SEKa-TYP‘IHG EN CHELIS‘REN AS A FUNCTION OF THE PARENW‘ SEXe-RQLE EDENNHCATEQN mus-:2 for {in Dogma a? M. A. MCHEGAN S'FM‘E MIME??? :‘v’éjyma Lane 19-6-8 F . ' A - . . why; . LIBRARY 3 : q . ; Ivhchigan Stan: «Igmmygmjmmnmmmmu I L, - f {a 10714 6676 J , _ o 7 M; L m A, CLT L9 1:378 1-942 ~-—---':- / I ;. _Hl$l_7_Q... ' ‘ ' ' _‘ :fMH-l ’ ABSTRACT SEX-TYPING IN CHILDREN AS A FUNCTION OF THE PARENTS' SEX-ROLE IDENTIFICATION By Myrna Lane The present study was designed to examine the relationship between the child' s sex-typing and his or her parents' covert and overt sex-role identification. Both psychoanalytic and learning theories of personality hold that children tend to identify with, or model their behavior after that of, their parents. It was therefore hypothesized that the child! 5 sex role would be influenced by that of his parents and that his adaptation or lack of adaptation to his sex- role will correspond to, and be in the same direction as, that of his parents, especially the same—sex parent. This study was divided into two phases. In the first, the Parent Activity Scale (PAS) was developed to assess the parents' overt sex-role identification. This scale consists of a list of activi- ties commonly performed by parents. It was given to 100 intro- ductory psychology students who rated each item as to which family Myrna Lane member should perform it most often. On the basis of the differ- ences in proportions between these ratings, items were assigned to one of three categories: Mother, Father, or Both. A scoring system was then developed in which an agreement with the assigned category received credit and a disagreement lost credit. In the second phase, 15 boys and 14 girls in the first and second grades were given the Rosenberg and Sutton-Smith Game Scale to obtain measures of their sex-typing. They were also given the PAS to assess their perceptions of their parents' overt behavior. Their parents' overt sex-role identification was measured by the PAS and their covert identification was measured by the Fe Scale of the CPI. The parents also filled out the Game Scale. The results of this study both support and refute the stated hypothesis. In relation to the parents' covert identification, there was a tendency for both the boys and the girls to become more appro- priately sex-typed as the same-sex parent did. This supports the major hypothesis. There was also some indication that the same- sex parent was more important than the opposite-sex parent for the child' s identification. Regarding the parents' overt behavior, the results were contradictory. While parts of the data indicated that as the parents became more appropriately sex-typed so did the child, other parts Myrna Lane indicated the opposite. Thus the parents' overt behavior did not seem systematically related to the child' 5 sex identification. Also, the conflicting results may be due to some deficiencies in the PAS and thus may not actually reflect the correct picture regarding the parents' identification. Another finding was that there was an inverse relation- ship between the fathers' overt and covert sex-role identification while for the mothers there was a direct relationship between the two. The results were clearer when the child' s perceptions of the parents' overt behavior were considered. Here, as the children saw the parents as more appropriately sex-typed, they themselves tended to become more appropriately sex-typed. This would suggest that the child' s perceptions may be a key factor in sex-typing. Also, there was some indication that the same-sex parent was more important for identification than the opposite-sex parent. In summary, the results tended to support the modelling theory and also to suggest that the same-sex parent is more important for the child' s sex-typing than the opposite-sex parent. Some methodological problems relating to the small sample size Myrna Lane and the instruments used were discussed with suggestions for future research. Approved: Luca“ '4, JIM" W Datezw.$0, 1T6? Thesis Committee: Lucy R. Ferguson, Chairman J. P. McKinney D. W. Thornton SEX-TYPING IN CHILDREN AS A FUNCTION OF THE PARENTS' SEX-ROLE IDENTIFICATION By Myrna Lane A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Psychology 1968 To my husband, Irv ii ACKNOWLEDGMENTS I would like to express my deepest appreciation to Dr. Lucy R. Ferguson, chairman of my thesis committee, for her guidance in this research, her encouragement and her assistance in the prep- aration of this manuscript. Gratitude is also extended to Dr. J. P. McKinney for his help in finding a setting for this research and his reading of the thesis and to Dr. D. W. Thornton for his participa- tion and reading of the thesis. To the principals and teachers of Saint Thomas Acquinas school, I wish to express appreciation for their cooperation and assistance in the administration of the instruments. Iwould also like to thank Dr. J. E. Hunter, W. G. Berger, G. M. Gillrnore and S. H. Cohen who provided statistical consulting. Finally I wish to extend a special thanks to my husband, Irv, without whose help and patience this thesis would not now be completed . iii TABLE OF CONTENTS Chapter I. INTRODUCTION Three Theories of Sex—Typing Research . . Modelling . Research Hypotheses II . ME THOD Phase I: The Parent Activity Scale Subjects Procedure . . . . Phase II: Employment of the Sflcale' Test of the Hypotheses . Subjects Procedure III. RESU LTS Phase I: Some Characteristics of the PAS Phase 11: Employment of the Game Scale, CPI and PAS . Part 1: Relationship between the Game Scale given to children and the Game Scale given to their parents . Part 2: The relationship between the Game Scale and the CPI. Part 3: Relationship between the PAS as given to the children and the PAS as given to their parents . Part 4: Relationship between the Game Scale and the PAS iv Page 11 13 17 18 18 18 19 21 21 22 25 25 27 28 33 37 40 Chapter Part 5: Relationship between the Fe Scale of the CPI and the PAS IV. DISCUSSION . Part 1: Relationship between the Game Scale as Given to Children and the Game Scale as Given to their Parents . Part 2: Relationships between the Game Scale and the CPI Fe . Part 3: Relationship between the PAS as Given to the Children and the PAS as Given to their Parents . Part 4: Relationship between the Game Scale and the PAS. . Part 5. Relationship between the CPI Fe and PAS Part 6: Some Additional Considerations and Criticisms V. CONCLUSION REFERENCES . Appendix I. CHARACTERISTICS OF THE SAMPLE II. LETTERS AND INSTRUMENTS USED . III. ASSIGNMENT OF PAS ITEMS TO CATEGORIES . IV. WEIGHTED SCORES FOR GAME SCALE ITEMS . Page 44 49 49 50 56 58 65 67 7O 73 75 77 86 88 Table LIST OF TABLES Correlations between the scores on the Game Scale for children and parents T-test scores comparing the number of responses of parents and children on the Game Scale Correlations between the scores on the children' 3 Game Scale and the parents' CPI Fe scores Correlations between the scores of the fathers, mothers and children for the Father scores on the PAS Correlations between the scores of the fathers, mothers and children for the Mother scores on the PAS Correlations between scores on the PAS as reSponded to by the fathers and the children' 5 scores on the Game Scale Correlations between scores on the PAS as responded to by the mothers and the children! 3 scores on the Game Scale Correlations between the children' 5 scores for the PAS and the children's Game Scale . Correlations between the fathers' CPI Fe scores and the Father scores on the PAS for the fathers of boys and girls . vi Page 29 31 34 38 38 41 41 43 45 Table Page 10. Correlations between the mothers' CPI Fe scores and the Mother scores on the PAS for the mothers of boys and girls . . . . . . 45 11. Correlations between the children' 3 responses to the PAS and their parents' CPIFescores 47 vii CHAPTER I IN TRODU C T ION A large number of studies in the area of sex-typing have been devoted to tracing the child' s increasing identification with the same-sex parent as he grows older. Another block of studies has dealt with the child' s identification with the parent possessing cer- tain characteristics, for example, nurturance, punishment, or power. However, few studies have dealt with the effects of the parents' identification on that of the child. This would seem to be an important area of investigation I, as the child does model his behavior after that of his parents and, therefore, the parents' own identification might be expected to influence that of his child. This study will concern itself with the effects of the parents' identification on that of the child. More specifically it will be con- cerned with the relative influences of the parents' overt and covert sex-role identity on the sex-typing of the child. Before describing this study in greater detail, some basic theories and research in the area will be discussed. There would seem to be two basic concepts involved in sex- typing: identification and modelling. The term identification has been used to refer to a wide variety of behaviors ranging from an initial perception of similarity between the child and his parents, through the process by which a child becomes more similar to his parents, to the end product of such a process (Kagan, 1964). Often identification encompasses modelling, which is a process of imita- tion that occurs as a result of observing the actions of others (Bandura and Walters, 1963). However, the two concepts are not entirely synonymous. There are several distinctions which are relevant to this study. One is that the process of identification concentrates more on the child' s adoption of a wide range of behaviors and attitudes, on the introjection of a set of attitudes or a "standard, " rather than on a specific set of behaviors. A standard, as used here, refers to "the culturally approved characteristics for males and females" (Kagan, 1964, p. 138). Thus, in the case of sex-typing the child adopts his parents' attitude as to what is appropriate role behavior. In this sense, the parents may or may not actually display this appropriate behavior for the child to want to adopt it; the main con- cern is with the values the parents hold. This, however, does not mean that the specific behaviors of the parents are unimportant, rather they are of secondary importance. Identification, then, empha- sizes covert similarity. Modelling, on the other hand, concerns itself more with specific behaviors. Thus the child learns to copy certain attitudes or actions because they are rewarded while he learns not to copy others because they are punished. The concept of modelling is more experimentally oriented than that of identification and it tends to emphasize overt similarities. Another difference is that identification has been used more to answer the question "Why does a child copy his parents?" while modelling is usually directed towards "What does the child copy and what facilitates or inhibits the process?" Thus many studies in identification, such as those by Mussen and his colleagues (Mussen and Distler, 1959; Mussen and Rutherford, 1963; Mussen, 1961), deal with the type of parent the child is most apt to identify with and why. From this point of view, three theories of identification have been derived and empirically tested. This is discussed below. Modelling deals more with the second part of the develop- ment of sex-typed behavior. Once the child has achieved the desire to be like the model, in this case the same-sexed parent, through a process of perceived similarity and identification, the questions arise as to what type of model does he have, what range of behaviors does he have to c0py from and how can his behavior be shaped? These are some of the questions that research on modelling attempts to answer. Thus there are two major contributing factors to the child' s sex-role development. This study is an attempt to gain more infor- mation as to the relative contributions of each. The main question to be answered is whether it is the specific behaviors of parents or the perceptions they have of themselves as masculine or feminine that influences the child' 8 sex role the most. Before describing the study in greater detail, the basic theories involved and some relevant research will be discussed. Three Theories of Sex-typing There are three main theories concerning the development of sex-typed behavior. The three are the developmental theory, the defensive identification or identification with the aggressor theory, and role theory. All deal with the process of identification in that they assume that the child will copy the behavior of the parent he identifies with although the reasons for this c0pying differ for the three'theories. Although this paper will deal mainly with the copying, or modelling aspects of the child' s behavior, it maybe helpful to look at these three theories to see why, in their views, the child begins to copy the behavior of his parents. All three theories have a common basis from which they soon depart. They all agree that identification is a belief that some of the attributes of the model belong to the self (Kagan, 1958). They also all agree that initially both boys and girls identify with the mother since she is the first source of gratification and affection but that sometime in his early years the boy must shift to a masculine identificand if he is to develop a normally masculine personality (Mowrer, 1950; Sears, Maccoby and Levin, 1957). It is in the causes for this shift in identificand that the theories diverge. (Although most of the theories focus on the development of identification in boys, the basic ideas, sometimes in modified form, apply to girls as well. ) The three theories may be summarized as follows: Mowrer formulated the developmental identification theory (Mowrer, 1950). He believed that when a boy is three or four years of age, the father begins to accept responsibility for him and the mother relaxes her domination. Now if the father is an important sourceof reward and nurturance, his responses and