, since such coalitions would be engineered by the coun selor and in this sense he is in control of the conjoint session. For the 40% that resulted in ineffective counsel-! ing it could be postulated that in these instances the counselor failed to initiate an alternating coalition pat tern and therefore the cases conform to the theoretical prediction of premature termination. Although this reasoning may be speculative, it is nevertheless clear that a counselor/spouse coalition is not as detrimental to the progress of therapy as is a husband/ wife coalition. In fact, it is possible, that in addition to the previously discussed strategies of single interviews and co-therapists, the conscious engineering of alternating; counselor/spouse coalitions may be a further strategy for coping with negative coalitions between spouses in the con joint situation. As will be discussed in the final chapter of this study, these are questions that could be answered through an investigation of the sequential patterns in co alition formation. The final area of discussion concerns the failure of the Leary dominance index as a predictor of coalition formation and the relative success of ego strength. While i the Leary index seemed reasonably effective in predicting the existence of a coalition, it appears weak in predicting a "no coalition" pattern. This may indicate that it is more; powerful in predicting submissive behavior rather than dom-| inant. It will be recalled that the dominance index was derived by subtracting the standard M.M.P.I. scores of Ma from D and Hs from Pt, and summing the result. One of the ! scales that has received some criticism within this index is Hs or psychosomatic scale. Leary and Coffey (1954) state that: the M.M.P.I. scales selected for the original (on the basis of clinical experience) stands up fairly well, with the exception of one unhappy choice. The Hs scale, which was nominated to predict dom inance actually is related (insignificantly) with submission. This at least in part, provides some explanation as to why the index did better in predicting coalitions than no co alitions. A further point that may serve as an explanation for the failure of this index concerns the Ma scale. If a counselor had high Ma, but the other scores resulted in a composite index which indicated low or medium dominance, then he or she would be classified as "low counselor poten-; cy." However it is possible that the high Ma in the coun seling situation, operated in such a way as to negate the submissive traits i.e., high D and high Pt. The rationale behind this suggestion is that the person having a high Ma : 125 score would tend to verbally dominate the session and thereby assert control. If this is valid then it appears as a reasonable explanation as to why six cases in which the counselor was classified as low potency, did not result in a coalition. A final point that may contribute some clarity in this area is provided by Leary (1957). In discussing the use of this index to predict behavior, he asserts (Leary 1957, p. 108) that "there were many cases, however, in which level I-M did not predict actual behavior in the group." What Leary is asserting is that dominance and af fection indices derived from the M.M.P.I. are satisfactory for predicting behavior in individual sessions, but not in a group. He bases his argument on the observation that in a group situation responses can be evoked in a patient that are different from his symptomatic behavior. The present study had assumed, and it now appears erroneously, that the conjoint interview, although a small group, would not in hibit the expression of symptomatic behavior to the extent that the predictive power of the index would be rendered insignificant. It is possible that some combination of all three of these reasons put forward in this section explain the failure of the Leary index to predict the pattern of "no coalition." The findings concerning ego strength would appear to confirm Phillips' (1970) hypothesis that 126 | ego strength should be the highest or the second highest scale on the M.M.P.I. profile of successful counselors. Summary The four major hypotheses of the study were con firmed and this leads to the conclusion that counselors who are in high potency structures where they are invulnerable to revolutionary coalitions, tend to conduct effective therapy. Conversely, if the counselor is in a low potency structure, coalitions tend to form which result in ineffec-; tive therapy. The prominance of the husband/wife coalition pattern was discussed