SEX, MATRIARCE‘W MD ACADEMG ACHlEVEflENT 0? BLACK SWDENTS Dissertation for the Degree of Ph. D. mam-um S's‘ATE Ui‘iiVERSlTY FRANCIS X. MULVIHILL 1974 ' THEQN Date 0-7659 May I3, I974 ' - e ' Thisis to certify ihat the thesis entitled Sex, Matriarchy and Academic Achievement of Black Students presented by Francis X. Mulvihill has been accepted towards fulfillment of the requirements for Ph.D. Sociology (1931-99 in Major professor .. ,i ‘_ r\ ABSTRACT SEX, MATRIARCHY AND ACADEMIC ACHIEVEMENT 0F BLACK STUDENTS BY Francis X. Mulvihill A substantial amount of literature on Black self-concept asserts that the matriarchal family structure is particular- ly detrimental to the self-concept of male children. In these families, the economic failure of the father is generalized by the mother to all males, including her chil- dren, with the result that she evaluates her daughters at a higher level than her sons. Through the process of social- ization this lower evaluation causes males to have lower self-concept than females. Some of the literature, however, offers no evidence to support the assertions, and the re- mainder bases its assertions on psychoanalytical case stud- ies. Studies which have compared male and female academic self-concept and achievement have not treated family struc- ture as a variable. As a result, the influence of the matriarchy on differences between sexes in regard to these variables cannot be ascertained. This study is designed to determine whether the assertions are true for a sample of eleventh and twelfth \I (V Francis X. Molvihill grade students in four medium size Michigan cities. There were 173 students of whom twenty males and sixteen females reported that their mothers are the primary breadwinners in their families. The instruments used to measure the con- cepts are: 1. academic achievement--total grade point average, 2. self-concept of academic ability, 3. perceived parental evaluation of academic ability. There were no differences between males and females whose mothers are primary breadwinners in any of the statis- tics (the means and standard deviation of each variable and the correlations between each pair). The only signifi- cant differences were: 1. the male standard deviation in grade point average was higher than the female, and 2. the grade point average of males whose mothers are primary breadwinners is lower than that of males whose fathers are primary breadwinners. It appears, therefore, that males whose families are matriarchal do suffer some ill effects, at least in compari- son with other males, but the cause does not seem to be linked to parental evaluation of ability. An alternative Francis X. Mulvihill explanation may lie in the absence of a successful male model, which is advanced by some writers as the cause of the failure of males whose families are matriarchal. Even here the effect does not operate through the self-concept, since there is no difference between the male groups in parental evaluation and self-concept. Research on the effects of family structure on perceptions of probability of economic success is necessary to explain the findings of this study. SEX, MATRIARCHY AND ACADEMIC ACHIEVEMENT OF BLACK STUDENTS By Francis X. Mulvihill A DISSERTATION Submitted to Michigan State University in partial ful illment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Sociology 1974 ACKNOWLEDGEMENTS First, I wish to thank Dr. John T. Gullahorn for his guidance, kindness and understanding throughout the course of the doctoral program. I also thank Drs. Jeanne E. Gullahorn, Ruth Hamilton and Donald Olmsted for serving on the Dissertation committee and for their thoughtful sugges- tions. Drs. Wilbur Brookover and Kevin Kelly also reViewed the work and offered helpful suggestions. A special thanks goes to Dr. James E. Hedgebeth, with whom I spent many enjoyable hours in developing the study. I also wish to thank Drs. Duane L. Gibson and Robert Anderson for their moral support at critical points along the way. I now realize why writers of books and theses, even those of modest size, acknowledge the help and support of their families. To my wife, Beverly, and my sons, Stephen and Daniel, I give my thanks and my pledge to compensate them for the intermittent periods in.which they went with- out husband and father. I thank my wife also for her many hours of typing and re-typing. ii TABLE OF CONTENTS Chapter Page I 0 THE PROBLEM o o o ' o o o o o o o o o o o o o o 1 II. REVIEW OF LITERATURE . . . . . . . . . . . . 5 Theoretical Orientation Prevailing Theories of Black Family Structure Summary and Discussion III. DESIGN AND METHODOLOGY . . . . . . . . . . . 30 The Sample The Questionnaire Additional Data Data Collection IV 0 HYPOTHESES O O O O O O O O O O O O O O 0 O O 34 V O RESIJLTS O O O 0 O 0 O O 0 O O O 0 O O O O 0 0 41 VI. DISCUSSION AND CONCLUSIONS . . . . . . . . . 49 APPENDIX QUEST IONMIRE O O O O O O O O O O O O O O O O O O 63 LISTOFREFERENCES................. 76 iii LIST OF TABLES TABLE PAGE 1. Per cent Families with Male Heads 18 2. Means and Standard Deviations from Epps 21 3. Correlations among Variables 23 4. Mean Values of Variables 42 5. t-values of Differences between.Means for Com- parison Groups 43 6. Standard Deviations of Variables 43 7. F-values of Differences between variances for Comparison Groups 43 8. Pearson Product-Moment Correlation Coefficients and F-values 48 iv CHAPTER I THE PROBLEM In the past ten to fifteen years, there have been sev- eral references in the literature to the negative conse- quences of a matriarchal family structure for the self- concept and/or academic achievement of Black males. There is a general consensus that the societal barriers which prevent the Black male from securing a stable and adequate economic base for his family are the cause of the relative prevalence of this family structure among Blacks. This economic and family structure causes a preference for daughters among Black mothers. Virtually all of this literature is based on what seems to be conjecture or on psychoanalytic case studies. None of the conjectural works cite any systematic research which links family structure to children's self-concepts or academic achievement. The psychoanalytic studies have the weakness of inferring properties of a general population (in this case, the Black male p0pulation) from properties of a select and non-representative sample (Black males who 2 have been treated for psychiatric disorders). Another weakness in the conjectural type is that the self-concept is not specified. The impression left by these sources is that the self-concept in some global sense is affected by the nature of the parental relations. The psychoanalytic studies are also concerned with general feelings of self-worth and with sexual consequences. In several of the conjectural sources, academic achievement is cited as evidence of the consequences of matriarchy, which at least implies that self-concept of academic ability is a part of the general self-concept which is affected. There is also the implication that the matriarchal family per se has negative consequences for the male. There are virtually no expressions implying degrees of likelihood or probability. The statements are highly general