characteristics acquire secondary reward value and the boy imitates his father to reproduce parts of the loved object, including those parts concerned with masculine activities. The mother, on the other hand, continues to dominate over the girl and her behavior thus has secondary reward value for the girl, if she is rewarding and nurturant. Thus, a nur- turant relationship between the child and the same-sexed parent is the key to the child' 8 identifying with and copying the parent accord- ing to developmental theory. Identification with the aggressor was originally formulated by Anna Freud (1937) and has subsequently been adopted by Mowrer (1950). Eording to this theory, the boy, at about four or five years, begins to see his father as a rival for his mother' s affection and as a result he develops hostility towards his father. This hostility creates an expectation of counter-hostility from the father and the fear of retaliation. Classically this fear was thought to be more specific, in the form of fear of castration by the father. Less dramatically it may be thought of as a fear of punishment which causes the boy to adopt, or introject, the "father role" as a defense (Sears, Maccoby and Levin, 1957). This defensive process is thought to have a violent, crisis-like nature (Bronfenbrenner, 1960). The process for the girl lacks this violent nature and is thus not true identification with the aggressor (Bandura and Walters, 1963). For the girl, the father becomes the love object and hostility is directed against the mother. Fearing counter-hostility from the mother, the girl introjects the mother' s role in defense. However, there is no castration anxiety and the final identificand, the mother, is the same as the initial one; the shift does not occur as it does for the b0; . Here, fear of punishment, rather than fear of loss of love, prompts both boys and girls to introject the personality and characteristics of the same sex parent. A major consequence of this is the acquisi- tion of behavior patterns appropriate to one' 8 sex. Role theory combines aspects of both the developmental and defensive identification theories. Role theorists equate identification with the father and father role—playing. According to role theory, a child will assume the role of a parent if the parent is perceived as a powerful individual, one who controls both rewards and punishments (Sears, Maccoby and Levin, 1957). The child observes what the parent does and then practices it by pretending to be the parent. This practice is often covert. The child takes on the role of the salient parent with all its feeling and actions, including those related to sex-typed behavior. Es, according to developmental theory, the child will identify with the nurturant parent gccording to the defensive identi- fication theory, the child will identify with the punitive parentjmd according to role theory, the child will identify with the parent who is both rewarding and punitive, that is, the parent who is the most salient. Research A good deal of the research concerning the accuracy of these three theories has been carried out by Mussen and his associates. In the first study, Payne and Mussen (1956) compared twenty adolescent boys considered the most highly identified with their fathers with twenty boys considered the least identified with their fathers. The measure of identification was the number of father-son agree- ments minus the number of mother-son agreements on a fifty question true-false test taken from three scales of the California Psychologi- cal Inventory. The children and each of their parents responded to this scale. For the forty boys selected, the father-son relationships were assessed from the boys' completion of five incomplete pro- jective stories. The results indicated that the most highly identified boys perceived their fathers as more nurturant and powerful than did the least identified boys. This supported the developmental and role theories but not the defensive identification theory. A second study (Mussen and Distler, 1959) compared pre- school age boys. Here, masculinity was measured by the IT Scale for Children and the father-son relationship was assessed from . structured doll play sessions. The data revealed that the group with high identification had higher father-punishment, higher father- nurturance and therefore higher father-power scores than the low identification group. While high father-punishment supports the psychoanalytic hypothesis and high father-nurturance supports the developmental hypothesis, it was assumed that each contradicted the other as they are diametrically opposed. It was therefore con- cluded that no specific aspect of the father-son relationship leads to identification but rather that the cause is father saliency which supports the role theory, as this accounted for the greatest amount of the data. Thus the importance of the model would seem to lie in his ability to control rewards and punishments. A later experiment (Mussen and Distler, 1960) found that only the nature of the father—son relationship was important for the boy' 8 sex-typing. Here it was found that the personality of the father, the father' 3 degree of sex identification and parental en- couragement of sex—appropriate activities were not important nor was the boy' 5 interaction with his mother. The IT Scale for Children and mother interviews were used to collect these data. Also, it is important to note that the results were based on the boy' 3 percep- tions of his parents, not on their actual behavior. Freud felt, too, that identification is based on the image the child has of the parent rather than on the parent as he or she actually is (Bronfenbrenner, 1960). This study brings up another important idea as well, and one that will be returned to below, that of the father' 3 sex identifi- cation and its influence on the child. 10 There is one other relevant study by Mussen and Rutherford which will be discussed shortly. First, there are some criticisms to be noted in relation to the above studies: The first is that in two of the three studies, the child' 8 sex- typing was measured by the IT Scale for Children which Sears, Rau and Alpert (1965) have shown does not correlate well with other mea— sures of sex-typing and which is not an independent measure of masculinity or femininity since choice is involved. A second is that these findings, although they have been replicated by some research- ers (Bandura, Ross and Ross, 1963), have not been replicated by others (Sears, Rau and Alpert, 1965). A third problem is that in one case the parents' behavior was measured by retrospective data which is not always accurate. As mentioned above, most of these studies relied on the child' s perception of the parents rather than on the parents' actual behavior, on covert attitudes rather than overt actions. In other words, little information was obtained as to what the parents actually do and what the child sees. It cannot be assumed that the parent who accepts his role and has a proper role identification will necessarily display this proper role in front of the child. Other factors may enter into the picture and alter the conceptions that the child actually gets. For example, even if both parents are psychologically well 11 identified with their respective roles, if the mother works all day, who cooks supper and does the dishes at night? This is beimvior that the child actually sees and which may be quite influential. Modelling theory attempts to deal with the overt behavior of the identificand rather than his or her attitudes and it is to this that this paper will now turn. Modelling The child who is identified with his parents is likely to imitate their behavior or model himself after them. This concept of modelling is very important. It is the cornerstone on which the criticism of the defensive identification theory is based. This theory assumes that identifica- tion begins during the Oedipal period, during the fourth or fifth year of life. Several authors have disagreed with this. Lynn (1963), Hartley (1959), and Sears, Rau and Alpert (1965) have suggested that the parent is an important model for the child and that as early as age three the child begins to imitate the behavior of his or her parents. Also, Lynn and Sawrey (1959) demonstrated that the same- sexed parent is important for adequate sex-role adjustment in their study of father-absent boys. Rau' s idea that the parent serves as a model and an influence on motivation is in accord with this (Rau, 12 1960). This would suggest that where the development of sex roles is concerned, the identification established during the Oedipal period may not be of primary importance (Sears, _e_t_a_l_. , 1965). According to these same authors (Sears, _e_t_al. ), the model- ling process is far more important in the establishment of gender roles and this is the process which leads to the development of appro- priate behavior so early in life. It should be noted here that the child' s assumption of sex-appropriate roles is much more a function of his or her identification with and imitation of the same-sex parent since this is the parent displaying the appropriate characteristics. In the child' 5 early years, the members of his family con- stitute his available models (Bandura and Walters, 1965). The child is more likely to copy behavior which he sees his parents perform than those that he is told about. In this sense, an actual performance seems to provide better cues for the child than does a description. Perhaps that is one of the reasons why it is easier for the girl to imitate her mother' s behavior than it is for the boy to imitate his father' s, since the girl sees her mother most of the day while the father is usually away at work and consequently the boy does not get to view his activities as much. Also, play materials for girls are closely related to mater- nal activities-—toys such as dolls, kitchen appliances and so forth 13 which foster imitative play and help the girl practice those activities which she will use later on in life. Play materials for boys, however, do not foster typical paternal activities to a comparable degree. While some boy's toys do foster an imitation of masculine occupa— tional roles—-toys such as doctor kits and dump trucks--there is a greater disparity between a boy' 3 play materials and the role activi- ties he will adopt later on than there is for the girl. Hence, it is more difficult for the boy to practice adult role behaviors in play. This assumption, in play, of the parent role is an important step in the identification process. It is also a place in which the activities of the parents can influence to a considerable extent the development of role behavior in their children. Research The experiment by Mussen and Rutherford mentioned above will now be considered (Mussen and Rutherford, 1963). This study explores the effects of the same-sex parent who serves as a model for the child. One of the major hypotheses tested in this study was that the like-sex parent with high self-acceptance would promote masculinity in boys and femininity in girls. The assumption behind this was that the self-accepting parent would be more satisfied with his or her own sex-role and would therefore be more likely to reward 14 the child for imitation of this behavior. This would lead to the acquisition of the behavior of the same—sex parent. The results of the study showed that the mothers of high-feminine girls were more self-accepting but not more feminine than the mothers of low-feminine girls and that the fathers of high-feminine girls were more mascu- line but not more self-accepting than the fathers of low-feminine girls. Thus it seemed that for girls the characteristics of the par— ents were important along with the nature of the relationship, with the more feminine girls seeing their mothers as more nurturant and powerful than the less feminine girls. For the boys, however, only the nature of the father-son relationship was important with the high- masculine boys seeing their fathers as more nurturant, slightly more punitive and more powerful as in previously noted studies. This would seem to indicate that differential factors affect the sex- typing of boys and girls so that while the nature of the parent-child relationship was important for both, the attributes of the parent were important only for the girl. Heilbrun (1965) designed a study to show that modelling theory alone could account for sex-role development. He predicted that maximal differences between males and females would be found in families where there was identification with a high-masculine father. The subjects in this study were normal college students. 