and it was argued that this pattern is the one most likely to result in ineffective counseling unless the counselor utilizes the strategy of single ses sions or brings in an additional therapist. It was also discussed that the existence of a counselor/spouse coali tion is not such a threat to therapeutic progress as the model developed in Chapter II had predicted. It was argued! that if a counselor is aware of the existence of this co alition then he may well be able to initiate corrective be-i havior in the form of alternating coalitions to prevent premature termination. Overall the model developed in Chapter II and elaborated in Figures 1, 2, and 3 appears to I be confirmed and the relevance of coalition theory for con- joint counseling established. CHAPTER V SUMMARY AND CONCLUSIONS Although the use of conjoint interviews in marriage- counseling has become a well established clinical proce dure, little attention has been given to developing an ade quate theoretical model of the process of conjoint inter views. Conceptualizing the phenomena as a triad this study utilized the framework of coalition theory, as explicated by Simmel and Caplow to develop a theoretical model and empirically investigate its validity. The major independent variable of the study was power and the major dependent variable coalition formation. Power was operationalized by two separate methods that pro vided a measure of power independent of the actual flow of interaction in the conjoint session. The first procedure utilized Barron's (1953) ego strength research scale of the Minnesota Multiphasic Personality Inventory. The second method was based on Leary's (1957) index of dominance de rived from four of the clinical M.M.P.I. scales, specific ally Ma, D, Hs, Pt. Individual members of each conjoint triad were classified as either high, medium, or low on interpersonal power. Each triad was then classified as 128 129 either high or low in terms of counselor potency. The data for the study of the existence of coali tions was obtained by coding audio tape recordings of the first conjoint interview. Coalitions were measured through the use of Bales' (1970) interaction process analysis meth od. Specifically the categories that refer to positive socio-emotional acts and negative socio-emotional acts were used. A coalition was deemed to exist in any particular dyad, within the triad, when the percentage of positive or negative acts from one member was in excess of 30% of the total positive and negative acts of this person and further there existed a reciprocal number of acts in excess of 30% of the same sign, from the other member of that dyad. In the second phase of the study, concerned with the outcome of counseling, coalition patterns constituted the independent variable and counseling outcome the depen dent variable. Outcome was measured by administering a questionnaire to the counselors six weeks after the initial conjoint session which obtained information on the current therapeutic status of the case. Twenty counselors who were, with one exception, graduate students doing internships in marriage counseling, participated in the study. Thirty five couples who were drawn from both a university setting and the community com-; prised the client sample. 130 Pour major hypotheses derived from coalition theory were tested in this study. As postulated in hypotheses one and two there was a significant relationship between coun selor power, as measured by ego strength and coalition for mation. Specifically, those conjoint sessions in which the counselor had high potency tended not to form coalitions and those in which the counselor was in a low potency posi tion demonstrated a significant trend to form coalitions. Using the Leary index of dominance the same hypotheses were not confirmed. Hypotheses three and four postulated that there would be a significant relationship between coalition for mation and counseling outcome. These hypotheses were con firmed and this led to the conclusion that if a coalition is present in the first conjoint session then there is a beyond chance probability that counseling will be ineffec tive. Conversely, a significant relationship was estab lished between no coalition present and effective counsel ing . Some additional data concerning the particular types of coalitions formed was obtained and this indicated that negative coalitions between the husband and wife were the most prevalent. The study also established that these particular coalitions were the ones most likely to result in clients prematurely