in nature giving the impression of a deterministic rather than probabilistic relationship. There have been two studies which use probability samples and which compare Black male and female youths in regard to academic self-concept and achievement, but neither of these control for type of family structure. They do provide some basis, however, for testing assertions of the other studies. If the assertion is true that matriarchal structure has negative consequences for males 3 but not for females, and since the total samples may be expected to consist of a substantial number of youths whose families are matriarchal, academic self-concept and achieve- ment of the male samples should be lower than that of the female samples. There are, therefore, two distinct sets of literature: the first set includes consideration of family structure and its effects on self-concept but does not provide support for its assertions through the use of measuring instruments administered to a sample of a general population; the second set utilizes measuring instruments administered to a sample of a general population but does not include family struc- ture as a variable. This study will combine the two ap- proaches by including family structure as a data item in a measuring instrument administered to a sample of high school students. This study, therefore, will empirically test the assertions contained in the first set of literature described above. It will also replicate some of the findings con- tained in the second set of literature. The design of this study is relatively uncomplicated, since it is a first step toward the empirical testing of statements which are quite general in regard to both the concepts and the populations discussed. It is to be 4 expected that the results of the study will suggest other variables which should be taken into account. The second purpose, therefore, is to develop a research design which will more adequately explore the factors which influence academic self-concept and achievement of Black youth. In addition to its importance as a theoretical problem in social science, the academic performance of minority groups is a problem for those engaged in the operation of educational systems. Since this study explores some correlates of academic achievement, its results may have something to contribute to the knowledge of educational practitioners. The design is not experimental and there- fore cannot signify causes of academic achievement, but in cases where no correlation is found between achievement and a given variable, that variable can be said to have no effect on achievement. Otherwise the correlation would be non-zero. The utility of this type of finding for the prac- titioner is to indicate those points of possible interven- tion in the learning process which are not likely to produce any change, since they are not related to the characteris- tics to be changed. CHAPTER II REVIEW OF LITERATURE Theoretical Orientation Following Mead (1934) and Sarbin (1954), self-concept may be defined as that organization of qualities that the individual attributes to himself. In the frame of refer- ence of symbolic interaction theory, the self-concept develops in interaction with significant others. These "significant others" are those persons directly responsible for the internalization of norms. In the course of social interaction, the individual takes on those values, beliefs, and evaluations, including evaluations of himself, held by significant others. Through the responses of those others to him, he builds and modifies a conception of self. The concept of the "other" used here is closely related to the concept of "orientational other" used by Manford Kuhn (1964, p. 18). He lists four defining attributes: (l) The term refers to the others to whom the indi- vidual is most fully, broadly and basically committed, emotionally and psychologically; 6 (2) it refers to the others who have provided him with his general vocabulary, including his most basic and crucial concepts and categories; (3) it refers to the others who have provided and continue to provide him with his categories of self and other and with meaningful roles to which assignments refer; (4) it refers to the others in communication with whom his self-conception is basically sustained and/or changed. It is this last attribute which will be most relevant here. Through "orientational others" (in this case parents), the concept of self (in this case academic ability) is sus- tained and/or changed. Since it is the individual's percep- tion of the other, rather than the "actual" other in some objective sense, which operates on the individual, the particular aspect of the other investigated here is the student's perception of how his parents evaluate his aca- demic ability. The "reference group” concept is closely related to this. Of particular interest here are the functions of reference groups in the develOpment of self. Shibutani (1955) distinguishes three types of refer- ence groups: 1. those which serve as a basis of comparison for the self; 2. those in which the individual aSpires to gain or maintain acceptance; and 3. those whose perSpective constitutes the frame of refer- ence for the actor. The evaluation of various aspects of the individual's en- vironment, including the self as object, occurs through the reference group in this last meaning. Turner (1956) also cites one use of "reference group" to mean a group or an individual which acts as a source of perSpectives. This use is identical with that type of role-taking which he labels as "identification," i.e., adopting the standpoint of the other in the role-taking process. When one takes the role of an other who is the member of a group one adopts the group's Standpoint as one's own. The terms "reference group" and "relevant other" refer to essentially the same phenomena. Reflexiveness refers to the characteristic of self as an object to itself. When the role of the other is employed as a mirror reflect- ing the expectations or evaluations of the self as seen in the other role, we may speak of reflexive role-taking. ‘When role-taking involves identification and is reflexive the self becomes specifically an object evaluated from.the standpoint of the other. The attitudes of the other which are adopted as one's own are the attitudes toward one's self rather than toward external objects and values in the O 8 environment. His own self-esteem is the adoption of the estimate of himself which he infers from the standpoint of the role of the other. The bonds of intimacy and prestige or the absence of alternative standpoints determine that the evaluation of relevant others will become the self- evaluation of the individual. Although Merton (1957) has done extensive writing on the reference group concept, his work has been largely an embellishment of concepts which are narrower in scope than those presented by Turner. Merton distinguishes two types of reference groups: those which serve as a source of expectations or norms and those with which an individual compares himself. This study is concerned with referent others who serve as a basis for evaluation. While this is closely related to the normative reference group, the process of evaluation is