Mutual influences among american parents for their values placed on children's conformity in the 1970s and 1980s

Mutual Influences Among American Parents for Their Values Placed on Children’s Conformity in the 1970s and 1980s Frans W. P. Van der Slik Geert W. J. M. Driessen ABSTRACT. Relative effects of both spouses’ educational levels on the value parents place on children’s conformity were examined. Eight General Social Survey samples, covering the 1970s and 1980s, containing information on 3,005 mothers, 2,634 fathers, and their spouses were analyzed simultaneously. Application of “diagonal reference models” showed symmetric influence. Although own educational level had the larger effect on conformity, the effect of spouse’s educational level, particularly the father’s, was substantive. Among mothers, interdependence was moderated by mother’s employment, and marital happiness. Education of mothers who are the sole breadwinners had a smaller effect on own child-rearing values, than education of mothers who are not the sole breadwinners. In addition, education of happily married mothers had a Frans W. P. Van der Slik is affiliated with the Department of Applied Linguistics, Radboud University, Nijmegen, the Netherlands. Geert W. J. M. Driessen is affiliated with the Institute of Applied Social Sciences (ITS), Radboud University, Nijmegen, the Netherlands. Address correspondence to: Frans W. P. Van der Slik, Radboud University, Department of Linguistics, P.O. Box 9103, 6500 HD Nijmegen, The Netherlands (E-mail: f.v.d.slik@let.ru.nl). The authors would like to express their gratitude to Jacqueline Goodnow and Nan Dirk De Graaf for helpful comments on an earlier version of the manuscript. Marriage & Family Review, Vol. 37(4) 2005 Available online at http://www.haworthpress.com/web/MFR © 2005 by The Haworth Press, Inc. All rights reserved. doi:10.1300/J002v37n04_06 95 96 MARRIAGE & FAMILY REVIEW smaller effect on own child-rearing values, than education of unhappily married mothers. [Article copies available for a fee from The Haworth Document Delivery Service: 1-800-HAWORTH. E-mail address: <docdelivery@ haworthpress.com> Website: <http://www.HaworthPress.com> © 2005 by The Haworth Press, Inc. All rights reserved.] KEYWORDS. American spouses, child-rearing values, dominance, education, employment, marital happiness INTRODUCTION Within sociology there is a growing interest in the degree to which wives and husbands influence each other’s viewpoints and behavior. In early studies, it was assumed that the wife’s viewpoints and behavior was completely dependent on her husband’s social characteristics, and that only the latter needed to be considered as explaining variables. This “conventional view” has been severely criticized because of its paternalistic assumptions (Acker, 1972; Sørensen, 1994). However, in predicting both a wife’s and husband’s voting behavior and class identification, the husband’s social characteristics turned out to be more important than the wife’s are, although the wife’s social background cannot be neglected (Baxter, 1994; Davis & Robinson, 1988; De Graaf & Heath, 1992; Sobel, De Graaf, Health, Zou, 2004). The conventional approach, then, seems to reflect the paternalistic character of Western societies quite adequately. It remains open to question, however, whether patriarchy is restricted (a) to particular content areas, and/or (b) to particular familial or contextual conditions. The first type of restriction (to particular content areas) is suggested by the observation that the areas in which asymmetrical patterns of influence by De Graaf and Heath (1992) labeled as “male dominance” have been particularly observed. For example, voting behavior and class identification might be expected to be areas in which the husband’s position carries particular weight (Van Berkel, 1997). When it comes to areas where women’s interests and involvement might be expected to be stronger, however, “male dominance” might be minimal or not seen at all. Child-rearing could be such an area, since mothers even today spend more time raising their children than do fathers. In the present study the value placed on a child’s conforming to external authority, as measured by Kohn (1989), will be considered. Other Frans W. P. Van der Slik and Geert W. J. M. Driessen 97 child-rearing values might be studied, as well as parental behavior. Within socialization research it is well known that parents in lower social positions differ from parents in higher social positions with respect to their child-rearing practices (e.g., Bronfenbrenner, 1958; Gecas, 1979). Lower educated parents are found to demand more conformity than do the higher educated parents. Because the parental child-rearing value conformity could offer a meaningful interpretation for this relationship (Kohn, 1989; Gecas, 1979), many researchers have tried to clarify the precise nature of the link between social position and this particular parental child-rearing value (e.g., Alwin, 1984). We will elaborate on this matter, though we want to emphasize that