15 Child—parent similarities and the instrumental-expressive orienta- tion of the parents were assessed from questionnaires. The child' s personality was rated according to those items of the Adjective Check List he or she selected as being the most self-descriptive. The results supported the hypothesis ; 88% of the students identified with high-masculine fathers were judged to be appropriately sex- typed, 58% of those identified with high-feminine mothers, 47% with low-feminine mothers, and 43% identified with low-masculine fathers were so judged. This study would seem to suggest that the identifi— cation of the model is quite important for the sex-role development of the child. The discussion of modelling behavior is not meant to imply that the relationship between the parents and the child is not impor— tant, nor even that it is of lesser importance than what the child sees. It may well be that a nurturant, punitive or salient relation- ship is the most crucial factor in the child' s sex-typing as attitudes and emotions are sometimes more important behavior elicitors than deeds. However, the roles actually displayed by the parents should not be ignored as they often have been in the past. It would seem, therefore, that there are two basic concepts involved--identification and modelling--and that the child models his behavior after the parent he identifies with. Therefore, it is 16 necessary to look not only at the identification the child forms but also at the models he has available to copy and their roles. In examining the behavior of the parent models it is thus necessary to assess the degree of their sex-role identification both in terms of self-perceptions (covert behavior) and in terms of what they do that can be seen (overt behavior). This will give an indica- tion of how the parents' ideas and actions related to sex-typed be- havior influence the behavior of their children. This study will attempt to answer the question, "How is the child! s sex-role identification influenced by that of his parents?" The underlying assumption here is that the child will be more prone to develop appropriate sex-typed behavior and sex-identification if his parents assume the appropriate sex roles and demonstrate these. Hypothe s e s The major hypothesis to be tested is that a child' s role identification will be influenced by that of his parents and that his adaptation or lack of adaptation to his sex role will correspond to and be in the same direction as that of his parents. In other words, if the parents are highly role identified and display accurate and clear-cut sex-typed behaviors, so will the child. Conversely, if the parents are not role identified and do not display accurate and clear-cut role behaviors, neither will the child. 17 An additional hypothesis is that the child' s role identifica- tion will be influenced more by the same-sex parent than by the opposite-sex parent. The parents' identification, for this study, will be opera- tionally defined as the degree to which their behavior corresponds to that which has been judged to be appropriate for their sex. CHAPTER II METHOD The Method and Results sections of this paper will each be divided into two parts, as there were essentially two parts to this study, the first being the development of the Parent Activity Scale and the second being the employment of that scale in order to test the stated hypotheses. Phase I: The Parent Activity Scale The Parent Activity Scale (PAS) is a list of seventy activi- ties which are often performed by parents in their everyday lives and which their children usually see them doing. The items were arranged alphabetically to minimize the formation of sets by the subjects. (See Appendix II.) Subjects The Parent Activity Scale was administered to one hundred introductory psychology students. 18 19 P rocedure Each student was to rate each item according to whether he thought it would usually be performed more often by the mother or the father in a family. If the student felt that the activity was one that was usually performed equally by both the mother and the father, he was to rate the item as Both. On the basis of the differences in proportions between these ratings, items were assigned to one of three categories: Mother, Father or Both. After administering the PAS to the introductory psychology students, the frequency of Mother, Father and Both reSponses to each item was tabulated. A t-test of proportions was then performed in order to determine into which of these three categories an item fell. For an activity to be classified as either Mother or Father or Both, at least 44 subjects had to indicate that the activity belonged in that category. Forty-four subjects was chosen as the cut-off point because that was the number needed to reach the .01 level of significance. For all but one of the 70 activities, one category reached this level of significance. (See Appendix III.) The activity for which no category reached significance was item #26, "listens to radio more. ” One can see from looking at Appendix IH that the responses were fairly evenly divided for this item. This would suggest that the 20 activity should be performed about as often by mothers as by fathers. For this reason it was considered to be a shared, or Both, item. The distribution that resulted after classifying the items was as follows: 30 Mother activities, 25 Father activities and 15 Both activities. Thus, while the re are several more Mother activi- ties than Father activities, the two sets of items are of approximately similar lengths. This should facilitate comparability of Mother and Father scores so that a higher score obtained by one parent will not be merely a reflection of the differences in scale lengths but, rather, it will reflect a real difference in the degree of identification. The Both activities, which are generally of a non-sex-typed or shared nature, will serve as buffer items and thus ’will not affect the scores in any way. A scoring system was developed for the PAS by which a person receives credit for performing an activity of the appropriate sex and loses credit for performing an activity of the opposite sex. For example, for every Mother activity that a mother performs she receives one point and for every Father activity she performs she loses one point. Thus, it is possible for mothers' scores to range from 30 points to -25 points and for fathers' to range from 25 points to ~30 points. 21 Phase II: Employment of the Scale: Test of the Hypotheses Subjects Fifteen boys and fourteen girls* in the first and second grades in a parochial elementary school were interviewed. This age was chosen so that the children will have developed a sex-role identification. First and second grade children were chosen as they have not been influenced by other members of society besides their parents (i. e. , teachers and peers) as much as older children. The parents of these children were included in the study. All partici- pants had volunteered after a form letter was sent to their homes describing the study. (See Appendix II for the letter. ) Additional information describing the sample may be found in Appendix I. Briefly, all the children had at least one sibling with the family sizes ranging from two to twelve children. Only three of the 29 children did not have a same-sex sibling while four did not have an opposite-sex sibling. The mothers were predominantly home- makers with only four engaged in any other form of occupation, two working part-time and two full-time. The fathers were generally engaged in professional, managerial and white-collar jobs, with but *Originally fifteen girls were interviewed but one had to be omitted from the final analysis when it was learned that her father had passed away. 22 two exceptions. Most of the families reported that child-rearing practices have not changed in recent years, nor have the patterns of living. Where changes were reported, they were usually of a minor nature being due to the growing up of the children or the birth of new siblings. Procedure Each child was administered the Rosenberg and Sutton-Smith Game Scale individually (Rosenberg and Sutton-Smith, 1959). The scale may be found in Appendix II. This list consists of a number of toys and games frequently played by children and for which sex-role assignments have already been established. Each child was told that he or she would be asked which of a number of games he played. The child was asked, "Tell me which of these games you play. Do you play ?" If a child appeared concerned that he did not play many of the games, he was reassured that there were a lot of games on the list and that no one played all of them and that he was only expected to play a few of them. From the subjects' choices, a weighted score was obtained according to the weights and sexual assignment of the items as they appear in the original scale. (A copy of the weighted scale will be found in Appendix IV.) Independent masculine and feminine scores 23 were obtained for each child depending on the number of Boys' and Girls! games he or she played. To check on the similarity between the children' 5 choices and the parents' perceptions of how mascu- line or feminine their children are, the parents also filled out the Game Scale. A set of papers was sent home to the parents of each child in the study. This included an Information Form with several ques- tions about the family, a Game Scale, two copies of the PAS and two copies of the Femininity Scale of the CPI (labeled Attitude Question- naire). These forms may be found in Appendix II. The parents' sex-role identification was obtained from two sources. One of these was the Parent Activity Scale which was dis- cussed above. Each parent rated this scale as to whether that parent or the Spouse performs the activity the most or whether they share the activity fairly equally. A score was then obtained for each par- ent according to his or her agreement with the preassigned catego- ries. An agreement was scored as a plus for that parent while a disagreement was scored as a minus. For example, each time that a mother reported that she performed a Mother activity she received a plus one and each time she said she performed a Father activity she received a minus one. The opposite held for the fathers. No credit was added or deducted for Both responses. Here, the focus 24 is on what the parent actually does, not on what sex he thinks the item should be assigned to. Each parent was to fill out this form independently. It was hoped that in this way any covering-up done by one parent to cast a more favorable light on the division of labor in the family would be revealed by the other parent in the form of disagreement in the rating of an item. As a check on what the child perceives, each child was also given the PAS to see what he thought his parents did. The other scale given to each parent was the Femininity Scale of the California Psychological Inventory. (Referred to as the Attitude Questionnaire in the Appendix.) This scale was used as a measure of covert sex-role identification. Each parent filled out this scale so that a score could be obtained for each. Thus, two scores were obtained for each child, one from the child himself and one from his parents. Five scores were ob- tained for each set of parents: two representing what they report they do, another two representing their more covert sex identity and one reflecting what the child sees them do. CHAPTER III RESULTS Phase I: Some Characteristics of the PAS One other analysis was performed on the data obtained from the introductory psychology students and that was the determination of a reliability coefficient for the PAS. This was done to assess the internal consistency of the scale. It was based on the Mother and Father scores derived according to the procedure outlined in the Method section. There were two methodological problems connected with this analysis. The first was that as the items were not randomized, a split-half reliability, such as that obtained by comparing odd versus even items, could not be used. The second problem was that this was a three-choice rather than a two-choice test. This made it impossible to use the more conventional reliability test, the Kuder- Richardson. To overcome these problems, Cronbach' 3 Alpha reliability test was used. This test gives a reliability for all possible halves, 25 26 thus taking into account the non-randomized nature of the items in the scale. Also, this test permits the analysis of a scale whose items can be responded to in more than two ways. The only assump- tion necessary is that the choices lie on a continuum. To meet this assumption it was necessary to postulate a continuum ranging from feminine (Mother) activities to masculine (Father) activities. Those activities which were shared (Both activities) were postulated as lying in the center of the continuum. Since the items did, in fact, receive different numbers of endorsements for the three categories it is likely that they do lie on a continuum. The Cronbach' 3 Alpha obtained for the PAS in this phase of the study was .687. This is a moderate degree of reliability. A factor which may have lowered this reliability coefficient somewhat is that it was found, upon examining the answer sheets of the subjects, that several subjects seemed to be responding in terms of who performs the listed activities in their homes rather than who they thought should perform the activities. This may have reduced the internal consistency as there would tend to be more variance in who actually does something than in who should do it. As the task presented to the students was for them to rate who should perform the activities on the PAS, it is possible that the instructions were not clear enough to accomplish this aim. Further clarification may thus be desirable. 27 Phase II: Employment of the Game Scale, CPI, and PAS The purpose of this study was to consider the effects of the parents' sex—typing on that of the child. Thus the degree of the child's sex-role identification is the dependent variable in this study and this was measured by the Rosenberg and Sutton-Smith Game Scale (Rosenberg and Sutton-Smith, 1959; see Appendix I). When the masculine and feminine scales are scored separately, independent measures of masculinity and femininity can be obtained as mentioned in the Method section and this was the scoring procedure used here. For most comparisons, girls and boys will be treated separately to assess the relative effects of same-sex versus opposite-sex parents. Both masculine and feminine scores were computed for boys as well as girls. Kuder—Richardson reliabilities were computed for the Game Scale as responded to by the subjects in this sample. The reliabilities for the children are as follows: boys' responses to Boys' Scale, .79, boys' responses to Girls' Scale, . 68, girls' reSponses to Boys' Scale, . 63 and girls' responses to Girls' Scale, .87. These show a moderate to high degree of internal consistency, with the higher reliabilities being for the same-sex scales. Thus the children tended to be more consistent when reSponding to the same-sex scale than to the opposite-sex scale. 28 For the parents the reliabilities are: boys' parents' re— Sponses to the Boys! Scale, .70, to the Girls' Scale, . 87, girls' parents' responses to the Boys' Scale, . 