prior to the process of norms and expectations and may not even lead to that state. The focus will be on the subject's perception of his parent's evaluation of his ability. Norms or expectations held by his parents concerning the child's actual performance will will not be included. The parents, therefore, do not constitute a normative reference group in this study. In addition to the role-taking concept mentioned above, "role" as a set of expectations associated with a 9 status, or social position, will also be important in the model presented here. First, the parent-child role rela- tionship involves the element of intimacy mentioned by Turner as a basis for the relevance of a significant other in the role-taking process. Second, this role relationship presumably includes the institution of education as one of its contents. Finally, status plays an important part in the socialization process in two ways: it determines the content and process of socialization according to where, in social and cultural subsystems, it is located and it serves as the goal toward which parents direct their chil- dren's socialization. In the first case it is an existent, operating condition of socialization; in the second it is an anticipated goal. The general orientation of this dissertation is based on Brookover, et. a1. (1967). In addition to the symbolic interactionist orientation outlined above,this study includes as a working hypothesis the contention that al- though the behavior of significant others is influential in the formation, maintenance and change of the indivi- dual's self-concept, the individual's perception of that behavior is also an important factor. Brookover also notes that an individual's self-concept may vary according to the "person or persons to whom he 10 is referring himself." This is conshstent with Merton's concept of the comparative, as distinct from the normative, reference group. A comparative reference group is one which serves as a referent to which one can compare and thus evaluate his own position. One of the research problems connected with the reference group as a basis of comparison is the fact that different individuals in the same setting may select different groups as a comparative referent. In a study whose purpose is to compare groups according to some structural category, holding constant as many variables as is feasible, it is necessary to hold constant the compara- tive reference groups employed. This, of course, applies as long as the process of reference group selection is not the object of the study. Prevailing Theories of Black Family Structure There is a substantial body of literature which con- tends that the prevalent form of family structure in the Black population is the matriarchy. These writings also contend that this structure is dysfunctional for the self- concept and behavior of the Black male. Although there are slight differences among authOrs in their orientations to the subject, all seem to agree that the matriarchy is the result of the economic insecurity of the Black male. ll Dai (1953), Deutsch (1962; 1963), Ausubel (1963), Pettigrew (1964), Billingsley and Billingsley (1965), Grambs (1965) and Proshansky and Newton (1968) all refer to the unstable economic position of the Black male as the cause of the prevalence of the matriarchy. Because the male is in an insecure economic position, his position in the family is weak if he has any position in the family at all. The female, therefore, is the only dependable adult on whom the children can rely. In some writings, the dependability of the female is stated as being economic, but in most cases the implication seems to be that she furnishes social- emotional support. The types of statements made are all general in nature, e.g., "Even in intact homes the Negro family tends to be matriarchal," (Deutsch, 1963, p. 66), "The father frequently deserts the family and in any case tends to be an unreliable source of economic and emotional security.” (Ausubel, 1963, p. 39), "Most Negro children, then inherit a family which is economically insecure from the very start," (Grambs, 1965, pp. 18-19), "The most salient feature of the lower class Negro family life is its characteristic matriarchal pattern," (Proshansky and Newton, 1968, p. 204). The contexts in which the statements are made range from the specific to the general, with some remaining unspeci- fied, but virtually all of them refer to segregated, 12 low-income urban areas. The general picture presented by these authors is that the Black male has little or no economic security and provides little or no emotional support for his family, with the result that the female provides virtually all the emotional support and perhaps even the economic support. The female is therefore the stable element in the family. In several of the writings, however, the im- plication can be made that in spite of the fact that the female is stable, the family is unstable. This occurs by virtue of the absence or weak position of the father, The term "stable" is not clearly defined in those references, but it seems to imply a continuous state over time, or at least a high probability of remaining in the same state over time. There is also the implication that the state must be secure, either economically or emotionally or both. In other words, a family which has a bare subsistence eco- nomic level with little chance of improvement would not be called "stable." Implied in all of this is a standard of what a family structure should be. Pettigrew (1964, pp. 15- 16) states that, I "such conditions obviously limit the ability of lower class Negroes to follow the typical American pattern-- that is, a stable unit with the husband providing a steady income for his family." 13 This general standard also seems to be implied in the other references cited, with the further implication that because the matriarchal family deviates from the norm of the "typi- cal American pattern" it has negative consequences for its members. All of the sources cited above proceed from a discussion of the matriarchal family structure to a discus- sion of its negative consequences for the self-concept and behavior of male children. The processes through.which these negative consequences occur follow from two conditions: the preference of mothers for daughters over sons and the absence of successful male role models. Deutsch (1962, 1963) contends that the failure of males as adults is due to the absence of successful male models to emulate in their youth. Dai (1953), Ausubel (1963), Pettigrew (1964), Grambs (1965), and Proshansky and Newton (1968) attribute the failure of males to the damage done to their self- concepts through the preference of mothers for their daugh- ters. Grambs (1965, p. 21), even states that, "We cannot lay the major blame for the way Negro boys develop on the lack of adequate male models. It is highly probable that the trauma suffered by Negro females is passed on and displaced upon the males in the situation." This "preference for daughters" argument has the fol- lowing line of reasoning: 14 l. the Black female perceives the role failure of the adult Black male, 2. she forms an expectation of failure for Black males in general, at least for those