the nature of the relationship between parental attitudes and behavior remains subject to debate. In addition, conformity to external authority has been associated with a body of past studies and a history of debate about the assumption and the presence of asymmetrical effects (Maccoby, 1961; Meijnen, 1977; Spade, 1991). It also has the virtue of having been measured in eight large-scale General Social Surveys (GSS) in the United States over the period of 1973 to 1986 (Davis & Smith, 1989), allowing to test if patterns of influence have changed over time.1 The 1970s and the 1980s have been characterized as decades in which the process of women’s emancipation has accelerated, resulting in, for example, increasing educational and occupational perspectives for females. The eight GSS surveys offer, therefore, a good opportunity to examine potential changes in asymmetrical influences among spouses in this particular episode of American history. Since the educational and occupational outcomes of males and females continue to converge (Freeman, 2004), the outcomes of the present study may, therefore, serve as a reference point regarding the power balance within contemporary American families. The second possible type of restriction (to particular familial conditions or contextual conditions) is suggested in part by the finding of “independence” in influence patterns, i.e., each spouse’s level of education being the main predictor of own values held (Spade, 1991). It yielded evidence that some particular family conditions–in this case, the employment of both spouses–has moderated the degree of asymmetry. The step now needed is for a check to be conducted on the effects for this particular value in the United States, taking into account a variety of familial and contextual conditions. The opportunity to do so is provided by the eight GSS surveys. 98 MARRIAGE & FAMILY REVIEW Theory The main objective of the present study is to investigate the degree to which spouses’ different educational backgrounds affect both husbands’ and wifes’ child-rearing values. In addition, the aim is to clarify the conditions that affect the strength of these patterns of influence. Attention is given to the extent to which mutual effects are moderated by different family conditions (spouses’ employment status, age, having been divorced, being married unhappily) and contextual conditions (birth cohort experiences, period effects, those: experiences associated with events occurring in a particular period). The choice of level of education as the primary background variable perhaps may seem surprising. The expected variable when predicting child-rearing values might be the particular conditions of work that Kohn and his colleagues have emphasized (conditions such as degree of autonomy or level of complexity of the job; e.g., Kohn, 1989). These are, however, conditions traditionally considered as part of the father’s occupation, leaving the bridge to the mother’s values unaccounted for. If mothers are involved in paid work, the conditions for both spouses could be considered, but that circumstance would restrict the sample to dual earner couples. Considering the relatively low rate of female employment during the 1970s and 1980s, such a restriction appears to be rather inappropriate. By contrast, a person’s level of education can be determined for all wives and all husbands. In addition, restricting a husband’s occupation as predictor for both spouses’ child-rearing values is paternalistic, and leads to a conceptual gap. How, for instance, does the nature of a husband’s occupation influence the values of an unemployed wife? Specifically, a person who is not experiencing a particular occupation in any direct fashion and in fact may be engaged in work that has quite different conditions, may be governed by different influences. In addition, all wives, whether employed or unemployed, have their own histories that are not shared with their husbands. Consequently, it may be presumed that their own past experiences exert at least some influence on their current values and behavior. There are at least two reasons why education might explain that more highly educated parents are more liberal in regard to child rearing than parents who are less well educated. First, according to psychodynamic theory, education facilitates the development of a more secure personality, enhancing the capacity to develop more autonomous, and likely, less conventional values and behaviors (Adorno, Frenkel-Brunswik, Levinson, & Sanford, 1950; Lipset, 1960). Second, socialization theory Frans W. P. Van der Slik and Geert W. J. M. Driessen 99 suggests that the extent to which people are exposed to the educational system has an enduring effect on the development of liberal values with regard to a variety of domains (Gabennesch, 1972; Hyman & Wright, 1979; Vogt, 1997). Because one’s educational development takes place during one’s transition from childhood to adulthood, it might be expected that the influence of education exceeds that of later vocational choices. It is for this