terminating from counseling. The 131 same conclusions could not be drawn concerning counselor/ spouse coalitions which appeared to have an ambiguous sta tus with respect to counseling outcome. Limitations of the Present Study The most obvious limitations of this study are the small size of the sample and the fact that all except one of the counselors were graduate students. However in com parison with previous studies on the effectiveness of mari tal therapy that have been extensively reviewed by Gurman (1973), the sample can be considered large. Gurman indi cates that in a total of 26 outcome studies between the years 1950 - 1972 only two of these studies had a sample size in excess of ten therapists. While this is not a jus tification for small sample size it nevertheless estab lishes a perspective within which the sample of the present study can be viewed. The limitations arising from using graduate student counselors does however present a more serious problem. It is possible that as counselors gain experience they become more astute in their ability to per-; ceive coalitions and also develop a wider range of tech niques for handling coalitions. Although the present study did not explore the question, it is possible that those cases in which effective counseling occurred under the con 132 dition of a counselor/spouse coalition, were in fact cases in which the counselor was more experienced. A further limitation of the present study is the use of audio tape recordings. Although the main coder was an experienced therapist and hence had developed listening skills, the problem of directionality of some responses and non-verbal communication still remained. Despite the pre cautions that were taken to ensure accuracy some loss of information must inevitably occur when using audio tapes alone. The problem could of course be corrected through the use of video tapes, however for this study such facili ties were not readily accessible. Another major limitation of this study is that it did not examine the sequential effects of coalition forma tion. By only taking the initial conjoint interview valu able data on shifting alliances was ignored. Such data would have allowed a more detailed examination of the dif ferential effects of types of coalitions on counseling out come. It would have provided data on the counselor/spouse pattern and may have led to an increased understanding of why an equal percentage of these cases resulted in effective therapy. Such a sequential analysis would have also pro vided the opportunity to study the stability of certain systems such as the husband/wife negative coalition and counselor strategies for coping with this. 133 The conceptualization of power as the relative dis- i tribution of dominance is a further limitation since it ig nores the role of warmth, congruence and empathy in the counseling process. A comprehensive index of power that included these variables in addition to dominance may re sult in a more powerful predictor of coalition formation. Finally the procedure for measuring counseling out come is in many respects inadequate. Because it relies on the counselors' subjective assessment of progress in all cases except those where the clients prematurely termi nated, it is open to the criticism of bias. Secondly the use of the global measure of outcome is a doubtful proce dure for assessing change accurately, in such a complex process as counseling. Some more comprehensive measure that took account of behavioral, emotional and perceptual aspects would have enhanced the findings of the present study. Suggestions for Future Research These suggestions follow directly from the limita tions. The first is to increase the sample size of both counselors and clients. This would result in a more de tailed and reliable analysis of the data and facilitate an examination of the various sub-hypotheses that could not be 134 tested in the present study. Research into coalition formation in conjoint coun seling would be enriched by the use of video tape equip ment. This would give the coders access to both verbal and non-verbal behavior and thus render the coding more reli able. It would also be useful, as a by-product, to examine the differences between coding based on audio versus video tapes. Finally and perhaps most importantly, further stud ies should be designed to test the effects of coalition formation over a series of interviews. This is the major suggestion that emerges from this study and in view of the significant results it would appear to be a worthy area for