in a lower-class environ- ment, 3. she passes this expectation on to her sons, 4. her sons internalize this expectation and make it part of their self-concept, and 5. the resulting negative self-concept produces low levels of achievement. In other words, it is a case of the "self-fulfilling prophecy." The mother predicts that her sons will fail and, by the prediction itself, assures that the failure will occur. This does not imply any intent by the Black female to destroy her son's self-image; she may be unaware of the consequences of her interaction with her sons and even un- aware, in a reflective sense, of the nature of the inter- action itself. She may say to him that he is incompetent without realizing that she has said it or even realizing that she believes it. The general piece of evidence presented to verify the result of this process is the fact that Black females have higher levels of academic achievement than do Black males. No evidence is presented to verify the existence of the other steps in the process. In the 15 absence of such evidence, one must assume that an alterna- tive explanation is possible. The logical process followed by these researchers seems to start with two propositions: l. the level of academic achievement is lower for males than for females, and 2. the matriarchy is a prevalent form of family structure among lower class Blacks. They then attempt to link the two through the socialization process asserting that an aspect of the matriarchy is the mothers' preference for daughters over sons. The result is that this preference is perceived as a reflection of self-worth.which in turn becomes internalized as negative self-concept. The statements made by the authors are not always as detailed as this; they generally make the con- ceptual jump from."preference" to behavior without expli- cating the intervening steps of perception and internaliza- tion. Grambs (1965) cites communication of preference and consequent lower self-concept as part of the process and Proshansky and Newton (1968) writing about the adult male, cite lowered self-esteem through the internalization of negative judgements of females. The other authors, however, cite preference for daughters as the explanation for the daughters' higher levels of achievement. It should also be pointed out that where academic achievement is used as a l6 measure of achievement, it is stated explicitly that this type of achievement is only an example. This implies that behavior in some global sense, and therefore self- concept in some global sense, is negatively affected by this process. Deutsch (1962, 1963), on the other hand, believes that matriarchal family structure causes lower levels of achieve- ment in males through the absence of successful male role models in that type of setting, rather than through the maternal preference for daughters over sons. The nature of the evidence which he presents (Deutsch, 1963) is con- siderably more detailed and specific than that presented by the proponents of the "preference for daughters" explana- tion. His subjects were fourth, fifth and sixth grade chil- dren from two schools. One school was 99 per cent Black (the experimental school) and the other was 94 per cent White (the control school). Research results presented here pertain to the experimental school. On four of six tests of academic achievement or ability Black boys have significantly lower scores than girls. Since the differ- ence between the sexes for Blacks on one of these, the Stanford Reading Subtest, is similar to that for the predominantly White control group, the results of this 17 test are not relevant to the propositions considered here. There were therefore three out of six tests in which the Black male performed at a considerably lower level than the female. The results are presented in terms of percen- tages within each sex group whose scores were above a threshold score. The ratios of the percentage of females to males above the threshold on the three tests were 1.67, 1.94, and 1.54, i.e., there were 54 to 94 per cent more females above the threshold. Two other variables measured were Family Atmosphere and General Self-Image. These were measured by sentence completion items, seven for Family Atmosphere and twelve for Self-Image. The ratio of male to female percentages above the threshold on the Family Atmos- phere variable, indicating a more negative atmosphere for boys, was 1.38. The per cent of boys on the negative side of the threshold of the Self-Image index was 47, while the per cent of girls was 42. Although this difference is statistically significant, it is a considerably smaller difference than those found on the other variables. This indicates that male self-concept, although lower than tha: of females, is not lower in the same degree as the other variables measured. This study indicates that both the self-concept and academic performance of Black males are lower than those 18 of Black females and that the sex discrepancy is greater in academic performance than in self-concept. Although Deutsch attributes these differences to the absence of suc- cessful male models, at least in part, he does not directly test for the effect of family structure on behavior. There is little in the literature to counter the writings cited above. Bernard (1966) calls attention to the fact that the most common type of Black household is the husband-wife. Herzog (1966, p. 10) notes that, "Not all fathers are absent fathers among the poor--in fact, about two thirds of them are present among low- income Negro families." The 1970 Census of Population (1972, table 85) shows that in 1970 approximately 73% of Black families and 91% of White families had male heads. Table 1 shows the per cent of families with male heads by income level (U. S. Bureau of the Census, 1973, table 250). TABLE 1. PER CENT FAMILIES WITH MALE HEADS Income in 1969 of Families _ Race $O-$4,999 $5,000-$9,999 §10,000+ Total Black 55 81 91 73 White 78 9O 96 91 Although the percentage of families with male heads is 19 higher for Whites than for Blacks at all income levels, the difference narrows considerably as income increases. The sex of the family head, therefore, is more a function of socio-economic level than of race. These sources are restricted to the father-absent type of matriarchy. There are apparently no current empirically based sources on the extent of matriarchies where the father is present but is in a weak economic position. There is a definite need for more knowledge about the structure and processes of the family, the antecedent conditions of variations in types, and the social and psychological effects of various types. Regarding the matri- archy in particular, research is needed to delineate types of matriarchies, the conditions which produce matriarchies of various types and the effects of various types on family members and on society. An initial step might be to delineate matriarchal types according to some of the func- tions of the family (Goode, 1959), i.e. the degree of dominance of the female in providing biological maintenance, status placement, socialization, emotional maintenance and social control. The proportion of all families which are matriarchies and the distribution of matriarchies by type needs to be studied, controlling for social class, race and culture. 