PATTERNS OF COMMUNICATlON IN NORMAL AND CLINIC FAMILIES Thesis for the Degree of M. A. MICHIGAN STATE UNWERSITY LENNARD A. LEIGHTON 1968 L [BR :1 R Y Michigan 5cm Univaliy ”mumHImullflumuiwwgngufln 3 1293 i | ABSTRACT PATTERNS OF COMMUNICATION IN NORMAL AND CLINIC FAMILIES by Lennard A. Leighton This study was undertaken to determine whether normal and deviant families could be differentiated on the basis of their communicative interaction. Concrete measures such as number of times each member Spoke. average duration of Speech. total length of time each member spoke. number of interruptions. frequency of simultaneous speech. and number of double messages communicated were employed. Eight normal and seven deviant families (four or five mem- bers each) whose interactions were recorded by Moore (1966) were used. Normal families had no history of psychiatric disorder and were obtained through labor unions and church groups. Deviant families had been referred to the Michigan State Psychological Clinic because a male child between the ages of eight and thirteen was an underachiever and/or be- havior problem in school. No clinic family received any treatment during the course of the experiment. 1 Lennard A. Leighton It was hypothesized that the two groups of families would show significant differences on the above stated var- iables. Further. differences were expected between father and mother within the same group. More specifically. normal fathers spoke more often. for a greater total length of time and for a longer average duration. than clinic fathers. The reverse was true when normal and clinic mothers were compared. Clinic children spoke more often and for a greater total length of time than normal children. whereas average dura- tion of Speaking was approximately the same. Normal fami- lies showed fewer interruptions. fewer instances of simul- taneous speaking. and fewer numbers of double messages than did the clinic families. Comparisons within families revealed that the normal father spoke more often and for a greater total duration than normal mothers. while average duration of speech was equivalent. Clinic mothers spoke more often. for a greater total length of time. and for a longer average duration than did clinic fathers. Interruption data showed that there were no differences between the normal father and the normal mother. In the clinic families. the father interrupts more often than the mother. There is no difference in the number Lennard A. Leighton of times the normal father and mother are interrupted. while the clinic mother is interrupted more often than the clinic father. The latter difference is maintained even when num- ber of times speaking is held constant. The results led the author to infer patterns of dom— inance and submission in the two groups of families. The normal family is characterized by father-dominance which appears to be accepted by the other members of the family. The clinic family is characterized by mother-dominance when frequency and temporal measures of Speaking are considered. However. data on interruptions suggests that this is a rela- tively unstable power hierarchy which leads to a recurring struggle on the part of the mother to maintain her position. Finally. the effects of a pathological style of in- teraction on the children. and directions for further re— search are discussed. PATTERNS OF COMMUNICATION IN NORMAL AND CLINIC FAMILIES BY Lennard A; Leighton A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Psychology 1968 .’7 {/4/9 A " I7 if.) .Z> ’ / I... -" U. " .5-‘( DEDICATION To my parents, without whose constant encouragement and support I might never have come this far. ACKNOWLEDGMENTS I wish first to express my profound gratitude to Dr. Gary Stollak who provided the basic ideas from'which this project developed, and to Dr. Dozier Thornton for his helpful advice and criticism. A special word of apprecia- tion and thanks is due to Dr. Lucy Ferguson, my committee chairman, without whose continuing assistance this study could not have been accomplished. In addition, I would also like to thank Marv Moore for permitting me to use his tapes, and Michael Weiss for his invaluable help in rating them. iii TABLE OF CONTENTS Page ACKNOWLEDGMENTS. . . . . . . . . . . . . . . . . . . . iii LIST OF TABLES . . . . . . . . . . . . . . . . . . . . v LIST OF APPENDICES . . . . . . . . . . . . . . . . . . Vii Chapter I. INTRODUCTION . . . . . . . . . . . . . . . . . . 1 Statement of the Problem. . . . . . . . . . Relevant Literature . . . . . . . . . . . . u3H II. METHOD . . . . . . . . . . . . . . . . . . . . . 19 Subjects. . . . . . . . . . . . . . . . . . l9 Interviewing Procedure. . . . . . . . . . . 21 Rating Procedure. . . . . . . . . . . . . . 23 III. RESULTS. . . . . . . . . . . . . . . . . . . . . 25 IV. DISCUSSION . . . . . . . . . . . . . . . . . . . 43 V. SUMMARY. . . . . . . . . . . . . . . . . . . . . 51 REFERENCES . . . . . . . . . . . . . . . . . . . . . . 54 APPENDIX . . . . . . . . . . . . . . . . . . . . . . . 58 iv Table 10. LIST OF TABLES Comparison of Normal and Clinic Family Groups on Several Composition Criteria. . . . . . . Means and T Ratios for Total Number of Times Speaking, Normal versus Clinic Families. . . Means and T Ratios for Total Duration of Speech, Normal versus Clinic Families. . . . Means and T Ratios for Average Duration of Speech, Normal versus Clinic Families. . . . Means and T Ratios for Number of Times Each Member of the Family Interrupts Another Member, Normal versus Clinic Families. . . . Means and T Ratios for the Number of Times Each Family Member is Interrupted, Normal versus Clinic Families . . . . . . . . . . . Means and T Ratios for Total Number of Family Interruptions, Normal versus Clinic Families Means and T Ratios for Total Number of Times Speaking, Normal Father versus Normal Mother; Clinic Father versus Clinic Mother . Means and T Ratios for Total Duration of Speech, Normal Father versus Normal Mother; Clinic Father versus Clinic Mother . . . . . Means and T Ratios for Average Duration of Speech, Normal Father versus Normal Mother; Clinic Father versus Clinic Mother . . . . . Page 21 25 27 28 3O 32 34 35 36 37 LIST OF TABLES Table Page 1. Comparison of Normal and Clinic Family Groups on Several Composition Criteria. . . . . . . 21 2. Means and T Ratios for Total Number of Times Speaking, Normal versus Clinic Families. . . 25 3. Means and T Ratios for Total Duration of Speech, Normal versus Clinic Families. . . . 27 4. Means and T Ratios for Average Duration of Speech, Normal versus Clinic Families. . . . 28 5. Means and T Ratios for Number of Times Each Member of the Family Interrupts Another Member, Normal versus Clinic Families. . . . 30 6. Means and T Ratios for the Number of Times Each Family Member is Interrupted, Normal versus Clinic Families . . . . . . . . . . . 32 7. Means and T Ratios for Total Number of Family Interruptions, Normal versus Clinic Families 34 8. Means and T Ratios for Total Number of Times Speaking, Normal Father versus Normal Mother; Clinic Father versus Clinic Mother . 35 9. Means and T Ratios for Total Duration of Speech, Normal Father versus Normal Mother; Clinic Father versus Clinic Mother . . . . . 36 10. Means and T Ratios for Average Duration of Speech, Normal Father versus Normal Mother; Clinic Father versus Clinic Mother . . . . . 37 List of Tables (continued) Table 11. 12. l3. 14. 15. Page Means and T Ratios for Number of Interruptions by the Normal Father versus Normal Mother; Clinic Father versus Clinic Mother . . . . . 38 Means and T Ratios for Number of Times Inter- rupted, Normal Father versus Normal Mother; Clinic Father versus Clinic Mother . . . . . 40 Means and T Ratios for Number of Instances of Simultaneous Speech, Normal versus Clinic Families . . . . . . . . . . . . . . . . . . 41 Means and T Ratios for Number of Double Mes- sages Communicated, Normal versus Clinic . . 41 Means and T Ratios for Number of Times Inter- rupted Divided by Number of Times Speaking. Clinic Father versus Clinic Mother . . . . . 42 vi LIST OF APPENDICES Appendix Page A. Characteristics of Families Sampled in This Study. . . . . . . . . . . . . . . . . 