investigation. Not only would it provide data on the sta bility of particular coalitions but also information on various counselor strategies for managing coalitions. Ul timately this would facilitate a closer examination of the effectiveness of the strategies for preventing therapeutic failures. In conclusion this study, has, for the author, at tested to the validity of Stryker's (1972, p. 339) state ment that, given three persons, two can gang up on the third; given three nations, two can join together to combat the one remaining. It is the ubiquity of the possibilities for and the fact of coalitions 135 in social life that makes systematic inquiry into such behaviors worth pursuing. Such a statement, in view of the established relationship between coalition formation and counseling outcome is an apposite comment on the importance of this phenomenon for the field of marriage counseling. REFERENCES 136 / REFERENCES Alexander, Francesca. The empirical study of the differ ential influence of self-concept on the profes sional behavior of marriage counselors. Unpub lished doctoral dissertation, University of South ern California, 1968. Amidjaja, I. R., and Vinacke, W. E. Achievement, nurtur- ance and competition in male and female triads. Journal of Personality and Social Psychology, 1965, 2, 447-450. Anderson, R. E. Status structures in coalition bargaining games. Sociometry, 1967, 3£, 393-403. Bales, R. F. Interaction process analysis: A method for the study of small groups. Cambridge: Addison Wesley, 1950. Bales, R. F. Personality and interpersonal behavior. New York: Holt, Rinehart and Winston, 1970. Barron, F. An ego strength scale which predicts responses to psychotherapy. Journal of Consulting Psychol ogy, 1953, 17, 327-333. Bergin, A. E. Some implications of psychotherapy research for therapeutic practice. Journal of Abnormal Psychology, 1966, 71_, 235-246. Blau, P. Exchange and power in social life. New York: John Wiley and Sons, Inc., 1967. Bond, J. R., and Vinacke, W. E. Coalitions in mixed sex triads. Sociometry, 1961, 24, 61-75. Burris, J. C. and Frye, R. L. The effects of initial re sources of individuals upon their selection of a partner in the formation of coalitions. Paper presented at the Southeastern Psychological Asso ciation Convention, 1966. Caplow, T. A theory of coalitions in the triad. American Sociological Review, 1956, 21, 489-493. 137 138 CaploW/ T. Further development of a theory of coalitions in the triad. American Journal of Sociology, 1959, 64, 488-493. Caplow, T. Two against one; coalition in triads. Engle- wood Cliffs, N.J.: Prentice-Hall, 1968. Chaney, M. V., and Vinacke, W. E. Achievement and nur- turance in triads varying in power distribution. Journal of Abnormal and Social Psychology, 1960, 60, 175-181. Chertkoff, J. M. Social psychological theories and re search in coalition formation. In S. Groennings, E. W. Kelley and M. Leiserson (Eds.), The Study of Coalition Behavior. New York: Holt, 1970, pp. 297-322. Emerson, R. M. Power dependence relations— Two experi ments. Sociometry, 1964, .27, 282-298. Francis, W. L. Influence and interaction in a state legislative body. American Political Science Re view, 1962, 56, 953-960. Gamson, W. A. A theory of coalition formation. American Sociological Review, 1961, 2£, 373-382. Gamson, W. A. An experimental test of a theory of coali tion formation. American Sociological Review, 1961, 26, 565-573. Gamson, W. A. Experimental studies of coalition formation. In L. Berkowitz (Ed.), Advances in experimental social psychology, 1, New York: Academic Press, 1964, pp. 82-110. Goffman, E. The presentation of self in everyday life. Garden City, New York: Doubleday Anchor, 1959. Gouldner, A. W. The coming crisis of western sociology. N.Y.: Basic Books, Inc., 1970. Gurman, A. S. The effects and effectiveness of marital therapy: A review of outcome research. Family Process, 1973, 12, 145-170. 