90, to the Girls' Scale, .80. Thus, for the parents the internal consistencies were also moderate to high, but here the high reliabilities were associated with the opposite-sex scale. Therefore the parents tended to be more consistent when responding to the opposite-sex Game Scale than to the same-sex Game Scale. Part 1: Relationship between the Game Scale Given to Children and the Game Scale Given to their Parents The Rosenberg and Sutton-Smith Game Scale was responded to by both the children and their parents. This was done in order to compare what the child reports he does with the parents' perceptions of the child' 3 activities. Although in this study the child's scores will be used primarily, the two will first be compared to assess their similarity. This may be seen as a parallel to the comparison of the child! s perception of what the parents do to the parents' actual performance, both as measured by the PAS (to be discussed below). The same procedure as that used for arriving at independent masculine and feminine scores for the children' s reSponses was 29 used for the parents' responses. The scores for the children's and parents' responses were then correlated. For same-sex activities it can be seen from Table 1 that the correlation between the Boys' Game Scale as responded to by the boys and the Boys' Game Scale as responded to by their parents is .232. For the girls, the correlation between their reSponses to the Girls' Game Scale and their parents' responses is —. 027. Neither of these correlations approaches significance. For opposite-sex activities, the situation is essentially the same. The correlation between the Girls' Game Scale as responded to by the boys and the Girls' Game Scale as responded to by their parents is . 049. For the girls, the correlation between their responses to the Boys' Game Scale and their parents' re5ponses is -. 030. TABLE 1. -- Correlations between the scores on the Game Scale for children and parents Scales Compared On Boys' Scale Girls' Scale Boys . 232 . 049 Girls -.030 -.027 These correlations indicate that, for the present sample at least, there is little or no correlation between what the child says he 30 does and what his or her parents think he does. This applies to both same-sex and opposite-sex activities. There would seem to be two major reasons for these low correlations. The first, and perhaps more important, is that the parents are not always with the child and therefore do not always know what activities the child is engaged in. The second is that a number of parents seemed to respond to the scales in a defensive manner. Both these factors could account for the discrepancies in the scores. The second reason is especially relevant in regard to the Girls' Game Scale as reSponded to by the boys and their parents. By computing t-scores for the differences between the scores on the parents' scales and those on the sons' scales (see Table 2), it can be seen that many parents tended to see their sons as playing signifi- cantly fewer games than they report they play. Mostly, the mothers saw their sons as playing fewer feminine games than the sons re- ported playing (a t-score of 2. 98). However, this difference is not consistent so that some parents saw their sons as playing more feminine games than they play and this contributes to the low corre- lation. This same pattern holds for the responses of the girls and their parents to the Girls' Game Scale except that the sizes of the discrepancies were considerably smaller. However, the inconsis- tencies were present here also. 31 TABLE 2. -- T—test scores comparing the number of responses of parents and children on the Game Scale T-test Score Boys: Boys' Scale 2.56* Girls' Scale 2. 98** Girls: Boys' Scale 0.18 Girls' Scale 1.05 *Indicates test is significant at the .05 level **Indicates test is significant at the .01 level Further evidence of the parents' defensiveness is that they responded to fewer items for fewer total points for three of the four t—tests. Although part of this may be due to the fact that they do not see their children all the time, on examination of the actual proto- cols, some of the parents seemed to be going through the scales and trying to guess which items were appropriate for their child and then checking only those items. (Also several parents added items to the scale which they thought were appropriate to their child' 8 sex.) This would seem to be an additional source of variance among the scores and thus a factOr leading to a lower correlation. Another explanation for the discrepancies is that many of the games included in this list are games that are played in school. 32 Thus they are played when the parent is not around. The larger dis- crepancies for the boys may indicate that this is especially true for them. The inconsistency in the discrepancies (i. e. , the fact that some parents saw their children as playing about as many, or more, of the girls' games while many others saw their children as playing fewer of these games) may stem from two sources: (1) some parents may be more aware of the games played at school than others or have better communication with their children about what they are doing and (2) some children may play more of these games at home, in the presence of their parents, than others. In regard to the Boys' Game Scale, the boys' parents tended to see them as playing fewer masculine games than they actually play. Again, however, the pattern was inconsistent but not as inconsistent as for the Girls' Game Scale. As a result, this correlation was larger, although still not significant. For the girls, although the overall comparison indicates that the parents saw their daughters as playing about as many masculine games as they actually play (a t-test score of 0. 18), the pattern is very inconsistent with six par— ents seeing their daughters as playing more masculine games and seven seeing them as playing fewer masculine games. This could account for the low correlation. 33 Part 2: The Relationship between the Game Scale and the CPI The Femininity Scale (Fe) of the California Psychological Inventory was filled out by both the mother and the father of each subject. This yielded a measure of each parent' s covert sex-role identification which then could be compared with the degree of sex- typing of the child as measured by the Rosenberg and Sutton-Smith Game Scale. Thus the influence of the parents' covert role identi- fication on the sex—typing of the child could be assessed in accordance with the major hypothesis of this study. On the Fe scale of the CPI a higher score indicates greater femininity. This is true for males as well as females. In order to assess whether the degree of the mother's sex-role identification is related to that of the father, these scores were correlated with each other. For the boys, the correlation between the mothers' CPI scores and the fathers' CPI scores was .410. This score approaches significance at the .05 level. It indicates that, for this sample, as the mothers of the boys became more feminine, so did the fathers. This was also the case for the girls except that the relationship was not as pronounced. Here, the correlation between the mothers' and fathers' CPI scores was .108. Thus for both boys and girls, as one parent became more or less feminine so did the other. 34 Turning to comparisons between the children and their par- ents, they will be compared first on the basis of the child' 8 same- sex identification (based on the same-sex scale of the Game Scale). Table 3 presents the results of these comparisons. TABLE 3. -- Correlations between the scores on the Children' s Game Scale and the parents' CPI Fe scores Fathers' Fe Mothers' Fe Boys: Boys' Scale -. 405 -. 391 Girls' Scale -. 063 -. 334 Girls: Boys' Scale -. 204 -. 312 Girls' Scale -. 061 . 126 It can be seen from this table that the correlation between the boys' responses to the Boys' Game Scale and the fathers' Fe scores is -. 405. This correlation approaches significance at the .05 level. This indicates that, for this sample, as the father became less feminine, the son became more masculine. Comparing the boys' responses to the Boys' Game Scale with the mothers' Fe scores, the correlation is -. 391, a correlation which approaches significance at the .05 level. This indicates that as the mother became more feminine, the son became less masculine. 35 Looking at the correlations between the girlsI responses to the Girls' Game Scale and their parents' Fe scores, neither of the correlations approaches significance. Comparing the girls' re— sponses with their mothers! Fe scores yields a correlation of . 126, which, although it is not significant, is in the predicted direction and suggests that as mothers became more feminine, so did their daugh- ters. The correlation between the girls' responses to the Girls' Game Scale and their fathers' Fe scores is -. 061, which indicates _ no relationship between the two. It can be seen here that there are lower correlations between girls and their parents than between boys and their parents. One other point to' be noted is that, although the differences between the correlations are small, there is a trend for both boys and girls to identify more with the same—sex parent than with the opposite-sex parent. Table 3 also presents the correlations between the parents! CPI scores and the opposite-sex Game Scale scores. It can be seen that the correlation between the boys' responses to the Girls' Game Scale and their fathers' CPI Fe scores is -. 063. This indicates that there is little relation between the two. . Comparing the boys' re- sponses to this scale with their mothers' Fe scores, the correlation is -. 334. This indicates that as the mothers became more feminine 36 the boys became less feminine. Thus the correlation is higher between the boys' and mothers' scores than between the boys' and fathers' scores. For the girls the opposite trend was noted with the girls becoming less masculine as the parents became more feminine. Thus the correlation between the girls' responses to the Boys' Game Scale and the mothers' Fe scores is -. 312 while the correlation between the girls' responses and the fathers' Fe scores is -. 204. Here again, there is a larger correlation with the mothers' scores than with the fathers' . It had been intended that the parents' responses to the Game Scale would be compared with their responSes to the Fe scale of the CPI. However, the Game Scale was filled out by only one parent rath- er than by both. Also, it could have been filled out by either parent as no stipulation was made for a specific parent to fill it out. Thus, in some families it may have been filled out by the mothers and in others by the fathers. This makes it impossible to attribute the perceptions to one set of parents rather than the other. Since it is not known which set of parents the perceptions belong to, mean- ingful relationships between the parents' perceptions of their children's sex-typing and their own sex-identification cannot be made. This is because it cannot be known with which parent' s identification 37 to compare the perceptions of their children without knowing whether it is the mothers' or fathers' perceptions that are being considered. Part 3: Relationship between the PAS as Given to the Children and the PAS as Given to their Parents The PAS was responded to by every child and every parent. Each responded in terms of his or her perceptions of who did what at home. From each protocol a Mother and a Father score was obtained. This yielded three Mother and three Father scores for each couple: one from the child, one from the mother and one from the father. As the six scores were based on the activities of the same two people, it was predicted that they should be similar. Cronbach' s Alphas were computed for the parents and chil- dren. They were as follows: boys' parents, .654; girls' parents, .685; boys, .416; and girls, .585. This shows that the test has a moderate degree of internal consistency for the parents and a lesser degree of internal consistency for the children. The first set of comparisons to be considered in this section is that between the different scores for each couple. This will be done to compare the perceptions each person has in regard to the division of labor in the family. Tables 4 and 5 present the correla- tions between the mothers, fathers and children for the Mother and Father scores of the PAS. TABLE 4. -- Correlations between the scores of the fathers, mothers, and children for the Father scores on the PAS Fathers Mothers Children Boys' Fathers 1. 000 Boys' Mothers . 573* 1. 000 Boys Themselves .236 .246 1. 000 Girls' Fathers 1 . 000 Girls' Mothers .445 1 . 000 Girls Themselves . 559* . 372 1 . 000 *Indicates significance at the .05 level **Indicates correlations significant at the .01 level TABLE 5. -- Correlations between the scores of the fathers, mothers, and children for the Mother scores on the PAS Fathers Mothers Children Boys' Fathers 1. 000 Boys' Mothers . 533* 1 . 000 Boys Themselves -. 035 . 191 1. 000 Girls' Fathers 1. 000 Girls' Mothers .682** 1. 000 Girls Themselves . 610* .571* 1 . 