58 vii PATTERNS OF COMMUNICATION IN NORMAL AND CLINIC FAMILIES INTRODUCTION Statement of the Problem The current trend in much psychological research is one that is moving away from the detailed study of the individual, and instead is moving toward the study of in— terpersonal relationships. This new approach has studied both agihgg small groups and family interaction situations. This trend in research is paralleling a similar movement toward conjoint therapy techniques such as group therapy and family therapy. In the latter case, mother, father, problem child, and other siblings are often seen together by one therapist. Concurrent with this move toward interaction re- search is a growing interest in communication theory; specifically its application to the study of interper- sonal relationships. Ruesch, Block, and Bennett (1953) l state that the vast majority of terms used in psychiatry refer to the communicative behavior of patients. “All psychopathology can be viewed as a disturbance of commun- ication." (p. 59) By focusing on communication we are describing observable ongoing events rather than the end products or end stages of processes. Riskin (1963) describes the family as an ongoing system; one in which certain repetitive and enduring tech- niques or patterns of interaction are developed over time. These techniques are developed and maintained as a means of regulating the equilibrium of the family. He further states that the family's manner of communication will elu— cidate the underlying interactional patterns of behavior. Farina and Dunham (1963) and others have shown that normal and abnormal families interact differently in similarly structured situations. If this is the case, an important step toward improved family treatment and diag- nosis would be the precise delineation of the differences in interaction techniques exhibited by these two types of family units. This study attempted to support the notion that normal and abnormal families differ significantly in their patterns of communication as reflected in measures such as the number of times each family member speaks, dura- tion of speaking, frequency of simultaneous speech, and number of interruptions. More specifically, it will be shown that the mother is the dominant member of the clinic family, while the father assumes this role in the normal family. Relevant Literature Jay Haley, a pioneer in communication and family interaction research, has stated that, There are increasing attempts to classify and describe the functioning of married couples and families as well as ongoing groups in in— dustry, military organizations, different psy- chotherapy situations, and other groups with a history. (1962, p. 265) These new variables being studied are precisely those which older forms of psychological research tried to elim- inate, i.e. interactional variables. The crucial differences between families would seem to reside in the sorts of trans- actions which take place between family mem- bers; the study of differences becomes a classification of communications patterns in the family. (p. 267) Haley suggests four basic assumptions of family study: (1) family members deal differently with each other than they do with other people, (2) the millions of responses which family members meet over time within a family fall into definite patterns, (3) these patterns persist within a family for many years and will influence a child's eXpectation of and behavior with other people when he leaves the family, and (4) the child is not a passive recipient of what his parents do with him, but an active co-creator of family patterns. For example, Bateson (1956) and Wynne (1958) have shown that families of schizophrenics have many similarities in common other than the schizophrenic behavior of one member. The way the parents deal with each other and with the children fall into certain basic patterns. The ultimate goal in family interaction research is to transfer these descrip- tive patterns into quantifiable variables that can be sta- tistically measured, and then to use these findings in im- proving our therapeutic techniques. The first half of this goal is the purpose of the present study. Haley goes on to say that the current trend is toward bringing families together and recording their in— teraction on tape and film. The study of communication between people, however, is exceedingly difficult. People communicate not only with their words, but also with vocal inflections and certain specific non-verbal means such as bodily movements. Bateson, in the previously cited study, placed schizophrenic as well as non—schizophrenic families in a standard interview situation. His procedure included leaving the family alone to talk together, in order to note similarities and differences in response to the sit— uation. This experiment was the prototype for many later studies in the field. Experimentation with a family provides unique problems that go beyond those encountered in research both with individuals and artificial groups. When exper- imenting with individuals, it is customary to eliminate the influence of interpersonal factors. In the case of family research, exactly the opposite goal is desired: it is the interpersonal factor which is to be measured. Experimentation and testing of individuals involves the exposure of the subject to some nonhuman stimulus and the measurement of his response to it. For experimentation using groups, one must create a standard context and place two or more people within it while measuring their re— sponses to each other. Then we must place two or more other people involved in some different form of relatedness in the same situation and measure their responses. Some clues to an adequate family research schema can be gained from the literature on small group experi- ments. There are, however, marked differences between these two types of interactional systems. The usual small group experiment consists of arranging a standard situa- tion and placing several unrelated people in it. The ef- fect of this particular context on their behaviors is then measured. The groups are carefully chosen so that the in- dividual members are not acquainted. This is done in order to isolate the effects of the particular setting on their behavior. In a family interaction experiment the goal is quite different. Here the problem is to measure how the members of a "group with a history" typically respond to each other, while attempting to eliminate as much as pos— sible the effects of the particular setting on their per- formance. A major problem in family experimentation is reducing the effects of the experimental situation on the typical patterns of interaction. Haley, in a later article (1964), discusses some of the methodological issues of family interaction research. He says that the ideal way to study a family system is to record it in operation. Numerous problems arise from such a procedure. The intended goal of this type of eXperimen- tation is to make reliable measurements of typical family interchanges while introducing as little bias as possible into the system being studied. Measuring the family in its natural habitat probably introduces the least amount of bias, but the measurements can only be the subjective reports of the experimenter. Placing families in struc- tured situations so that their communication with each other is in some way restricted, for example limited to pushing buttons (Haley, 1962), can lead to accurate mea— surement of the interaction. However, such measurement is greatly affected by the difference between the experi— mental situation and the family's typical means of commun- ication. Since family members normally communicate by conversation, it is most desirable to make measurements of that conversation. The ideal data in family interac- tion research should be the "recording of observable events which are accurately measurable, so that compari- sons and contrasts can be made." (p. 42) Ideally there should be no guesswork or inference in the data itself, even though the evaluation and interpretation of the data can, and often must, involve such inference. For example, the raw data may show that the mother in a clinic family speaks more often than the mother in a normal family. Such measurement involves no inference. The possible conclu— sion that the clinic family is mother-dominant and father— passive does indeed involve some inferential processes: that the number of times speaking is an index of dominance in the family, and that the experimental situation is rep- resentative of the normal patterns of communication that take place within the family. In his 1964 study, Haley analyzed a sample of eighty families (40 normal and 40 clinic) borrowed from Ferreira and Winter's 1965 study. He scored the frequency of all possible sequences of "who speaks after whom." He then proposed answers to two basic questions. First, "can we demonstrate that the family is an organization follow— ing repetitive interactional patterns?" Haley showed this to be true by demonstrating that in all families, both normal and clinic, the variations of conversation sequences differed significantly from random expectations. The sec- ond question he posed was "can we on some scale differentiate a disturbed family from a normal one?" He showed this to be possible by analyzing the patterned sequences of inter- action for the two groups. He found that the normal fam- ilies tended to make greater use of the possible interac- tion patterns than did the clinic families. In conclusion, Haley discussed data showing the effects of family therapy on the interaction patterns of the clinic families. After a number of therapy sessions, the clinic families showed a definite change toward the greater use of varied sequences of interaction. The earliest studies of family interaction are those of Bishop (1951) and Moustakas, Sigel, and Schalodk (1956). These experimenters limited their investigations to the direct observation of the mother—child dyad in a free play setting. A basic limitation of these studies was of course the omission of the effects of other family members on the particular dyad studied. Strodtbeck (1951) began the serious study of husband—wife interaction patterns by developing the Re- vealed Differences Technique. He asked each couple to choose three families with whom they were familiar. He then separated the couple and had each spouse state which 10 of the three families best met each of twenty-six differ— ent criteria. He then brought the couple together and asked them to discuss and reconcile their differences. Strodtbedk made use of Bales' (1950) interaction categor- ies and subsequently found that the partner who said the most tended to win more of the final decisions, while the partner who said the least generally passively agreed while making more overt signs of frustration and aggres- sion. In a later study (1954), Strodtbedk made use of three member families: parents and adolescent sons. He obtained a series of disagreements from each member's re- sponse to alternative solutions to parent—son conflicts. Again using the Bales interaction analysis, he attempted to compare the power relations within the families to those of 2Q.§2£ groups. Strodtbedk found that families in obvious disagreement tried to give the experimenter the impression that they never really disagreed. There seems to be some relationship between this phenomenon and Wynne's (1958) description of pseudo—mutuality. Follow-up studies using the concept of power were conducted by Farina (1960). He hypothesized that the characteristic interaction patternof parents of schizo- phrenic patients is that of dominance of one over the 11 other to a greater extent than would be expected in the case of parents of non—schizophrenic offspring. His study demonstrated that the sex of the dominant parent and the pattern of adjustment of the son are also important var— iables. Farina made use of the structured situation test developed by Rodnick and Garmezy (1957), and he found that maternal dominance was related to poor premorbid adjust- ment of the schiZOphrenic son, and father dominance was related to a generally good premorbid adjustment. He used thirty—six pairs of parents divided into three groups of twelve each. Of the two experimental groups, one had sons characterized by isolation and asexuality, and the other had sons who had been married and had friends. The con— trol group consisted of families whose sons were hospital- ized for tuberculosis. Interaction analyses derived from indices of dominance and conflict led Farina to conclude that good premorbid patients had fathers who were more assertive than those of poor premorbids; and that the in- teractions of the parents of poor premorbids were gener- ally characterized by more conflict and aggression than the interaction of the parents of good premorbid sons. Farina and Dunham (1963) studied the relationship between 12 the family and the patient's illness by including the schizophrenic son in the interaction. The findings were the same as in the earlier study. Farina and Dunham point out the necessity of establishing empirically valid scales of measurement that can be used to determine such descrip— tions as dominance and conflict. Truly meaningful research can only be carried out once such scales have been estab— lished. Levinger (1959) used interaction patterns (scored by Bales' technique) with self-perceptual data from Leary's Interpersonal Checklist (1957). The tasks consisted of problem situations and joint TAT stories, and required the families to reach an agreement on solutions and interpre— tations. His results show that mothers in clinic families participated most often and also exhibited significantly more negative emotional behavior. This data lends support to the notion that a reversal of the normal male-female roles tend to have an adverse effect on the children. Levinger also found that marital satisfaction (low dis- crepancy score between each of the spouse's "real and ideal self" descriptions of their partner on the Interper— sonal Cheeklist) was positively related to the partner's satisfaction with himself. 13 Ferreira (1963) continued the technique of using normal and pathological families. He had each member of the family reach a decision about three emotionally neu- tral items. Then he had the entire family try to reach a decision while considering the preferences of the indi- vidual members. He described four possible types of de- cisions: (1) unanimous decisions (family choice corre- sponded to the individual choice of every member), (2) majority decisions (family choice corresponded to the in- dividual choice of two members), (3) dictatorial decisions (family choice was that of only one member's preference), and (4) chaotic decisions (family choice corresponded to none of the individual preferences). The results show that there is significantly more agreement in the normal families than in the pathological families. A later study by Ferreira (1965) used a larger sample to replicate the earlier findings and also added new variables. He found that his earlier results were repeated, that abnormal fam— ilies took longer to reach a decision, and that they also showed a lower degree of appropriateness in their decisions than did normal families. Fisher, Boyd, Walker, and Sheer (1959) compared the interaction approach with the individual approach in 14 family research. Parents of twenty normal, twenty neu- rotic, and twenty schizophrenic men were compared on mea— sures assessing individual functioning as well as patterns of spouse interaction. As expected, the parents of the normal men were less disturbed than the parents of either of the two pathological groups. It was found, however, that only the interactional analysis differentiated the parents of neurotics from the parents of schizophrenics. As a result of these findings, the authors concluded that, Schizophrenia results from the combined malad- justment of both parents as they interact with the child. If a husband and wife combined forces in a relatively congruent manner, they tended to compensate for their individual pathologies. (p. 165) Framo makes the important point that the "lack of clarity in the communication between parents and between parent and child is more important in the etiology of schizophre— nia than the amount of open parental disagreement and con- flict." (p. 429) Caputo (1963) provides us with a different type of comparison. He investigated the dominant—mother and passive-father notion in families with schizophrenic sons. He proceeded to demonstrate the superiority of direct ob- servation of the interaction situation over the use of 15 paper-and~pencil tests. Using Osgood's Semantic Differen- tial (1957) and the Parent Attitude Inventory, Caputo found relatively nonpathological interactional patterns for the parents of schiz0phrenic sons. However, when he used the Bales interaction analysis, he discovered a con- siderable degree of antagonism and hostility between the parents. These two studies give us a strong indication of the necessity of analyzing the ongoing interaction within families. Neither the assessment of individual members nor the use of paper-and-pencil measures of inter- action is sufficient. In a continuation of Haley's communication analy— sis, Lennard, Beaulieu, and Embrey (1965) studied the com- munication sequence in twenty 3-member families (ten normal and ten with a schizophrenic son). The families were re- quired to discuss three topics such as "when a boy needs a helping hand with his homework, is it better for the mother or father to help out?" Sequential analysis of the interactions revealed that significantly less communica- tion flowed from son to father and vice versa in the case of the abnormal families. This was also true for mutual communication between mother and father. These results 16 lend further credence to the idea of the passive—father and dominant—mother in schizophrenic families. Riskin (1963 and 1964) used skilled clinicians to rate family members' speeches on several dimensions: com— munication clarity, topic shifts, agreement with previous speeches, and affective intensity. When an experienced clinician then listened to the previously rated tapes he missed a significantly large amount of detail that was ob— tained by microscopic analysis. Thus we can see the im- portance of restricting our investigations to easily spec- ifiable and concrete variables. Moore (Ph.D. dissertation, 1966) added the longi- tudinal dimension to the study of interaction in normal and clinic families. He posed two hypotheses: 1) judges' ratings of family interaction observed in the standard in- terview would be relatively similar in two interaction ses— sions ten weeks apart, and 2) these interaction ratings would reveal differences between the normal and clinic families. He utilized trained raters to score each fam- ily on the Family Rating Scale developed at Michigan State University. The families were required to perform various tasks including planning an activity, discussion of desired 17 changes within the family, and discussion of problem sit— uations. Moore found that normal families could be differ— entiated from clinic families by an overall pathology score and on a number of items on the Family Rating Scale. In comparison to the clinic families, the normal families were characterized by more inter-member agreement, a greater capacity for reaching common decisions in an equal- itarian fashion, less overall anger but greater tolerance of individual independence in thought and action, more in- terpersonal warmth, less manifest tension, and a greater degree of happiness. Both parents in the normal family displayed more overall satisfaction and effectiveness within their various roles. Regarding consistency, he found that "there exists a core of interaction consistency over time for both experimental groups." (p. 57) This finding lends support to the notion of permanence in fam— ily interaction patterns. An unexpected finding showed that normal families were rated as significantly less pathological after the second interview, while no such differences were_found for the clinic families. This re- sult may show a basic difference between normal and clinic families: that of being able to profit from experience 18 and perfect smoother ways of carrying out experimental tasks when encountered for a second time. METHOD Subjects Tape recordings of the family interaction sessions recorded by Marv Moore (1966) were used. The families for the study consisted of four or five member units (both par- ents and two or three children) all meeting the following criteria: 1) every family lived together for at least four uninterrupted years previous to participation in the research, 2) children ranged in age from eight to seven- teen, 3) all families included at least one male child be- tween the ages of eight and thirteen, and 4) they met the criteria listed below for inclusion in one of the two ex- perimental groups. The normal group consisted of eight families none of whose members had ever received or was recommended to receive any type of psychiatric treatment for an emotional or nervous disorder. Normal families were obtained from two sources. Three families volunteered as a result of a call for subjects at local labor union meetings, and five 19 20 volunteered after being recommended by their minister as representing the "most emotionally mature" families in his congregation. For their cooperation all normal families received $10 for each of two interviews. The clinic group consisted of seven families waiting for psychotherapy at the Michigan State University Psychological Clinic. No family received any treatment during the course of the ex- periment. All families initially contacted the clinic be- cause a male child between the ages of eight and thirteen had been referred for underachievement and/or lack of be- havior control in school. The clinic families received no remuneration for their part in the project, because partic- ipation in ongoing research was part of the treatment agree— ment. Moore used a total of sixteen families in his study (eight normal and eight clinic), but due to an incomplete collection of recordings only eight normal and seven clinic families were used in the present study. Table 1 shows that the two groups are essentially the same in composition except for the mean level of fathers' education which was 1.2 years higher for the clinic sample. This difference was not significant. Complete breakdowns of characteris- tics of each family are shown in Appendix A. 21 Table l.—-Comparison of Normal and Clinic Family Groups on Several Composition Criteria Mean Mean Mean years of number of age of Family completed children children education per family per family Father Mother Normal 12.8 12.2 2.3 10.6 Clinic 14.0 12.1 2.6 10.4 Interviewing Procedure At the beginning of each interview, each family was told that they would be observed by two raters through a one-way mirror, and that the purpose of the experiment was to increase our skills in helping families. Prelimin- ary remarks before the actual session began were designed to put the family at ease. Although Moore used nine sep- arate experimental tasks, the present study focused only on the first three. Task 1: The interviewer saw each family member separately just long enough to ask him the following 22 question: "At this point in time what changes would you like to see made in your family, as a whole or in any par— ticular members?" While the experimenter received this information, the remaining family members waited in an ad- joining room with instructions not to discuss the question among themselves. After each member had been seen, the family met as a whole in the experimental room, and was then asked to carry out the following instructions: "Dis- cuss among yourselves the question I have just asked each of you separately. You may discuss any aspect of the ques- tion you wish. The only specific request I make is that at some point you talk about specific steps you might take as a family to bring about any of the desired changes. You will have about four minutes, or more if you need it. I will not take part in your family discussion, but will re— main quietly in the room." It was at this point that the rating in the current study began. Task 2: The entire family was instructed: "Plan an activity you could all do together; it should be some— thing you might actually do. I will leave the room for four or five minutes. Choose one person to summarize your plans for me when I return." 23 Task 3: The parents received the following prov- erb: "While the cat's away the mice will play." The ex- perimenter asked them to discuss this proverb and then to plan how they would teach it to their children. Upon the parents' request, the interviewer brought the children badk into the room and they were taught the proverb. The rating for this study began once the parents started teach— ing the proverb to the children. Rating Procedure The tapes were scored for the following items: 1) total number of times each family member spoke, 2) total length of time each family member spoke, 3) average dura- tion of speech for each family member, 4) total number of times any one family member interrupted another, 5) num— ber of times each family member was interrupted, 6) num- ber of instances of simultaneous speech (two or more fam— ily members speaking at once), 7) total number of double messages communicated. The two temporal measures were ob- tained by using a stop watch and were calculated in seconds. Totals and means were derived for each of the three 24 experimental tasks as well as overall means and totals (across tasks). Due to the difficulty of determining which child was speaking, the data for all children in each family were combined for these analyses. There were a total of eighteen children in each experimental group. The data for double messages were the combined scores of two raters, as this particular variable was also part of another study carried on concurrently using the same sample. Interrater reliability was .89, as measured by the Pearson r. The rater did not know whether the family being rated was a normal or clinic family. This information was available only for the purposes of data analysis. RESULTS The first hypothesis stated that there would be a significant difference between normal and clinic families in the total amount of time that each family member spoke. Means and t ratios are presented in Table 2. Table 2.--Means and t Ratios for Total Number of Times Speaking, Normal versus Clinic Families l a========= Variant Normal Clinic T ratio P level Father: Task 1: 11.3 8.4 1.02 .20 Task 2: 31.4 18.0 2.20 .05 Task 3: 17.9 7.7 2.96 .02 Overall: 60.6 34.1 3.03 .01 Mother: Task 1: 5.2 12.3 4.01 .01 Task 2: 15.6 23.4 2.30 .03 Task 3: 5.5 10.6 1.61 .10 Overall: 26.3 46.3 3.13 .01 25 26 Table 2 (continued) Variant Normal Clinic T ratio P level Children: Task 1: 15.0 17.6 1.79 .10 Task 2: 27.3 32.7 2.25 .05 Task 3: 9.8 10.8 1.03 .40 Overall: 52.1 61.1 2.27 .05 The results show that the normal father speaks more often than the clinic father, with all t ratios except for task 1 significant at least at the .05 level. The clinic mother speaks more often than the normal mother with all t ratios except for task 3 significant at least at the .05 level. Clinic children speak more often than normal children with only task 2 and overall means significantly different at the .05 level. The second hypothesis stated that there would be a significant difference between normal and clinic families in the total duration of speaking time for each member. Means and t ratios are shown in Table 3. 27 Table 3.-—Means and t Ratios for Total Duration of Speech, Normal versus Clinic Families Variant Normal Clinic T ratio level Father: Task 1: 54.9 44.7 1.01 .20 Task 2: 78.3 66.0 1.31 .20 Task 3: 48.9 32.3 1.75 .10 Overall: 182.1 143.0 3.15 .01 Mother: Task 1: 22.9 55.4 3.25 .01 Task 2: 50.4 65.9 1.69 .10 Task 3: 30.0 44.1 1.54 .20 Overall: 103.3 165.4 5.68 .001 Children: Task 1: 21.6 48.9 3.03 .01 Task 2: 63.6 100.6 3.25 .01 Task 3: 32.5 40.3 2.37 .05 Overall: 117.7 189.8 4.89 .001 28 The results show that the normal father speaks for a greater length of time than the clinic father, with only the overall t ratio significant at the .01 level. Al- though not reaching this level, differences for the three tasks are all in the eXpected direction. The clinic mother speaks for a greater length of time than the normal mother, with t ratios on task 1 and overall significant at least at the .05 level. Again, tasks 2 and 3 show differences in the expected direction. Clinic children speak for a greater length of time than normal children, with all t ratios significant at least at the .05 level. The third hypothesis stated that there would be a significant difference between normal and clinic families in the average duration of speaking time for each member of the family. Means and t ratios are shown in Table 4. Table 4.——Means and t Ratios for Average Duration of Speech. Normal versus Clinic Families .- ‘ _.-— - Variant Normal Clinic T ratio P level _,_ Father: Task 1: 7.6 4.1 2.47 .05 29 Table 4 (continued) 1" Variant Normal Clinic T ratio P level Task 2: 3.4 4.0 No Difference* Task 3: 8.9 4.5 2.67 .02 Overall: 19.9 16.6 2.26 .05 Mother: Task 1: 4.5 6.1 1.09 .30 Task 2: 3.1 6.3 1.12 .30 Task 3: 7.9 10.9 1.55 .20 Overall: 15.5 23.3 2.44 .05 Children: Task 1: 3.2 2.8 No Difference Task 2: 2.4 5.1 1.26 .30 Task 3: 4.3 4.5 No Difference Overall: 9.9 12.4 1.31 .30 *For all tables, "No Difference" means a t ratio with a probability greater than .50. It can be seen from Table 4 that the normal father speaks significantly longer each time he speaks than does the clinic father, with all t ratios except task 2 significant 30 at least at the .05 level. The clinic mother speaks longer each time she speaks than does the normal mother with only the overall t ratio significant at the .05 level. Differ- ences for the three tasks are in the expected direction. Clinic children tend to speak for longer average durations in task 2 and overall. Neither t ratio is significant at the .05 level. The fourth hypothesis is that there will be a sig— nificant difference in the number of times each member of the family interrupts another when normal and clinic fam- ilies are compared. Means and t ratios are shown in Table 5. Table 5.--Means and t Ratios for Number of Times Each Mem- ber of the Family Interrupts Another Member, Normal versus Clinic Families Variant Normal Clinic T ratio P level Father: Task 1: 0.62 1.00 1.75 .20 Task 2: 1.25 2.30 2.60 .02 Task 3: 0.63 1.75 2.18 .05 Overall: 2.50 5.05 2.21 .05 ‘mfi~. Table 5 (continued) 31 Variant Normal Clinic T ratio P level Mother: Task 1: 0.00 1.00 2.86 .02 Task 2: 0.88 1.30 1.09 .30 Task 3: 0.38 0.71 1.27 .20 Overall: 1.26 3.01 2.62 .03 Children: Task 1: 0.38 1.70 2.20 .05 Task 2: 2.40 5.10 2.25 .05 Task 3: 0.25 1.40 1.25 .30 Overall: 3.03 8.20 3.14 .01 The results show that the clinic father interrupts more often than the normal father, with only the task 1 t ratio failing to reach the .05 level of significance. The clinic mother interrupts more often than the normal mother, with task 1 and overall t ratios significant at least at the .05 level. Tasks 2 and 3 show differences in the expected di— rection. Clinic children interrupt more often than normal 32 children, with only the task 3 t ratio failing to reach the .05 level of significance. The fifth hypothesis is that there will be a sig- nificant difference in the number of times each family mem- ber is interrupted by another family member when normal and F3 clinic families are compared. Means and t ratios are shown in Table 6. f Table 6.——Means and t Ratios for the Number of Times Each Family Member is Interrupted, Normal versus Clinic Families t;========= Variant Normal Clinic T ratio P level Father: Task 1: 0.38 0.14 No Difference Task 2: 1.60 3.00 1.51 .20 Task 3: 0.00 0.58 1.49 .20 Overall: 1.98 3.72 1.56 .20 Mother: Task 1: 0.37 1.10 1.28 .30 Task 2: 1.50 1.60 No Difference Task 3: 0.50 0.60 No Difference Overall: 2.37 3.30 1.39 .20 33 Table 6 (continued) Variant Normal Clinic T ratio P level Children: Task 1: 0.38 1.70 2.27 .05 Task 2: 1.40 4.90 3.93 .01 Task 3: 0.75 1.90 1.26 .30 Overall: 2.53 8.50 4.65 .001 Comparisons between normal fathers and clinic fathers, and normal mothers and clinic mothers regarding number of times they are interrupted show no t ratios to be signifi- cant at the .05 level. The trend in both cases, however, seems to be toward a greater number of times interrupted in the clinic families. Clinic children are interrupted significantly more often than normal children, with all t ratios except task 3 significant at the .05 level at least. The sixth hypothesis is that there will be a sig- nificant difference in the total number of family interrup- tions in the clinic and normal families. Means and t ra- tios are shown in Table 7. 34 Table 7.-—Means and t Ratios for Total Number of Family Interruptions, Normal versus Clinic Families Variant Normal Clinic T ratio P level Task 1:» 1.00 3.70 2.27 .05 Task 2: 4.53 9.60 2.73 .02 Task 3: 1.25 2.86 1.84 .10 Overall: 6.78 16.16 2.89 .02 The results indicate that the clinic families exhibit a greater number of total interruptions than do normal fam- ilies, with all t ratios except task 3 significant at least at the .05 level. Hypothesis seven states that there will be signifi- cant difference in the total number of times speaking for the normal father compared to the normal mother; and the clinic father compared to the clinic mother. Means and t ratios are presented in Table 8. It can be seen from the results that the normal father speaks significantly more often than the normal mother, with all t ratios significant at least at the .05 level. Clinic mothers speak signifi- cantly more often than clinic fathers, with all t ratios F" 35 Table 8.-—Means and t Ratios for Total Number of Times Speaking, Normal Father versus Normal Mother; Clinic Father versus Clinic Mother . Normal Normal . Variant Father Mother T ratio level Task 1: 11.3 5.2 2.19 .05 Task 2: 31.4 15.6 3.07 .01 Task 3: 17.0 5.5 3.24 .01 Overall: 60.6 26.3 3.69 .001 Clinic Clinic Father Mother Task 1: 8.4 12.3 2.33 .05 Task 2: 18.0 23.4 2.71 .02 Task 3: 7.7 10.6 2.85 .02 Overall: 34.1 46.3 2.69 .02 also significant at least at the .05 level. These findings seem to indicate the reversal of the "normal" role relationship in the clinic family. male—female Hypothesis eight states that there will be a sig- nificant difference in the total duration of speech of the normal father compared with the normal mother; and the 36 clinic father compared with the clinic mother. Means and t ratios are presented in Table 9. Table 9.--Means and t Ratios for Total Duration of Speech, Normal Father versus Normal Mother; Clinic Father versus Clinic Mother Normal Normal Variant Father Mother T ratio P level Task 1: 54.9 22.9 2.34 .05 Task 2: 78.3 50.4 3.11 .01 Task 3: 48.9 30.0 3.07 .01 Overall: 182.1 103.3 6.26 .001 Clinic Clinic Father Mother Task 1: 44.7 55.4 2.17 .05 Task 2: 366.0 65.9 No Difference Task 3: 32.3 44.1 2.54 .05 Overall: 143.0 165.4 2.74 .02 The results show that the normal father speaks for a longer time than does the normal mother. All t ratios are signif- icant at least at the .05 level. The clinic mother speaks for a longer amount of time than does the normal mother, 37 with all t ratios except task 2 significant at least at the .05 level. Hypothesis nine states that there will be a signif- icant difference in the average duration of speech for the normal father compared with the normal mother; and the clinic father compared with the clinic mother. Means and t ratios are shown in Table 10. Table 10.-—Means and t Ratios for Average Duration of Speech, Normal Father versus Normal Mother; Clinic Father versus Clinic Mother Variant :::::: 32:22: T ratio P level Task 1: 7.6 4.5 1.4 .20 Task 2: 3.4 3.1 No Difference Task 3: 8.8 7.9 No Difference Overall: 19.9 15.5 1.17 .30 Clinic Clinic Father Mother Task 1: 4.1 6.1 1.89 .10 Task 2: 4.0 6.3 1.84 .10 Task 3: 4.5 10.9 2.23 .05 Overall: 16.6 23.3 3.75 .01 38 It appears that there is no significant difference in aver- age duration of speech for the normal parents. Task 1 and overall t ratios show a trend toward the normal father speaking longer each time he speaks. The clinic mother's average duration of speech is significantly longer in task 3 and overall, with t ratios significant at least at the .05 level. Tasks 1 and 2, although not significant at the .05 level, are in the expected direction. 1 The tenth hypothesis states that there will be a ! significant difference in the number of times the normal father interrupts another family member compared with the normal mother; and the clinic father compared with the clinic mother. Means and t ratios are presented in Table 11. Table 11.-—Means and t Ratios for Number of Interruptions by the Normal Father versus Normal Mother; Clinic Father versus Clinic Mother , Normal Normal . Variant Father Mother T ratio P level Task 1: 0.62 0.00 No Difference Task 2: 1.25 0.88 No Difference Task 3: 0.63 0.38 No Difference Overall: 2.50 1.26 No Difference 39 Table 11 (continued) Variant Clinic Clinic T ratio P level Father Mother Task 1: 1.00 1.00 No Difference Task 2: 2.30 1.30 2.47 .05 Task 3: 1.75 0.71 1.39 .20 Overall: 5.05 3.01 2.20 .05 The results show no difference in the number of times the normal father interrupts compared with the number of times the normal mother interrupts. The clinic father interrupts more often than the clinic mother, with t ratios for task 2 and overall significant at the .05 level. The ratio for task 3 is in the expected direction. Hypothesis eleven states that there will be a sig- nificant difference in the number of times the normal father is interrupted compared to the normal mother; and the clinic father compared to the clinic mother. Means and t ratios are presented in Table 12. The results show no difference in the number of times the normal father is interrupted com— pared to the normal mother. The clinic mother is interrupted more often than the clinic father, with all t ratios except task 3 significant at least at the .05 level. 40 Table 12.--Means and t Ratios for Number of Times Inter- rupted. Normal Father versus Normal Mother; Clinic Father versus Clinic Mother Normal Normal Variant Father Mother T ratio P level Task 1: 0.38 0.37 No Difference Task 2: 1.60 1.50 No Difference Task 3: 0.00 0.50 No Difference Overall: 1.98 2.37 No Difference Clinic Clinic Father Mother Task 1: 0.14 2.10 2.66 .02 Task 2: 1.12 3.60 2.19 .05 Task 3: 0.58 0.60 No Difference Overall: 1.84 6.30 3.76 .01 Hypothesis twelve states that there will be a sig- nificant difference in the number of instances of simultan- eous speech when normal and clinic families are compared. Means and t ratios are shown in Table 13. The results show that the clinic families demonstrate a greater incidence of simultaneous speaking than do normal families. All t ratios are significant at least at the .05 level. 41 Table l3.--Means and t Ratios for Number of Instances of Simultaneous Speech, Normal versus Clinic Families A Variant Normal Clinic T ratio P level Task 1: 0.50 2.10 2.22 .05 Task 2: 2.00 3.70 2.17 .05 Task 3: 0.25 2.10 2.69 .02 Overall: 2.75 7.90 3.07 .01 Hypothesis thirteen is that there will be a signif— icant difference in the number of double messages communi- cated when normal and clinic families are compared. Means and t ratios are presented in Table 14. Table 14.--Means and t Ratios for Number of Double Messages Communicated, Normal versus Clinic Families Variant Normal Clinic T ratio P level Task 1: .06 .79 No Difference 1 Task 2: .06 .36 2.50 .05 Task 3: .06 .00 No Difference Overall: .18 1.15 3.03 .01 fl __V. 42 The results show that the clinic families communicate more double messages than do normal families, with t ratios for task 2 and overall significant at least at the .05 level. The fourteenth hypothesis is that the difference in the number of times interrupted for the clinic father and clinic mother will remain when interruptions are divided by number of times speaking. Means and t ratios are presented in Table 15. Table 15.——Means and t Ratios for Number of Times Inter- rupted Divided by Number of Times Speaking, Clinic Father versus Clinic Mother Clinic Clinic Variant Father Mother T ratio P level Task 1: .04 .29 2.25 .05 Task 2: .11 .34 2.39 .05 Task 3: .01 .13 2.19 .05 Overall: .16 .76 3.31 .01 The results show that the hypothesis is confirmed, with all t ratios significant at least at the .05 level. DISCUSSION The results obtained from this study show that when normal and clinic families are placed in identical situa— tions and are asked to perform the same tasks, clear differ— ences in their manner of communication can be demonstrated. The implications of this very broad statement are exceed— ingly important and will be discussed later. More specif- ically, certain basic differences in the interactive styles of the two groups of families are suggested. The results may be broken down into two basic areas: who is the domin- ant member of the family, and clarity of communication be— tween family members. The problem of dominance within the family is a complex one to discuss. Much of the literature presented in the introduction to this paper dealt with the reversal of the customary father-mother role relationship in the families of disturbed children. Most of these earlier studies utilized families that included a schizophrenic child or young adult. These studies showed the dominant— mother and passive-father relationship to be the typical 43 44 pattern in the schizophrenogenic family. The results of the present study show that the clinic mother speaks more often, for a greater total length of time, and longer on the average than does the clinic father. Such concrete measures of interactive and communicative styles can be ex— F tended in meaning to indicate the nature of dominance and submission in the family. If we can assume that these mea- sures of speaking time and frequency are indications of dominance, which this experimenter and others have done, then the results seem to indicate the reversal of mother— father roles in the family extends beyond the schizophrenic families, and includes families whose sons are underachiev— ers and/or behavior problems in school. This result points the way to further research in the area of family interac- tion. Different pathological families need to be rated using the same techniques to determine whether this role reversal is common to all types of abnormal families, and if so, why the reversal produces adverse effects on the children of these families. In the case of the normal family, a definite pat- tern of dominance is also demonstrated, but it seems to be less clear-cut than in the clinic family. The normal father 45 speaks more often and for a greater total length of time than does the normal mother, but their average duration of speech is approximately the same. This result might tend to indicate a more democratic power relationship in the normal family than is the case in the clinic family. This tentative conclusion is in line with those of earlier work in the field. The data regarding the number of interruptions and number of times interrupted reveal further important dif— I? ferences between the two groups of families, and it also gives us an indication of how the two groups accept the prevailing conditions of dominance and submission. The clinic father interrupts more often than the normal father, perhaps in an attempt to "fight" his passive position in the family. The clinic mother, in spite of her apparent dominant position, interrupts more often than the normal mother. It seems as if the clinic mother must constantly struggle to maintain her dominant position, while the nor- mal mother is content with her respective role in the fam- ily hierarchy. The normal father and normal mother are interrupted the same number of times in spite of the fact that the normal father speaks for a greater total length 46 of time and more often. The clinic mother, on the other hand, is interrupted more often than the clinic father even when differences in number of times speaking are equated. Again, we see a dissatisfaction in the clinic family re- garding the relative positions of the family members. The data comparing normal and clinic children are just as revealing. Results show that the clinic child speaks more often and for a greater length of time than does the normal child. However, there is little difference E in the average duration of speech for the two groups of children. Interruption data show that the clinic child is interrupted more often and interrupts more often than the normal child. It seems as though the clinic child must interrupt in order to have his views considered by the fam- ily. He may also be more involved in the conflict between the parents. In conclusion, we see that the father is the dom— inant member of the normal family, and this dominant role is accepted by the other members of the family. The clinic mother, on the other hand, appears to be the dominant member of the clinic family on the basis of speaking time and fre- quency of speaking. However, the interruption data shows 47 that this state of affairs is not acceptable to the other members of the family, and that the clinic mother must con- tinually exert herself in order to maintain her unstable position of dominance. Caputo's (1963) study showed that paper-and-pencil techniques of assessment are not suffi— cient to gain an accurate picture of the interactive styles of families. In a similar manner, the present study sug- J- gests that one or two measures of dominance alone are [I equally lacking when trying to present the total picture of family interaction. Future research must make use of multiple measures of significant variables in order to be most meaningful. The second major area of difference between the two groups of families is the clarity of communication be- tween family members. Studies cited earlier showed that the clinic families exhibited less communication clarity, less inter-member agreement, and fewer democratic decisions than did normal families. ‘The results from the present study add support to these findings in a number of ways. u _ . _A.__ 4___4._l The clinic families show a significantly greater number of instances of simultaneous speaking than do normal families. This finding lends credence to earlier results showing 48 greater amounts of conflict and less communication clarity in clinic families. The members of the clinic families seem less able to follow democratic techniques in reaching family decisions. The interruption data add further evi— dence of this inability to reach democratic decisions. The clinic family exhibits a significantly greater number of total interruptions than does the normal family. The clinic family also communicates a significantly greater number of , double messages than is the case for the normal family. I‘— These findings seem to reveal an interrelationship that is quite important. An unacceptable power hierarchy and a comparatively large number of double messages may lead to a high incidence of interruptions and simultaneous speech. Only in this way can the members of the clinic family attempt to make their preferences known. Once the democratic process breaks down (as seems to be the case in the clinic family), individual members must resort to dis- ruptive techniques in order to gain a voice in family de- cisions. Once the necessity of interruptions is established, a vicious circle is begun in which one interruption leads to another, with the breakdown in communication as the net result. 49 The results of the present study give us some in- dication of what led the abnormal family to seek help at the clinic. From the data we see that there tends to be more conflict and fewer means of resolving these conflicts because of the breakdown in communication in the clinic families. Moore's (1966) data revealed that the normal families showed a greater tolerance for independence in thought and action than did the clinic families. One might conclude that the children, who are most susceptible to the adverse effects of the pathological interactive style of the clinic family, must seek alternative ways of eXpressing themselves. The inability of the clinic family to tolerate the independence strivings and attempts at self- expression of the children, forces the child to make these attempts in the school setting. This particular situation may be as rigid and constricting as the clinic family and the child's behavior may be seen as more inappropriate there, thus leading to the child being referred to the clinic as a behavior problem. This interpretation can only be a tentative one until further research can show a correlation between the specific problem of the child, and its relation to the interactive style of the family in which he is raised. ’ 1"“ f. 9 [T 50 We may now return to the broad statement made ear- lier regarding the clear demonstration of differences in the communicative styles of the two groups of families. The implications of such a demonstration are crucial to the area of family interaction research, diagnostics, and therapy. The present study is really only one of many pilot studies that are indicating that much work needs to be done in this area in order to gain a full understanding of the dynamics of family interaction. These many and varied studies have shown that normal and clinic families can be differentiated on a number of variables ranging from aspects of communication to techniques of decision- making. The research, however, cannot end here. Once the critical variables have been isolated, the job becomes one of applying these variables to diagnostic work in family treatment clinics and to perfecting our therapeutic tech- niques. Haley's (1964) pilot work on teaching clinic fam- ilies to clarify and broaden their scope of communication among the members of the family is such an attempt. Only when our strictly theoretical findings can be translated into techniques of helping families will the true goal of family interaction research be reached. 