139 Haley, J. whither family therapy? Family Process, 1962, 1, 69-100. Haley, J. Strategies of psychotherapy. New York: Grune and Stratton, 1963. Haley, J. Uncommon therapy: The psychiatric techniques of Milton H. Erickson, M.D. New York: W. W. Norton and Company, Inc., 1973. Hathaway, S. R. and McKinley, J. C. MMPI manual (revised). New York: The Psychological Corporation, 1951. Hoffman, Lynn. Deviation-amplifying processes in natural groups. In J. Haley (Ed.), Changing families: A family therapy reader. New York: Grune and Stratton, 1971, pp. 285-311. Hoffman, P. J., Festinger, L., and Lawrence, D. H. Tenden cies toward group comparability in competitive bargaining. Human Relations, 1954, 7, 141-159. Homans, G. C. Social behavior: Its elementary forms. New York: Harcourt, 1961. Kelley, H. H., and Arrowood, A. J. Coalitions in the triad: Critique and experiment. Sociometry, 1960, 23, 231-244. Lamb, R., and Mahl, G. F. Manifest reactions of patients and interviewers to the use of sound recording in the psychiatric interview. American Journal of Psychiatry, 1956, 112, 731-737. Leary, T., and Coffey, H. The prediction of interpersonal behavior in group psychotherapy. Journal of Group Psychotherapy, 1954, 7 _, 7-51. Leary, T. Interpersonal diagnosis of personality. New York: The Ronald Press, 1957. Lerner, D., and Aron, A. France defeats EVC. New York: Frederick A. Praeger, 1957. Luce, R. D., and Rogow, A. A game theoretic analysis of congressional power distribution for a stable two- party system. Behavioral Science, 1956, 1, 83-95. 140 Luce, R. D., and Raiffa, H. Games and decisions. New York: John Wiley, 1957. Manus, G. I. Marriage counseling: A technique in search of a theory. Journal of Marriage and the Family, 1966, 28, 449-353"; Mills, T. M. Power relations in three person groups. American Sociological Review, 1953, 18^ 351-357. Mills, T. M. Coalition pattern in three person groups. American Sociological Review, 1954, 19, 657-667. Neubeck, G. Toward a theory of marriage counseling: A humanistic approach. Family Coordinator, 1973, 22, 117-122. Olson, D. H. Marital and family therapy: Integrative review and critique. Journal of Marriage and the Family, 1970, 32, 501-538. Parsons, T., Bales, R. F., and Shils, E. A. Working papers in the theory of action. Glencoe, Illinois: Free Press, 1953, Chapter V, See p. vii. Parsons, T., and Bales, R. F. Family, socialization and interaction process. Glencoe, Illinois: Free Press, 1955. Phillips, C. E. A study of marriage counselor's M.M.P.I. profiles. Journal of Marriage and the Family. 1970, 32, 119-130. Plummer, N. A. Patient-therapist need compatibility and expectation of psychotherapeutic outcome. Disser tation Abstracts, 1966, 27B, 1628-1629. Riker, W. H. The theory of political coalitions. New Haven, Conn.: Yale University Press, 1962. Schelling, T. C. The strategy of conflict: Prospects for a reorientation of game theory. Journal of con flict resolution, 1958, 2^, 203-264. Schubert, G. Judicial behavior. Chicago: Rand-McNally, 1964. 141 Shapiro, R. J., and Budman, S. H. Defection, termination, and continuation in family and individual therapy. Family Process. 1973, 12, 55-68. Shapley, L. S., and Shubick, M. Method for evaluating the : distribution of power in a committee system. American Political Science Review, 1954, 48, 287- 1WF. Simmel, G. The sociology of Georg Simmel, K. H. Wolff (Ed.), Glencoe, Illinois: The Free Press, 1950. Strodtbeck, F. L. The family as a three person group. American Sociological Review, 1954, 19, 23-29. Stryker, S., and Psathas, G. Research on coalitions in the triads: Findings, problems and strategy. Socio metry, 1960, 23, 217-230. Stryker, S. Coalition behavior. In C. G. McClintock (Ed.), Experimental social psychology. New York: Hall, Rinehart and Winston, 1972, pp. 338-380. Swensen, C. H. Psychotherapy as a special case of dyadic interaction: Some suggestions for theory and research. Psychotherapy, 1967, £, 7-13. Thibaut, J. W., and Kelley, H. H. The social psychology of groups. New York: Wiley, 1959. Torrance, P. E. Some consequences of power differences on decision making in permanent and temporary three- man groups. In A. P. Hare, E. F. Borgatta and R. F. Bales (Ed.), Small groups: Studies in social interaction. New York: Alfred A. Knopf. 1955, pp. 482-492. Truax, C. B., and Mitchell, K. M. Research on certain therapist interpersonal skills in relation to process and outcome. In A. E. Bergin and S. L. Gargield (Eds.), Handbook of Psychotherapy and behavior change: An empirical analysis. New York: John Wiley and Sons, Inc., 1971, pp. 299-344. Turk, T., and Turk, H. Group interaction in a formal setting: The case of the triad. Sociometry, 1962, 25, 48-54. 