reason that the past experiences of women have to be taken into consideration as well. Effects of Own and Spouse’s Education There is indeed evidence showing that level of education is a stronger predictor for the value placed on children’s conformity than is current occupation of parents. Within the US, Alwin (1984) found over time that the effect of social class on child-rearing values declined between the years 1958 to 1983, while the predictive power of education increased. Even in 1976, however, Wright and Wright (1976) found that educational level was a stronger predictor than social class for the value placed on a child’s conforming to authority. These findings have been repeatedly replicated (Alwin & Jackson, 1982; Alwin & Krosnick, 1985; Krosnick & Alwin, 1987), and they have been found to be stable across countries. Slomczynsky, Miller, and Kohn (1981) found that education is a stronger predictor than occupational conditions for the value placed on a child’s conformity in both the U.S. and Poland. Education also predicts such child-rearing values in the Netherlands as conformity to parental rules for mothers and fathers, but both occupation and income are not predictive (Meijnen, 1977). The level of a spouse’s education also has effects on the child-rearing value conformity. Bronfenbrenner (cited by Maccoby, 1961) and Meijnen (1977) found that the husband’s education is a strong predictor of both own as well as spouse’s child-rearing values, while the wife’s education has no effect on her husband’s values. In line with that trend of asymmetrical effects, Spade’s (1991) study of dual earner couples found that the child-rearing values of husbands and wives are best predicted by their own educational backgrounds. Spouse’s education had no significant effect. However, a trend existed in which the educational level of husbands had a stronger influences on wives’ child-rearing values than the wives’ education effecting the husband’s values. Such asymmetrical effects may have several bases, the first being the general presence of a widely accepted cultural norm that the position and the viewpoints of husbands should be dominant (Ganong & 100 MARRIAGE & FAMILY REVIEW Coleman, 1992; Tynes, 1990; Sidanius, Pratto & Bobo, 1994). Marriages in which wives have less education than her husband are particularly susceptible to the culturally accepted norm that husbands should be dominant, a premise that is not violated nor often even recognized. However, compliance to this norm could become potentially problematic when the wife has a higher education than her spouse. For husbands in such a mixed marriage, in turn, an “adjustment” to values corresponding with his wife’s education could imply a sacrifice he is not able or willing to make. Such asymmetry is referred to as the “sexual sociology of adult life” and implies that a status violation occurs when the husband has a lower education than his wife (Brines, 1994; Chodorow, 1974). Being married to a higher educated wife is one thing, being ridiculed by his friends and acquaintances for his loss in his “masculinity” is quite another (Chodorow, 1974; Williams, 1989). For a wife with more education than her husband, an “adjustment” to her spouse’s position would imply a sacrifice that would have several “advantages.” First, she may have a less problematic relationship. A study by Hornung, McCullough, and Sugimoto (1981) has found that in marriages in which husbands has less education than his wives, husbands engaged in more. Second, such an adjustment would be positively reinforced by her spouse and her surroundings as wives and husbands “do gender” (West & Zimmerman, 1987). Specifically, wives “do gender” by accommodating to their spouse’s position, whereas husbands refuse to do so in the face of opposition. Following De Graaf and Heath (1992), therefore, we formulate the “male-dominance” hypothesis, by stating that the educational level of wives is less important than their husbands’ educational level for influencing the child-rearing value of conformity. Moreover, this will be true both for her own and for her husband’s degree of valuing conformity to parental rules. Such a prediction is expected to hold, even when the wife is more highly educated than her spouse. When we speak of “male dominance,” we mean that wives adjust more to their husband’s position than the other way around, irrespective of the level of the wife’s education. This means that the probability that wives adopt values, which are in accordance with the husband’s educational level, is greater than the probability that they will retain values that are in line with own educational level (Van der Slik, De Graaf & Gerris, 2002). Alternatively, what may count is the difference in educational attainment between spouses. The asymmetrical pattern may be reversed in families in which the wife’s level of education is higher that the hus- Frans W. P. Van der Slik and Geert W. J. M. Driessen 101 band’s (Davis & Robinson, 1988). “Resource theory” (Blood & Wolfe, 1960) predicts that the person with the most resources within the relationship will be the most powerful. Blood and Wolfe consider spouses’ education, occupation, and income as potentially relevant resources. Given that parents’ education is a main predictor of the value that mothers and fathers place on children’s conformity, this resource, rather than parents’ occupation, is used in the present study. It may be assumed that the one with more education has more status outside the family than the one with lesser education. In mixed couples, the parent with the lower level of education, therefore, may be expected to move toward the child-rearing values of a parent with higher education. Consequently, we hypothesize that within educationally heterogamous couples, spouses adjust to whoever has the higher educational level, an expectation referred to as the “status maximization” hypothesis. Interaction Effects of Family Conditions In addition to levels of education, effects may result from conditions that are unevenly distributed throughout families. Consequently, the present study considers four conditions that may bring about such effects: employment, age, the occurrence of divorce, and the degree of marital unhappiness. Spouses’ employment status is considered because it is known to affect power relationships within the family (e.g., Komter, 1989; Szinovacz & Harpster, 1993). The results of a study by Spade (1991) support the hypothesis that a wife’s employment increases her power base within the family. In a sample of dual-earner couples, independence between spouses was reported. For both husbands and wives, their own educational level was a strong predictor of child-rearing values, whereas their spouse’s education had no significant effect. Moreover, the wife’s employment seemed to affect the authority of her husband even more when the husband had retired (Szinovacz & Harpster, 1993). The comparisons that should be made (and are made in the present study) are between families in which (1) only the husband has a paid job outside the home, (2) both husband and wife are employed, (3) only the wife is employed, and finally, (4) both husband and wife are unemployed.2 If the mother’s employment does, indeed, affect her power base in the family, it might be deduced that male dominance is typically observed in circumstances where fathers are the sole breadwinners. Mutual dependency occurs, on the other hand, in case where both par- 102 MARRIAGE & FAMILY REVIEW ents are involved in paid work, or both parents are unemployed. Finally, female dominance is expected to occur in circumstances where mothers are sole breadwinners. However, a violation of the culturally accepted norm that the position and the viewpoint of the husband are dominant, might neutralize or even reversed, in circumstances such as the family being unemployed. In this particular case of male status violation, the mother might, willingly or unwillingly, preserve her husband’s “masculinity” by taking less account of own educational level, implying that the relative effect of the father’s education on the mother’s child-rearing values will be stronger. Thus, in accordance with the “status violation thesis,” male dominance is expected to occur both when fathers are the sole breadwinners or in case where fathers are unemployed. In contrast, mutual dependency will typically be observed in circumstances where both parents are employed. The age of parents is considered because this may also affect the power balance within the family. Hill (1965), following Blood and Wolfe (1960), has found that the power of wives increases across the life span. Szinovacz and Harpster (1993) have found that the husband’s dependency on his wife increases with age, while no such aging affect is observed for wives. Wink and Helson (1993) also have reported that husbands’ and wives’ competence and emotional dependence tend to converge as spouses grow older, suggesting that more egalitarian relationships develop among older couples. Thus, older wives may be expected to adjust less strongly to their husband’s child-rearing values than would younger wives. In contrast, older husbands may be expected to adjust more strongly than younger husbands to their wife’s child-rearing values. We term this the “aging” hypothesis. Next, the pattern of influence between husbands and wives may be affected by the degree of marital happiness that couples experience. As can be deduced from literature on the “marriage gradient” (Peplau, 1983), unmarried men and women have asymmetrical expectations about their future partners. Females expect their future partners to be ‘more intelligent, better able to handle things, more successful, higher paid, and to have higher educational achievements’ (Ganong & Coleman, 1992: p.60) (or, in short, to be dominant), however, males expect their future partners to be comparable to themselves. Occasionally, however, these expectations are not fulfilled in the marriage, and, given their higher expectations, it might be argued that wives are more easily disappointed with their spouse’s lack of abilities Frans W. P. Van der Slik and Geert W. J. M. Driessen 103 than husbands are disappointed with deficiencies in their wives’ abilities. As a consequence, wives may report being unhappily married more often (Tichenor, 1999; Twenge, Campbell & Foster, 2003), and at the same time as they are becoming more dominant within the family. Hence, the “unhappily married” hypothesis predicts that unhappily married wives will adjust less strongly to their husband’s child-rearing values, and will affirm their own child-rearing values more strongly than will happily married wives. We do not expect to find differences between unhappily married husbands as compared to happily married husbands. Finally, having been divorced could also make a difference in child rearing values. For example, mothers have more influence on the child rearing of their children than do stepfathers (Crosbie-Burnett & Giles-Sims, 1991), a finding suggesting that wives are apparently not expected to adjust to the child-rearing values of their new husbands. Since, in most cases the children within stepfamilies are not the father’s biological children, it might be hypothesized that a husband, married to a wife who has been divorced will adjust to his wife’s child-rearing values, (Marks & McLanahan, 1993). The relevance of this “having been divorced” hypothesis emerges from the rapid increase of divorce and remarriage rates from the late 1960s until the second half of the 1980s, and the accompanying increase of American children who experienced the divorce of their parents. Interaction Effects of Contextual Conditions General contextual conditions may also be expected to be relevant. There is much evidence for an increase of educational attainment for women; that is to say that the educational gap between men and women in the US has diminished in recent times. In 1955, for example, 43% of males 18 and 19 years of age were enrolled in school, compared with 23% of young adult females. By 1985, however, 52% of males, and 51% of females in this particular age group were enrolled in school (US Bureau of the Census, 1998). Recent reports of the Census Bureau show that girls’ school enrollment rates have even surpassed boys’ (US Bureau of the Census, 2002). As a result of their increased educational opportunities, women’s self-assertion may have increased, as well. Consequently, the asymmetrical pattern of a husbands’ education influencing both spouses’ values more than the wife’s education, might be expected to occur more frequently in birth cohorts that experienced greater educational inequalities between males and females, than those 104 MARRIAGE & FAMILY REVIEW that have experienced more equal opportunities for both sexes (the ‘birth cohort’ hypothesis). To differentiate between successive birth cohorts, we used differences in school enrollment rates for males and females when respondents were 18 years of age because most young adults start looking for a steady relationship at this age. Gender differences in school enrollment rates of birth cohorts are preferred for at least two reasons. First, if birth cohort effects exist, theoretically meaningful hypotheses about the nature of these effects can be tested (Rodgers, 1982). Second, birth date, age and period are linear dependent, making it impossible to disentangle their unique effects (Glenn, 1977). Instead of a relaxation of gender based educational differences, the women’s liberation movement during the late 1960s and 1970s may have increased the self-consciousness of females (Markson, 1984). Assuming that some time delay is required before feminist ideas have permeated much of American society, the asymmetrical pattern of husband’s education influencing their own as well as wife’s viewpoints, this process might be less likely to occur during the 1980s than it did in the 1970s. Since these changes are expected to affect all people equally, we term this the “period” hypothesis. METHOD Data Analyses are based on the 1973, 1975, 1976, 1978, 1980, 1983, 1984, and 1986 data sets of the General Social Surveys (GSS), simultaneously. In the 1972, 1974, 1977, 1982, and 1985 data sets the questions about child rearing were not asked, and there were no GSS surveys in 1979 and 1981. Unfortunately, more recent GSS surveys could not be used since the dependent variable for the present study “the value placed on a child’s conformity to external authority,” as measured by Kohn (1989), has not been included. In the 1973 survey and for half of the 1975 and 1976 surveys block, quota sampling was used. Full probability sampling was employed in half of the 1975 and 1976 surveys and in the 1978 through 1986 surveys. “Each survey is an independently drawn sample of English-speaking persons 18 years or over, living in non-institutional arrangements within the United States” (Davis & Smith, 1989: p. 1). Frans W. P. Van der Slik and Geert W. J. M. Driessen 105 The selected data sets have information from 12,035 respondents. Given the aim of the present study, only the information for parents who were married at the time of interview