w m H LEE .1 L \ WI 1 \ MI W -—Lb_x Imo il 'moooo \ AGE AND.- SEX TREKDS 5N ASE LJEMEE‘ STE/{I [EA EETY TE?- TE‘ E EAR/LE SE2 EAEENI AS A *~LJ[~{"T*Q“% Gift-x; T {‘E §-§\ ELCPE EN'L‘V ’7’: '8ng L: I" Thesis Eav fins Dogma of D5. D. MECHEGAN STATE UNIVERSETY R E‘E‘GSE BLHLLL E967 L ~. L— Luau n v ""i Michigan Stan. University THESIS A I D L; This is to certify that the thesis entitled AGE AND SEX TRENDS IN ASSUMED SIMILARITY TO THE SAME SEX PARENT AS A FUNCTION OF THE DEVELOPMENT OF IDENTITY presented by Ernest Bruni has been accepted towards fulfillment of the requirements for Ph. D degree in Psychology 8466” (WI Major professor Date January 3, 1967 0-169 ABSTRACT AGE AND SEX TRENDS IN ASSUMED SIMILARITY TO THE SAME SEX PARENT AS A FUNCTION OF THE DEVELOPMENT OF IDENTITY by Ernest Bruni This study was an attempt to demonstrate that assumed similarity to the same sex parent varies with both age and sex. More specifically, it was proposed that assumed simi- larity is affected by identity diffusion and would conse- quently reflect age and sex trends in the development of identity. Seven age groups from nine to twenty-one years of age with 30 males and 30 females in each group were adminis- tered a form of the semantic differential. Subjects rated themselves and their parents, and a similarity score between self and the same sex parent was derived. It was eXpected that first born children would'per- ceive themselves more similar to their parents than would non-first born children in agreement with an earlier study. No consistent pattern was found between first born and non- first born children and the prediction was not confirmed. Ernest Bruni Since identity in both males and females is thought to be more diffuse during middle adolescence than at any other time in life, it was predicted that both males and females would see themselves less similar to their same sex parent at this time than would younger or older subjects. Fifteen and 17 year old males do perceive less similarity between themselves and their fathers than do other males, but the prediction was not confirmed for females. The trend may be curved or spiked for males, but appears not to vary consistently for females. It was predicted that late adolescent females would perceive themselves less similar to their mothers than males see themselves similar to their fathers, although the sex difference was not expected to emerge before late ado- lescence. This hypothesis is based on the notion that fe— male identity remains more diffuse because it involves more choices and is in general more complex and difficult to achieve in middle class American culture than is male iden- tity. The prediction was not confirmed, but the mean dif- ference scores of 6 of the seven age groups studied tend to support the hypothesis. The results give some support to the hypothesis that there is an association between perceived similarity to the same sex parent and trends in the development of identity. Implications of assumed similarity trends with regard to age Ernest Bruni and sex differences were discussed and special attention was given to the particular sample as a factor in the inter- pretation of results. Specific questions that warrant further inquiry were suggested. [3) M f KW" Committee Chairman Date: /—'{Ci\é7 AGE AND SEX TRENDS IN ASSUMED SIMILARITY TO THE SAME SEX PARENT AS A FUNCTION OF THE DEVELOPMENT OF IDENTITY By Ernest Bruni A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Psychology 1967 DEDICATED To Dianne ACKNOWLEDGMENTS The author wishes to express sincere gratitude to his committee chairman, Dr. Bill Kell, and to his committee members, Dr. William Mueller, Dr. Norman Abeles and Dr. Dozier Thornton, not only for valuable suggestions and criticism, but particularly for the manner of their participation that helped make the task more pleasant and meaningful. The author wishes to express further thanks to Dr. Bill Kell for the major role that he played in the progress and con- structive change experienced by the author during the course of his graduate study. The staff members of the Okemos, Michigan, school system were most cooperative and the author is indebted for their helpfulness. Appreciation is expressed to Harry Groulx, Howard Splete, John Waldo, Ed Kiley, Ed Banfield, Warren Dolsen, and the many teachers and pupils who were involved in the study. Thanks is expressed to Dr. Terrence Allen for his assistance in analysis of the data, and to Gerald Musgrave, Richard Fremon, and David Kline whose help with computer pro- grams and computer laboratory routines was most valuable. And to my wife, Dianne, for giving her time and sharing both the many hours of tedious work and the final accomplishment, a very special thanks. iii TABLE OF CONTENTS Page ACKNOWLEDGMENTS . . . . . . . . . . . . . . . . . . iii LIST OF TABLES . . . . . . . . . . . . . . . . . . v LIST OF FIGURES . . . . . . . . . . . . . . . . . . vi LIST OF APPENDICES . . . . . . . . . . . . . . . . vii INTRODUCTION . . . . . . . . . . . . . . . . . . . 1 Review of the Literature . . . . . . . . . . . l The Problem . . . . . . . . . . . . . . . . . 11 METHOD . . . . . . . . . . . . . . . . . . . . . . 13 Subjects . . . . . .y. . . . . . . . . . . . . l3 Instrument . . . . . . . . ... . . . . . . . . 15 Procedure . . . . . . . . . . . . . . . . . . 17 RESULTS . . . . . . . . . . . . . . . . . . . . . . 19 DISCUSSION . . . . . . . . . . . . . . . . . . . . 25 Research Implications . . . . . . . . . . . . 30 SUMMARY . . . . . . . . . . . . . . . . . . . . . . 33 REFERENCES CITED . . . . . . . . . . . . . . . . . 35- RELATED BIBLIOGRAPHY . . . . . . . . . . . . . . . 38 APPENDICES . . . . . . . . . . . . . . . . . . . . 42 iv Table LIST OF TABLES Summary of the analysis of variance of similarity scores with respect to age, sex, and birth order . . . . . . . . Summary of the analysis of variance of similarity scores across age samples for males . . . . . . . . . . . . Summary of the analysis of variance of similarity scores across age samples for females . . . . . . . . . . . . . Summary of the analysis of variance of similarity scores for preadolescents, middle adolescents, and late adoles— cents for male and female subjects . Mean differences of similarity scores between males and females in seven age samples . . . . . . . . . . . Page 19 21 21 22 24 LIST OF FIGURES Figure Page 1. Mean similarity scores for each age group with males and females graphed separately . . . . . . . . . . . . . . 20 2. Mean similarity scores for three age samples representing preadolescent, middle adolescent, and late adoles- cent males and females . . . . . . . . 23 vi LIST OF APPENDICES Appendix Page A. Questionnaire used with subjects from fourth grade through high school . . . . 42 B. Questionnaire used with college subjects . 45 C. Stimulus concepts and rating scales . . . 48 D. Instructions and semantic differential . . 50 vii INTRODUCTION Review of the Literature Modeling or imitation has long been recognized as an effective mode of learning new behaviors and attitudes. Although modeling is sometimes conceived simply as learning selected segments of behavior, some relationships foster the modeling process by more intense interpersonal involve- ment at least on the part of the imitating person. He may consciously wish to be like the model and he will tend to perceive himself like the model. Kagan (196A) emphasized the perception of similarity as a motive for modeling when a parent is the model, but Freud implied that it was a more general characteristic. Commenting on identification in this general sense, he wrote that "It (identification) may arise with every new perception of a common quality shared with some other person who is not an object of the sexual instinct" (1922, p. 65). The perception of similarity to a model is a behavioral manifestation of the modeling process. It is usually conceded that parents are the most significant early models and that modeling after one's parent, particularly the same sex parent, is important in psychological development. Researchers interested in parental modeling have sometimes focused on perceived similarity in the