lllllliHli HHIIHMIHIHHIIM ll U'illlmmfllllll mm m“... v--w~ 6 5 (3M 3 1293 10573 2535 ‘ L 15" A AK ' Liichigan Stats , 1 University F5 Mi,,w‘ “*4 ...;s.._- 4.1.41.1 y.‘ This is to certify that the thesis entitled PERCEPTIONS OF PARENTAL BEHAVIOR IN HIGH- AND LOW—AUTONOMY COLLEGE STUDENTS presented by Sally W. Haithman has been accepted towards fulfillment of the requirements for M .A. degree in Jamming / / Mgior professor Datefi—Z/Z/ l/(TO/ 0-7639 arr- OVERDUE FINES : 25¢ per day per item RETURNIKS LIBRARY MATERIALS: Place in book return to remove charge from circulation records war W"“‘ i 5 i: PERCEPTIONS 0F PARENTAL BEHAVIOR IN HIGH-ANDlLOW-AUTONOHY COLLEGE STUDENTS By Sally W. Haithman A.THESIS Submitted to Michigan State university - in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Psychology 1981 ABSTRACT PERCEPTIONS OF PARENTAL BEHAVIOR IN HIGH- AND IDW-AUTONOMY COLLEGE STUDENTS By Sally W. Haithman Based on theories of aversive maternal control, and Lidz's (1965) concepts of family skew and schism, it was hypothesized that perceived parental Dominance and Submission would interact with sex of the subject in differentiation failure or success. It was further hypothesized that persons who were less autonomous would perceive their parents as more Hostile and Dominant than would more autonomous persons. Using self-ratings from Benjamin's (1979) Chart of Social Behavior, groups of relatively more and less autonomous college students were identified. Neither hypothesis was confirmed. However, an interaction was feund between mothers' Affiliation and sex of the subject. Males describing themselves as less autonomous, and females describing them- selves as more autonomous, perceived their mothers as being more Friendly. Between-group differences in variance also suggested that the groups varied in the character of their perceptions. ACKNOWLEDGEMENTS I would.like to thank Dr. John.Hurley for his patient assistance with numerous revisions of this thesis. ii TABLE OF CONTENTS Page IntroduCtion OOOOCOOOCOOOOCOOO0.00000000000000000000000000.0... 1- Review of Relevant Literature: I. Differentiation Failure as Described on the Chart Of 30cm BahaV1or OOOOOOOOOOOOOOOOOOIOOO000... 3 II. Validity of Benjamin's Measure 10 III. Research Supporting the Entrapment Hmtheses O.OOOOOOOOOOOOOOOOOOOOOOOO00.00.00000000000011 IV. Parental Dominance and Submission in Realtion to Differentiation Failure ....................14 Methods: I. Data Collection .........................................21 II. Data Analysis ..........................................22 Discussion ....................................................33 conCImions .OOOOOOOOCOOOOOOOOOOOOOOOOOOOOOOOOOIOOO0.0...0.0.0.37 iii Table 1. Table 2. Table 3. Table h. Table 5. Table 6. Table 7. LIST OF TABLES Page Intercorrelations of Dependent Variables .............28 Means and Standard Deviations (in parentheses) Of mpndent Vuiables CCOOCCCOOOOOCOOCCCCCO...0.0... 31 Multivariate Tests of Significance, MS Affiliation andHSAUtonomy OOOOOOOOOOOOOOOOOOOOOO 30 Univariate F-Tests, MS Affiliation mdMSAutonomy OOOOOOOOOOOOOOOOOOOOO00.00.000.000... 32 Multivariate Tests of Significance, FS Affiliation md FS Autonomy OOOOOOOOOOOOOOOOOOOOOOO 32 Multivariate Tests of Significance, MF and FM Affiliation and Autonomy, MS and FS Consistency, and M and F Control .................... 33 Univariate Tests for Homogeneity of varimce OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO 35 iv LIST OF FIGURES Figure 1. Two Dimensions of Interpersonal Behavior . . ....Pag. , Figure 2. Complete Chart of Social Behavior .................. 6 Figure 3. Surfaces ofChartof Social Behavior............... 8 Figure 4. Mean Scores on Introject Plane ..................... 23 Figure 5. Example for'Computation of M Control ............... 26 Figure 6. Cluster Diagam of Intercorrelations . . . . . . . . . . . . . . . 30 INTRODUCTION The purpose of this study is to clarify the nature of students' perceptions of their parents, and the relationship of these per- ceptions to success in the individuation process. College students are in a transition period from adolescence to adulthood, in which previously untested perceptions about self and others are being subjected to scrutiny in the ”real world". Some find that their approach to new events allows them to assimilate and learn from their experiences, while others find themselves trapped within old perceptions of the world which become increasingly inaccurate and maladaptive. Benjamin (1979) has described individuation success or failure in terms of interpersonal perceptions. It was preposed that these perceptions are largely determined by interactions with parents and significant others, in two ways. First, previous interpersonal experience creates similar expectations for other situations. Benjamin also proposed that parental behavior is introjected, there- by creating either a self-sustaining or self-destructive internal environment. Persons who individuate successfully were found to have two characteristics: an accepting, friendly attitude toward.them- selves, and faith in their ability to exist independently. The apposite was true for persons who were less successful. They were 2 found to be hostile towards themselves and others, and permitted themselves to be controlled by others. Much of’Benjamin's theory is based.on observations of extreme differentiation failure, or schizophrenia. Benjamin's subjects were feund to perceive their parents as both hostile and dominant, and also as engaging in confusing doubleébind control tactics. It is not known whether normal college students who are relatively low in autonomy have similar perceptions of their parents. The present study will apply Benjamin's theory to a normal student sample, in order to determine if the same relationship exists outside of psychiatric populations. In addition, the study will test hypotheses based on.Lidz et al.'s concept of skewed and schismatic families. Lidz et a1. hypothesized that the relative dominance or submission of each parent interacts with the sex of the child in determining his/her eventual ability to differentiate from the family. This theory was also developed based on observations of schizophrenics, and the present study will test its applicability to other populations. REVIEU OF RELEVANT LITERATURE I. Differentiation Failure as Described on the Chart of Social Behavior Before describing differentiation failure it is necessary to discuss the concepts underlying Benjamin's (1979) Chart of Social Behavior. This instrument is based on the assumption that inter- personal behavior can be classified along two dimensions: Affiliation and,Autonomy. In this respect it is similar to models preposed by other researchers (Leary, 1957: Schaefer, 1959: Carson, 1969). Earlier models have arranged behaviors in circular fashion around vertical and horizontal axes representing, respectively, the Dominance-Submission and LoveeHate dimensions. Insert Figgrg 1 about here The proportions of Dominance or Submission and Love or’Hate present in a given behavior can be determined by its position on the circle. For example, managerial, responsible behavior falls close to the Dominant end of the vertical axis. Dominance decreases as we move downward around the circle, until we find the behavior "following orders" near the Submissive end of the axis. Since the Dominance- Submission and Love-Hate dimensions are orthogonal, behaviors which fall near the horizontal axis (such as seeking friendship or'hostile QOMINANLE I , .MANAGERIAL, RESPONSIBLE FOLLOWING. ORDERS .FIGURE I. TWO DIMENSIONS OF INTER'PERSONAL BEHAVIOR 5 attack) are assumed to contain no elements of’Dominance or Submission. Behaviors containing various proportions of the feur basic qual- ities are ordered between the ends of the axes. Another important concept in describing interpersonal behavior is complementarity. In early models, Dominant behavior was assumed to stimulate a response containing a similar proportion of Submissive- ness. Along the Love-Hate dimension, behavior is assumed to stim- ulate a response from the same end of the continuum. Benjamin's model replaces Dominance and Submission with the Autonomy dimension, on the grounds that persons who are compelled to dominate others have no more freedom than those who are domi- nated. In addition, it incorporates three planes representing active initiating (or Parentlike), passive responding (or Child- like), and introjected behaviors. The Introject plane represents the subject's attitudes toward him/herself, and the characteristic mode of'relating to others. A complete diagram of the Social Behavior Chart is seen in Figure 2. Igsert Eiggrg 2 about here In addition, behavior is placed on one of several surfaces. Surface 1, Introject, has only one plane representing the self. The other surfaces each contain two planes, representing Parentlike and Childlike behaviors. There are three of these surfaces, as shown in Figure 3. In art e about here Figure 2. Complete Chart of Social Behavior smuomaw-cwu . 5 aunt-o Ara-7 m ”Mi/UT 3:5 ......245 mM’“ Maser eff/MB 7 :,___l.a3f7 ‘5’ Q“ 5"" “mm ”474”; Mill” J, 3," MARIO ”Mam #62464? new femur/A: 4g . one“. acme n: was: sax-— waif/”£0, UNI/vow era; ‘9 .zwewm. .souo use “(5 M Jfilc £15.32) 3/; 9Z0“, m “If (after, W sea-’32, I [I L'WTERTM, (my 551; (our; (0625!! M500“ 7'0 «550' as! 3w sees seer For sens 35‘" a” sum, ovens/com 551.! 3:3. .302. mm. lesion set; VEMF’N 5&2! hMII_ as: - 19.3 PRerta‘ 5st! fifl/Vfilfl'lrxlf .334 an Mug/mam: au- ‘WLT, am up .352! as .345 "Adler-3 mam! sewer!» ”or.” .562; now” 3.16 £16 sec:- m, M (GsMA, MD «at .550 331 17.4w BEMBWIENY EYE M: SELF Met news/m .331 4% Palm? ion. mourn ‘ ’ ' INTROJ ECT “- mm,mmmrsw :zo camouflage: SELF ova e- omens-z: I10 8pm VM”“,’ (ET ‘0“1 my), r--r seam Inseam “WW 7"me IrFW ; I6MIE, ’KFTE'M ”7' W31; /16 wtéllcr 14756.: 73', MAP“, ’45 MIN/Mt ”"me I“ Mr MAMA. lave m we“ 33 I OTHER ,, many, new ewe/pa #215000 misc 5W alum/c W CONF/M at A: a I!“ m. seems. our sun/E, cor arr/4L; r—Aa‘” ’I“35,MJ£‘TI1_I 1/] “Ml-Y “Ml ”nodal MIIA¢IM¢PIJI mm M PK’MV [M 5M” I’D fit/ff, ltd/fl 13.1 mm Mt! evens: m saw, aw” mauve u? “£035, J45” ICU for run, Aer “WEI/m 1.36 .14 kt, 1.07mi, lfs‘fllc‘f 1.37 (More! comment W U! LWIHMTWO Arr-M no M4“, cement Ho 22o may cowsm so so can “mam: 04 r 1“ .... 1 I' mm, Jo error/fl mu of}, mine”; 3‘ us? #771 new rlmé 4’5 mum/u!” (Glens/v ‘3 Jamal MP4: M: 1/ (1,055 “1/57”“: (W ‘ mesa” , «VI/{MIN J! ROWDE FIR, MW N3 Re 7227', MK 0! stmsm: M4! 795 I?! can snot-7706 :r/mm I'l‘ ”DR , WWW [‘77 ”W 4W1!“ ) am In my W73 our [ INTERPEBSQNAL [ ow laser/Iv, mum» 1 546950704! new lfi'v'vr «(as aw we use“ 1/ 01954147 amuse, strut 1 «EM: pores in rflususiru. was ‘ #5441: use), Fmor (Y: ’34!!!“ 21/ JACl/FICF 64547: f 2.11 Affine, an!” 2.53 memes/luau an: .117 was, skim,va rt” 15! Jock, Aer for ”on 116 our»: 71¢ cosh/Mt: 157 Ana-m («(3, ”or” m W7? M0153? 23° View, Saaw‘r, one I» .290 Figure 2. 3” .70?” AP “‘74"- 4/ Found, 4:4er WM W' ”a ”1 11¢er 173 Ask, mar. Nimrod :99 Acts” new 195m 14!, mm m 1’5 64m. ammo OER-'K, evacuees»: 111405460: zero (at: SELF mo_>§V~ “(be 33 3 xi 3% New; §\/ 3 ...V\ \7kwko! g :3ng Mafia. s..\/§ Maw... flzonmmm E15 59... “am . . .n 8&st 3 see $.93. Imam: 8253 a; g Quad: 3 5mm 9 5:5 no 5m 352. ... moimom 5%“ 08% JOBRU $§\/ / 63313 \ Kuhn uksxexso.‘ @353“ .g ._ m5... 35d. . .N wufimbm 3 use rites terse Lewes. serrate. / hoist scenes“ < \Kevo‘n‘hmg 3 0)? s§ veg» .hw‘g $934“ > VV‘k“ «($333 tantrum < ikaxxsxx: Squawk 388$ 9 To rate interpersonal behavior, subjects samplete a behavioral checklist. The list is comprised of behaviors from each point on the perimeter of the chart; subjects rate each of these for fre- quency and/or intensity on a scale of 0 (completely tales) to 100 (completely true). Based.on the subject's ratings, it is possible to compute Affiliation and Autonomy scores, which estimate the subject's overall behavior in the relationships being rated. In this study, these two scores will form the basis for hypothesis testing. Some scores on individual tracks (or behaviors between the axes) will also be examined. According to this model, persons with some degree of differ- entiation failure should have a behavioral average falling within the lower half of the chart. In Benjamin's observations of schizo- phrenics, the behavioral average was within the Hostile-Submissive quadrant. Complementary behavior was also evidenced.by the parents, placing their behavioral average within the Hostile-Dominant quadrant. These subjects also had a Hostile-Submissive introject, reflecting internalization of the parents' hostile, demanding attitudes. Benjamin stated that hostile parental control may also exist in the fern of a double bind. In this case parental behavior would come from two quadrants: Hostile-Dominant and Friendly- Dominant. Benjamin viewed this as evidence supporting the entrapment h thesis, in which she proposes that the children of schizogenic families are ”held captive under the guise of love". Stated in terms of the circumplex model, this means that the parents engage in Friendly-Dominant behavior, but only unless the child complies with their demands. Hostile-Dominant behavior is always 10 present as a threat if the child tries to make independent decisions. II. validity of’Benjamin's Measure Internal consistency of the Social Behavior Chart was feund to be high for normal subjects (Benjamin, 197b), Autocorrelations among points on the circle were large and positive for adjacent points, zero for orthogonal pairs, and large and negative for pairs of opposites. Internal consistency was also thought to mean that descriptions would be consistent over time. Internal consistency was often lower for psychiatric popu- lations (.81 vs. .97). This was thought to represent the conflictual and ambivalent nature of their relationships. High, positive correlations between pairs of opposites were frequently present. This finding is valuable because it demonstrates that conflictual relationships are as accurately portrayed as unifbrmly positive or negative ones. It also means that double bind relationships can be represented on the chart. The subject's inability or un- willingness to acknowledge the conflicts does not affect their accurate description by this method. These data are based on an earlier version of the measure. In a 1979 article, Benjamin states that the version used in the present study has greater reliability, although no details are given. 