20 While these sources offer some qualifications of the prevalence of the matriarchy, they do not address the question of its possible differential effect on self-concept and academic performance of males and females. Just as Deutsch provides evidence in support of the differential consequences of the matriarchy, Epps (1969) provides evidence in refutation of these differential consequences. In both studies, the evidence is based on children's self- concepts and academic performance without consideration of underlying structural conditions and processes, e.g. family structure. In a study of the correlates of academic achievement of northern and southern high school students in urban areas, Epps investigated the correlation of self-concept with achievement. The self perception scale was Brookover's (1967) "self-perception of ability" scale, with an additional two items. School achievement was measured by school record grades for as many semesters of high school work as were available in English, Social Studies, Arith- metic and Science. There were 400 males and 566 females, selected from the 9th through the 12th grades of four schools in a large northern city. Since geographic region is likely to be a significant factor in achievement and self-concept of Black youth, and since this study is 21 limited to subjects in the north, the southern data will be omitted here. The statistics for self-concept of ability and grade point average were as follows: TABLE 2. MEANS AND STANDARD DEVIATIONS FROM EPPS VARIABLES MALES FEMALES SE s .D. 3? s .D . GRADE POINT AVERAGE 2.7 .6 3.0 .7 SELF-CONCEPT 34.3 6.0 34.2 6.1 The difference between the sexes in self-concept is not significant. The difference in GPA, significant at the .001 level, may be a particular case of the usual findings regarding sex differences in academic achievement. On the basis of a review of research on sex differences in academic achievement of adolescents, McCandless (1970, 1972) concludes that teachers generally give higher grades to girls than to boys. The findings in regard to GPA are consistent with Deutsch's findings of lower levels of test performance among boys than girls. The absence of a difference be- tween males' and females' self-concept does not agree with Deutsch's findings. It must be remembered, however, that the difference in Deutsch's study, although statistically 22 significant, was small: 47% of the boys yg, 42% of the girls fell above the threshold score on the Index of Negative Self-Image. Deutsch was interested in general self-concept. He cites three of the twelve sentence completion items composing his index: 1. "If someone makes fun of me..." 2. ”When I look in the mirror I..." 3. "If I could be an animal I would most like to be...". These do not reflect the types of concepts necessary for investigating academic self-concept and may be based more on evaluations of the societal generalized other than the more specific parental significant other. While there is probably a selection factor operating, whereby the higher male self-concept in high school can be explained by the fact that those with lower self-concepts and academic performance leave school before graduation, it seems unlikely that all students from matriarchal families are in this group. Also, measures of academic achievement for males are lower than those of females in both the primary and high school samples, indicating that not all of the lower achievers have "dropped out." Correlations among variables found in three studies are presented in Table 3. 23 TABLE 3. CORRELATIONS AMONG VARIABLES VARIABLES AND SOURCES MALE FEMALE Achievement x Self-Concept .50 .60 Epps ' _ Achievement x Self-Concept .36 .64 Green and Farquhar ' Self-Concept x Parental Evaluation .47 .47 Kleinfield Kleinfield (1972), in a study which included 166 Black eleventh and twelfth graders in an inner city northeastern school, found no difference between males and females in that subsample in the correlation between self-concept of academic ability and perceived parental evaluation of academic ability. Self-concept was measured by three items, two of which were taken from the eight item self-concept of academic ability scale developed by Brookover, et.a1. (1967). Perceived parental evaluation was measured by three of the five items from the perceived parental evalua- tion of academic ability, also developed by Brookover, et. a1. Finally, Epps (1969) and Green and Farquhar (1965) found that the correlation between academic achievement and self-concept of academic ability was lower for males than females. 24 Summary and Discussion There are two fairly distinct sets of literature. The first type claims that the self-concepts and achieve- ments of the Black male child whose family is matriarchal are lower than those of his female counterparts. But these writings either offer no evidence to support the claims or they base their evidence on psychoanalytic case studies, which may tend to give an unrepresentative view when the evidence is generalized to apply to an entire population. The samples, in these cases, consist of males whose families are matriarchal and who have had psychoanalytic treatment. The population to Which the findings are generalized are all Black males whose families are matriarchal. The second type of literature is based on more repre- sentative samples of the population, but family structure is not treated as a variable. It is therefore impossible to differentiate male subsamples whose families are matriarchal from those whose are not. In general, these studies indicate that: 1. male academic achievement is lower than that of fe- males, 2. there is no sex difference in self-concept of academic ability, 3. the correlation between self-concept of academic ability 25 and academic achievement is lower for males than for females, 4. there is no sex difference in the correlation between self-concept of academic ability and perceived parental evaluation of academic ability. Even though these studies were not based on samples whose families are matriarchal, and do not distinguish that type of subsample from the total sample, there are strong indica- tions that the unsupported and psychoanalytical literature is incorrect. First, there is no difference between sexes in self-concept. The possible argument that males come to reject their parents as significant others and thus reject their low evaluations is countered by the finding that there is no sex difference in the correlation between self- concept and parental evaluation. If males rejected their parents but females did not, the correlation for males would be lower. The lower GPA of males does not necessarily support the impressionistic and psychoanalytic literature. High self-concept is a necessary but not a sufficient con- dition for high achievement, i.e., it acts as an upper limit for achievement. One cannot infer, therefore, that if achievement is low, self-concept must also be low. The reason for lower achievement may be outside the factors of self-concept and parental evaluation. The fact that the correlation between GPA and self-concept is lower for males 26 than females is compatible with this argument. The lower correlation indicates that it is more difficult to estimate performance on the basis of self-concept fiar males than it is for females. If average GPA increased with increasing self-concept for both sexes, but at a slower rate for males, the correlation for males would be smaller. This would reflect the presence of factors other than self- concept which act to hold achievement below the limit imposed by self-concept. At least one of these factors is included in the rea- soning of some of the sources cited above. Pettigrew (1964), Grambs (1965), Deutsch (1963) and Proshansky and Newton (1968) all state that access to opportunities in the economic system is greater for Black females than for Black males. If the male perceives that economic rewards for the effort of acquiring an education are not likely to be achieved, he is less likely than the female to expend that effort. This has direct implications for social programs directed at improvement of the economic situation of Black males. The cause of the condition does not appear to lie within the individual; it more likely lies in the struc- ture of society. Efforts to improve the condition, there- fore, must be directed at the social structure rather 27 than at the individual. Higher levels of education and training will achieve little if the economic institutions of society remain closed. The review of literature presented here reveals two hazards which the social or behavioral scientist faces in recommending solutions to social problems. The first is the generalization of experience with a select subset of a population to the larger population of which it is a part. The clinical psychologist's experience with Black patients cannot be generalized to all Blacks without an unbiased examination of the population. Second, untested hypotheses cannot serve as the basis for recommending social action. The unsupported assertions of some of the references cited above could lead to a program oriented toward the improve- ment of the self-concept of Black males. Empirical research, however, indicates that self-concept is not the problem. The only major piece of information excluded from the empirical studies is the level of perceived parental eval- uation. It is tempting to infer from the combination of equal self-concept and equal correlation between self- concept and perceived parental evaluation, that perceived parental evaluation is also equal. But this is not necessar- ily the case. If for every given value of self-concept, there is a sex difference in average parental evaluation 28 and that difference is constant for all values of self- concept, the correlation between the two variables will be the same for both sexes but the average parental evalua- tion will differ by the amount of the difference for each value of self-concept. This can be conceptualized in terms of lines of least-squares regression through scatter diagrams for the two samples. The equal correlations mean that the slopes of the lines are equal and therefore the lines are parallel. There is no information available, however, as to whether these lines also coincide. If they do not, the average values of perceived parental evaluation are not the same. In addition to the problems of inadequate information in the empirical studies is the more serious problem of the composition of the sample. The proportion of subjects whose families are matriarchal is unknown. There is the possibility that there is a bi-modal distribution involved for males, so that lower scores and correlations for matri- archal males and higher scores and correlations for non- matriarchal males combine to produce average numbers equal to those of females. There is only one clue in the data which counters this possibility. If this negating process is occurring for males but not for females, then the stan- dard deviation of GPA and self-concept in the Epps study 29 should be larger for males than females. This would reflect the wider variation of scores in a bi-modal distribution. But the standard deviations presented by Epps are slightly lower for males than females. The possibility still exists that the female scores also represent a bi-modal distribu- tion but there is no way to determine whether this is actually happening. In summary, where there are direct statements regarding the negative effect of the matriarchal family on male self- concept, there is no evidence presented to support them, or the evidence is biased and not generalizable to the total population. In those studies which utilize more represen- tative samples to compare Black male and female self-con- cepts, family structure is not considered. Although some estimates regarding the effect of matriarchy can be made on the basis of these studies, the conclusions are indirect and tenuous. CHAPTER III DESIGN AND METHODOLOGY The Sample The sample consisted of 499 eleventh and twelfth grade students from four high schools, each of which is located in a different Michigan city of 100,000 to 250,000 popula- tion. Of the 173 Black students in the sample 83 were males and 90 were females. Twenty males and 16 females reported that their mothers are primary breadwinners in their fami- lies. The data was collected in May and June, 1969. The Questionnaire (Appendix) Sex: ReSpondents were asked to check the appropriate Space in the following item: "Sex: M F " Primarnyreadwinner: Respondents were asked to indicate the appropriate choice for the following item: "Who is the main breadwinner in your family?" 1. Father 2. Mother 3. Other" 30 31 In terms of the functions of the family (Goode, 1959), this item identifies the family member who fulfills the major part of the re3ponsibility of providing biological mainte- nance. Academic Achievement (GPA): The students' Grade Point Averages for the current year in English, Mathematics, Social Studies and Science were taken from the school re- cords. All schools used the 4 point grading system. Self-Concept of Academic Ability (SCA): The eight- item.Michigan State General Self-Concept of Ability Scale was used to measure self-concept of academic ability. This is a Guttman-type scale with five reSponse categories for each item which is designed to measure the individual's self-concept of academic ability. One of the eight items specifies "close friends" as the comparative reference group; three Specify a school class, two of which are actual ("your class in school") and one projected ("your class in college"); three Specify no reference group and one Specifi- cally excludes reference groups ("Forget for a moment how others grade your work"). The type of reference group ex- cluded in this last case, however, is the evaluative rather than the comparative type. In general, four items Specify the comparative reference group to be employed and four do not. 32 The score used is the sum of the scores for all eight items. In the analysis, a value of "5" represents the high- est rating of self-concept for each item and "1" represents the lowest. The Score range therefore is 8 to 40. Perceived Parents' Evaluation of Academic Ability (PPE): The five item scale developed by Brookover (1967) to "deter- mine the effects of a Student's perceptions of the evaluation of his academic ability" by his parents on his self-concept of academic ability was used to measure the parent's evalua- tion of academic ability. Two of these items Specify comparative reference groups; three do not. This scale is also a Guttman-type scale with five response categories for each item and with wording Similar to five of the eight SCA scale items. The score used is the sum of the scores for all five items. As in the scoring for the SCA scale, the value "5" represents the highest rating and "1" represents the lowest. The score range is 5 to 25. Additional Data Rage: The race of the respondents was classified as Black or non-Black by the investigators, using the general physical features of the respondents as the basis of classification. 33 Data Collection The investigators arranged with the school administra- tors a mutually acceptable time for collection of the data. The maximum number of junior and senior class sessions being held was a factor in selection of the time. Those teachers whose classes were in session at the time were asked by the administrators if they would allow the investigators to conduct the study. There were no refusals. Time for completion was restricted to one fifty-minute class period. The investigator distributed a copy of the questionnaire (Appendix) to each Student and read the directions to the class. Each completed questionnaire was picked up by the investigator, at which time the race of the student was noted on the questionnaire. CHAPTER IV HYPOTHESES There are two alternate approaches to the development of the hypotheses. One is to state the assertions of the un- supported and psychoanalytic literature in the form of testable hypotheses. This is a somewhat deductive approach, starting with assumptions concerning the perspectives and behavior of Black mothers in matriarchal families and apply- ing symbolic interaction theory to deduce the consequences for children in those families. The second approach is to estimate possible values of and relations among variables for matriarchal subjects, based on studies of subjects un- differentiated by family structure. This is the more induc- tive approach, Since more reliance is placed on observed events as the source of hypotheses. The first alternative will be taken in this study. The empirical references cited will be used as an indirect check on the results of the analysis. The essence of most of the literature reviewed in Chapd ter II is that, because of the insecure economic position of 34 35 the Black male and his consequent economic failure, mothers in lower class Black families expect little from males in general and prefer daughters to sons. As a result, self-concept, and therefore achievement, are lower for males than females. The most important implied assumption in this reasoning is that the low expectation for success held by the female toward the male is based, at least in part, on a low evalua- tion of the male's ability to achieve. This in turn implies that the female perceives that the reason for failure inheres, at least in part, in the individual. If the female perceived that it was the male's position in the cultural and social system which caused his failure, the basis of her expectation would lie in that system. As a consequence, evaluation of ability Should be independent of expectations of achievement. Another important assumption is that the female's experiences with adult males become generalized to all males. If these experiences result in low expecta- tions, because of the male's social position, it is rea- sonable to assume that they would be transferred to male children, since they will be in the same position. This of course assumes that the female does not perceive any immediate change in the social system. It is not at all clear how low evaluations of ability can be transferred 36 merely on the basis of sex, but this assumption is necessary to infer from the male's economic failure that his self- concept of ability suffers. An alternative assumption is that the male perceives that his failure and the low expectations held for him imply that he does not have the ability to succeed. If the female does base her expectations on the social system and if the socialization process does not make clear that low expectations are not based on low evaluation, but are based on lack of opportunities for success, the male could assume that his ability is the rea- son for his failure. In this case it is the son's percep- tions, rather than the mother's, which lead to lower self- concept. This alternative assumption, however, does not seem to be implied in the literature. The derivation of hypotheses based on the literature is as follows: 1. Economic opportunities for the Black male are highly restricted. 2. As a result, the Black male is economically insecure and unable to provide an adequate and continuous income for his family. 3. The Black female perceives this failure as due to the male's lack of ability. 4. The female generalizes this failure as characteristic 37 of most males, including her children. 5. The female, in the role of the mother, is a significant other for the male. 6. The female communicates her low evaluation of her male children to them. 7. Through the socialization process this evaluation is accepted by the male and adopted as his own. 8. This low self-evaluation acts as a limit on the male child's ability to achieve. 9. Achievement of Black male children whose families are matriarchal is therefore correSpondingly low. 10. This process (Statements 1-9) applies only to males whose families are matriarchal; other children are evaluated on an individual basis. Statements 5-10 can be phrased as testable hypotheses. Since they apply directly to the children in matriarchal families they can be tested by analysis of data gathered from the children themselves, without the need for testing parents. Although Statement 10 is intended to specify that the process applies only to males whose families are matriarchal, it contains the implication that these males are evaluated primarily on the basis of their sex whereas other Black children are evaluated on the basis of a greater number of 38 factors, some of which include personality and other attributes of the individual. This, in turn, implies that the degree of variation among these children in regard to parental evaluation, self-concept and achievement is greater than it is among matriarchal males, whose scores on these variables are tied more closely to their sex. Hypothesis 1, therefore, is: H1: The variation in levels of perceived parental evalua- tion of academic ability, self-concept of academic ability and academic achievement is lower among Blade males whose families are matriarchal than among Black females whose families are matriarchal. Statement 6 implies that children are made aware of their mother's low evaluation of them. But to test the assertion, some standard against which to measure the level of evaluation is necessary. Since the argument is that mothers evaluate daughters higher than sons, the evalua- tion of female children whose families are matriarchal is the logical standard of comparison. The second hypothesis therefore is: H2: Black males whose families are matriarchal perceive that their parents evaluate their academic ability at a lower level than do Black females whose families are matriarchal. 39 The consequence of Statement 7 is that: H3: Black males whose families are matriarchal have lower self-concepts of academic ability than Black females whose families are matriarchal. The consequence of Statement 9 is that: H4: Academic achievement of Black males whose families are matriarchal is lower than Academic achievement of Black females whose families are matriarchal. Statement 5 can be interpreted as the proposition that there is a relationship between parental evaluation and self-concept, which can be stated as: H5: The correlation between perceived parental evaluation of academic ability and self-concept of academic ability of Black males whose families are matriarchal is greater than zero. Finally, Statement 8 implies that there is a relation- ship between self-concept and achievement, which can be stated as: H6: The correlation between self-concept of academic ability and academic achievement of Black males whose fami- lies are matriarchal is greater than zero. A stronger test of Hypotheses 5 and 6 would be to assert that there is no difference between males and females in these correlations. There is no basis in the literature 40 for this strong a hypothesis, but the difference between groups will be analyzed in Chapter V without hypothesizing their relationships. CHAPTER V RESULTS The following notation will be used to identify the variables and Samples groups: GPA: grade point average SCA: self-concept of academic ability PPE: perceived parental evaluation of academic ability MM: ‘males whose mothers are primary breadwinners g females whose mothers are primary breadwinners MF: males whose fathers are primary breadwinners FF: females whose fathers are primary breadwinners MT: all males .FT: all females Three statistical techniques will be used in testing the hypotheses: l. F-test for differences between variances: 3% / SE , where 8% . variance of a variable for sample 1 $3 = variance of a variable for sample 2 with 41 42 N1 degrees of freedom in the numerator and N degrees of freedom in the denominator 2 where Nx = Size of sample x. 2. t-test for differences between means: the form of the t-test depends on whether the samples have equal variance; the forms of the tests are: Equal Variances t = (E1 - 322) / (82(N1 + N2) / N1N2)1/2, where 82: ( (N1 ~ 1)sf + (N2 - 1)s§)/(N1 + N2 - 2) Unequal Variances t = (:51 - 332) / (Sf /N1 + 5% / N2)1/2 3. F-test for non-zero values of cornelations: F = r2 (N -2) / (l - r2). Tables 4 through 7 present the means and standard deviations of the three variables and the values of sta- tistical tests for differences between groups. TABLE 4. MEAN VALUES OF VARIABLES. Sample Group GPA 1.6 1.8 1.9 1.9 1.9 1.9 SCA 26.6 27.1 27.5 27.1 27.3 27.1 PPE 18.3 18.2 18.5 18.6 18.5 18.6 43 TABLE 5. t-VALUES OF DIFFERENCES BETWEEN MEANS FOR COMPARISON GROUPS Groups VARIABLE MM-FM MMAMF MT-FT GPA -.38 -1.77* -.11 SCA -.14 -.94 1.31 PPE .94 -.28 -.54 *Differences significant at .05 level. TABLE 6. STANDARD DEVIATIONS OF VARIABLES Sample Group VARIABLE MM FM MF FF MT FT GPA .6 .4 .7 .5 .6 .5 SCA 3.9 3.8 4.0 3.5 4.0 3.6 PPE 3.4 3.1 3.1 3.0 3.2 3.0 TABLE 7. F-VALUES OF DIFFERENCES BETWEEN VARIANCES FOR COMPARISON GROUPS Groups VARIABLE MMeFM MM4MF MT-FT GPA 2.08 1.25 1.76* SCA 1.05 1.06 1.26 PPE 1.22 1.25 1.08 *Differences significant at .05 level. 44 Hypothesis 1: The variation in levels of perceived parental evaluation of academic ability, self-concept of academic ability and academic achievement is lower among Black males whose families are matriarchal than among Black females whose families are matriarchal, Letting, 8% = variance of each variable for the MM group, and S3 = variance of each variable for the FM group, the test statistic is: F=S§IS§. The null hypothesis is: F = 8% / Si = l for each variable. Tables 6 and 7 Show that there is no difference between the MM and FM samples in any of the three variances. This indicates that the male group is evaluated on as wide a range of criteria as females. The hypothesis is therefore rejected. Hypothesis 2: Black males whose families are matriarchal perceive that their parents evaluate their academic ability at a lower level than do Black females whose families are matriarchal. Letting, ii = mean value of PPE for the MM group and x2 = mean value of PPE for the FM group, 45 the test statistic is: /2 t = (Ei - E5) / (82(N1 + N2) / N1N2)1 . The null hypothesis is: tELO. Tables 4 and 5 show that there is no difference between the MM and FM samples in mean PPE. Male perception of parents' evaluation of their academic ability is therefore not lower than that of females. The hypothesis is, there- fore, rejected. Hypothesis 3: Black males whose families are matriarchal have lower self-concepts of academic ability than Black females whose families are matriarchal. Letting, E, mean value of SCA for the MM group and MI I 2 - mean value of SCA for the FM group, the test statistic is: t = (321 - 352) / ($2011 + N2) / N1N2)1/2. The null hypothesis is: t3;0. Tables 4 and 5 Show that there is no difference between the MM and FM samples in mean SCA. Self-concept of academic ability is no lower for males than for females. The hypothesis, therefore, is rejected. 46 Hypothesis 4: Black males whose families are matriarchal have lower levels of academic achievement than Black females whose families are matriarchal. Letting, Ei = mean value of GPA for the MM group and Eé = mean value of GPA for the FM group, the test statistic is: t = (2i -'§é) / (32011 + N2) / N1N2)1/2. The null hypothesis is: t310. Tables 4 and 5 Show that there is no difference be- tween the MM and FM sampkas in mean GPA. Academic achieve- ment is not lower for males than for females. The hypothesis, therefore, is rejected. Hypothesis 5: The correlation between perceived parental evaluation of academic ability and self-concept of academic ability of Black males whose families are matriarchal is greater than zero. Letting, r = Pearson product - moment correlation coefficient be- tween PPE and SCA for the MM group, the test statistic is: F = r2(N-2) / (1-r2). 47 The null hypothesis is F = 0. Table 8 (next page) shows that the correlation between SCA and PPE for the MM sample is significant at the .05 level. There is, therefore, a relationship between parental evaluation and self-concept for males whose families are matriarchal. The hypothesis is accepted. Hypothesis 6: The correlation between self-concept of academic ability and academic achievement of Black males whose families are matriarchal is greater than zero. Letting, r = Pearson product moment correlation coefficient between SCA and GPA for the MM group, the test Statistic is: F = r2(N-2) / (l-rz). The null hypothesis is F = 0. Table 8 shows that the correlation between SCA and GPA for the MM sample is significant at the .05 level. There is, therefore, a relationship between academic achievement and self-concept fiar males whose families are matriarchal. The hypothesis is accepted. 48 .Ho>oH mo. um uaoowmwawfim oon>s swm.m mm. soH.¢ ea. raw.m mm. «mm.m on. wo.H Ha. wm.a mm. E .E as as E :2 $85 33.6. mafia; mMDA<>um 92¢ mHzmHonmmoo ZOHH