142 Uesugi, T. T., and Vinacke, W. E. Strategy in a feminine game. Sociometry, 1963, 26, 75-88. Vinacke, W. E., and Arkoff, A. An experimental study of coalitions in the triad. American Sociological Review, 1957, 22, 406-414. Vinacke, W. E. Sex roles in three-person game. Socio- metry, 1959, 22, 343-360. Vinacke, W. E., and Gullickson, G. R. Age and sex differ ences in the formation of coalitions. Child Development, 1964, 1217-1231. Von Neumann, J., and Morgenstern, O. Theory of Games and Economic behavior. Princeton: Princeton Univer- sity Press, 1944. Zuk, G. Family therapy: A triadic based approach. New York: Behavioral Publications, 1971. APPENDIX I 143 CASE # APPENDIX I Individual BALES' INTERACTION PROCESS ANALYSIS (Revised 1970) ACT Directed Toward #1 ACT Directed Toward #2 ACT Directed Toward #3 To Group or Unclear 1. Seems Friendly 2. Dramatizes 3. Agrees 4. Gives Suggestions 5. Gives Opinion 6. Gives Information 7. Asks for Information 8. Asks for Opinion 9. Asks for Suacrestion 10* Disagrees 11. Shows Tension 12. Seems Unfriendly it* A P P E N D I X II I 145 146 APPENDIX II COUNSELING OUTCOME EVALUATION 1) Presenting Problem: 2) Which Partner Initially Sought Help? (a) Wife (i.e.) initiated counseling. (Please check) (b) Husband (c) Both 3) Pattern of Interviewing Following First Conjoint Inter view. (Please check) Date Single Inter- Single Inter- Conjoint Conjoint view with view with Interview interview wife husband with one with two counselor counselors and both and both spouses spouses 4) Any Cancellations or Postponements During This period following the First Conjoint Interview? (Please check) Yes No 5) If "yes/" by whom? (Please check Date of husband Wife Both Counselor appropriate cancellation person) 6) Are Both Spouses still Attending Counseling? Yes No 147 7) 8) If "no," which spouse is not attending? (Please check) (a)Wife (b)Husband (c)Both If counseling is still continuing, what is your assess ment of the Clients current therapeutic Progress? (a) WIFE: Considerable Improvement Moderate Improvement Some Improvement (please No Change check one) Mild Deterioration Considerable deteriora tion (b) HUSBAND: Considerable improvement Moderate Improvement Some Improvement (please No Change check one) Mild Deterioration Considerable deteriora tion (c) THE MARITAL RELATIONSHIP: Considerable Improvement Moderate Improvement Some Improvement (please No Change Check one) Mild Deterioration Considerable Deterioration 9) If Counseling has discontinued, is it because: (a) Clients failed to keep appointments (i.e.) out? Dropped (b) Counseling terminated, because problem was solved? re- (c) Counseling terminated by counselor because improvement? °f (d) Counseling terminated by counselor because clients were referred elsewhere or to another counselor? (e) Any other reason? Please state... 148 10) Were you conscious of any feelings of being overall more sympathetic to one client than the other? (Please check) Yes No 11) If "yes," which one? (Please check) Husband Wife 12) Were you conscious of any resistance on one or both spouses part to counseling? (Please check) Yes No 13) If one spouse resisted, which one? (Please check) Husband Wife Thank you for participating in the study, your co operation is greatly appreciated. 149 APPENDIX III ♦ 150 APPENDIX III DISTRIBUTION OF COALITION PATTERNS (In Percentages)* Case Number Counselor-Wife Dyad Counselor-Husband Dyad Husband-Wife Dyad c->w w->c C->H H+C H+W W-HH 1 +29 0 +12 2 +23 5 +10*7 -50*9 -37*6 2 +64 0 +57 5 - 5 8 +31*0 - 3*0 -20*0 3 - 5 1 +27 5 +42 0 + 51*6 -18*0 -49*0 4 + 34 4 +19 5 + 19 6 - 3*2 -56*8 -54*3 5 +34 9 +22 2 + 5 1 - 8*7 -43*6 -38*8 6 +44 0 + 30 4 + 37 0 + 45*4 +36*3 +34*7 7 +20 0 - 5 1 + 8 5 - 2*0 -51*1 -43*3 8 +45 9 +39 3 +29 7 +17*9 -34*0 -28*6 9 + 50 0 +26 0 +45 6 +20*4 +25*8 +21*0 10 - 2 0 -23 0 +24 2 +12 » 6 -32*3 -39*6 11 + 35 7 +40 6 +15 7 +21*7 -34*6 -45*7 12 + 57 1 +30 5 +39 2 +39*1 +34*7 - 8*3 13 +18 7 +15 7 +28 7 +18*7 -51*0 -37*8 14 +17 4 +19 0 +12 8 + 1-1 -40*7 -56*5 15 +41 8 +36 5 +33 1 +30*0 +33-1 +12 • 3 16 +38 6 +50 7 +4 5 4 +40*0 - 1-5 - 5*0 17 +59 3 +34 8 +20 3 +13*9 +37*2 +31*2 18 +27 7 +10 0 +30 0 + 6 • 6 -35 • 5 -40*0 19 +34 5 +45 1 +44 3 +44 • 2 +25-1 +23 • 3 20 +33 3 +53*4 +43 9 +54-5 M <<• M < • 1 -27*2 M <4 -11 * 6 n A A ^These percentages are derived from tne tormuia discussed in Chapter III using Bales (1970) categories. 151 Case Number Counselor-Wife Dyad Counselor-Husband Dyad Husband-Wife Dyad c-*w W+C C->H H->C H+W W-HH 21 +37*6 +54 5 +46*7 + 65*9 + 2 1 +15*5 22 +40*4 +60 9 +32 • 8 +53*5 - 4 2 - 9*5 23 +54 * 6 + 4 7 +26*3 +13 • 8 -58 4 -57*1 24 +18*0 + 30 5 + 62*8 +51* 4 + 0 5 +13*6 25 +32*0 +21 2 + 50*6 +48*6 -26 9 -31*3 26 +17*8 + 2 7 +17*8 + 6*5 -32 2 -54*8 27 +48*3 +36 3 +38*7 +36*9 + 22 8 +14*7 28 +52*3 +54 8 -12*1 +10*8 -12 6 -12*7 29 + 8*5 +29 6 +61*7 +74*5 +10 0 + 4*6 30 ' +51*5 + 81 2 +26*1 +46*6 - 8 8 - 4*6 31 +63 • 5 +89 2 +24*0 +61*0 +19 3 + 6*4 32 +23*8 + 4 1 +40*0 + 8*2 -71 1 -58*6 33 +51*7 +46 6 +32*9 +28*5 - 4 8 + 1*6 34 +29* 6 +16 3 +37 • 0 +25*5 - 6 2 - 9*0 35 + 7*1 + 4*0 +43*0 + 7*2 -67 7 -70*3