14 S UMMARY This study was undertaken to determine whether normal and deviant families could be differentiated on the foi- basis of their communicative interaction. Concrete meas- ures such as number of times each family member spoke, total length of time each member spoke, average duration of speech, number of interruptions, frequency of simultan- : eous speech, and number of double messages communicated were employed. Eight normal and seven abnormal families (four or five members each) whose interactions were re- corded by Moore (1966) were used. Normal families had no history of psychiatric disorder and were obtained through labor union and church groups. Deviant families had been referred to the Michigan State Psychological Clinic because a male child between the ages of eight and thirteen was an underachiever and/or behavior problem in school. No clinic family received any treatment during the course of the ex— periment. 51 52 It was hypothesized that the two groups of families would show significant differences on the above stated var- iables. Further, differences were expected between father and mother within the same family. More specifically, normal fathers spoke more often, for a greater total length of time, and for a longer average duration than clinic fa- thers. Clinic mothers showed the same pattern when compared with normal mothers. Clinic children spoke more often and for a greater total length of time than normal children, whereas average duration of speaking time was approximately the same. Normal families showed fewer interruptions, fewer instances of simultaneous speaking, and fewer numbers of double messages than did the clinic families. Comparisons within families revealed that the normal father spoke more often and for a greater total duration than normal mothers, while average duration of speech was equivalent. Clinic mothers spoke more often, for a greater total length of time, and for a longer average duration than did clinic fathers. Interruption data showed that there were no differences between the normal father and the normal mother. In the clinic families, the father interrupts more often than the mother. There is no difference in the number 53 of times the normal father and mother are interrupted, while the clinic mother is interrupted more often than the clinic father. The latter difference is maintained even when number of times speaking is held constant. The results led the author to infer patterns of dominance and submission in the two groups of families. The normal family is characterized by father-dominance which appears to be accepted by the other members of the family. The clinic family is characterized by mother- dominance when frequency and temporal measures of speaking are considered. However, data on interruptions suggest that this is a relatively unstable power hierarchy which leads to a recurring struggle on the part of the mother to maintain her position. Finally, the effects of a pathological style of interaction on the children, and directions for future research are discussed. REFERENCES REFERENCES Bales, R. F. "Some uniformities of behavior in small so- cial systems," Readings in Social Psychology. Ed. G. E. Swanson, T. M. Newcomb, and E. L. Hartley. New York: Holt, 1952. Bateson, G., Jackson, D. D., Haley, J., and Weakland, J. H. "Towards a theory of schizophrenia," Behavioral Sciences, 1, 251-264, 1956. Bishop, B. M. "Mother-child interaction and the social behavior of children," Psychological Monographs, No. 11, 65, 1-43, 1951. Caputo, D. V. "Parents of schizophrenics," Familerrocess, 2, 339-356. 1963. Farina, A. "Patterns of role dominance and conflict in parents of schizophrenics," Journal of Abnormal and Social Psychology, 61, 31-38, 1960. Farina, A., and Dunham, R. N. "Measurement of family re- lationships and their effects," Archives of General Psychiatry, 9, 64—73, 1963. Ferreira, A. J. "Decision making in normal and patholog— ical families," Archives of General Psychiatry, 8, 68-73. 1963. Ferreira, A. J., and Winter, W. D. "Family interaction and decision making," Archives cf General Psychiatry, 13. 214-223. 1965. Fisher, 8., Boyd, 1., Walker, D., and Sheer, D. "Parents of schiZOphrenics, neurotics, and normals," AE- chives of General Psychiatry, 1, 149—166, 1959. 55 56 Framo, J. L. "Systematic Research on Family Dynamics," Intensive Family Therapy. Ed. Boxzormeny-Nagy, I. and Framo, J. L., Hoeber Medical Division, Harper and Rowe, New York, 1965. Haley, J. "Family experiments: a new type of experimenta— tion," Family Process, 1, 265-293, 1962. Haley, J. "Research on family patterns: an instrumental measurement," Family Process, 3, 41-65, 1964. Leary, T. Interpersonal Diagnosis of Personality. New York: Ronald Press, 1957. Lennard, H. L., Beaulieu, M. R., and Embrey, N. G. "Inter- action in families with a schizophrenic child," Archives of General Psychiatry, 12, 166-183, 1965. Levinger, G. "The assessment of family relationships: a progress report," Western Reserve University, Cleveland, Ohio (mimeo), described in chapter 11, Intensive Family Therapy, Ed. Boszormenyi-Nagy and Framo, New York: Hoeber Medical Division, Harper and Row, 1965. Moore, M. "Consistency of interaction in normal and clinic families," Ph.D. Dissertation, Michigan State Uni— versity, 1966. Moustakas, C. E., Sigel, I. E., and Schalock, H. D. "An objective method for the measurement and analysis of child-adult interaction," Child Development, 27, 109-134, 1956. Osgood, C. E., Suci, G. L., and Tannebaum, P. H. The Meas— urement of Meaning. Urbana: University of Illinois Press, 1957. Riskin, J. "Methodology for studying family interaction," Archives of General Psychiatry, 8, 343-348, 1963. Riskin, J. "Family interaction scales: a preliminary re- port," Archives of General Psychiatry, 11, 484-494, 1964. 57 Rodnick, E. H., and Garmezy, N. "An experimental approach to the study of motivation in schizophrenia," Ng- braska Symposium on Motivation, Ed. Jones, M. R. Lincoln: Univ. of Nebraska Press, 109-184, 1957. Ruesch, J., Block, J., and Bennett, L. "The assessment of communication: a method on the analysis of commun- ications," Journal of Psychology, 35, 59-80, 1953. Strodtbeck, F. L. "Husband-wife interaction over revealed differences," American Sociological Review, 16, 468-473, 1951. Strodtbeck, F. L. "The family as a three-person group," American Sociological Review, 19, 23-29, 1954. Wynne, L. C., Ryckoff, I. M., Day, J., and Hirsch, S. I. "Pseudo-mutuality in the family relations of schizophrenics," Psychiatry, 21, 205-220, 1958. APPENDIX 59 APPENDIX Characteristics of families sampled in this study; families 1-8 are the normal group and families 9-15 are the clinic group 1 m T = Father Mother Children Family - Number Occupation Education Occupation Education Sex Age 1 Tool & Die Housewife 12 M 11 Maker 12 M 14 2 IBM Pro- Housewife 14 M 5 grammer 14 M 9 3 Personnel Housewife 12 M 7 Director, M 7 Mich. Dept. F 12 of Social Service 12 4 Labor Union Housewife 12 M 8 Leader 12 M 9 F 11 5 Bricklayer 12 Housewife 12 M 10 F 15 6 Accountant 16 Housewife 12 M 12 M 17 7 Postal Clerk 12 Secretary 12 M 12 F 16 8 Insurance Housewife 12 F 7 Salesman 12 M 9 60 Appendix (continued) Father Mother Children Family Number Occupation Education Occupation Education Sex Age 9 Machinist 12 Housewife 12 M 7 10 Graduate Housewife 13 M 10 Student 17 F 9 M 12 11 Pet Store Housewife 12 M 8 Manager 15 M 12 12 Mechanical Housewife 14 M 8 Engineer 17 M 12 13 Cartographer 13 Housewife 11 M 7 F 12 M 14 14 Graduate Secretary 12 M 8 Student 16 M 12 M 15 15 Factory Housewife 12 M 7 Worker 12 F 8 ." K‘- x IVQ/US MICHIGAN STATE UNIV. LIBRARIES 293103590786