000 >I‘lndicates significance at the .05 level **Indicates correlations significant at the .01 level 39 It can be seen from these tables that, for both the parents of the boys and girls, three out of the four correlations between the mothers' and fathers' perceptions of what each other does are significant at the .05 or .01 level and that the fourth approaches sig— nificance. Thus the parents tend to perceive the division of labor in their homes quite similarly. This is true for girls and both their mothers and fathers also. Here, again, three out of four of the cor- relations are significant at the .05 level and the other is in the pre- dicted direction, which again indicates similarity of perceptions. For the boys and their mothers and fathers, although none of the correlations is significant, three out of four are in the predicted direction. The only one that is not is that between the boys' fathers' perceptions of what the mothers do and the boys' perceptions of this. This correlation implies that there is little relationship between the two sets of perceptions. This relationship may be attenuated by the low reliability of the boys' scores. Therefore, for all but one of the comparisons, the percep- tions that the family members have of what the parents do are fairly similar with six being significantly related at the .05 or .01 levels and five others being in the predicted direction although not signifi- cant. Also, the perceptions of the mothers, fathers and daughters are more similar to each other than are those involving the sons. 40 One final point is that both boys and girls tended to agree more in their perceptions of the same-sex parent than in their perceptions of the opposite-sex parent. Part 4: Relationship between the Game Scale and the PAS Both the Game Scale and the PAS were used as measures of sex-role identification, the Game Scale for the children and the PAS for the parents. Here, as in Part 2, the two are compared to assess the influence of the parents' role-identification on the sex-typing of the child. The difference between the two is that here the degree of the parents' overt identification is being considered whereas in Part 2 the focus was on the parents' covert identification. Tables 6 and 7 present the correlations between the parents' PAS scores and the children's Game Scale scores. It can be seen from these tables that for both the mothers' and fathers' responses to the PAS, five out of the eight correlations were in the same direc- tions so that the difference in the relationships was one of degree only. Those five relationships will be considered first. It can be seen from. the tables that as the boys' mothers became more feminine, the boys became both less masculine and less feminine. Also, as the fathers became more masculine, the boys became both less 41 masculine and less feminine. The femininity results are in the pre- dicted direction; the masculinity ones are not. TABLE 6. -- Correlations between scores on the PAS as responded to by the fathers and the children' 3 scores on the Game Scale Mother Score Father Score Boys: Boys' Scale -. 357 -. 495* Girls' Scale -. 306 -. 166 Girls: Boys' Scale -. 462 -. 559* Girls' Scale -. 013 -. 141 *Indicates correlation is significant at the .05 level TABLE 7. -- Correlations between scores on the PAS as responded to by the mothers and the children's scores on the Game Scale Mother Score Father Score Boys: Boys' Scale -. 345 -.258 Girls' Scale -. 330 -.455 Girls: Boys' Scale -. 299 .037 Girls' Scale .141 .432 42 In regard to the girls, the one set of correlations that were in the same direction for both parents was that between the mother's' PAS scores and the girls' Boys' Scale scores. These implied that as the mothers became more feminine, the daughters became less masculine, which is consistent with the prediction. The other three correlations present inconsistent and con- flicting data. One possible prediction, however, is that there is little relation between the mothers' overt femininity and the daugh- ters' femininity with perhaps a slight tendency for the daughter to become more feminine as the mother does. Regarding the fathers' scores, the larger correlations would suggest that as the fathers became more masculine the daughters would become less masculine and more feminine. This would be consistent with theory. However, there is no reason to believe that the larger correlations are more accurate than the smaller ones so that the reverse could be true. Thus, in connection with the girls' identification, the only tendency to be reported is that girls tended to become less masculine as their mothers became more feminine. The other results reveal, if nothing else, shortcomings in the administration of the PAS. Now that two sets of scores have been examined, those of the mothers and fathers, it is time to turn to the third set, those reflecting the children' s perceptions of the division of labor in the home. Table 8 presents these results. 43 TABLE 8. -- Correlations between the children' 5 scores for the PAS and the children' 8 Game Scale Mother Score Father Score Boys: Boys' Scale -. 281 .460 Girls' Scale -. 340 -. 243 Girls: Boys' Scale -. 254 -. 220 Girls' Scale .282 .195 The correlations between the boys' perceptions of their fathers' activities and their own responses to the Boys' and Girls' Game Scales are .460 and -. 243 respectively. These imply that as the boy saw the father as becoming more masculine, he became more masculine and less feminine. In regard to the mother, the correlations suggest that as the boy perceived her as more feminine, he became less feminine and less masculine. The first three of these are consistent with the modelling theory while the last is not. For the girls, all the correlations, while low, are in the predicted directions. Thus the correlations between the girls' per- ceptions of their mothers' activities and their own responses to the Boys' and Girls' Game Scales are, respectively, -. 254 and .282. These suggest that as the girls saw their mothers as becoming more feminine, they themselves became less masculine and more feminine. 44 For the fathers the corresponding correlations are -. 220 and . 195. These imply that as the girl saw the father as becoming more mascu- line, she tended to become less masculine and more feminine. All these results, while not significant, are consistent with the model- ling theory and with each other. Here, as with the Fe scale of the CPI, it was intended that the parents' Game Scale scores would be correlated with their PAS scores. However, due to the methodological problem noted in that section, these correlations have been omitted. Part 5: Relationship between the Fe Scale of the CPI and the PAS It has already been stated that in this study the Fe scale of the CPI was used as a measure of the parents' covert sex-role identification while the PAS was used as a measure of the parents' overt sex-role identification. It would be expected that the two forms of identification would be positively related, that is, as one became more appropriate so would the other. To see if they were, the sets of scores for each were correlated with the other. Tables 9 and 10 present these correlations. It should be recalled that for the fathers and mothers alike a higher CPI Fe score indicates greater femininity while on the PAS a higher Father score indicates greater masculinity. Thus, for the father, there should be negative correlations between his femininity on the CPI and his masculinity on the PAS. 45 TABLE 9. -- Correlations between the fathers' CPI Fe scores and the Father scores on the PAS for the fathers of boys and girls Boys Girls Fathers' Father Score .288 . 177 Mothers' Father Score .068 .023 The correlations in Table 9 imply that as the fathers became more feminine on the covert level, they tended to become slightly more masculine on the overt level. This was more the case for the relationships involving the fathers' perceptions of his own overt behavior than for those involving the mothers' percep- tions of the fathers' overt behavior. In the latter case the corre- lations were near zero. These results are not in the predicted direction. TABLE 10. —- Correlations between the mothers' CPI Fe scores and the Mother scores on the PAS for mothers of boys and girls Boys Girls Fathers' Mother Score . 589* -. 024 Mothers' Mother Score .348 .378 *Indicates the correlation is significant at the .05 level 46 For the mothers there was a tendency for the overt sex-role identification to increase as the covert sex-role identification in- creased although only one of the correlations reached significance. Also, one of the correlations was not in the predicted direction. The correlation between the girls' fathers' perceptions of the mothers' overt behavior and the mothers' Fe scores is -. 024; however, this relationship is negligible. Thus, for the mothers, three out of four correlations were in the predicted direction and of these one was significant. The highest relationship was between the boys' fathers' perceptions of the mothers! overt behavior and the mothers' reported covert behavior. In the last section it was discovered that the children' s per- ceptions of their parents' behavior was more important to their sex- typing than were the parents' perceptions. For this reason the children' 3 responses to the PAS will now be compared with the par— ents' CPI Fe scores to see if the children' s perceptions of the parents' overt behavior was related to the parents' covert behavior. Again, it should be remembered when looking at Table 11 that a higher CPI score means greater femininity. It can be seen from the table that there is a sex difference in these correlations so that the correlations for the boys are all positive while those for the girls are all negative. For the boys, the 47 correlations suggest that as the fathers became more covertly feminine there was a tendency for their sons to see the mothers as more overtly feminine (.328). As the mothers became more covertly feminine there was a tendency for the sons to see them as more overtly feminine and the fathers as more overtly masculine. Thus, as the fathers became less appropriately sex-typed and the mothers became more appropriately sex-typed on a covert level, the sons tended to see both the fathers' and mothers' behavior as more appropriately sex-typed. TABLE 11. -- Correlations between the children' 8 responses to the PAS and their parents' CPI Fe scores Fathers' Fe . Mothers' Fe Boys: Father Score . 032 . 340 Mother Score . 328 . 116 Girls: Father Score -. 155 -. 501* Mother Score -. 253 -. 387 *Indicates the correlation is significant at the .05 level The opposite is true for girls. Here, as the fathers became more covertly feminine there was a tendency to see them as less overtly masculine and to see the mothers as less overtly feminine. 48 Too, as the mothers became more covertly feminine the daughters tended to perceive them as being less feminine and the fathers as less masculine in overt behavior. For the fathers this relationship was significant at the .05 level with a correlation of .501. There- fore, for the girl, as the fathers became less appropriately sex- typed and the mothers became more appropriately sex-typed, the daughters tended to see them as engaging in less appropriate sex- typed overt behavior. . Thus, for boys and girls, as the same-sex parent became more appropriately sex-typed on a covert level, the children tended to perceive their parents as less appropriately sex-typed on an overt level while as the opposite-sex parent became more appropriately sex-typed so did the children' s perceptions of both parents. CHAPTER IV DISCU SSION This section will concentrate on the implications of the relationships between the Game Scale, the CPI and the PAS. The focus for the discussion will be the hypotheses stated in the Introduc— tion, that is: (1) that a child' s sex-role identification will correspond to and be in the same direction as that of his parents and (2) that this identification should be influenced more by the same-sex parent than the opposite—sex parent. It will be recalled from the Introduction that the more highly sex-typed parent should provide a better model for his or her children, both in attitudes and actions. Part 1: Relationship between the Game Scale as Given to Children and the Game Scale as Given to their Parents The scores on the Game Scale as given to the children were compared with the scores on the Game Scale as given to their par- ents to assess the similarity between the parents' perceptions of what the child does and the child' s own perceptions. 49 50 In the Results section it was noted that all the correlations between the two scales were quite low, suggesting that the two sets of perceptions have little or no relation to each other. Two reasons were given for these low correlations: (1) the parent is not always with the child and therefore is unaware of much of what the child does and (2) several of the parents seemed to be responding defensively to this scale. These reasons, along with the data on which they were based, have already been presented and will not be repeated here. They were included in the Results section because of the quantitative nature of the data used to support them. This summary is included here to provide continuity between the sections. Part 2: Relationships between the Game Scale and the CPI Fe In the Results section it was noted that, for this sample, as one parent became more feminine, so did the other. This result was obtained by comparing the CPI Fe scores of mothers and fathers. Considering that the Fe scale of the CPI is largely a measure of a person' s interests, this result would tend to indicate that people with similar interests marry each other. This means that the child is presented with two parents with similar role identifications, that is, with two models that resemble each other. This child would then get a relatively good idea of the particular interest pattern and role identification displayed in the family. 51 A possible implication of this is that if one parent is highly sex-typed and presents a good role model, the other will tend to be less sex-typed and present a rather poor role model. Thus, if the mother is highly feminine, she will present a good model, but at the same time, the father will tend to be feminine and thus a poor model. Children in a family like this would tend to get a fairly clear picture of feminine interests with little idea of masculine ones. The oppo- site would be true in families where both parents were low in femininity. Of course, these are extreme cases. However, except when both parents present moderate patterns, one will tend to be more appropriately sex-typed than the other if the two are similar. Turning to the relationship between the children and the parents, as the father became less feminine, the sons became more masculine. As more masculine fathers would possess more appro- priate attitudes which should be conveyed to their sons, this result is consistent with the present theory. It also supports the first hypothesis. The relationship between girls and their mothers par- allels this in that as the mothers became more feminine so did their daughters. Again, one can conceive of the feminine mother as the good model who fosters appropriate role attitudes in her daughters. This, too, supports hypothesis one. Thus, as the same-sex parents became more appropriately sex-typed, so did their children. This is consistent with both hypotheses one and two. 52 In regard to boys and their mothers, the indication was that as the mother became more feminine the son became less masculine. Theory would predict that the reverse would be true because a femi- nine mother would possess appropriate sex-role attitudes and should convey these to her children. However, as the mothers became more feminine there was a tendency for the fathers to become more feminine so that both parents presented feminine models to their sons which the sons may have copied. Thus, the more feminine the mother, the more feminine the father and son. In this case, as the parents became more feminine, there would be an increasing lack of masculine attitudes to convey to the son so that he would have little to model his masculinity after. Also, as the mother, and there- fore the father too, became less feminine, so would the son because he would have less feminine models to.copy. Here it is difficult to say whether it was the mothers' or fathers' femininity (or lack of it) that affected their sons the most. The slightly greater correlation between the boy and his father than between the boy and his mother might suggest a trend toward greater identification with the father which would lend some support to hypothesis two. Comparing the girls and their fathers it would seem that there was no relationship between the two. However, the re was a 53 slight tendency for the girls to become more feminine as their mothers did. This is consistent with the modelling theory that a more feminine mother would present more appropriate attitudes for her daughter'to imitate so that the daughterwould then become more appropriately sex-typed. This result also lends some support to hypothesis one. As with the boys, the correlation for the girls was greater for the samelsex parent (in this case the mother) than for the opposite-sex parent. This suggests a trend towards greater identi- fication with the mother which would support hypothesis two. One other point may be noted in regard to these comparisons and that is that the correlations between the girls and their parents are lower than between the boys and their parents. There are two possible and related explanations for this. One is that in our society the roles of the female are not as well-defined as those of the male and thus more variance is permitted. That is, the female is free to choose from a wider range of activities than is the male. This makes it harder for any two patterns to be comparable and this may enter into the lower correlations. The other explanation revolves around the psychoanalytic proposition that, due to the less intense nature of the Oedipal situation for the girl, she does not become as sex—typed as the boy. Here, too, there would be more variance in the girls' identification and thus lower correlations. 54 To summarize this section, the childI s same-sex identifi- cation tended to increase as the same—sex parents' role-identification increased. This supports hypothesis one. However, in regard to the opposite-sex parent, the data does not support the hypothesis that as the parent becomes more appropriately sex-typed so does the child. One reason for this may be that the child tended to identify more with the same-sex parent than with the opposite-sex parent. This would support hypothesis two. Next to be considered are the relationships between the children' s opposite-sex identification, as measured by the Opposite- sex Game Scale and the parents' covert identification as measured by the Fe scale of the CPI. It was found that there was little relationship between the boys' femininity and his fathers' femininity. However, as the mothers became more feminine, the boys became less feminine. This latter result is consistent with the hypothesis that as the model becomes more appropriately sex-typed so will the child. In this case, as the mothers became more appropriately sex—typed, that is, more feminine, the boys tended to become more appropriately sex- typed in becoming less feminine. This lends support to hypothesis one. The former result, however, is inconsistent with this hypoth- esis in that it would have been expected that as the fathers became 55 more feminine so would their sons. Thus, hypothesis one would not be supported here. Also, these results suggest that the boys were identifying more with their mothers in regard to femininity than with their fathers. Although the mother may be regarded as a more appro- priate model for femininity, it had been predicted that boys would identify more with their fathers since they are the same-sex parents. Thus, hypothesis two is not supported here. For the girls the trend is reversed in that as the parents became more feminine the girls became less masculine. Here the parents may be seen as becoming better feminine models for the girls as they become more feminine. Also, as the correlation is higher between the girls and their mothers than between the girls and their fathers, this would suggest that the girl is identifying more with the mother which supports hypothesis two. It also suggests that the increasing appropriateness of the mothers' attitudes with increas— ing femininity is more important than the decreasing appropriateness of the fathers' attitudes. This also lends support to hypothesis two. Thus for both the boys and girls in this sample, the mother seems to be the dominant identificand with regard to opposite-sex behavior. This tends to support hypothesis one in that the sex- appropriate attitudes on the part of the parent, particularly the mother, fosters sex-appropriate attitudes in the child. In the case 56 of the girls, it also lends support to hypothesis two in that the mother is the major identificand. However, for the boys, hypothesis two is not supported. As mentioned in the Results section, it was intended that the parents' reSponses to the Game Scale would be compared with their CPI Fe scores. However, due to the methodological problems men- tioned above, this was not done. Part 3: Relationships between the PAS as Given to the Children and the PAS as Given to their Parents The PAS was given to each member of the family that was included in the study. This was done to compare the perceptions they have of who performs what tasks at home. It was predicted that these correlations should be positive and relatively high as the per- ceptions are based on the same two people, the mother and the father, performing relatively overt and observable behaviors. The data support the prediction with six out of the twelve correlations being significant at the . 05 level or better and five others also being in the predicted direction. The perceptions of the mothers, fathers and daughters were fairly similar. Those of the sons, however, were not as similar as those between the other re Spondents . 57 One possible reason for this may be that the sons take less interest in household activities than do the participating members, namely the mother and father, and the daughters. When one con- siders sex-role behavior as being a preparation for what will be done later in life, it can be seen that it is more important for the girl to know about household activities than it is for the boy. The boy, then, should be concerned mainly with the father' 3 activities. This is sup- ported by noting that the boys' perceptions of what the fathers did were more similar to the parents' perceptions than were the boys' perceptions of what the mothers did. For the girls also, there was greater comparability between the girls' perceptions of the mothers' activities and the parents' perceptions of these than between the girls' perceptions of the fathers' activities and the parents' perceptions of these. Thus, the child seemed to be more aware of the same-sex parent' s activities (as reported by the parents) than of the opposite- sex parent' s activities, and girls seemed to be more aware than boys. The fact that the correlations between the parents' reports are fairly high would tend to suggest that these may be reasonably accurate measures of the behaviors, or that where covering—up is being done, both parents are doing it. However, it should be noted that, even though the correlations are high, the scores are not the same and these differences may be important. 58 Part 4: Relationships between the Game Scale and the PAS The first set of comparisons in Part 4 of the Results section compared the children! 8 scores on the Game Scale with the parents! scores on the PAS. As three out of the four correlations between the mothers! and fathers! PAS scores were significant it was expected that the results would be similar no matter which parent! 3 scores were used. However, this was not the case. One reason for this is that a correlation of about .5, although significant, only accounts for twenty-five percent of the variance which means that the two sets of scores are still quite different. Thus only five out of the eight pairs of comparisons were in the same direction for both mothers and fathers. In regard to the boys, all four pairs of comparisons were in the same direction. The correlations for the mother suggested that as the mother became more feminine, the sons became both less masculine and less feminine. As more feminine mothers tended to be married to more feminine fathers, it would be expected that their sons would be less masculine since they would have only feminine models to identify with. Conversely, as the mothers became less feminine, the fathers and the sons would then become more mascu- line as this is the type of models they would have. This supports 59 hypothesis one. In this case it seems that as the father became more or less appropriately sex-typed, so did the son, so that the appro- priateness of the son's sex-typing would seem directly related to that of his father. This would support hypothesis two. It would also be consistent with the results obtained with the CPI for boys. However, looking at the fathers! scores, it seems that as the fathers became more masculine, the sons became less masculine and less feminine. The masculinity result is inconsistent with the theory that more appropriately sex-typed fathers would be better models for their sons. It is also inconsistent with the results for the mothers. Even if one considers the more masculine father as being married to a more masculine mother, an inappropriate model, the two results are still contradictory. One possible reason for this is similar to that given above: that a correlation of about . 4 between the identifications of the two parents only accounts for sixteen per- cent of the variance so that although several more feminine mothers may be married to more feminine fathers (and the same for less feminine mothers and fathers), others may not be. Thus, the fathers and mothers might be considered separately. In that case, only the relationships between the mothers! and sons! femininity and fathers! masculinity and son! 3 femininity would be in the predicted directions and therefore in support of hypothesis one. The other results would refute this hypothesis . 60 Another possible reason for the conflicting results may be some imperfections in the PAS or the way the subjects are respond- ing to it. Thus the scale may need modification. Whatever the reason, the results for the boys are inconsis- tent with each other and while two parts of the data lend support to the theory, the other two parts do not. This would suggest that, for the boys in this sample, the relationships between the parents! overt behavior and the boys! sex-typing are, in part, inconsistent with modelling theory. Instead of the boys becoming more appropriately sex-typed as the parents! overt behavior became more appropriate, the opposite seemed to occur, at least regarding their masculinity. For the girls, only one pair of correlations was in the same direction for both parents so that both predicted that as the mother became more feminine, the daughter became less masculine. This is consistent with hypothesis one in that the more feminine mother would display more appropriate behavior which the daughter would tend to copy. It is also consistent with the results obtained with the CPI. Too, as the girls seemed to be modelling their behavior after their mothers, hypothesis two receives support. The other three pairs of correlations contradict each other. As the two correlations between the mothers! PAS scores and daugh- ters! Girls! Scale scores were quite low, it may be suggested that 61 there is little relationship between what the mother did and the daughter! s femininity. Regarding the fathers! scores, little can be said except that there is no way of knowing which of the correlations are a better representation of the real world. The larger correla- tions would suggest that as the father became more appropriately sex-typed, so did the daughter, which lends support to hypothesis one. However, it can be seen that the parents differ somewhat in their perceptions of the division of labor in the families so that one cannot say which parent is more accurate. It seems, from this, that further development of the PAS, or the way it is responded to by the parents, is necessary so that one relatively accurate measure of the parents! roles can be arrived at. Perhaps if the parents conferred on their reSponses this might be achieved. Thus, for both boys and girls, there is considerable contra- diction in the results obtained by comparing the parents! responses to the PAS to the children! s responses to the Game Scale. One major factor that may account for this has already been discussed and that is that the PAS may need refinement. Another explanation is that children! s sex-typing may not be related in any consistent manner to the parents! overt sex-role identification. It may be that only the parents! attitudes or the child! s perceptions of the parents! roles are important and that any results obtained with the parents! PAS 62 are due to chance factors. This explanation cannot be either ruled out or confirmed without additional evidence. It has just been suggested that the child! s perceptions of the parents! overt sex—roles may be an important factor. This implies that it may not be as much what the parents think they do as what the child thinks they do that influences the child! s sex-role development. To see if this was the case the children! 3 responses to the PAS and their responses to the Game Scale were compared. For the boys, three out of the four correlations were in the predicted direction. That is, as the son perceived the father as be- coming more masculine, he tended to become more masculine and less feminine; and as he perceived the mother as becoming more feminine, he tended to become less feminine. This is consistent with modelling theory in that as the boys saw the parents as being better models, i. e. , more appropriately sex-typed, they tended to become more appropriately sex-typed. This supports hypothesis one. The only correlation inconsistent with this is that as the boys saw their mothers as becoming more feminine, they tended to become less masculine. If here one considers that more feminine mothers are sometimes (but not always) married to more feminine fathers, it 63 is plausible to assume that the boys may be modelling themselves after the more feminine father. (Even if the more masculine father is married to the more masculine mother, the boy would still be modelling himself after the father and the dominant mode in the family if he becomes more masculine as stated above.) These re- sults would also lend support to hypothesis two as well as hypothesis one. For the boys then, the results seem much clearer and lend more support to the hypotheses when considering the boys! percep- tions of what the parents do rather than the parents! perceptions. For the girls all the correlations are in the predicted direc- tion. That is, as the girls perceived their mothers as becoming more feminine, they tended to become more feminine and less mas- culine and as they perceived their fathers as becoming more mascu— line they also tended to become more feminine and less masculine. This is consistent with the modelling theory in that as the girls per- ceived their parents as becoming more appropriately sex-typed, and therefore as better models, they themselves became more appro- priately sex-typed. Here, all the results support hypothesis one. Also, as the correlations with the girls and their mothers tended to be slightly higher than those with the fathers, this may suggest a slight tendency for the girls to identify more with their mothers than with their fathers. This would support hypothesis two. 64 Thus, for the girls as well as the boys, the relationships between their own perceptions of what the parents do and their own sex-role identification are much clearer than the relationships between the parents! perceptions of what they, themselves, do and the children! s sex-role identification. Two explanations may be cited here: (1) The child! s perceptions of his parents may be more important to his sex-typing than the parents! perceptions as it is in terms of his own perceptions that he responds and (2) The PAS may be responded to differently by children than by parents. To test this hypothesis, additional study with the scale would be nec- essary. It should also be noted that the results of this section, while not reaching significance, are more clear-cut and consistent than those in preceding sections. Thus it might be that the child! s perceptions of his parents! overt behavior is a major factor in his sex-typing. As noted in the Results section, it was intended that the parents! Game Scale scores would be compared with their PAS scores. However, due to methodological difficulties in the admin- istration of the Game Scale, as mentioned in Part 2 of the Results, this comparison was omitted. 65 Part 5: Relationships between the CPI Fe and PAS The scores obtained by each parent on the CPI and PAS were correlated with each other to ascertain the relationship between the parents! covert and overt sex—role identification. It was pre- dicted that the two sets of scores would be positively related. This prediction was confirmed only for the mothers in which case three out of four of the correlations were in the predicted direction. Thus the mothers! overt sex-role identification increased as their covert identification increased. This would seem to suggest that either the mothers! sex-role attitudes influence their overt behavior or that there is a more basic underlying factor influencing both of them. Whichever the reason the results do imply that whether the child is picking up cues from the mother! s overt or covert behavior, there will be some consistency between the two sets of cues. For the fathers the situation was quite different with all the correlations being in the opposite direction. They tend to suggest that as the fathers became more covertly feminine, they tended to become more overtly masculine. However, these relationships were of a low magnitude. They may, though, suggest that covertly femi- nine fathers tend to compensate for this by becoming overtly more masculine. These results are inconsistent with the prediction stated above. 66 The relationships between the children! s responses to the PAS and the parents! CPI scores were also considered as the pre- vious section indicated that the child! s perception of the parentS! overt behavior was more important for his sex-role development than were the parents! perceptions. Here the results differed for boys and girls. For the boys, as the father became less appropriately sex-typed and the mother became more appropriately sex-typed on a covert level, the sons tended to perceive both parents as becoming more appropriately sex-typed overtly. Considering that the com- parisons between the fathers! own PAS and CPI suggested that as the father became more covertly feminine, they tended to become more overtly masculine, the son may be responding more to the more masculine overt behavior when he perceives the more covertly feminine fathers as more overtly masculine. If this is the case then these results suggest that as the models become more appropriately sex-typed in one way or the other, so do the sons! perceptions of them. Otherwise, if only the covert behavior is considered, the statement applies to mothers only. For the girls on the other hand, as the fathers became less appropriately sex-typed on a covert level, the daughters tended to perceive them as becoming less appropriately sex-typed in overt behavior. This would suggest that, for the girls, as the fathers 67 became better models they were seen as better models while as the mothers became better models they were actually seen as poorer models, at least on an overt level. For the boys, this trend was reversed when only the parents! covert identification is considered. Thus, for each child, the perception of the appropriateness of the sex-role identification of the parents was positively related to the parents! covert identification in the case of the opposite-sex parents and negatively related to the parents! covert identification in the case of the same-sex parents. This is inconsistent with the hypothesis that children are more similar to the same-sex parent if one considers that they tend to have the wrong image of that parent. Part 6: Some Additional Considerations and Criticisms It can be noticed that, although the major hypotheses received some support, throughout this study there are few signifi- cant results. Most of the correlations tend to be quite low. Also, in several sections it was demonstrated that a number of the results conflicted with each other. The lack of significant results and the occurrence of contradictory results may be due, in part, to two general considerations along with the more Specific ones discussed under individual sections. These two general considerations are the sample used and the instruments employed. 68 In regard to the sample, the subject number was quite small. This increases the size of the correlation needed for significance. It also makes the correlation more subject to the influence of indi- vidual variations so that only a few deviant cases are necessary to produce a marked change in the correlation. It had been hoped to include more subjects in this study but the time necessary for individual testing and the situation in which the testing was done did not permit this. However, any study of this type-to be carried out in the future should include more subjects. Also a greater vari— ance in the population would be desirable. The subjects in this study came from a rather restricted range: they were all middle- to upper-middle class Catholics. This tended to restrict the range of attitudes and actions sampled and thus cut down on the Spread of the scores. Therefore a larger-and more varied population would be desirable. Some of the problems encountered in this study may be directly related to the small sample size and restricted population. Concerning the instruments, the CPI and PAS yielded scores within very narrow ranges. This tended to cut down on the magni- tudes of the correlations. To counteract this, one of two courses should be taken: (1) scales with larger ranges should be used, (2) weighted scores which yield larger ranges should be devised. Another criticism of the instruments is that they were subject to 69 faking and defensiveness. At times questions were not answered with the choices provided so that objective scoring became difficult. Also, it is not known to what extent these factors influenced the accu- racy of these scores. Before one can be confident that his scores are reasonably accurate he should try to devise a system to mini- mize the faking and defensiveness. One final criticism is that, as has been noted, the PAS seems to need revision either in its compo- sition or instructions, especially for use with parents. These, then, are some general considerations that apply to the low and sometimes inconsistent results in all the sections of this paper. CHAPTER V CONC LU SION This study was designed to assess the influence of the par- ents! covert and overt sex-role identification on the sex-typing of the child. Previous studies, concentrating on the parents! covert identification, had shown that as the parents became more highly sex-typed, so did the children. However, the effects of the parents! overt identification on the sex-typing of the child has received little attention. For this reason, the Parent Activity Scale (PAS) was developed to assess the parents! overt sex-role behavior. The major hypothesis of this study predicted that as the parents became more appropriately sex-typed on both covert and overt levels, so would the child. Also, it was expected that the same-sex parent would be a more important influence on the child than the opposite— sex parent. The results of this study both support and refute these hypotheses. Regarding the parents! covert identification, there was a tendency for both the boys and girls to become more appropriately 70 ~'. 71 sex-typed as the same-sex parent did. Sex-appropriate attitudes on the part of the parents seemed to foster sex-appropriate behavior on the part of the child. This was especially so for the mother. This, then, supports the first hypothesis and is consistent with past research. There was also some indication that the same-sex parent was more important for the child! s identification than the opposite-sex parent, which supports hypothesis two. In relation to the parents! overt behavior, the results were contradictory. When the parents! perceptions of their overt behavior were considered, conflicting results were obtained. Several of the results that could be interpreted tended to refute-the hypothesis. That is, as the parents became more appropriately overtly sex- typed, the children tended to become less appropriately sex-typed. This might suggest that the parents! overt behavior had a negative influence on the child. Another explanation is that the PAS may need further refinement for use with parents. Thus the inconsistent results may be more a function of the instrument than of the parents! actual identification. The children! s perceptions of the parents! overt behavior were also compared to their own sex—typing. The results here differed markedly from those utilizing the parents! perceptions in that as the children perceived their parents as more appropriately 72 sex-typed on an overt basis, the children tended to become more sex-typed. This supports hypothesis one. It also suggests that the children! s perceptions are more important for their sex-typing than the parents! perceptions. There was also some support for hypothe- sis two in this section. Regarding the PAS, which was developed for this study, the results would imply that the scale needs further development for use with parents. However, the scale seems to be useful with children. In summary, although few of the results were significant, the trends were for the modelling theory to be supported. Some methodological criticisms of this study were men- tioned which should be taken into account in future research. That is, a larger and more varied sample would be desirable along with instruments yielding wider ranges of scores. REFERENCES REFERENCE S Bandura, A., Ross, Dorothea, and Ross, Sheila A. A comparative test of the status envy, social power, and the secondary- reinforcement theories of identificatory learning. J. abnorm. soc. Psychol. 1963, 6_7_, 527-534. Bandura, A. , and Walters, R. H. Social Learningfiand Personality Development. New York: Holt, Rinehart and Winston, Inc. , 1963. 'Bronfenbrenner, U. Freudian theories of identification and their derivatives. Child Developm. 1960, 31, 15-40. Freud, Anna. The E go and the Mechanisms of Defense. London: Hogarth Press, 1937. Hartly, R. E. Sex-role pressures and the socialization of the male child. Psychol. Repts. 1959, 5, 457-468. Heilbrun, A. B. , Jr. An empirical test of the modeling theory of sex-role learning. Child Develpm. 1965, 36(3), 789-799. Kagan, J. The concept of identification. Psychol. Rev. 1958, _6_5_, 296-305. Kagan, J. Acquisition and significance of sex typing and sex role identity. In Hoffman, M. L., and Hoffman, Lois W., Review of Child Development Research, Vol. I. New York: Russell Sage Foundation, 1964. Lynn, D. B. Learning masculine and feminine roles. Mar. and Fam. Living. 1963, 2_5, 103-105. Lynn, D. B., and Sawrey, W. L. The effects of father-absence on Norwegian boys and girls. J. abnorm. soc. Psychol. 1958, Q, 258-262. ' 73 74 Mowrer, O. H. Identification: a link between learning theory and psychotherapy. In Learning Theory and Personality Dynamics. New York: Ronald Press, 1950. Pp. 573-616. Mussen, P. H. Some antecedents and consequents of masculine sex— typing in adolescent boys. Psychol. Monogr. 1961, E, No. 2 (Whole No. 506). Mussen, P. H., and Distler, L. Masculinity, identification, and father-son relationships. J. abnorm. soc. Psychol. 1959, 53, 350-356. Mussen, P. H. , and Distler, L. Child rearing antecedents of mas- culine identification in kindergarten boys. Child Develpm. 1960, §_1_, 89-101. Mussen, P. H., and Rutherford, E. Parent-child relationships and parental personality in relation to young children! 8 sex- role preferences. Child Develpm. 1963, 34, 589-607. Payne, D. E., and Mussen, P. H. Parent-child relations and father identification among adolescent boys. J. abnorm. soc. Psychol. 1956, 5_2_, 358-362. Rau, Lucy. Parental antecedents of identification. Merrill-Palmer Quarterly. 1960, _6_, 77-83. Rosenberg, B. G., and Sutton-Smith, B. The measurement of mas- culinity and femininity in children. Child Develpm. 1959, 32, 37 3-380. Sears, R. R. ,1 Maccoby, Eleanor E., and Levin, H. Patterns of Child Rearing. Evanston, 111.: Row, Peterson, 1957. Sears, R. R., Rau, Lucy, and Alpert, R. Identification and Child Rearing. Stanford: Stanford University Press, 1965. APPENDICES APPENDIX I CHARACTERISTICS OF THE SAMPLE APPENDIX I CHARACTERISTICS OF THE SAMPLE I. OCCUPATION OF FATHERS Occupation Frequency Professional 10 Owner of Business 1 Managerial 1 Other White Collar 15 Skilled 1 _Unskilled 1 II. OCCUPATION OF MOTHERS Frequency Occupation Full Time Part Time Professional 1 1 White Collar 1 1 Housewife 25 0 III. NUMBER OF CHILDREN Number 123456789101112 Frequency 013834330001 75 VI. 76 POSITION OF SUBJECT Position Oldest In Between Ymflest Frequency: Boys 0 9 6 Girls 0 10 4 HAVE CHILDREARING PRACTICES CHANGED RECENTLY IN YOUR FAMILY ? Response Yes No Frequency 5 24 HAVE PATTERNS OF FAMILY LIVING OR ACTIVITIES CHANGED RECENTLY? Response Yes No Frequency 7 2 2 APPENDIX II LETTERS AND INSTRUMENTS USED 77 Dear Parents: I am a graduate student in psychology at Michigan State University. I am doing some research on identification in school children and would appreciate your cooperation very much. I am interested in seeing what influences parents! behavior has on that of their children. For this reason, I hope to interview as many first and second grade students as I can. Each child will be asked whether or not he plays several games and also some questions about daily household activities. Also, the parents of participating children will be sent three short questionnaires to fill out and return with their child. One of these questionnaires is a checklist of family activities, another concerns several attitude questions and a third pertains to some basic family information. This is an anonymous study in the hopes of getting accurate, candid information. If you decide to participate, you and your child will be given a number and will be referred to by that number. All information will remain confidential. Please fill out the bottom of this form and return it with your child. Thank you, Mrs . Myrna Lane I will allow my child to participate. 78 Dear Parents, Last Spring you volunteered to participate in a study about identification in school children. Your child has already been inter- viewed and you have received a number. As you were told at that time, this is an anonymous study and all information will be kept confidential. There are several forms which I would like you to fill out. The first of these, entitled Information Form, may be filled out by either parent. The second, the Game Checklist, may also be filled out by either parent. There are two forms of which there are two c0pies each, the Parent Activity Scale and the Attitude Questionnaire. Each parent is to fill out one of these. For example, the mother fills out the Parent Activity Scale labeled Mother and the father fills out the copy labeled Father. The same holds for the Attitude Questionnaire. Please fill out these forms separately. Do not consult each other. Instructions for each form are included on the forms. I would appreciate it if you would complete these forms and return them to school with your child by this coming Wednesday. Thank you for your cooperation. Mrs . Myrna Lane 79 INFORMATION FORM Father! 3 occupation Full time Part time Mother! 8 occupation Full time Part time Please list the ages and sexes of all the children in your family. Place a check next to the child who has brought this home. Age Sex Have child rearing practices changed recently in your family? If yes, please describe the changes. Have patterns of family living or activities changed recently? If yes, please describe the changes. plays . child plays . 80 GAME CHECKLIST The following is a list of games some of which your child Simon says soldiers dolls dressing up cops and robbers houses store school spacemen bows and arrows red rover fox and geese stoop tag ring around the rosey throw snowballs London Bridge darts farmer in the dell in and out the window drop the handkerchief huckle buckle beanstalk wrestling football . , mulberry bush jump rope hopscotch jacks shooting follow the leader using tools crack the whip hunting boxing Please put a check next to those games which you know your Please do not ask your child for help. leap frog name that tune musical chairs hide the thimble Mother may I make model airplanes statues scrapbook making dance toy trains bandits sewing cooking see saw doctors blind man! s buff cars marbles build forts I! ve Got a Secret puzzles cowboys Clue wall dodgeball pick up sticks King of the Mountain build snowmen bowling Spin the bottle I Spy soccer clay modeling 81 ATTITUDE QUESTIONNAIRE Following is a list of statements. If you agree with a statement, or feel that it is true about you, put a E on the line next to it for TRUE. If you disagree with a statement, or feel that it is not true about you, put an E on the line next to it for FALSE. I am very slow in making up my mind. I think I would like the work of a building contractor. I think I would like the work of a dress designer. I become quite irritated when I see someone spit on the sidewalk. It is hard for me to start a conversation with strangers. I must admit that I enjoy playing practical jokes on people. I get very tense and anxious when I think other people are dis- approving of me. A Windstorm terrifies me. I think I would like the work of a clerk in a large store. I get excited very easily. I like to boast about my achievements every now and then. I think I would like the work of a garage mechanic. I like adventure stories better than romantic stories. I prefer a shower to a bathtub. The average person is not able to appreciate art and music very well. The thought of being in an automobile accident is very frightening to me. 82 At times I feel like picking a fist fight with someone. Sometimes I have the same dream over and over. I think I am stricter about right and wrong than most people. I think I would like to drive a racing car. I like to be with a crowd who play jokes on one another. I am somewhat afraid of the dark, I think I could do better than most of the present politicians if I were in office. I always tried to make the best school grades that I could. I am inclined to take things hard. I would like to be a soldier. I liketo go to parties and other affairs where there is lots of loud fun. I very much like hunting. In school I was sometimes sent to the principal for cutting up. I think I would like the work of a librarian. Sometimes I feel that I am about to go to pieces. I would like to be a nurse. If I were a reporter I would like very much to report news of the theater. I like mechanics magazines. I want to be an important person in the community. I must admit I feel sort of scared when I move to a strange place. 83 I! m pretty sure I know how we can settle the international prob- lems we face today. If I get too much change in a store, I always give it back. I‘HH life. _F_ on the line next to it. 84 PARENT ACTIVITY SCALE The following is a list of activities concerned with family If you think that the activity is performed by the father more often in your home, please put an Please read each item carefully. If you think it is performed by the mother more often in you rh,ome please put an M on the line next to it. If it is an activity which is usually equally divided between you and your spouse, please put a _B_ on the line next to it for both. ||||||||||||||||||||||||||||||| answers telephone calls relatives on phone changes light bulbs cleans table after meals cooks meals decides on menues decides on T.V. programs does the laundry drives car dusts expresses affection expresses hostility feeds the dog (or other pet) feeds young children fixes car gets the mail goes to ball games goes to gas station goes to hardware store goes to PTA meetings hangs up clothes host or hostess at parties host or hostess to relatives involved in girl/boy scouts irons things listens to radio more listens to records more makes decisions about going out makes the bed mends clothes more athletic more physically active mows the lawn opens cans opens doors paints house pays bills plays ball with children plays cards (in group) plays with children indoors plays with children outdoors praises children prepares snacks puts children to bed rakes leaves reads nSWSpaper more reads novels more reads to children repairs appliances repairs house scolds children sets the table shops for clothes shops for food Shovels snow swears more sweeps the floor takes the children on picnics takes the children to the movies air-L; .r. .— _— 85 takes children to school throws out garbage walks dog washes clothes washes dishes washes the car watches T. V. works outside the home works to earn money works with tools vacuums the carpet APPENDIX III ASSIGNMENT OF PAS ITEMS TO CATEGORIES (Dm'qmmI-pWNH 00000.... APPENDIX III ASSIGNMENT OF PAS ITEMS TO CATEGORIES Item answers telephone calls relatives on phone changes light bulbs cleans table after meals cooks meals decides on menues decides on T.V. programs does the laundry drives the car dusts expresses affection expresses hostility feeds the dog (or other pet) feeds young children fixes car gets the mail goes to ball games goes to gas station goes to hardware store goesto PTA meetings . hangs up clothes . host or hostess at parties . host or hostess to relatives involved in girl/boy scouts irons things listens to radio more listens to records more . makes decisions about going out makes the bed . mends clothes 86 Number of Responses Mother 89* 80* 17 96* 97* 87* 15 98* 5 99* 32 19 64* 88* 0 61* 2 5 5 62* 77* 27 29 46* 100* 39 38 16 97* 97* Father 2 ll 70* 0 0 2 33 1 52* 0 8 31 Both 9 9 13 4 3 11 52* 1 42 1 60* 50* 13 11 3 25 23 30 10 35 21 72* 71* 38 0 32 45* 54* 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. Item more athletic more physically active mows the lawn opens cans opens doors paints house pays bills plays ball with children plays cards (in group) plays with children indoors plays with children outdoors praises children prepares snacks puts children to bed rakes leaves reads newspaper more reads novels more reads to children repairs appliances repairs house scolds children sets the table shops for clothes Shops for food shovels snow swears more sweeps the floor takes children on picnics takes children to the movies takes children to school throws out the garbage walks dog washes clothes washes dishes washes the car watches T.V. works outside the house works to earn money works with tools vacuums the carpet 87 Number of Responses Mother 46* 32 72* 70* 47* 67* 27 98* 88* 81* 2 8 100* 17 37 63* 34 15 97* 95* 4 14 2 1 1 94* Father 87* 54* 93* 15 54* 80* 54* 80* 18 15 67* 7 5 2 74* 36 14 4 92* 90* 15 89* 77* 10 14 45* 46* 90* 24 69* 77* 91* Both 11 38 5 15 38 13 23 19 65* 39 30 61* 23 28 22 61* 39 27 5 8 58* 1 12 16 9 12 0 75* 52* 23 21 30 0 4 5 62* 27 21 5 3 *Indicates significance at the . 01 level APPENDIX IV WEIGHTED SCORES FOR GAME SCALE ITEMS APPENDIX IV WEIGHTED SCORES OF GAME SCALE ITEMS* Boys Girls Weight 3 soldiers dolls cops and robbers dressing up spacemen houses bows and arrows store throw snowballs school darts red rover wrestling fox and geese football stoop tag boxing ring around the rosey shooting London Bridge using tools farmer in the dell hunting in and out the window make model airplanes toy trains drop the handkerchief huckle buckle beanstalk mulberry bush jump rope hopscotch jacks follow the leader crack the whip leap frog Simon says *From Rosenberg and Sutton-Smith, 1959 88 Boys Girls Weight 3 name that tune musical chairs hide the thimble Mother, may I statues scrapbook making dance sewing cooking see saw cartwheels Weight 2 bandits cars marbles build forts doctors blind man! 3 buff I've got a secret puzzles Clue pick up sticks build snowmen Weight 1 cowboys wall dodgeball King of the mountain bowling soccer racing Spin the bottle I Spy clay modeling "IIIIIIIIIIIII