belief that it reflects a process between the child and his parent which is thought to occur quite early in develop- ment--that is, the modeling of the child after his own par- ticular parent. It has been demonstrated, however, that models change with age. Winker (19A9) and Havighurst, Robinson and Dorr (1946) in separate studies asked youngsters whom they would "like to be like." Although the age sequence was not rigid, the general choice moved outward from the family circle. Parents, glamorous adults, attractive and visible young adults and composite imaginary persons were chosen as objects as age increased from childhood to late adolescence. Bronfenbrenner (1958) pointed out that a child may actually be modeling himself after one of these latter figures and the fortuitous result may be increased, or decreased, assumed similarity to his parent. This progression of models suggests that perceived similarity to the presumed first (parent) model may be over- shadowed or perhaps in some other way affected by later modeling. Since models change with age, assumed similarity to the parent may also vary with age. Most research has not attended to age as a variable. One reason for this is that it has been assumed that later modeling would not negate the perception of similarity to a parent because it is dependent upon and correlated with parental modeling. That is, parental modeling precedes all other modeling, and the readiness with which a child accepts these non-parental models and incorporates the cultural sex role will be determined by the extent to which he has learned to model himself after his parent. Following this line of reasoning, Kagan (196A) commented that the child, in modeling after a parent, learns also how to "identify," that is, learns how to become similar to a model and what models are appropriate. But this approach to the integration of early and later modeling in the achievement of identity is oversimpli— fied. The process not only fosters similarity with the model, but the child must recognize differences as well. Seward (195A) observed that this is necessitated partly by age-role discontinuity. She says, The three-year—old boy cannot identify with his 30 year old father. A more reasonable interpretation would be . . . that in an atmosphere of parental acceptance the little boy can accept his own masculinity, that is, the social sex role of a three year old boy (195A, p. 231). What is suggested here is that the result of modeling is both a sense of similarity and difference, and that these together lead to a "sense of who one is" in terms of a social role or a sense of individual identity. This sense of identity becomes increasingly important as new models are provided in society because each new model or new as- pired role requires a re-definition of the self. Assumed similarity is directly related to this sense of individual identity. The judgment of similarity-— or difference--made by an individual has implicit two separ- ate kinds of judgments; one is how the individual defines or identifies himself and the other refers to his definition of his parent. Whatever social and intrapsychic factors affect individual identity also affect the perception of similarity. Individual identity is distilled from a variety of social experiences. Modeling, as noted above, is most important. There are other ways that peOple learn to de- fine themselves; by interacting with others they come to know their distinctive abilities and limitations, i.e., how they are defined by others. The end product is a more or less integrated definition of self shared by an individu- al and significant others. We are interested here in variations, particularly age trends in identity formation. In his discussion of ego identity, Erikson (1950, 1959) introduced the concept of identity diffusion as a significant variable in identity development. As with many other concepts denoting develOp- mental events, he found it most useful to characterize it in terms of its pathological forms. Identity diffusion is the lack of clear individual identity; there is confusion and uncertainty in an individual's sense of who he is. But identity diffusion is not necessarily a patho- logical characteristic. Indeed, it is esentially an exagger- ated form of the identity crisis of normal adolescence. Dignan (1964) described diffusion as a sense of aimlessness, futility, feeling of boredom and lack of purpose. It is frequently marked by conscious doubts about sex roles and vocational choice. These feelings in milder form during normal ado- lescence are the result of both social and intrapsychic factors. On the one hand, the adolescent's sense of identity is disturbed by the experiences of growth changes and the advent of sexual maturity with their attendant cultural de- mands. Whereas the pre-pubertal child has learned to be relatively comfortable in a period of relative quiescence, the adolescent must adjust to changes within himself. Social reaction to the "gangly" adolescent is not lacking either; he is generally defined as somewhat peculiar--and whatever ails him will pass when he emerges from "this stage." Soc- ietal expectations change when a child reaches adolescence; he is expected to become aware of his future role and must begin to make vocational commitments. He is urged with in- creased insistence to define himself publicly by declaring what he will be. Conversely, the adolescent is also defined in this culture as someone who is nobody yet. Erikson called this a period of moratorium, that is, a period of psycho-social suspension. The various aspects of identity and their related conflicts are present in every stage of psychological growth from birth to old age, but they are particularly crucial during adolescence. At this stage there are converging psy- chosocial pressures that demand attention: (1) physical changes; (2) changes in reactions from others; (3) insistence that new responsibilities and future plans be declared, and that, since this is a period of moratorium; (A) decisions and roles are on trial basis only and not to be taken very seriously. Thus it is during adolescence more than at any other period of life that the individual is called upon to integrate a variety of conflicts, roles, model images, abil- ities, and aspirations into a unitary sense of self. During this period in normal deveIOpment, identity is more diffuse than at any other time. We may conclude that judgments about oneself will be more difficult at this time. Inconsistencies within himself and in society must result in a less clear sense of similarity or difference from others, but the cultural consensus would emphasize that the adolescent define himself mostly "different." We expect this feeling to be reflected in assumed similarity measurement. We may also find that the conflicts of identity resolution may provide significant insight into the dif- ferences between sexes with regard to the certainty and clarity of a sense of self. Erikson suggested that the inability to make a vocational choice is a significant in- dicator of identity diffusion. Galinsky and Fast (1966) found concurring evidence in case studies. Douvan and Adelson (1966) point out, however, that while boys seem to emphasize vocational choice as an aspect of ego identity, girls do not-—at least not as frequently. It is important to note that a sense of identity is derived from the role that is ascribed by society. Who a child is to be will be influenced and in some cases determined by what the environ- ment permits and encourages. Douvan and Adelson observe, Our culture's expectations for the girl are less simple than for the boy; they are both more ambiguous and less consistent, perhaps because of a recognition of the complexities of feminine development (1966, p. 33). The result is that, Girls tend to keep identity diffuse and misty. The boy is made to feel (however much he may doubt it, deep down) that his identity is in his own hands, that the choice of vocation, and with it, of a life style, will define him. The girl cannot count on this degree of active preferment in identity; her identity is bound up not so much in what she is as in what her husband will be (Douvan and Adelson, 1966, p. 18). The emphasis throughout identity theory is that a person's sense of self, particularly during the crisis phase, is very largely dependent upon the sense of the future. For a girl the future is necessarily hazy, and identity formation must remain so also. Since a person's identity is derived primarily from interpersonal definitions of himself, it follows that diffi- culty in identity is the result of inadequacies in social role definitions. Cultural discontinuities are perhaps the single most important contributing factor to identity dif- fusion. If role diffusion is significantly related to the judgments in assumed similarity measures, then a systematic factor other than parental modeling will be reflected in the scores. Assumed similarity research has generally focused on the relationship between similarity and some other tested variable, for example, pathology or anxiety level. There has been no systematic empirical study of variations in similarity related to age or other deve10pmenta1 stage con- cepts. There are two studies, however, that report inci- dental data that show trends of similarity between 10 and 14 years of age. Using a technique similar to the adjective checklist, Brodbeck (195A) showed that for both boys and girls, there is no significant change in perceived similarity to the same sex parent. Gray (1959) studied grades 5 through 8 using the semantic differential to measure assumed similar- ity. She found that there was a very slight tendency for similarity to the same sex parent to increase. The trend of scores reported by these authors was not supported in a study by Bruni (1965). Using the semantic differential, three age groups were compared; 11-12 year olds, 15-16, and 21-22 year olds. Assumed similarity was markedly decreased for the middle group. That is, a curvilinear relationship was suggested with a satisfactory degree of reliability (p. < .025) for the male subjects, although the trend was not significant for the females. The disagreement of the latter study with the former two is most likely due to the broader age range sampled. Although there has been little interest in study- ing age trends in assumed similarity, this is not the case with sex differences. In general, there is a relatively high consensus that among college student subjects, males tend to see themselves more similar to their fathers than females see themselves like their mothers. Concurring data are reported by Lazowick (1955), Shell e£_al. (196A), Heil— brun (1965), Bieri gt_al. (1959), and Dyal (in Osgood gt_al., 1957). One study only, by Gray and Klaus (1956), showed the Opposite pattern of female subjects seeing themselves more similar; the reason for contradiction in this study is not apparent. Studies utilizing younger subjects, however, do not agree with the consensus of these data. Gray (1959) in the study already cited, observed that preadolescent girls generally see themselves more like their mothers than boys see themselves like their fathers. This trend, however, was not statistically significant. But in a similar study by Carlson (1963), in which she tested sixth graders, simi- larity (she used a real similarity measure) of girls with mothers was greater than boys with fathers. Carlson sug— gested that this trend was due to a difference in her 10 similarity measure. She claimed that the scores denoted specifically "developmental identification"; also, she felt that biases are reflected in other similarity studies that were controlled in her experiment; eg., social desirability and a tendency to be biased by favoring male stereotype concepts. In the study by Bruni (1965), these age by sex variations were generally supported. Among the younger subjects (11-12 and 15-16 year olds), sex differences in similarity to the same sex parent were not significant, but college males saw themselves more similar than college fe- males. The interpretation offered here is that the differ- ence is not due to a variation in measures as suggested by Carlson (1963), but that the tendency toward greater iden- tity diffusion among females in late adolescence and early adulthood is the significant variable. Only one study has been reported in which simi- larity scores have been related directly to ego identity. Dignan (196A) developed an inventory to measure ego identity based on Erikson's conceptualization. Using freshman and sophomore college women, she found that assumed similarity to mother as measured by the semantic differential, was positively related to scores on her ego identity scale. However, in that study, she assumed that the two measures were independent and denoted independent constructs. Her data support the contention we propose here, although we 11 would interpret the results as due to interdependence of the constructs she measures. The Problem If the rationale presented here is accurately de- scriptive of the processes reflected in assumed similarity, then adolescents, who are experiencing greater identity dif- fusion, will perceive themselves less like their parent models than either younger or older individuals. In addition to this general expectation for both sexes, after the period of identity crisis, young women will see themselves less like their parent models than will young men, because identity integration is more difficult for them. This difference should not emerge during childhood when identity is not as urgently linked to a clear sense of the future, and it may decrease after early adulthood when female identity may be- come as stable as male identity. These general hypotheses are to be investigated in the present study. The data cited above by Bruni (1965) were sugges- tive that such an age by sex interaction in assumed similarity to the same sex parent does exist. The results were incon- clusive, however, because the sample was too small and the age range insufficiently broad. This study is an extension of that research. Two major predictions are investigated. Hypothesis I: Hypothesis II: 12 Middle adolescents see themselves less like their same sex parent than do either younger or older subjects. A general curvilinear relationship is predicted for both sexes. From childhood through middle adolescence there is no significant difference between male and female subjects in assumed similarity to the same sex parent; from this period on, however, males see themselves more like their fathers than females see themselves like their mothers. METHOD Subjects The total sample consisted of 420 subjects. There were 7 age-grade groups: Nth, 6th, 83h, 10th, and 12th grade children and two groups of late adolescents, one com- posed of college freshmen (or lower division students), and the other was composed of college juniors and seniors (upper division students). In each age group there were 60 subjects that consisted of 15 first born males, 15 first born females, 15 non—first born males, and 15 non-first born females. To determine the apprOpriate age for the youngest subjects to be included it was necessary to consider two conditions: (1) the youngest subjects must be below adoles- cent age; and (2) they must be able to use the test instru- ment as reliably as the older subjects in the study. The youngest children used in previous similarity studies were eleven year olds; these youngsters are perhaps sociologically too close to early adolescence and may already be influenced by the adolescent identity image. Regarding the second con- dition, Lilly (1966) demonstrated that fourth grade children could comprehend the instructions of the semantic differential and that re-test reliabilities were comparable to those of 13 1“ adults. Fourth grade children (about nine years old), there— fore, appear to be the appropriate group for the youngest age subjects in the range to be studied here. Clarity of sex role was considered a relevant control variable because it may be different for lower class subjects as compared to middle class subjects, and this may have bearing on assumed similarity scores. An estimate of social class position was determined on the basis of educa- tion and employment of father, and only middle class subjects were used in order to emphasize homogeneity in the sample. For the youngest children, information about father was taken from school records; for other subjects, information was sup- plied in a questionnaire by the subject himself (see Appen- dices A and B). The criteria outlined by Hollingshead and Redlich (1958) were used for classification, and only class II and III subjects were included. Palmer (1966) found that first born children see themselves more similar to the same sex parent than do non- first born children. Birth order as a possible confounding variable was controlled in the present study by including an equal number of first born and non-first born subjects in each age—sex category. Only subjects who had lived with both parents to the time of testing or through high school were used. Fourth,6th,8th,10th,and 12th grade subjects were testedin their classrooms in a small, predominantly middle class residential community. (Okemos, Michigan, has two 15 elementary schools, one junior high and one senior high school.) In each grade, classrooms were arbitrarily drawn until a sufficiently large sample was obtained so that each age-sex-birth order category contained the pre-determined number of middle class subjects. Out of the pool of quali- fied subjects tested at each age level, subjects were drawn at random until each group was completed. The college stu- dents were tested at Michigan State University; the freshmen were in an introductory psychology class, and the junior and senior students were tested in two upper division courses in abnormal psychology. Instrument Each subject rated himself and his parents as he perceived them on a series of 7—point scales, i.e., a form of the semantic differential. The index of perceived simi- larity between a subject and his same sex parent is Osgood's D statistic--which is a measure of the similarity between the two sets of ratings or descriptions given by the subject. The technique was introduced by Osgood et_al. (1957) and it is described in detail by Beitner (1961). Q is derived in the following manner: The difference between the two ratings on each scale is squared, and the square root of these differences summed over all scales is the raw similar— ity score. 16 The 2 statistic is a very stable measurement. Norman (1959) reported a test—retest coefficient of .97. Nine concepts were chosen as either relevant to modeling after parents or other persons or roles. Six con- cepts were included for later analysis and three concepts were used for this study. They were, Me, My Father, and My Mother. The complete list of concepts is given in Appendix C. The only score used in the present study, how- ever, was the similarity between Mg and the same sex parent. Eighteen bipolar scales were chosen from those used in previous studies by Lazowick (1955), Beitner (1961), Osgood gt_al. (1957), and Mueller (1965). Only those scales were included that appeared to be comprehensible to most fourth graders. Although the Q statistic is not dependent upon the scale factors commonly reported in semantic dif- ferential research, an equal number of scales was chosen 'to represent each factor: Evaluation, Potency, and Activity. The list of scales is reported in Appendix C. Scales were presented in a random order for each concept and the direction of the positive and negative pole of each scale was also left to a random process. The three concepts used in this study (Mg, Father, and Mother) were never presented first, so that subjects could become ac- quainted with the procedure and practice at least one other concept before the test concepts for this study were presented. 17 The order of presentation was left to a random process with this one exception. A sample test booklet is given in Ap- pendix D. Procedure The semantic differential was administered to 27 separate classes. It was introduced as part of a research study on the meaning of certain words, but no further ex- planation was given other than the comments on the instruc- tions page of the questionnaire itself. For grades 4, 6, and 8, the instructions were given orally in addition to the written instructions in the text booklet (see Appendix D). The students read the instructions while the examiner read or recited identical instructions and c0pied the il- lustrations on the blackboard. Fourth and sixth graders ocasionally asked the meaning of some of the scales. They were told first to use the words with whatever meaning made the most sense to them. If a child insisted that he still did not know how to pro- ceed, he was told that the word meant "about the Opposite of that word," referring to the other pole of the scale. They were then encouraged to make guesses. This was usually adequate, although a few children were observed using the neutral rating consistently on such difficult scales. Few children expressed such excessive difficulty with the test. Since the majority of youngsters were able to ascribe meaning 18 to the scales and concepts without hesitation, it is the judgment of the Experimenter that the test was suitable for the fourth grade children. All of the subjects completed the test in class, with the exception of one of the two classes of college juniors and seniors. In that class, subjects took the tests with them, and returned them completed in the following class period. t-tests for both males and females showed that there was no significant difference between subjects who completed the test in class as contrasted with those who completed the test out of class. RESULTS An analysis of variance was done to evaluate the effect of age, sex, and birth order on perceived similarity to the same sex parent. The summary of the analysis is re- ported in Table 1. It can be seen from the table that the control variable, birth order, was not related to the depen— dent variable, and it was therefore drOpped from further analyses. Table 1. Summary of the analysis of variance of similarity scores with respect to age, sex, and birth order. Mean Sum of Source Squares df Squares F P Age 62.06 6 10.34 2.50 .025 Sex .02 l .02 Birth Order 2.44 l 2.44 Age x Sex 60.54 6 10.09 2.44 .025 Age x Birth Order 22.12 6 3.69 Sex x Birth Order 1.14 l 1.14 Age x Sex x Birth Order 26.39 6 4.40. Error 1,622.76 392 4.14 TOtal 1,797.47 “19 In Hypothesis I, it was predicted that similarity scores would be curvilinear with respect to age, and that the peak of large differences would occur during middle ado- lescence for both males and females. Mean similarity scores 19 20 do vary significantly with age, but interaction between age and sex (see Table 1) indicates that the trend is not iden- tical for both sexes. The data are reported graphically in Figure l where it can be seen that the trend is curvilinear for males only. Figure 1. Mean similarity scores for each age group with males and females graphed separately. .80- .40« .00- Females .60- .20- .80- U‘IUTOONNN .40- Mean Similarity Scores 5.001 T I ' I —T Grade 4 6 8 10 12 l3 16 Mean Age 9.49 11.47 13.58 15.55 17.52 18.50 21.06 Analysis of variance for males and females separ- ately shows that the age variations for males is significant (p. < .001), while that for females is not (p. < .20). Sum- maries of those analyses are reported in Tables 2 and 3. The hypothesis is strongly supported for males,but not for females. In Hypothesis II it was predicted that there would be no differences between males and females in assumed simi- larity to the same sex parent from childhood through middle 21 Table 2. Summary of the analysis of variance of similarity scores across age samples for males. Sum of Mean Source Squares df Squares F P Age 92.94 6 15.49 3.39 .001 Error 928.65 203 4.57 Total 1,021.59 209 Table 3. Summary of the analysis of variance of similarity scores across age samples for females. Sum of Mean Source Squares df Squares F P Age 29.66 6 4.94 1.35 .20 Error 746.20 203 3.68 Total 775.86 209 adolescence, and that from this period on, males see them- selves more like their fathers than females see themselves similar to their mothers. With reference to the data, this is a prediction of age by sex interaction such that there would not be significant sex differences in the first five age groups, but that there would be differences in the last two. The hypothesis was tested by an age by sex analysis of variance in which the last two age groups were combined to represent a later adolescent sample, the lOth and l2£h grade groups were combined to represent middle adolescents, and the 6th and 83h grade groups together represented 22 preadolescents. The first group in the total sample, 4th grade, was dropped from the analysis to allow equal sample size in each group. A summary of the analysis is reported below in Table 4. Table 4. Summary of the analysis of variance of similarity scores for preadolescents, middle adolescents, and late adolescents for male and female subjects. Sum of df Mean Source Squares Squares F P Age 56.79 2 28.40 6.86 .005 Sex .13 l .13 Age x Sex 21.36 2 10.68 2.58 .10. Error 1,464.67 354 4.14, Total 1,542.96 359 There is an age by sex interaction, but it is not statistically significant (p. < .10). When the data are presented graphically (see Figure 2), however, it is evident that the data approach the trend predicted. In the preado- lescent and middle adolescent groups, the mean sex differen- ces are .28 and .47 respectively; these differences are not statistically significant. The difference between males and. females in the last group is in the predicted direction (males perceive themselves more similar to the same sex parent than do females), and the difference is reliable (Mean difference = .64; p. < .05). The obtained age by sex interaction is not statis- tically significant and consequently the hypothesis is not substantiated. The trend in the data, however, warrants a 23 Figure 2. Mean similarity scores for three age samples representing preadolescent, middle adolescent, and late adolescent males and females. 8 § 7.40m m 7.004 >: :3. 6.60- fa H 6020.1 2 g 5.80- % 5.404 . . 1 2 Grade 6 & 8 10 & 12 13 & 16 more detailed analysis in order to isolate the source of the interaction that is present. In the initial analysis of variance reported above in Table l, the age by sex interaction was more pronounced than that discussed in the preceding jparagraph. Inspection of the data from all seven groups (see Figure 1) suggests that the sex difference in the upper division college stu- dent group is the largest, and is the major contributor to the interaction. Tests of significance of sex differences reported in Table 5 support this observation. College Junior and senior males preceive themselves significantly more like their fathers than females in the same group See themselves like their mothers (p. < .005), whereas there is no signifi- cant difference between males and females in any of the other groups. It was predicted that sex differences would not be significant in the first five groups and this prediction was 24 Table 5. Mean differences of similarity scores between males and females in seven age samples. Grade Males Females Diff. t P Fourth Mean 6.21 6.35 .14 .27 .801 S.D. 2.12 2.03 Sixth Mean 6.77 5.88 .89 1.70 .101 S.D. 2.21 1.61 Eighth Mean 6.25 6.57 .32 .61 .601 S.D. 1.55 1.60 Tenth Mean 7.54 6.75 .79 1.50 .201 S.D. 2.79 2.45 Twelfth Mean 7.15 6.99 .16 .29 .801 S.D. .2.14 1.94 College Mean 6.41 6.15 .26 .50 .352 Freshmen S.D. 2.21 1.72 College Mean 5.33 6.88 1.55 2.94 .0052 Jrsw-Seniors S.D. 1.73 1.94 lTwo-tailed tests. 2One—tailed tests. supported. It was further predicted that there would be a difference between sexes in the two college groups, and this difference did emerge in the upper division group. It appears that the college freshmen in particular did not conform to the prediction and this resulted in failure to confirm the hypothe- sis. DISCUSSION A number of factors must be considered in evalu- ating the failure to replicate Palmer's (1966) finding that first born children perceive themselves more similar to their same sex parent than do non-first born. The small sample size (15 Ss per group) in each age-sex-birth order category may have been insufficient to counter the effect of high between subject variability. Further, Palmer's difference scores were based on test items that were re- stricted to perceived similarity with regard to attitudes about inhibitory demands of parents and discipline, whereas the instrument used here measured a general factor of per- ceived affective distance. The subjects in the former study were paid volunteers, but this sample was composed of middle class students who took the test in class as they would do any other assignment. Any or all of these differ- ences may be important variables with regard to the similar- ity score patterns. First born children may not perceive themselves more similar to their parents in a general sense as Palmer implied, but the trend may be true if perceived similarity is measured in specific content areas or in a specific sub-group sample. 25 26 The curvilinearity of the trend of male scores does support the thesis presented here that scores vary as expected on the basis of identity theory. The data show that middle adolescent males see themselves less like their fathers than do either younger or older males. This suggests that the in- crease in a sense of dissimilarity reflects and coincides with an increase in diffuseness of identity during the same age period. The lack of curvature in the trend of female scores does not contraindicate identity as a factor in similarity scores in women. A more gentle sloping was predicted for fe- males, and the downward trend was expected or occur later than for males--perhaps later than the age range sampled in this study. This prediction is in accord with the notion that identity is more difficult to establish and is stabilized later for females. Twenty-one year old males who are ap- proaching college graduation may indeed have a stronger and more clear sense of the future and of personal identity than females in the same circumstance--particularly unmarried fe- males as were those in the present sample. Females face more choices than males with regard to the focus of identity; for example, junior and senior college women are not as close to resolution of important career decisions as are the males. If a vocational career has been chosen by a college woman, it still remains to be comfortably integrated with other as- pects of female identity, particularly marriage and the family role, whereas career decisions are less complex for males. 27 The data for females do not conform to the hy- pothesis as stated, but they are not inconsistent with the rationale underlying the hypothesis. The pattern of de— velopment of identity in females is different than for males, and this is reflected in the data. But the data do not show whether female identity remains more diffuse in— definitely (as suggested by Douvan and Adelson, 1966) in which case the trend of similarity scores would not be curvilinear regardless of how old the subjects were, or on the other hand, if identity becomes more clear eventually, but later than 21 years of age. It is interesting that the trend of scores for both males and females approaches the curvilinear quality expected when the data are clustered into three age samples as in Figure 2. The specific choices of age groups and the combinations of sub-samples is very influential on the out- come and interpretation of the data. In this regard also, it appears that the precise nature of the trend is not our- vilinear if all seven age groups are considered. For male subjects at least, the trend is spiked. Brodbeck (1954) and Gray (1959) both reported no variations with age from about 10 to 14 and there is no contradictory evidence in the data reported here. But there is a sharp increase in difference scores between ages 13 and 15 and a rather sharp decrease after 17. These observations suggest that the variations in identity that are involved do not change 28 gradually for males, but are age specific events, and that the changes for females, at least in the range studied here, may be insignificant or too gradual to be consequential. The hypothesis that there is no difference be- tween males and females in perceived similarity to the same sex parent until later adolescence was not confirmed. There were no significant differences from childhood through high school age, as predicted, but the late adolescents, defined in this study as college age subjects, did not differ con- sistently as expected. Two groups had been chosen to repre- sent the latter group, 1ower-division college students (freshmen) and upper-division students (juniors and seniors). The junior and senior men had a smaller mean similarity score than did the women, and this finding is in agreement with the consensus of previous studies with college students (Lazowick 1955, Shell e£_al., 1964, Heilbrun, 1965, Bieri §t_al., 1959, and Dyal (in Osgood, Suci, and Tannenbaum, 1957), but the freshmen males and females did not differ. It may be that freshmen who were tested only two months after the start of their first year in college are not yet comparable to the college samples used in previous studies in which the sex difference was obtained. "Middle adoles- cence" and "late adolescence" are not precisely defined age categories, and the choice of age-grade samples is somewhat arbitrary. The data suggest that freshmen have a more clear sense of identity than do high school seniors, but that later 29 in college, women return to their earlier more diffuse identity while men continue to solidity theirs. College students do experience many changes during the four years spent in college and this fluctuation in identity diffusion may be one of them. Obviously the data do not allow a con- clusive statement about college student identity changes, or sex differences in the changes; they are suggestive only. The finding that college junior and senior women see themselves less similar to their mothers than males see themselves like their fathers might lead to the speculation that college women are more masculine and aggressive in their orientation and that this accounts for the discrepancy between the sexes at that age. The data reported here do not support such an explanation (although they do not un- equivocally refute it). An explanation that seems more parsimonious with regard to the data is that college women have not attained an identity any less diffuse than that of younger females while college males have less diffuse iden- tity than do younger adolescent males. That is, junior and senior college women do not appear to be unique from other females as the hyper-masculine interpretation implies. Even a comparison of college women with non-college but equal age women would not necessarily resolve the question raised here. Non-college women between 18 and 24 years of age may have resolved some of the issues of identity earlier by marrying or starting careers sooner than college women 30 who usually wait until completion of college. There is a sense in which both males and females who are in a "diffuse identity" state are less masculine and feminine respectively than those who have resolved identity crisis problems. To be consistent with this line of reasoning we might speculate that the data imply that middle adolescent males are less masculine (and therefore more feminine) than other males. It seems more sound theoretically, however, to interpret this as identity diffusion rather than lack of apprOpriate sex role typing. Research Implications The purpose of this study was to demonstrate that perceived similarity to the same sex parent varies with age and sex in accord with differences in identity diffusion between sexes at various ages. Adequate supporting evidence was found for age trends in male subjects only; observed sex differences gave moderate support but were not conclusive. Two foci for continued research may be suggested to clarify unsettled questions. The first is aimed at clarification of sampling problems and sex differences encountered in this study, particularly with regard to females; the second would provide data of a different sort for converging vali- dation of the hypothesis that identity and perceived simi- larity are related. 31 First,it remains unclear whether the trend of per- ceived similarity to mother in females would show the curvi- linear quality predicted if an older sample were studied. The question of importance here is whether female identity changes during early adulthood as male identity seems to change between 18 and 21 years of age. It was noted above that male identity is associated with vocational choice while feminine identity may be more strongly related to marriage and the role of wife. These assumptions imply that perceived similarity may vary not only with age as studied here, but also with these non-age specific role aspirations and achievements. For example, young adult males who have jobs and support themselves presumably have a more clearly established sense of identity than students of the approximate same age. A similar inference could profit— ably be studied in young adult women who are married as con- trasted to similar age unmarried career women, or perhaps women who have combined marriage and career. The question was raised whether the college fresh- men used here are representative of late adolescence. It was suggested that important changes in identity may occur during the college years, which implies that some of the changes may be quite rapid. It would certainly be important for future research in perceived similarity and other aspects of identity to determine if such changes result in reliable variations in test responses between college classes. 32 Second, a more direct test of the hypothesis that identity and perceived similarity are related could be at- tempted if a reliable test of identity diffusion were de- veloped. If subjects within a single age-sex category who have clear and "un-diffuse" identity also tend to see them- selves more similar to the same sex parent than subjects who are diffuse in identity, then the agreement of different research strategies would greatly enhance the usefulness of the identity hypothesis. Dignan's (1964) study discussed above is a step in this direction, but her identity scale is useful only for freshman and sophomore college women. A replication and expansion of that study is needed. SUMMARY This study was an attempt to demonstrate that assumed similarity to the same sex parent varies with both age and sex. More specifically, it was proposed that as- sumed similarity is affected by identity diffusion and would consequently reflect age and sex trends in the de— velopment of identity. Seven age groups from nine to twenty-one years of age with 30 males and 30 females in each group were administered a form of the semantic differential. Subjects rated themselves and their parents, and a similarity score between self and the same sex parent was derived. It was expected that first born children would perceive themselves more similar to their parents than would non-first born children in agreement with an earlier study by Palmer (1966). No consistent pattern was found between first born and non-first born children and the pre- diction was not confirmed. Since identity in both males and females is thought to be more diffuse during middle adolescence than at any other time in life, it was predicted that both males and fe- males would see themselves less similar to their same sex parent at this time than would younger or older subjects. 33 34 Fifteen and 17 year old males do perceive less similarity between themselves and their fathers than do other males, but the prediction was not confirmed for females. The trend may be curved or spiked for males, but appears not to vary consistently for females. It was predicted that late adolescent females would perceive themselves less similar to their mothers than males see themselves similar to their fathers, although the sex difference was not expected to emerge before late adolescence. This hypothesis is based on the notion that female identity remains more diffuse because it involves more choices and is in general more complex and difficult to achieve in middle class American culture than is male identity. The prediction was not confirmed, but the mean difference scores of 6 of the seven age groups studied tend to support the hypothesis. The results give some support to the hypothesis that there is an association between perceived similarity to the same sex parent and trends in the develOpment of identity. Implications of assumed similarity trends with regard to age and sex differences were discussed and special attention was given to the particular sample as a factor in the interpretation of results. Specific ques- tions that warrant further inquiry were suggested. REFERENCES CITED Beitner, M. S. Word meaning and sexual identification in paranoid schizophrenics and anxiety neurotics. J. abnorm. soc. Psychol., 1961, 63, 289—293. Bieri, J., Lobeck, R., and Galinsky, M. A comparison of direct, indirect, and fantasy measures of iden- tification. J. abnorm. soc. Psychol., 1959, 58, 253-258. Brodbeck, A. J. Learning theory and identification: IV. Oedipal motivation as a determinant of conscious develOpment. J. genet. Psychol., 1954, 84, 219— 227. Bronfenbrenner, U. The study of identification through interpersonal perception. In R. Tagiuri and L. Petrullo (Eds.), Person Perception and Inter- personal Behavior. Stanford: Stanford Univer. Press, 1958, 110-130. Bruni, E. A semantic differential study of sex-role stereo- types and parental identification. (Unpublished manuscript). Carlson, Rae. Identification and personality structure in pre-adolescents. J. abnorm. soc. Psychol., 1963, g: 566-573 . Dignan, Sister Mary Howard. Ego identity, identification, and adjustment in college women. Dissert. Abstracts, 1964, 24, 5541-5542. (University Microfilm Order No. 64-2419). Douvan, Elizabeth A., and Adelson, J. The Adolescent Ex- perience. New York: Wiley, 1966. Erikson, E. H. Childhood and Society. New York: Norton, 1950. Identity and the Life Cycle. Psychol. Issues, 1959, 1, whole issue. 35 36 Freud, S. Group Psychology and the Analysis of the Ego. New York: Boni and Livewright, 1922. Galinsky, M. D., and Fast, Irene. Vocational choice as a focus of the identity search. J. couns. Psychol., 1966, Ii: 89-92. Gray, S. W., and Klaus, R. The assessment of parental identification. Genet. Psychol. Monog., 1956, 23, 87-11“. Perceived similarity to parents and adjustment. Child Develpm., 1959, 30, 91—107. Havighurst, R. J., Robinson, M. Z., and Dorr, M. The de- velOpment of the ideal self in childhood and adolescence. J. educ. Res., 1946, 40, 241-257. Heilbrun, A. B. Sex differences in identification learning. J. genet. Psychol., 1965, 106, 185-193. Hollingshead, H. B., and Redlich, F. C. Social Class and Mental Illness: A Community Study. New York: John Wiley and Sons, Inc., 1958. Kagan, J. Acquisition and significance of sex typing and sex role identity. In Hoffman, M., and Hoffman, Lois, (Eds.), Review of Child Development Re- search; Vol. I., New York: Russell Sage Founda- tion, 1964, 137-167. Lazowick, L. M. On the nature of identification. J. abnorm. soc. Psychol., 1955, 51, 175-183. Lilly, R. S. A developmental study of the semantic dif- ferential. Dissert. Abstracts, 1966, 26, 4063- 4064. Mueller, W. J. Anxiety level, inferred identification and response tendencies on a semantic differential. J. couns. Psychol., 1966, 13, 144-151. Norman, W. T. Stability characteristics of the semantic differential. Amer. J. Psychol., 1959, 1g, 581-584. Osgood, C. E., Suci, G. J., and Tannenbaum, P. H. The Measurement of Meaning. Urbana, Illinois: Uni- versity of Illinois Press, 1957. 37 Palmer, R. D. Birth order and identification. J. consult. Psychol., 1966, 88, 129-135. Seward, Georgene H. Learning theory and identification; V. Some cultural aspects of identification. 8. genet. Psychol., 1954, 88, 229-236. Shell, S. A., O'Mally, J. M., and Johnsgard, K. W. The semantic differential and inferred identification. Psychol. Rev.,l964, 88, 547-558. Winker, J. B. Age trends and sex differences in the wishes, identifications, activities and fears of children. Child Develpm., 1949, 223 191-200. RELATED BIBLIOGRAPHY Bandura, A., and Walters, R. H. Social Learninggand Person— ality Development. New York: Rinehart and Winston, 1963. Baxter, J. C., Lerner, M. J., and Miller, J. S. Identifica- tion as a function of the reinforcing quality of the model and the socialization background of the subject. J. per. soc. Psychol., 1965, 2, 692-697. Beier, E. G., and Ratzeburg, F. The parental identification of male and female college students. J. abnorm. and soc. Psychol., 1953, 88, 569-572. Bieri, J., and Lobeck, R. Self-concept difference in rela- tion to identification, religion and social class. J. abnorm. soc. Psychol., 1961, 88, 94-98. Block, J. An unprofitable application of the semantic dif- ferential. J. consult. Psychol., 1958, gg, 235- 236. Bronson, G. W. Identity diffusion in late adolescence. 8. abnorm. soc. Psychol., 1959, 82, 414-417. Couch, A., and Keniston, D. Yeasayers and Naysayers: agree- ing response set as a personality variable. 8. abnorm. soc. Psyghol., 1960, 88, 151-174. Cozens, W. R., and Jacobs, A. Empirically derived distribu- tions similar to those used in computing the semantic differential. J. Psychol. studies, 1961, 1g, 143-149. Cronbach, L. J. Further evidence on response sets and test design. Educ. Psychol. measmt., 1950, 18, 3-31. Processes affecting scores on "understanding of others" and "assumed similarity." Psychol. Bull., 1955: 23, 177-193- , and Gleser, G. C. Assessing similarity between profiles. Psychol. Bull. 1953, 88, 456-473. 38 Donahoe, Edwards, Ekman, G. 39 J. W. Changes in meaning as a function of age. J. genet. Psychol., 1961, 99, 23—28. A. L. Social desirability and the description of fitgers. J. abnorm. soc. Psyghol., 1959, 82, 434- 3 . Dimensions of emotion. Acta Psychol., 1955, ll, 279-285. Ford, LeRoy H., and Meisels, M. Social desirability and Friedman Gage, N. the semantic differential. Educ.,psy9hol. measmt., 1965, 2_5, 465-475. , C. J., and Gladden, J. W. Objective measurement of social role concepts via the semantic differen- tial. Psychol. Reports, 1964, l8, 239-247. L., and Cronbach, L. J. Conceptual and methodo- logical problems in interpersonal perception. Psychol. Review., 1955, 8;, 411-422. , Leavitt, G. S., and Stone, G. C. The intermediary key in the analysis of interpersonal perception. Psychol. Bull., 1956, 83, 258-266. Goldfried, M. R., and Kissel, S. Age as a variable in the Green, L. Gruen, W. connotative perceptions of some animal symbols. J. proj. Tech., 1963, 81, 171-180. Relationship between semantic differential meas- urement of concept meanings and parent-sex identi- fication phonomenon: A study of the construct validity of the semantic differential. Dissert. Abstracts, 1964, 25, 3100-3101. Rejection of false information about one's self as an indication of ego identity. J. consult. Psychol., 1960, 88, 231-233. Gulliksen, H. How to make meaning more meaningful. Contemp. Hand, J. Psychol., 1958, 8, 115-119. Measurement of response sets. Psychol. Reports., 1964. 14. 907-913. Hayman, Anne. Verbalizations and identity. International Heise, D. J. Psycho-Analysis, 1965, 88, 455-466. R. Semantic differential profiles of 1000 most frequent English words. Psychological Monogr., 1965, 19, Whole Number, 601. 4O Helper, M. M., and Garfield, S. L. Use of the semantic differential to study acculturation in American Indian adolescents. J. per. soc. Psychol., 1965, g, 817-822. Howard, Linda. Identity conflicts in adolescent girls. Smith Coll. Stud. soc. wk., 1960, 8;, 1-21. Jenkins, J. J. Degree of polarization and scores on the principal factors for concepts in the semantic atlas study. Amer. J. Psychol., 1960, 13, 274- 279. , Russell, W. A., and Suci, G. J. An atlas of semantic profiles for 360 words. Amer. J. Psychol., 1958, 11, 688-693. Jourard, S. M. Identification, parent-cathexis, and self- esteem. J. consult. Psychol., 1957, 8;, 375-380. Kagan, J. The concept of identification. Psychol. Rev., 1958. 92. 296-305. Krieger, Margery H. A control for social desirability in a semantic differential. Brit. J. soc. clin. Psychol., 1964, 3, 94-103. T Lichtenstein, E., and Bryan, J. H. Acquiescence and the MMPI: An item reversal approach. J. abnorm. soc. Psychol., 1965, 18, 290-293. Maes, J. L. Identification of male college students with their fathers and some related indices of affect expression and psychosexual adjustment. Unpub— lished Ph.D. Dissertation, Michigan State Uni- versity, 1962. Maltz, H. E. Ontogentic change in the meaning of concepts as measured by the semantic differential. Child. Develpm., 1963, 38, 667-674. Messick, S. J. Metric properties of the semantic differen- tial. Educ. psychol. Measmt., 1957, ll, 200-206. Miron, M. S. The influence of instruction modification upon test-retest reliabilities of the semantic differential. Educ. psychol. Measmt., 1961, 81, 883-893. Mordkoff, A. M. An empirical test of the functional an- tonomy of semantic differential scales. J. verb. Learn. verb. Behav., 1963, g, 504-508. 41 Mowrer, O. H. Learning Theory anngersonaligy Dynamics. New York: Ronald Press, 1950. Murray, J. B. The identity image of the college student. Psychol. Rep., 1964, 88, 267-271. Osgood, C. E. Studies on the generality of affective mean- ing systems. Amer. Psychol., 1962, 81, 10-28. Sanford, N. The dynamics of identification. Psychol. Rev., 1955, Q8, 106-118. Shulman, L. The adolescent and our culture: A search for identity. J. of Jewish Communal Service, 1965, El, ”la-“2’40 Slater, P. E. Toward a dualistic theory of identification. Merrill Palmer Quart., 1961, 1, 113-126. Solarz, A. R._ Perceived activity in Semantic Atlas words as indicated by a tapping response. Percept. mot. Skills, 1963, 88, 91-94. Sommers, Vita S. The impact of dual-cultural membership on identity. Psychiatry, 1964, 81, 332-344. Sopchak, A. L. Parental "identification" and "tendency toward disorders" as measured by the Minnesota Multiphasic Personality Inventory. J. abnorm. soc.Psychol., 1952, 81, 159-165. Stoke, S. M. An inquiry into the concept of identification. J.,genet. Psychol., 1950, 18, 163-189. Tabachinick, N. Three psycho-analytic views of identity. International Journal of Psycho-Analysis, 1965, 88, 467-473. Tanaka, Y., Oyama, T., and Osgood, C. E. A cross-culture and cross concept study of the generality of semantic spaces. J. verb. Learn. verb. Behav., 19633 2., 392-14050 Triandis, H. D., and Osgood, C. E. A comparative factorial analysis of semantic structures in monolingual Greek and American college students. J. abnorm. soc. Psychol., 1958, 81, 187—196. Weiss, J. H., Goldfried, M. R., and Bayroff, I. The levels hypothesis and the study of stimulus value of pro- jective test items. J. proj. Tech., 1965, 88, 445-453. APPENDIX A QUESTIONNAIRE USED WITH SUBJECTS FROM FOURTH GRADE THROUGH HIGH SCHOOL 10. 11. PLEASE ANSWER THESE QUESTIONS Your name: Phone number: Sex: Age: years months Do you live with both of your parents? (Check one) Yes No If you do not live at home, how long have you lived away from your parents? Do you have any brothers? Yes No Do you have any sisters? Yes No If you do have brothers or sisters, are you the oldest? Yes No How far did your father go in school? (a) Circle the highest grade he completed: 1 2 3 4 S 6 7 8 9 10 ll 12 (b) College year completed 1 2 3 4 graduate or professional school What kind of work does (did) your father do? Describe his work briefly: 12. Check the one that best describes your father's job: (a) (b) (C) (d) (e) (f) (s) (h) (i) Manaul work with no Special training Skilled work that required training Clerk, salesman Owns a small business Farm owner Manager Business executive Teacher Professional APPENDIX B QUESTIONNAIRE USED WITH COLLEGE SUBJECTS APPENDIX C STIMULUS CONCEPTS AND RATING SCALES APPENDIX C Stimulus Concepts and Rating Scales Concegts: 1. Me. 2. My mother My father 4. Women 5. Men 6. Females 7. Males 8. My best friend 9. My favorite adult Rating Scales: Evaluation Potency Activity 1. clean-dirty l. strong-weak l. active-passive 2. happy-sad 2. heavy-light 2. fast-slow 3. kind-cruel 3. rugged-delicate 3. hot-cold wise-foolish 4. large-small 4. sharp-dull 5. fair-unfair 5. thin-thick 5. moving-still 6. worthless-valuable 6. humble-proud 6. excitable-calm 49 APPENDIX D INSTRUCTIONS AND SEMANTIC DIFFERENTIAL INSTRUCTIONS This is a study of the meaning of certain words and the things they stand for. This questionnaire will help us to find out what people think about the words listed inside this booklet. This is not a test and there are no right or wrong answers. We want to know what the words mean to you. There are 9 pages. At the top of each page there is a word and below it there are 18 scales on Wthh to rate the word. Rate the words on the basis of what they mean to you. If you were rating the word EXPRESS TRAIN and came to the scale "fast- slow” it would look like this: EXPRESS TRAIN 0‘1. x] 1 2 3 4 5 fast : : : ° : : slow .—.~-»~ .....- - _ ..- v~~ WM --.-— You would probably consider an express train quite fast and so you would place an X on the "fast" end of the scale, perhaps like this for extremely fast: l 2 3 4 5 ’6 7 fast X : : : : : : vslow Or like this for ygry fasg'but not extremely fast: fast ____ : X : : : : : slow .. --—.— - ..-—.-. ...—..- ...- ..__-.. --- Be sure to mark between the dots. If next you were rating the word BICYCLE, you might think it is only 1 2 3 4 5 6 7 fast : : X : : : : slow -2- Or if you felt that a BICYCLE is neither fast nor slow, you would mark it like this: 1 2 3 4 5 6 7 fast : : : X : : : slow Or if you félt it is very slow, you would mark it like this: 1 2 3 4 5 6 7 fast : : : : : X : slow Mest of the ratings you are to make will not be as literal as these examples. For instance, if you were to rate the word AMERICAN on the scale "hot-cold," there is no obvious "correct" answer. Decide what you think the correct rating is and give your first answer. Work as quickly as you can. Be sure to put your marks between the dots, and fill in all scales. slow unfair valuable calm happy proud hot thin cruel heavy strong active still sharp clean small delicate foolish FEMALES Go on to the next page. fast fair worthless excitable sad humble cold thick kind light weak passive moving dull dirty large rugged wise unfair excitable cruel sad proud worthless dirty sharp wise thick strong small cold heavy delicate fast passive moving MEN Go on to the next page. fair calm kind happy humble valuable clean dull foolish thin weak large hot light rugged slow active still valuable calm clean thin active hot large fast sad foolish light moving proud dull kind fair delicate weak MY MOTHER O. Go on to the next page. worthless excitable dirty thick passive cold small slow happy wise heavy still humble sharp cruel unfair rugged strong unfair small clean moving cold rugged fast thick kind proud valuable active strong wise dull heavy sad excitable WOMEN Go on to the next page. fair large dirty still hot delicate slow thin cruel humble worthless passive weak foolish sharp light happy calm fair foolish rugged fast kind proud small passive sharp happy valuable calm hot moving thin light clean weak Go on to the next page. unfair wise delicate slow cruel humble large active dull and worthless excitable cold still thick heavy dirty strong moving cold strong worthless excitable fast clean heavy dull cruel humble sad rugged foolish fair thin large active MY FATHER O. Go on to the next page. still hot weak valuable calm slow dirty light sharp kind proud happy delicate wise unfair thick small passive fast valuable large weak unfair happy dull clean kind thick wise proud rugged moving excitable cold light passive MALES 4 5 6 7 Go on to the next page. slow worthless small strong fair sad sharp dirty cruel thin foolish humble delicate still calm hot heavy active cold cruel small delicate light thick still dull dirty excitable slow valuable happy active fair 'weak foolish proud MY BEST FRIEND hot kind large rugged heavy thin moving sharp clean calm fast worthless sad passive unfair strong wise humble age: l. 3 4 5 Give the sex of this person: Go on to the next page. .. .. hot small thin fair kind worthless passive foolish strong sharp clean proud slow still sad delicate heavy excitable MY FAVORITE ADULT (Not a parent) O. Give the sex of this person: cold large thick unfair cruel valuable active wise weak dull dirty humble fast moving happy rugged light calm age: Go on to the next page. M71711177111117!fllfllfll’lflllfllfllll“