11 III. Research Supporting the Entrapment Hypothesis A basic condition in the development of differentiation failure is the presence of a hostile, domineering approach to childrearing in at least one parent. Early research in this area has focused mainly on the mother, although some hypotheses about the father's role have also been advanced. Some of these studies have employed circunplex models similar to Benjamin's. Others, while not employing this model directly, have produced results in the same direction. The concept of hostile parental control originated in work with the families of schizophrenics (Bateson, Jackson, Haley'& Weakland. 1956), which led to the development of double bind communication theory. According to double bind theory, the child is exposed to contradictory communication from the mother. The mother overtly maintains an attitude of love and protectiveness, while her behavior covertly communicates rejection and domination. The child learns to perceive the world as dangerous and incompre- hensible, and him/herself as incompetent and unable to cope with it. Elements of hostile control and hostile dependence are evident in this theory, although it was not originally described in these terms. It was also hypothesized that the father could participate in this process, either by behaving in the same way as the mother, or by openly disagreeing with her. In the second case, the double bind is created when the child is always rejected by one parent for pleasing the other. The existence of aversive double bind control in schizogenic families was also documented by Laing and Esterson (1965). These researchers described the process of "mystification", in which 12 the identified patient begins to feel that he/ she has no control over his/her behavior, and no identity apart from the family. At this point, the person also begins to experience his/her own affect as external, in the form of hallucinations or paranoid delusions. Without exception, parents of these persons were found to engage in hostile control, which they attempted to disguise by bizarre distortions in perception and communication. More recently, aversive control has been separated into two classifications: binding and expelling (Stierlin, 1978). In the binding mode, the identified patient is bound to the family by regressive gratification and demands for total loyalty, usually in combination with disordered communication. The expelling mode is described as complete, overt rejection, which forces the child into precocious autonomy. The expelling mode is viewed as poten- tially less (la-aging. If the child does not succumb to total, emotional deprivation, he/she is at least free to develop corrective relationships outside the family. This is not true of the binding mode, since the child in this family lives by a set of rules which makes forming relationships difficult, if not impossible. Heilbrun (1973) found that normal adolescents with adjustment problems came from similar family environments. Using an interper- sonal circumplex measure (Schsefer'i Bell, 1958), Heilbrun identified college students who perceived their mothers as hostile and controlling. These students were found to have deficiencies in social skills and cognitive functioning. In addition, types of aversive control similar to Stierlin's binding and expelling modes were described. In one of Heilbrun's styles the mother maintained control by Open rejection and criticism. In the other, control was of a double bind, l3 guilt-inducing nature. There is some evidence suggesting that the aversive control element is the same in "normal" and schizophrenic disorders. The difference may lie in the presence or absence of communication disorder. Stierlin (1978) and Lidz (1978) have hypothesized that communication disorder (in the form of abnormal rules for interac- tion) is the factor which determines how fully a child will be able to separate from the pathogenic family system. Benjamin (1979) included hostile parental control among the conditions necessary for psychosis (or complete differentiation failure) to occur. The necessary conditions are: 1.. Irrational, double-binding parental control: 2. Symbiotic fusion; 3. Implicit sexual bonding: 4. Introjection of parental attacks on self-esteem: 5. Learned helplessness and dependency: 6. Experience of vicious retaliation by the parents. Benjamin also pointed out that the child interacts with the outside world as well as with the family. While the family has a powerful influence on the quality of other relationships, inter- vention by a benevolent teacher, friend, or other person may still promote normal individuation. In Benjamin's conceptualization, differentiation.failure exists along a continuum, with severity linked to the number of pathogenic factors. The present study will attempt to replicate Benjamin's findings about aversive parental control, the first of the necessary conditions for differentiation failure. Given that this will be done using a 14 normal student sample, it can be expected that not all of the other necessary conditions will occur. These would also require extensive testing and observation of the students' families, which are beyond the scope of the present study. No hypotheses will be tested concerning conditions 2 through 6. It is hypothesized that students who are less autonomous will per- ceive their parents as more Hostile and more Dominant. IV. PhRENTAL DOMINANCE AND SUBMISSION IN RELATION TO DIFFERENTIATION FAILURE As described in the last section, aversive control inhibits the child's individuation from the family. Within this framework, the relative dominance or submission of each parent also plays a part in determining individuation success or failure. Overall, research evidence suggests that the child is especially vulnerable to differentiation failure if his/her opposite-sex parent is dominant in the parents' relationship. There.is some disagreement on this point, but this may be due to the confounding effect of social class in some studies. Parental dominance and submission seem to have the greatest importance in middle-class families, where more emphasis is placed on acquiring traditional sex-role behaviors. But there is some indication that parental Autonomy has some influence beyond the learning of traditional sex roles. A major theory in this area was developed by Lidz, Fleck and Cornelison (1956). According to the theory there are two types of schizogenic families: schismatic and skewed. In schismatic 15 families there is open disagreement, with each parent trying to undermine the other's authority and personal worth. Usually the father holds a dominant position, while the mother tends to be the degraded underdog. Children in these circumstances have a number of alternatives, all of which are harmful to psychological deve10pment. One possibil- ity is for the child to become a go-between, and try to step the conflict by satisfying both parents' unmet needs. Another is for the child to attempt to widen the gap between the parents, in order to win one parent's affection for him/herself. The parental conflicts may also at times create a double bind situation, in which one parent is always displeased. Lidz et al. believed that this situation led to the internal- ization of two disparate objects, causing the child to experience perpetual inner conflict. In addition to the child's real distress, becoming 111 was also thought to serve a function fer the other family members by diverting their attention away from their own problems, maintaining the image of a stable family. The other type of discord was termed "family skew". In these families one parent (usually the mother) is clearly dominant, and the other is clearly submissive. The mother often engages in bizarre and intrusive childrearing practices, while the father does little or nothing to prevent it, Family life is often arranged to accomodate the mother's eccentricities. In both types of families disordered communication was thought to be necessary for schizoPhrenia to develop. Lidz et al. hypothesized that one way these families create psychosis is by interfering with norma1.sex role identification. 