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

The Parameters Of Singleness: An Inquiry Into Some Factors Influencing The Choice Of Singleness Over Marriage As A Way Of Life

PDF

The Attitude Of Resistance To The Adoption Of An Older Child

PDF

The Effect Of Counselor Gender And Sex-Role Attitudes On Change Of Female Clients' Sex-Role Attitudes

PDF

Strategies Of Marital Communication

PDF

Attitudes Toward Sex In Marriage And Patterns Of Erotic Behavior In Dating And Courtship Before Marriage

PDF

The Application Of Biostatistical Methodology To Personnel Classificationand Turnover

PDF

Personality And Interpersonal Factors Associated With The Duration Of Marriage Counseling

PDF

A Psychosociological Study Of Fertile And Infertile Marriages

PDF

Time Duration In Marathon Groups: Effects Upon Self-Concept And Self-Actualization

PDF

An Analysis Of The Bases For The Concept Of Equality

PDF

Passage Through The Expressive Student Subculture: The Initiation And Cessation Of Marijuana Use

PDF

The Psychological-Social Variables Contributing To Marital Cohesiveness And Classification For Graduate Counseling Students

PDF

Power Relationships In Marital Discord

PDF

An Exploratory Analysis Of Two Dimensions Of Family Separation

PDF

Organization, Conflict, And Change: A Test Of A Multivariate Model Within Two Types Of Simulated Social Systems

PDF

Communication Between Engaged Couples

PDF

A Study Of Relationships Between Occupational And Marital Roles And Marital Adjustment

PDF

Personality Characteristics: Ideal And Perceived In Relation To Mate Selection

PDF

A study of personality needs involved in the selection of specific leisure interest groups

PDF

The Deterrent Effect Of Criminal Justice Agencies On Felony Offenses

Asset Metadata

Creator O'Connor, Peter Andrew (author)

Core Title Coalition Formation In Conjoint Marriage Counseling

Degree Doctor of Philosophy

Degree Program Sociology

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag OAI-PMH Harvest,sociology, individual and family studies

Language English

Contributor Digitized by ProQuest (provenance)

Advisor Broderick, Carlfred Bartholomew (committee chair), Glaser, Daniel (committee member), Schrader, Don R. (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c18-742325

Unique identifier UC11356734

Identifier 7501077.pdf (filename),usctheses-c18-742325 (legacy record id)

Legacy Identifier 7501077.pdf

Dmrecord 742325

Document Type Dissertation

Rights O'Connor, Peter Andrew

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA