CURRENT RESEARCH IN SOCIAL PSYCHOLOGY Publication Date: November 16, 1995 First Submitted: October 30, 1995 Resubmitted: November 9, 1995 Accepted: November 10, 1995 THE COALITION STRUCTURE OF THE FOUR-PERSON FAMILY* Oscar Grusky Phillip Bonacich Cynthia Webster University of California, Los Angeles ABSTRACT Caplow's model of coalitions and power relations in triads is here extended to tetrads. Forty-eight four-person families were studied with equal numbers of each of the four sibling gender and birth position constellations: older boy-younger girl; older girl-younger boy; two boys; and two girls. A total of 673 coalitions were identified. It was found that arguments led to coalitions about 30% of the time, with spousal coalitions found to be the dominant type. Support was thus found for Caplow's model, maintaining that power counts in family decision-making. Family composition was shown to be related to the formation of conservative, revolutionary, and improper coalitions. [page 16] - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - [page 17] INTRODUCTION "Family life is fraught with the tension of conflicting emotions precisely because it is based on coalitions.... A family is sustained by the interlocking forces of love and hate in somewhat the same way that buildings are held up by the opposing forces of tension and compression" (Caplow, 1968). Although Caplow may exaggerate the significance of coalitions for families, his perspective encourages deeper examination of family coalition phenomena. Caplow's (1968) study of coalitions in triads with special emphasis on family organization remains one of the most sophisticated theoretical treatments to date. The present study examines four-person rather than three-person families, and in contrast to Caplow's work, is considerably more empirical. The first goal of this paper, then, is to describe the coalition structure of the four-person family. We describe in detail the methods used to measure coalitions in the family with particular focus on conservative coalitions and revolutionary coalitions as described by Gamson (1961a; 1961b) and Caplow (1968). Then we apply these definitions to four-person families to demonstrate that Caplow's triadic theory of coalitions can be usefully applied to four-person families and possibly to other tetrads. Utility theory underlies Caplow's model in that it is assumed that family conflict is governed by the rational assessment of benefits and costs, thus implying that family members initiate conflicts because the perceived benefits of conflict outweigh the perceived costs. The benefits of conflict may include higher self-esteem or less esoteric rewards, such as additional resources. The costs may also be the loss of valued resources and/or psychological losses. The application of utility theory to conflict has a long history (Rapoport, 1957; Schelling, 1960; McGinnis, 1991; Coleman, 1991.) Cook and Gillmore (1984) have pointed out that coalition theories have largely ignored the analysis of power struggles among actors; therefore, not much is known about coalition formation in situations (such as in the family) where power differences may have long-term consequences. By moving out of the laboratory and exploring family dynamics, a number of difficult but significant questions regarding power relations may be explored that can add to our knowledge of coalition dynamics. One question, for instance, concerns the frequency of coalitions in four person families. Families with two parents and two children, unlike those with three members, have an opportunity to form counter- coalitions (such as parents versus children). So, in addition to the question of how often coalitions form, there is raised the further question: what types of coalitions predominate in four-person families? [page 17] - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - [page 18] As indicated, power is the central concept in Caplow's theory of coalitions in the family triad. Caplow asserts that coalitions form because the parties seek power in order to obtain desired resources. This raises the question with regard to four-person families: how important is power in determining who wins or loses? Finally, one important aspect of family power structures is family composition or gender distribution. Hence, the fourth question examined concerns the relationship between the coalition structure and family gender composition in four person families. METHODS Sample Questionnaires were distributed to approximately 1500 children at a junior high school in a suburban Los Angeles community. From these questionnaires families were selected that satisfied these criteria: two children living with both their natural parents, the younger child between the ages of eight and twelve, the older between twelve and sixteen. The minimum age was set by the younger child's ability to be an effective interviewee when asked questions about family dynamics. The older child had to be sufficiently older so that there could be a significant power difference between the children, but not so old that he or she was about to leave the family. Sex was balanced, with half of the younger children and half of the older children being male. We contacted 78 families to get the 48 families in the study. Family members were interviewed together and separately for three to four hours using a variety of instruments designed to measure different aspects of family decision-making and attitudes of family members toward each other. The parents were married an average of 16 years. Only two parents, both males, had been married before. Neither had children by their previous marriages. Fathers' occupations were mainly business-related or professional. There were eleven attorneys, five professors or deans, and four engineers. The rest were businessmen. Seventeen wives reported full-time employment, and fourteen reported part-time employment outside the home. Because of the area in which the school was located, median income was high, $64,000. Fathers and mothers averaged 42 and 39 years of age, respectively. The median income of husbands fell into the $40- 80,000 range, while the median income of the wives was $10-12,000. In 86% of the 44 families in which both spouses answered our income question, the husband's income was higher than the wife's. Eighteen of the 48 husbands reported incomes of $75,000 or more, while no wife did. The picture was similar with respect to years of education. In 28 of the families, the husband had more years of education than the wife; in thirteen families, they were equal; and in seven families, the wife had more years of education than the husband. [page 18] - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - [page 19] Coalition Measures Coalitions exist when family members jointly use their power to control a decision. Coalitions are not the same as affective cliques of mutual attraction. Coalitions are not indicated by the absence of disputes among family members. Family members who do not argue are not in a coalition unless they support one another in disputes with other family members. Thus, for our purposes, coalitions are measured by the frequency with which family members support one another in arguments. This definition meets the strict criterion stated by Gamson (1961b:84) that "participation on the same side of an argument is sufficient justification for asserting that a coalition has been formed." Each family member was asked a set of questions about each of the six possible dyadic arguments in the family: father versus mother; father versus older child; father versus younger child; mother versus older child; mother versus younger child; and older child versus younger child. Small Fisher-Price dolls were used to represent each family member. These helped make clear, particularly to younger children, between which two family members each argument occurred. Family members were asked to recall the last important argument involving each two-person set of family members, what it was about, what each of the other two non-involved family members did during the argument, how the argument ended, and how often arguments between these two parties took place. We asked what each of the other family members did in arguments between a pair: "Think about what did during the argument. Which of the following comes closest to what s/he did?" The respondent was then presented a card with these alternatives: S/he did not know about the argument; tried to avoid taking sides; agreed with (one party to the argument); agreed with (other party to the argument); tried to settle the argument without taking sides, did not care. We did not attempt to assess the consistency of the reports from different family members because we did not require them to describe the same argument. We wanted a variety of situations and types of arguments between each pair of family members. Since there were 48 families and four members in each family, there were 192 reports of what other family members did in arguments between members of each dyad. Every family member was asked twelve coalition questions. Thus, the total possible coalitions that could be named was 2304. For each coalition question, six alternative responses were presented, only two of which were coalition responses. Overall, then, a total of 673 coalitions were named. [page 19] - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - [page 20] Conservative and Revolutionary Coalitions A conservative coalition is a coalition that does not alter the existing power structure; whereas a revolutionary coalition is a coalition that dominates the superior member of the family (i.e., the one with the most power), and an improper coalition is a coalition that is neither conservative nor revolutionary. The coalition measures used differed from the traditional game rules used in studying coalitions. All family members were not simultaneously given an opportunity to form coalitions with other family members. Only one family member at a time chose a coalition partner. Also, Caplow's (1968) assumption, "in a set of linked triads a coalition partner in one triad may not be an opponent in another," was not maintainable. Coalition partners in one triad could be opponents in another. The coalition models utilized all depend on the identification of the power structure of the family. Family power structure was determined by this question: "Now I'd like to return to consideration of the set of dolls representing each member of your family. Would you rank order the dolls in terms of which family member, in your opinion, has the most and which has the least control over the property and money that your family has?" The largest number of respondents, by far (70%), identified the family hierarchy as follows: F > M > O > Y. This is the classical patripotestal family. The remainder were about equally divided between F = M > O > Y (16%), the equipotestal family structure, and M > F > O > Y (14%), a matricentered family structure. The definition of conservative, revolutionary, and improper coalitions followed Gamson (1961a) and Caplow (1968): Conservative Revolutionary Improper Type 3 (A = B > C) AB AC, BC -- Type 5 (A < (B + C)) AB BC AC Type 6 (A > (B + C)) AB, BC -- AC Type 7 (A = (B + C)) AB -- AC, BC In order to distinguish between Type 5 and Type 6 coalitions, questions that asked about the outcome of arguments were used. The format for these questions was, as follows: "Suppose that (mother) were to argue with (father) and (older child). Who would be more likely to give in or agree, (mother) or (father) and (older child)?" These questions enabled us to determine which coalitions were likely to win and which were likely to give in or lose. [page 20] - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - [page 21] FINDINGS How Frequent Are Family Coalitions? Respondents were asked to think about what they did during a specific argument, and to select from several alternatives that may describe what they actually did. About three out of ten (29.2%) reported that they participated in a coalition, and these respondents reported forming a total of 673 coalitions in the 48 families studied. Hence, arguments precipitated coalitions in less than one- third of the cases. In over one-fourth (28.2%) of the incidents the respondent did not know about the argument, and in 8.7% the respondent reported that s/he did not care. On the other hand, in 7.2% of the incidents the participants tried to settle the argument without taking sides, and in 11.8% they avoided taking sides. No information was available for 14.8% of the cases. This information is relevant to the issue of whether or not the formation of coalitions is a common or not-so-common response to family conflict. One might point to the fact that arguments precipitated the formation of coalitions in only 29.2% of the arguments. This suggests that quite often dyadic arguments are simply resolved by the participating parties and that is the end of it. Alternatively, it might be asserted that, despite attempts to settle arguments by the parties themselves and the natural tendency of other family members to either avoid taking sides or stay out of the conflict, in about three out of ten arguments their scope was enlarged and coalitions were formed. We are unaware of reliable national sample data on how frequently family members argue. The definition of what constitutes an argument is problematic. Family members are prone to distinguish between disagreements, discussions, and arguments, and may disagree as to which is the appropriate label. Such differences in perception make it hard to estimate how often arguments take place and, therefore, how often they lead to the formation of coalitions. What Types of Coalitions Predominate in the Family? Table 1 displays the distribution of the thirteen different types of two-party coalitions in four-person families. Elsewhere we have elaborated and tested a status maintenance theory of coalition formation (Bonacich, Grusky, and Peyrot, 1985) which asserts that coalitions form to maintain the existing power structure. The finding in Table 1 shows the strong predominance of parental coalitions, which make up over 41% of the coalitions formed, are consistent with this approach, which stresses the significance of maintaining family solidarity and supporting the status difference between parents and children. [page 21] - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - [page 22] TABLE 1. DISTRIBUTION OF FAMILY COALITIONS Coalition Type N Per Cent Father + Mother 278 41.31 Mother + Older Daughter (61) (9.06) Father + Older Son (49) (7.28) Mother + Older Son (44) (6.54) Father + Older Daughter (36) (5.35) Parent + Older Child 190 28.23 Mother + Younger Daughter (36) (5.35) Father + Younger Son (35) (5.20) Mother + Younger Son (33) (4.90) Father + Younger Daughter (25) (3.72) Parent + Younger Child 129 19.17 Older Daughter + Younger Daughter (27) (4.01) Older Son + Younger Son (20) (2.97) Older Daughter + Younger Son (17) (2.53) Older Son + Younger Daughter (12) (1.78) Older Child + Younger Child 76 11.29 Total 673 100.00 The institutional significance of maintaining the status hierarchy is further demonstrated by the finding that the second greatest number of coalitions are between a parent and an older child (28%), followed by parent/younger child coalitions (19%), and finally by older child/younger child coalitions (11%). How Important is Power in Determining Who Wins and Loses? Coalition theorists see coalitions as a strategy that members use to attain their goals. Family members, like political party members in multi-party political systems, also prefer winning to losing and may form coalitions for that purpose. Table 2 is designed to answer the question as to what happens when there are disputes between family members aligned in coalitions or not aligned. [page 22] - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - [page 23] TABLE 2. COALITION WINS AND LOSSES WHEN OPPOSING AN INDIVIDUAL OR ANOTHER COALITION Opposing an Individual Coalition Wins Losses Total N Father + Older Child 70.8% 29.2% 100.0% 192 Father + Younger Child 68.2 31.8 100.0 192 Mother + Older Child 67.2 32.8 100.0 192 Mother + Younger Child 58.3 41.7 100.0 192 Chi Square DF Significance Min in E.F. Cells with E.F. < 5 7.58 3 0.056 65.00 None ________________________________________________________________ Opposing Another Coalition Coalition Wins Losses Total N Father + Older Child 54.2% 45.8% 100.0% 192 Father + Younger Child 41.1 58.9 100.0 192 Mother + Older Child 39.1 60.9 100.0 192 Mother + Younger Child 31.8 68.2 100.0 192 Chi Square DF Significance Min in E.F. Cells with E.F. < 5 20.65 3 0.0001 79.75 None The top half of the table presents the percentage of wins and losses when particular family coalitions are aligned against an individual opponent. The table shows that parent-child coalitions with the father included are more successful than those including the mother, and that parent-child coalitions with the older child are more successful than those with the younger child (Chi Square = 7.58, df = 3, p < .06). The lower half of the table shows a similar pattern of findings when the opponent is another coalition (Chi Square = 20.64, df = 3, p < .001). The fundamental finding is that family power structure remains the key to winning and losing. Coalitions that include the father are the strongest and, therefore, the most likely to win. By contrast, coalitions involving the younger child are the weakest and most likely to lose. Conservative Coalition Patterns The most common status order in a triad would be where A>B and B>C, and A<B+C. This is the familiar Type 5 pattern. The most likely coalition is an AB coalition because this facilitates A's maintainance of control and prevents a BC coalition (a revolutionary [page 23] - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - [page 24] one in that it upsets the existing power structure). The situation is somewhat different in the tetrad and in the family. Figure 1 presents two different conservative coalition structures in the four-person family. The first pattern shows a parental coalition dominating the family and opposing children. Since the parents are the two most powerful individual members of the system, a coalition between these two is virtually unopposable. The second pattern is quite different. This structure consists of two coalitions consisting of each of the parents and the older child. In this case, not only do the children oppose each other, but perhaps more significantly the parents co-opt the older child by forming a coalition that includes him or her. As noted by Selznick (1949), co-optation refers to the process of assimilating new elements into the policy-determining or leadership structure of a system. It is a policy which enables the group in charge of the social system to maintain its control. Hence, this structure as well as the structure shown in Figure 1, which consists of a simple spousal coalition, facilitate the maintenance of the existing status hierarchy. FIGURE 1. CONSERVATIVE COALITION PATTERNS IN FOUR-PERSON FAMILIES Parents Oppose Children A B Father * * * * Mother # # # # # # # # # # # # # # # # # # # # # # # # Younger Older Child Child D C TYPE 5 Conditions: 2 Parents, 1 Child Key A>B>C>D **** = coalition A<(B+C) #### = opponent A<(C+D) [page 24] - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - [page 25] Children Oppose One Another A B Father Mother * * * * * * * * * * * * Younger # # # # Older Child Child D C TYPE 6 Conditions: 1 Parent, Children A>B>C>D A<(B+C) A<(C+D) ______________________________________________________________ Adapted from Caplow (1968), p. 70. A Revolutionary Coalition Pattern In the triad, the most obvious revolutionary coalition is BC, which is an obvious threat to A, so much so that, as we noted above, it induces A to form a coalition with B to prevent a BC coalition. Again, things are not the same in tetrads or in families. Figure 2 presents one type of revolutionary coalition pattern that we found. In this diagram, we find that the second most powerful family member, the mother, forms separate coalitions with the older child and with the younger child, thereby isolating the powerful father. Thus, the father stands in opposition to all three of the other family members. FIGURE 2. A REVOLUTIONARY COALITION PATTERN IN FOUR-PERSON FAMILIES Father Opposes Others A B Father # # # # Mother # # * * # # * * # # * * # * # * # * # * # * # * Younger Older Child Child D C TYPE 5 Conditions: 2 Parents, 1 Child Key A>B>C>D **** = coalition A<(B+C) #### = opponent A<(C+D) ______________________________________________________________ Adapted from Caplow (1968), p. 71. [page 25] - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - [page 26] Is Family Composition Related to Coalition Structure? In order to enhance and maintain family stability, families develop norms that limit conflict in certain subsystems. Since the spousal subsystem is the most crucial one for family survival, conflict is least likely to be tolerated in that system. Indeed, solidarity in the spousal subsystem is essential for the survival of the system (Cousins, 1960). The finding that spousal coalitions were by far the most common type supports this perspective. Parent-child relationships are also important to family solidarity. Elsewhere (Grusky, Bonacich, and Peyrot, 1988) we have shown that male children are more involved in family conflict than female children. Conflict can contribute to family solidarity if it integrates the parents and ties them more closely to the family. We proposed that older sons enter conservative and avoid revolutionary coalitions to help maintain family solidarity. Table 3 shows that family composition is related to the average number of conservative coalitions. Specifically, it shows that there is a significant main effect: older son families are more likely than older daughter families to form conservative coalitions (df = 1, F = 7.23, P = .01). TABLE 3. FAMILY COMPOSITION AND AVERAGE NUMBER OF CONSERVATIVE COALITIONS Family Composition Mean Std. Dev. a. Older Daughter + Younger Son 7.00 2.04 b. Older Son + Younger Daughter 10.92 4.34 c. Two Girls 9.67 4.42 d. Two Boys 11.42 3.03 e. Older Daughter (a & c) 8.33 3.63 f. Older Son (b & d) 11.17 3.67 Table 4 provides additional support for this (and other alternative) formulations. Revolutionary coalitions and improper coalitions are much less frequent than are conservative coalitions. The mean number of conservative coalitions for the 48 families was 9.75, S.D. = 3.88; for the revolutionary coalitions, the mean was 1.56, S.D. = 2.75; and for improper coalitions, the mean was 2.70, S.D. = 1.94 (Conservative versus revolutionary coalitions, p < .001; and conservative versus improper coalitions, p < .001). Hence, the basic finding is that stable family organizations prefer conservative coalitions. Table 4 shows that older son families are less likely than older daughter families to form either revolutionary coalitions (df = 1, F = 5.55, P = .023) or improper coalitions (df = 1, F = 7.2, P = .01). [page 26] - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - [page 27] TABLE 4. FAMILY COMPOSITION AND AVERAGE NUMBER OF REVOLUTIONARY AND IMPROPER COALITIONS Revolutionary Coalitions Family Composition Mean Std. Dev. a. Older Daughter + Younger Son 2.08 2.68 b. Older Son + Younger Daughter .50 .67 c. Two Girls 2.83 4.45 d. Two Boys .83 1.12 e. Older Daughter (a & c) 2.46 3.61 f. Older Son (b & d) .67 .92 __________________________________________________________________ Improper Coalitions Family Composition Mean Std. Dev. a. Older Daughter + Younger Son 2.67 1.72 b. Older Son + Younger Daughter 3.42 1.83 c. Two Girls 1.33 1.16 d. Two Boys 3.42 2.32 e. Older Daughter (a & c) 2.00 1.59 f. Older Son (b & d) 3.42 2.04 DISCUSSION AND CONCLUSIONS This paper extends Caplow's theory of coalitions in triads to four-person groups, or tetrads. Organizationally, tetrads differ from triads in two major ways. First, tetrads are more complex and allow for greater opportunity for coalition formation. Willis (1962) has identified seventeen different types of coalitions in the tetrad and has predicted the most frequent kinds of two-way and three-way coalitions within each type. However, Willis did not apply his formulations to families. Second, in addition to their greater complexity, tetrads permit the possibility of counter-coalitions. Thus, we have applied Caplow's theory of power in triads to the study of four-person families, or tetrads, and have examined four questions: (1) How frequent are coalitions? We found that arguments led to coalitions in about three out of ten cases, leading to the formation of 673 coalitions. Although in this study we cannot answer the question as to whether or not coalitions are frequent or rare in American families, the data presented, at the very least, suggest that coalitions exist in many families, and consequently are worthy of study. [page 27] - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - [page 28] (2) What types of coalitions predominate in the family? We found that spousal coalitions were the dominant form. This finding is consistent with a theoretical approach that emphasizes the importance of maintaining family solidarity. (3) How important is power in determining who wins or loses? We found support for Caplow's model, asserting that power counts in family decision-making. Coalitions involving the father were the ones most likely to win; whereas those involving the younger child were the weakest and were most likely to lose. (4) Is family composition related to coalition structure? Some evidence was found that family composition is related to the formation of conservative, revolutionary, and improper coalitions. Older son families were less likely than older daughter families to form revolutionary or improper coalitions. ENDNOTE * This is a revised version of a paper presented at the Asian-Pacific Regional Conference of Psychology, International Union of Psychological Science, Guangzhou, China, 1995. Grusky and Bonacich gratefully acknowledge the support of National Science Foundation grant SOC-78-07131 and National Institute of Mental Health grant MH- 19127. REFERENCES Bonacich, Phillip, Oscar Grusky, and Mark Peyrot. 1985. "Family Coalitions: A New Approach and Method." Social Psychology Quarterly 48:42-50. Caplow, Theodore. 1968. Two Against One: Coalitions in the Triad. Englewood Cliffs, New Jersey: Prentice-Hall. Coleman, James. 1991. Foundations of Social Theory. Cambridge, Massachusetts: Belknap Press. Cook, Karen S. and Mary R. Gillmore. 1984. "Power, Dependence, and Coalitions." in Edward J. Lawler (ed.), Advances in Group Processes, Vol. 1, 27-58, Greenwich, Connecticut: Jai Press. Cousins, Albert N. 1960. "The Failure of Solidarity" in Norman W. Bell and Ezra F. Vogel (eds.), The Family, 403-416, Glencoe, llinois: The Free Press. Gamson, William A. 1961a. "A Theory of Coalition Formation." American Sociological Review, 26:373-382. [page 28] - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - [page 29] ------. 1962b."An Experimental Test of a Theory of Coalition Formation." American Sociological Review 26:565-573. Grusky, Oscar, Phillip Bonacich, and Mark Peyrot. 1988. "Group Structure and Interpersonal Conflict in the Family." in Edward J. Lawler and Barry Markovsky (eds.), Advances in Group Processes, Vol. 5, 29-51, Greenwich, Connecticut: Jai Press. McGinnis, Michael. 1991. "Richardson, Rationality and Restrictive Models of Arms Races." Journal of Conflict Resolution, 35:443-73. Rappoport, Anatol. 1957. Fights, Games, and Debates. Ann Arbor, Michigan: University of Michigan Press. Schelling, Thomas C. 1960. The Strategy of Conflict. London: Oxford University Press. Willis, Richard H. 1962. "Coalitions in the Tetrad." Sociometry 25:358-76. AUTHOR BIOGRAPHIES Oscar Grusky is Professor of Sociology at UCLA and Director, NIMH- supported AIDS research training program. His current research concerns the organization of managed mental health care systems for children and adults with emotional disorders and the role of the child in family coalitions. Email address: grusky@soc.sscnet.ucla.edu Phillip Bonacich is Professor of Sociology at UCLA. His current research interest is power and coalitions within exchange networks. Email address: bonacich@soc.sscnet.ucla.edu Cynthia Webster is Lecturer in Sociology at UCLA. She received her Ph.D. from the University of California, Irvine in 1993. Her current research interests are social networks and quantitative methods. Email address: webster@soc.sscnet.ucla.edu