16 The researchers held that male children must have a strong father, who will break the initial mother-child symbiosis and encourage aggressive ”masculine” traits. It was believed that females must be able to accept a passive, receptive role. To accomplish this the mother must provide a warm, accepting role model, and the father must be a suitable love object. Therefore skewed families were thought to be more pathogenic for'male children, and schismatic for females. The demeaned, low status position of mothers in schismatic families, combined with fathers' confusing mixture of seductiveness and disparagement, makes the female role unacceptable. In skewed families the weak, passive fathers do not ”rescue" their sons. They remain tied to their mothers rather than becoming independent. Recent research tends to confirm.Lidz' hypothesis, although the results are inconclusive. The father's role remains relatively unexplored, although there is abundant data concerning the mother's influence upon interpersonal style. There is also a scarcity of research concerning parental Dominance and Submission as they influence the child's individuation. However, some inferences may be drawn from studies of the father alone. Study of the father's role is complicated by several factors. Fatherless families are likely to be worse off financially than motherless families, and subject to increased stress on this ground alone. In addition, if the father is present, his usual role as family breadwinner undoubtedly influences his interpersonal behavior within the family. It is virtually impossible to study the father's influence separately from socio-cultural factors. There is also 17 a scarcity of research on the father's contribution to differentiation failure, which makes it necessary to draw conclusions based on other types of psychological disturbance. Hunt and.Hunt (1977) introduced social class as a factor in their study of father-absent girls. It was hypothesized that father-absent middle-classgirls would be better off, in terms of assertiveness and autonomy, than father-absent lower-class girls. The explanation given for this was that since middle-class fathers have a high enough income to support their families by themselves, there is more chance fer traditional sex role behaviors to exist. Middle-class girls from intact families are therefore more likely to observe their mothers in submissive, traditional female roles. This was thought to be less true for lower-class girls, who often observe their mothers as family breadwinners, closer to their fathers in status. The findin gs of this study are confounded by a tendency for Black subjects to be concentrated in the lower class, and White subjects in the middle and upper classes. However it was found that father-absent‘flhite (and middle-class?) subjects achieved better grades in school than father-presentJWhite subjects. This was not true for father-absent Black subjects, who showed only a a tendency to date more frequently than father-present.Black subjects. In this study, father absence had little or no impact when economic factors were taken into account. This finding tends to refute the hypothesis that both parents interact in determining the psychological adjustment of their female children. However it is possible that closer examination of subjects' 18 personality characteristics would reveal differences not detected by measuring school achievement. In addition, it appears that while father absence is not harmful to girls' development, the presence of a hostile, domineering father definitely is. Shaw (1977) ob- served that young women complaining of "interpersonal difficulties" all had some conflict with their fathers. In all cases the fathers had rejected or denied their daughters' developing sexuality. women with more severe difficulties described their fathers as critical and humiliating. Shaw concluded that since the father provides his daughter's first contact with men, his approval is crucial in the daughter's acceptance of her femininity. Other findings may have some bearing on the nature of the father's influence. It was found that women who were able to resolve their problems in short-term therapy had conflicts mainly with their fathers. Those requiring longer therapy had conflicts with both parents, and viewed their mothers as demanding, dependent and unreliable. This seems to support Lidz et al.'s hypothesis that parental interaction is an important contributor to severe psychological disorder. It is not known whether the same pathogenic dyad exists in the backgrounds of better-adjusted college students. Green (1976) discussed the father's role in relation to daughters. and also concluded that the father influences daughters' self-esteem and acceptance of femininity. Again, a cruel or domineering father was considered to be worse than no father at all. Green also corroborated the findings of Hunt and Hunt (1977), stating that father-absent girls may have a greater chance to become independent l9 ‘ and assertive. Small (1979) has also feund that the father's presence can be worse than his absence. Psychopathology was measured in terms of a discrepancy between real and ideal self, in adoles- cent girls. Father-present disturbed girls showed an even greater discrepancy than a comparable group of father-absent girls. For males, it is also clear that a bad father is worse than no father at all. The difficulty lies in deciding which paternal behaviors are harmful. Green (1976, p.93) noted that sons of distant, abusive or domineering fathers had numerous problems. Among these were impulsivity, overaggressiveness, and low self-esteem, On the other hand, some of passive, childlike fathers have also been found to have increased psychological problems and anxiety (Goldstein, 1977). The main consequence of father-absence for males seems to be a less successful heterosexual adjustment. Green (1976, p.80) prOposed that this may be due to lack of Opportunity to observe their fathers and mothers together. One other characteristic of father-absent males was described, which was a tendency towards intuitive, non-logical thinking rather than analytical detachment. Head and Rekers (1979) also described father-absent boys as more .effeminate, and having greater difficulty with sexual adjustment, These researchers also feund that father absence affects sexual adjustment in females, but to a lesser extent. In summary, the father's influence cannot be studied based on father-absence research alone. There appears to be greater potential fer both psychological health and disturbance in children when both parents are present and interacting. For females an overly dominant father seems to be most harmful, while fathers of males 20 apparently can be harmful if they do not strike a balance between dominance and submission. The father's influence in normal families seems to be mainly upon sexual adjustment, for both males and females. Research on father absence is often confOunded by socio-cultural factors. This is true in studies of'middle-class familieS, where "normal" sex role behaviors for women (submissiveness and depen- dency) might be considered somewhat maladaptive. Father absence appears to reduce this negative socialization, and this may lead to erroneous conclusions about the father's contribution to personality development. The other case in which economic factors may confound results is extreme poverty. In studies of extreme lower-class families (Kellam, Ensminger at Turner, 1977; Adams 5. .horovitz, 1980) it was found that father absence was less important than the families' economic security as a predictor of childrens' psychological well-being. In families where father absence caused extreme finan- cial hardship, it is quite possible that any alternative produced better adjustment, whether this meant living with a father or any significant other. However, there may have been differences which were obscured by, or seemed less important than, the struggle for basic survival. In light of the conflicting conclusions concerning fathers' influence upon their sons' psychological well-being, and the possible confounding effects of social class, no specific hypotheses will be formulated concerning Autonomy within the parental dyad and the students' success at individuation. These factors will be measured and examined for possible differences. METHODS I Data Collection Subjects were 1+8 male (xage 19.85) and 48 female (x8L8 19.41) s undergraduates, who received course credit for participating in the study. Six male and five female subjects were from Lansing Community College: the rest were from Michigan State University. Data were collected during the summer and fall terms of 1980. Subjects who arrived for testing completed a form providing demographic data (age, sex, parents' occupations and parents' education). They then received copies of the Benjamin questionnaire, which were completed at home and returned the following week. No instructions were given on how to answer the questions, other than those included with the questionnaire. More data were obtained than were actually used. Data on subjects' behavior toward others, and others' behavior toward the subject were obtained, but only the latter were used. Approximately 3% of the questionnaires were discarded due to subject errors in completing them. About twice as many females as males signed for the study, and data were collected until a sufficiently large male sample was obtained. Female data were then discarded at random. For both males and females, 47 of #8 subjects were from intact families, with both parents present at least through childhood 21 22 II. Data Analysis Data analysis was a 2 x 2 multivariate ANOVA, with Sex and Interpersonal Style as dependent variables. Interpersonal Style was determined using scores from the Introject plane, which repre- sents the subject's overall approach to others and him/herself. Nearly all subjects placed within the right half of the Introject chart, as shown in Figure 9. The sample as a whole also fell slightly on the submissive side of the Autonomy axis. The sample was found to be slightly more friendly than a student sample previously tested by Benjamin (Note 1): 7:- 107.52 vs. Benjamin's 97, 50. Male students in the present study were more Autonomous (X:= 9.29 vs -l0.00) and females less Autonomous (X:= -23.02 vs. -15.00) than Benjamin's sample, Two Interpersonal Style groups were identified: High and Low Autonomy. Insert Figgge 4 about hegg, It can also be seen that all subjects were relatively friendly, and that the main difference between Interpersonal Style groups was in Autonomy, This means that conclusions based on these data will probably hold more true for Autonomy than for Affiliation. However, Autonomy may be the more important dimension in the study of differentiation from the family. Since each student was rated along two dimensions, it was not feasible to use a median split to separate High and Low Autonomy groups. Instead, Low Autonomy was defined as the 24 subjects of 23 {so tgo 9o 30 2.0 1. mg ... ”0'16 90 m .e ’0 --Fa*zo~de so So (e 70 to yo lflflo no : \CE 00 .so ‘ a “r . 1,30 dMEAN-SCORE, ALL MALES Q MEAN SCORE, ALL FEMALES GQ Low AUTONOMY, MEANS @@ HIGH AUTONOMY MEANS FIGURE 4. MEAN SCORES ON INTROJECT PLANE 24 each sex falling closest to the lower vertical axis. This resulted in a dividing line which was slightly slanted, and close to the hori- zontal axis,as shown in Figtn‘e4. Autonomy and Affiliation did not prove to be significantly correlated (r = .0007) . It therefore seemed necessary to divide subjects based on both scores, rather than discarding the Affiliation scores altogether. Dependent variables were measured as follows: 1. figthgg'g Affiliation and Autgnogy Toward Subject (MS Affilia- tion, MS Autonomy) were taken from Surface 2. This surface describes the mother's behavior toward the subject, as perceived by the subject. 2. Father's Affiliation 52d Autopggy Toward Subject (FS Affilia- tion, FS Autonomy) were also taken from Surface 2 scores. 3. Eggher's and Eathgg's Consistency of Behavior (MS Consistency, FS Consistency) are measures of how regularly a parent's behavior toward the subject occurs in the same quadrant. This score is pro- duced based on autocorrelations between points around the circum- ference. Positive values indicate that behavior is consistent, since correlations between adjacent points are high. Zero values indicate unpredictabliity, while negative vaules indicate contra- dictory, bouble bind behavior. 4. Mother's Autonomy Toward Father, Father's Toward Mother (MF Autonomy, FM Autonomy)-These scores are taken from Surface 4 of the chart, which shows the subject's perception of the parents' behavior towards each other, and their responses to each other. 5. Mothgr's Affiliation Toward Father, Fapher's Toward Mother (MF Affiliation, FM Affiliation) scores are also taken from Surface 4. 6. Mother's and Father's Control (M Control, F Control) 25 were computed using scores from Surface 4, and reflect complemen- tarity along the Autonomy dimension. Control scores are the per- centage of behaviors in the lower half of the chart which are en- dorsed by both parents. A weighted sum of Dominant or Submissive behavior is computed for both planes of Surface 4, for both parents. For example, assume that the mother receives three endorsements in the lower half of’the chart on the Parent plane. 88 illustrated in Figure 5. v sert F about here The mother's Control score would therefore be [(.1 x 6)+(.4 x ?)+(.l x 6IL or 4. As shown in Figure 5, the father has received two endorsements in the lower half of the Child plane, representing his response to the mother. His Submission score is [(.1 x 6)+(.l x 8)], or 1.4. %f% equals .35. and.this is the M Control score. It indicates that approximately 35 percent of the mother's Dominant behaviors receive a complementary Submissive response. F Control was computed using ratings from the father's Parent and the mother's Child plane scores. Since subjects did not con- sider one parent's Dominance when rating the other's Submission, it was possible for a parent's Control score score to exceed 1.0. 7. Social glass was computed according to Hollinghead's (note 2) Two-Factor Index. Information about the head of household's educa- tion and occupation are combined to yield a social class score, with possible scores between 11 and 77. Smaller numbers indicate higher social class. Both education and occupation are placed in one of seven 26 [W'PAMSVEEFFSINOULGE P WN, . x . ILIJCTTDSfiJPERIOIRIE3 (‘M 4mm. Jam's swam) (Glade denim, 3 unit: 4096 BENEVOLENT MONITOR\ Acne/#7) REMIND {7am}! obedience. , .2 Mt?! frlutolllhs: ) - __9THER‘S OOMINANT BEHAV;I_QR TOWARO FATHER ‘ '0‘”. CLING, OEPENO 'O% (4 Md‘ Mains/ea, 3w$ FOLLOW RULES, PROPER. fuudliue“) (1 wads Mmsaon , I and Lecithin, ) . FATHER’S SUBMISSIVE RESPONSES TO MOTHER FIGURE 5. EXAMPLE FOR COMPUTATION OF M CONTROL 27 possible categories. The education category number is multiplied by 4, and occupation by 7. The two products are added to produce the final score. In all cases, the social class score was based on data from the parent designated head of household by the subject. RESULTS A preliminary examination of the data revealed significant correlations between many of the dependent variables, as shown in Table 1. Figure 6 depicts the intercorrelations of dependent variables in diagrammatic form, and.also illustrates clusters of variables which have the strongest relationships. Insert Table 1 about here Ipsert Figggg 6 about here Not surprisingly, students perceived their parents as behaving similarly toward their spouses and their children. In addition, subjects viewed their parents as tending to have partners who were more similar to themselves along the Affiliation than along the Autonomy dimension. It is also interesting to note that these inter- dimensional correlations (MS Affiliation vs. MS Autonomy, r ==.05: F3 Affiliation vs. FS Autonomy. r = .11, etc.) were generally not statistically significant. This finding supports the assumption 28 Table 1 . Intercorrelations of Dependent Variables mmmwmwm is; -- -09 -03 15 -11 18 07 g -09 -- 08 -05 -07 oo 06 m -03 08 -- -12 -07 -23° 00 mm 15 -05 ~12 - 05 59a 11 m -11 -07 -07 05 -- ~08 29" mm; 18 00 -23° 59a -08 -- 11 M1119 07 06 oo 11 29b 11 -- mmuggs'r 19 -17 -1o 62“ -ou uaa 14 g: consgs'r 15 11 -13 38a -12 5 53' on HF AFFIL 10 on -19 65a -04 uaa 1n M -0'+ ~19 -05 23° 33b 03 13 114 mm. 12 -o 5 -261’ 503' -ou 863' 02 w 23° -05 -06 15 13 22° 52‘?L 120m 07 -05 05 oo -03 -33"' 19 c b a 12m 03 -09 21 oo 10 -32 35 Decimal points omitted 3p Q, 001 bp 4. . 01 °pé.05 MfiflNSIST BCONSIST HF AFFIL HF AUTO FM AFFIL FM AUTO H CONTROL 20 15 10 '05 29 Table l . , continued -ou 12 '19 '05 -ou ~26b 23° 508 33° -OH 03 86a 13 02 03 47a -02 116° 32° 60a .. 07 O7 ‘- 23° 13 -014, -343 on -32 C 23 '05 -06 15 13 22° 52 17 06 a 11 23 13 C ‘05 O7 05 05 FCC _— ._ ...— .._.._......._- “—- ' ‘ 30 — “Feovfml “‘ ~ Mcontrol .;__- -'..-.-:. ,- . I . ' - o" > ' : o A; 7 A Sq: —fiFM auto A s: - Fauna...— jgétho 33- MF auto « l-nI- A33 . - - ..-; . — h h _ Figure 6. Cluster Diagram of Intercorrelations ..-mm-.. *~-~~'Ihick-arrowe-ehow-typal linkages (McQuitty. 1957). Double _ _ bondeidentify cluster cores. Additional linkages that reached stat- , ’ tisticel significance are shown by lighter lines. Host correlations ”"7" werepositiverbut inverse correlations are depicted by-broken or ‘ ’ .-adashed lines. ‘“._——— .—_..- . - A.-..__.. . . - ..-. .. - -'-‘ ...”.-.— 31 that these are the orthogonal basics of behavior. The group means and standard deviations for dependent var- iables are presented in Table 2. The test fer significant differences was a multivariate ANOVA, which was performed in stages due to the numerous significant correlations among the dependent variables. MANOVA was selected to control fer the fact that several observations were made for each subject. The observations are therefore not in- dependent, and are not suitable fer repeated univariate F-tests. Univariate F-tests were used to determine which dependent variables were responsible for significant multivariate effects. Insert Table 2 about here The first stage of the analysis revealed a significant interaction effect between Sex and Interpersonal Style upon subjects' perceptions of their mothers. univariate F-tests showed that this effect was due to variation in MS Affiliation. Neither Sex nor Inter- personal Style had a significant effect separately. Table 3. Multivariate Tests of Significance, MS Affiliation and MS Autonomy E_f,f_‘_e_§ Wilks'Jl Amximate F' Hmthesis DF Error DF Sigif, F Interaction .92 3.96 2 91 .02 Sex .97 1.37 2 91 .26 Interpersonal 1.00 .57 2 91 .61 Style 32 «N u 32 2d. Amo.ov so.o Ama.mv nH.H Asm.sv Hm.s Ame.ov ma.o Houpeoo a Amo.HV ms.a Amc.ov ms.o Aoo.sv mm.o Ama.av mo.a Houpeoo s Ams.sev na.s Amm.m~v mm.~au Ame.mmv es.o. Ase.flmv ms.s- seoeoosc as Amm.osv mo.mm~ Amn.mmv sm.ms Aoo.smv mm.n~a Aon.smv ms.m~H eosoosasee< :a Amm.mmv ma.m Asm.mmv mo.a Ao~.m~v ~m.a1 Awo.smv ~m.o ssoeopea a: Aso.sov No.3ms Aso.omv sH.mo Asm.msv mo.o~H Amo.snv mm.HeH eoseoaflaem< a: Aam.ov om.o Ase.ov so.o As~.ov mm.o Amm.ov ~m.o soeoaosoeoo ma Amo.ov sm.o Asm.ov ss.o AHm.ov am.o Amo.ov mm.o soeooosoeou m: AHH.omV om.H- Amo.smv o~.m~- Ae~.mmV sm.sa. Amn.o~v mm.~- saoeoue< ms Asm.msv mm.soa Aem.nwv Hm.mm Amm.smv os.oas Aos.swv ne.om cosposadmm< me Amm.omV mH.m- Amo.omv on.a~. Ano.nmv es.s~- Amm.mmv sm.s- saoeoes< m: Aso.smv no.0ma Aom.eav sa.as Amo.mmv sm.mHH Asw.nev mo.o~a eosoosasmsc m: Aflo.o~v sm.~m Amo.efiv matmm AmH.sHv so.o~ Ams.mflv as.o~ moose Hosoom egos 3o: llmwurfie LIF Fey! 3 \Hsosoes< swam hsosoas< so; mausm A assessmen mo Amomogvsowsm adv usedpsd>on unsussvm use made: .N manna 33 Table it. Univariate F-Tests, MS Affiliation and MS Autonomy (DP 1, 92) SS SS MS MS Variable Between Within Between Within F Sigif, F 118 Affiliation 2b929.1 369814 .9 24929.1 #01997 6.20 . 02 MS Autonomy 2181.9 100113.6 2181.9 1088.2 2.01 .16 In the next step, F‘Sr Affiliation and FS Autonomy were introduced as dependent variables, with MS Affiliation and MS Autonomy as covar- iates. F‘S Affiliatiow and PS Autonomy did not vary significantly between groups, beyond their correlation with the mother' a behavior. Table 5. Multivariate Tests of Significance, FS Affiliation and FS Autonomy Effect wilks'z Amximate F Hmthesis DF Error DF SigLif, F Interaction .96 1.62 2 89 .20 Sex .98 .9“ 2 89 .40 Interpersonal . 98 . 79 2 89 . 1&8 Style The remaining dependent variables (HF and FM Affiliation and Autonomy, MS and FS Consistency, M and F Control, and Social Class) were then introduced, with the preceding dependent variables as covar- iates. MANOVA did not reveal any significant effects, although HF Affiliation reached significance in univariate F-tests. This may well have been a false rejection of the null hypothesis, since F-tests are not reliable in the absence of significant multivariate effects. 34 Table 6. Multivariate Tests of Significance, HF and FM Affiliation and Autonomy, MS and FS Consistency, and M and F Control Effect wilks'zd Appggximate F Hypothesis DF Error DF Signif, F Interaction .88 1.35 8 81 .23 Sex . 97 . 33 8 81 . 95 Interpersonal .9“ .60 8 81 .77 Style DISCUSSION The correlations between the perceived Affiliation (r =-,59) and Autonomy (r =*.29) of mothers and fathers toward the subject has implications for one of Benjamin's hypotheses concerning differentia- tion failure. Benjamin stated that a child may be "rescued" from differentiation failure if he/she has one healthy parent. The may be possible in theory. But these data suggest that relatively few pairs of one pathogenic and one normal parent would be found. This is particularly true of Affiliation; only 12 of 96 subjects placed one parent on the Hostile side of the Social Behavior Chart, and the other on the Friendly side. It should be emphasized that these conclusions are based only on the subjects' perceptions of their parents, and might, therefore, be entirely different if the entire family were observed. Benjamin also stated that double binding, inconsistent behavior 35 must be present along with aversive parental control to cause differ- entiation failure. This is another theory which may be difficult to test in practice. There were substantial correlations between Affilia- tion and consistency of behavior for both mothers (r =6 .62) and fathers (r ==.55). This suggests that inconsistency may be an outgrowth of a hostile orientation, rather than a separate dimension of interpersonal behavior. It may also be difficult to ascertain which parent has more power in the marital relationship. The correlation of .46 between variables M Control and F Control (intended to measure power-taking) shows a substantial positive linkage where a negative correlation was expected. Apparently controlling behavior by one parent does not always produce a submissive response by the other. Or at least there may not be a static distribution of power in the parental relationship. Since power-taking increases simultaneously for both parents, there may be two marital styles for sharing power, rather than a clear division into dominant and submissive roles. In one style the parents would function independently. In the other, the parents might take turns trying to control each other, with the result that neither is autonomous. It should be noted that these variables concern only perceived interpersonal behavior within the parental dyad. It is possible that one parent might have more power than the other when measured by different standards. For example, one parent might be a scapegoat for the other, and appear submissive even though engaging in his/her own control maneuvers. There might also be a difference in social power, such as 36 a successful executive married to a ”mere housewife". The definition of power should be taken into account when applying these findings to Lidz et al.'s concepts of family skew and.schism. In view of the positive correlation (r'='.46) between maternal and paternal Control. it may be useful to characterize entire families as either power-taking or independent. F Control is also linked slightly and inversely (r’=I.21) to social class (high social class score indicates low social class). This shows a tendency for lower-class fathers to engage in more power-taking behavior. These fathers were also viewed as less friendly toward their wives and children. This is in disagreement with the hypothesis that middle-class fathers are more likely to play a dominant role in the family than lower-class fathers. However, the same points regarding the definition of power may be relevant here. Tests for homogeneity of variance were performed for all dependent variables. The results of these tests suggest that there may have been some difference in the character of subjects' perceptions, although the group means did not differ significantly. Table 7 shows Cochran's C and BartlettéBox F statistics for variables having significant differences in variance. Table 7. Univariate Tests for Homogeneity of variance variable Cochran's C Bartlett-Box F MS Affiliation .113 P=.02 6.10 P=. 00 MS Consistency .53 P-.00 36.87 P=300 M Control . 78 P s. 00 38 .148 P 3. 00 F Consistency . ’44 P =. 01 2 . 62 P =. 05 FControl .35 P=.18 3.11 P=.03 37 The most remarkable differences occurred for M Consistency, M Control, and M Affiliation. For all three variables there appears to be an interaction effect between Sex and Interpersonal Style. Males in the High Autonomy group perceived their mothers as less consistent, less friendly, and more controlling (see Table 2), and demonstrated greater variance in their perceptions. variance in the Low Autonomy group was significantly smaller. Low Autonomy males also perceived their mothers as more friendly, more consistent, and.less controlling. One explanation fer this might be that Low Autonomy males felt a need to deny negative feelings toward their1mothers, and tended to report an idealized version of them. This seems intuitively more acceptable than an interpretation that increased maternal friendliness and consistency leads to poorer adjustment in sons. However, this would require further testing to confirm. The reverse was true for female subjects. In this case it was , the High Autonomy group which had a combination of small variance and more positive perceptions. For females, an idealized perception appears to be an asset, if these data do in fact represent an idealized perception. It is possible to speculate from this that males must be able to perceive their mothers realistically to become psychologic- ally healthy, while females benefit more from the presence of an ideal female role model. An interaction effect also occurred for F Control. For this variable it is High Autonomy sons and Low Autonomy daughters who may have more idealized perceptions. This is consistent with the hypothesis that the same-sex parent serves as an ideal role model, while the opposite-sex parent must be perceived realistically. 38 However there appears to be no similar effect for the variance of F Consistency, and no significant effect was present for F5 Affiliation. This suggests that the effect for father perceptions is not as strong. Or the father may be important mainly as he is perceived as behaving toward the mother, rather than as a primary determiner of interper- sonal style. CONCLUSIONS Overall these findings added little to our knowledge of the family dynamics behind successful individuation in college students. This does not mean that family dynamics are unimportant. Nor does it mean that the mother has sole responsibility for the child's personality development, although only the mother's behavior emerged as to be significant in this study. Perhaps few pathogenic mothers exist within healthy families, as indicated by the substantial positive inter- correlations between students' perceptions of both parents. It is also likely that family dynamics are important in more subtle ways that were missed by the present approach of measuring basic interpersonal orientation. This sample is probably at least fairly representative of the college student population, and it is unlikely that significant results would be obtained by replication of the present study. 3.9., These data suggest that the character of perceptions varies between groups more than did the actual ratings. This conclusion is mainly speculative, based on differences in variance. In cases where variance is smaller, this may signify a more rigid, less reality-oriented perception of the parent involved. In these cases the subjects appear to be saying "Mother was always good to me", as opposed to ”Nether was not always perfect, but she was usually 0.x." If this is true, Lew Autonomy subjects may tend to idealize their perceptions of the opposite-sex parent, while High Autonomy subjects tend to idealize the same-sex parent. This is consistent with the Lidz theory discussed earlier: that male children show greater psychopathology in skewed families, while schismatic families have such effects on females. If an idealized perception of the same-sex parent promotes normal individuation, perceiving that parent as a devalued scapegoat would do much to prevent it. At the same time these data appear incongruent with numerous findings that linked aversive maternal control with differentiation failure. The present male subjects who described themselves most favorably also described their mothers as somewhat less friendly. It is possible that this is due to the nature of the questionnaire, which relies entirely on the subjects' self-reports. It is therefore only as accurate as the subjects are honest, and it is quite possible that less well-adjusted subjects made an attempt to conceal family difficulties. It is also possible that this was due to the subjects' own unawareness or denial of family problems, rather than to a conscious effort to appear well-adjusted. 40 It should be emphasized that these conclusions (apart from those concerning MS Affiliation) are based solely upon significant differences in variance. They require confirmation by further research, directed toward clarifying the nature of subjects' perceptions, as well as some external verification. A more extensive study of family inter- actions, with ratings made by outside observers, would be helpful in this regard. Along with observer ratings, it would also be helpful to clarify what is meant by parental dominance and submission. For example, is it most meaningful to define this in terms of one parent obeying the other? This definition was used in the present study, and no significant results were obtained. The researchers cited earlier have observed something unusual in the parental relationships, but it may well be something more complex than obedience. Thus, they may have observed one parent repeatedly discounting the other's feelings and perceptions without necessarily demanding obedience or subservience. In extreme cases the parents may have attempted to force each other into imaginary, need-determined roles, by psychotic distortion if necessary. In a way this is the ultimate in hostile domination: pretending the other person does not exist. However this might easily be missed or distorted by interpersonal ratings of dominance and submission. Much of the research cited was based on clinical observations that something was wrong, but quantifying this may be a difficult task. A study combining interpersonal ratings with some type of intra- personal data might also shed some light on the relationships in question. The present findings suggest that persons who describe hl themselves as relatively low in Autonomy may have a more rigid, idealized perception of their apposite-sex parents. Perhaps a study of ego defenses might help clarify this, by revealing the degree of flex- ible, realityhoriented functioning present in each of the four groups. This might also by useful in studying possible differences between Low Autonomy perceptions of the opposite-sex parent, and High Autonomy perceptions of the same-sex parent. Both of these might be idealized, but it is likely that some differences exist. One methodological change would also be helpful in future studies of interpersonal style. This would be selection of more extreme groups of well and poorly-adjusted persons. While it seems unlikely that differentiation failure is an all-or-nothing phenomenon, between-group differences might be more pronounced if more extreme groups were included. It is unlikely that significant differences in these data were camouflaged by unequal within-group variances, since analysis of variance is robust with respect to heterogeneity of variance. LIST OF REFERENCES REFERENCES Adams, P. L. a Horowitz, J. H. Psychopathology and fatherlessness in poor boys. Chgd Psychiatgy and Human Develoment, 1980, l_(_)_ 135'143. Bateson, 0.: Jackson, D.: Haley, J. 8: Weakland, J. Towards a theory of schizophrenia. Behavioral Sci ence, 1956, 1, 251-2614. Benjamin, L. S. Structural analysis of social behavior. Psychological Review, 1971}, §_l_, 392-425. Benjamin, L. S. Structural analysis of differentiation failure. Ps chiat , 1979, 3;. 2-19. Carson, R. 0. Interaction Concepts of Personality. Chicago: Adline, 1969. Coldstein, M. 7.. Fathering: A neglected activity. American Journal of P3 choanal sis, 1977, 31, 325-336. Green, M. Fathering. New York: McGraw-Hill, 1976. Heilbrun, A. B. Aversive Maternal Control: A Theo of Schizo enic Development. New York: John Wiley 8: Sons, 1973. Hunt J. G. 8: Hunt, L. L. Race. daughters, and father loss: Does absence make the girl grow stronger? Social Problems, 1977. 2.5.. 90-102. Kellam, S. 0.: Ensminger, M. E. a: Turner, R. J. Family structure and the mental health of children. Archives of General Psychiatry, 1977, 33, 1012-1022. Laing, R. D. & Esterson, A. Sanity, Madness, and the Family. London: Tavistock Publications, ltd. 1964-. Leary, T. Iptemgsofl Diagosis of Personality. New York: Ronald Press, 1957. Lidz, T.: Fleck S. 8: Cornelison, A. R. Schizophrenia and the Family. New York: International Universities Press, 19 5. Lidz, T. Egocentric cognitive regression and the family setting of schizophrenic disorders. In Wynne, Cromwell 8: Matthysse (Eds. ), The Nature of Schizoggenia: New Appr_q_arche§ to Research and Treatment. New York: John Wiley 8: Sons, 1978, 526-533. 42 43 Mead, S. L. a Rekers, G. A. Role of the father in normal psychosexual development. Psychological Repgrts, 1979, 55, 923-931. McQuitty, L. L. Elementary linkage analysis for isolating foth ortho- gonal types and typal relevancies. Educational and Psychological Mea- surement, 1957:.12: 207-229. Schaefer, E. S. A circumplex model for maternal behavior. Journal of Abnormal and Social Psychglogy, 1959,‘§9, 226-235. Schaefer, E. S. a Bell, R. Q. Development of a parental attitude research instrument. Genetic Psychology Monoggapgs, 1958, 29, 339-361. Small, P. A study of social problems in a group of young women treated with brief psychotherapy. British Journal of Medical Psy- chology, 1977' 5.9. 155-161 e Small, A. C. Sexual identity and personality variables in normal and disturbed adolescent girls. Adolescence, 1979, i&, 31-99. Stierlin, H The transmission of irrationality reconsidered. In Wynne, Cromwell & Mathysse (Eds.), The Nature of Schizophgenia: New Apppgaches to Research and Treatment, New York: John Wiley & Sons, 1978. REFERENCE NOTES Benjamin, L. S. Manual fer use pith SASB. Unpublished manuscript, University of Wisconson, Department of Psychiatry. Hollingshead, A. B. Two-factor index of social pgsition (copyrighted 1957). Privately printed: Yale Station, New Haven, Connecticut. 44 "ITITJ'THWE'LITWMEWTHELWET