can be used, or a total of 6,407 respondents who fit this criterion. Excluded, of course, were Childless persons and those who were not married at the time of the interview affects were excluded. These exclusions also effected the representativeness of the sample,3 which meant that the outcomes of the present study are restricted to married Americans with children. Cases with no valid information on the selected variables have been omitted. As a result, 5,639 respondents (3,005 mothers and 2,634 fathers, and their spouses) remain for the analyses. In response to this loss of cases, we tested whether the loss of 768 additional cases affected the representativeness of the sample of married persons. We found two differences: the reduction of cases affected the distribution of respondents’ education: those with less than a high school education are over-represented (χ2 = 190.49; df = 4; p < .001) as well as those who were interviewed in the 1970s (χ2 = 69.32; df = 1; p < .001). The latter difference occurs primarily because at least one third of the respondents interviewed in the 1980 and 1986 GSS surveys have been asked about the value they place on children’s conformity in a phrasing which is incongruous with the one used in the present study. Despite the rather large number of respondents involved, however, we found no significant differences for the remaining characteristics of the respondents. Consequently, the outcomes from the present study, though treated with some caution, appear to be representative of married Americans with children during the 1970s and 1980s. Variables Within the eight GSS samples selected, all respondents have been asked to rank the qualities they value in children in terms of statements adapted from Kohn’s national study (Kohn, 1989). The GSS versions of the Kohn-statements refer to a child in general, regardless of age or gender, whereas Kohn’s items were referring to a child of the same age and sex of a preselected child of one’s own (Kohn, 1989). The original phrasing of Kohn instrument was adopted in the 1973, 1975, 1976, 1978 and 1983 GSS surveys, as well as in part of the 1980 survey. The wordings were changed slightly in the 1984 and 1986 surveys and part of the 1980 survey. For example, instead of using the “he” form when a child was designated, these surveys speak of “a child.” Previous research has 106 MARRIAGE & FAMILY REVIEW demonstrated that this different phrasing does not have an impact on the mean rankings of the items (Schaeffer, 1982; Alwin & Krosnick, 1985). In addition, it has been found that neither children’s age (Alwin, 1989) nor children’s gender (Alwin, 1995) has a substantive effect on the qualities that parents value in children. In all but the 1984 and the 1986 GSS surveys, the child-rearing qualities were presented in the same ordering that Kohn used. However, in the 1984 and 1986 surveys, one-third of the respondents were presented the child-rearing statements in reversed order. Research indicated that these changes in response order do not affect the latent variable, conformity to external authority, which underlies the responses (Krosnick & Alwin, 1987; 1988). In sum, it appears that the information can be analyzed simultaneously. Following Kohn (1989), we performed a one-factor principal component analysis, using twelve of the thirteen child-rearing qualities and the listwise option for missing data.4 High scores on this component indicate conformity to external authority, whereas low scores point to autonomy as a child-rearing value (but see Alwin, 1989, and Alwin & Krosnick, 1985, for a shaded interpretation). We analyzed the conformity scores of male respondents and female respondents separately, the two dependent variables for this study. The results were very similar to those reported by Kohn, 1989). The independent variables are both spouses’ educational levels coded as: (1) less than high school; (2) high school; (3) more than high school (including associate/junior college) and less than bachelor’s degree;5 (4) bachelor’s; (5) graduate. Covariates are both spouses’ employment (coded as 1 = only husband has a paid job [reference category], 2 = both spouses paid job, 3 = only wife has a paid job, and 4 = both spouses no paid job); (b) respondent’s age;6 (c) respondents having been divorced; (d) the rated unhappiness of the marriage (coded as 1 = very happy; 3 = not too happy); (e) differences between school enrollment rates of females and males (i.e., when respondents were 18 years of age (US Bureau of the Census, 1998) have been imputed to measure successive birth cohort experiences7); and (f) period (1970s [= reference category], 1980’s). In addition, we controlled for family income (1 = lower than $1,000; 2 = $1,000 to $2,999; 3 = $3,000 to $3,999; 4 = $4,000 to $4,999; 5 = $5,000 to $5,999; 6 = $6,000 to $6,999; 7 = $7,000 to $7,999; 8 = $8,000 to $9,999; 9 = $10,000 to $14,999; 10 = $15,000 to $19,999; 11 = $20,000 to $24,999; 12 = $25,000 or over), because it has been found to have a negative effect on valuing children’s conformity (Alwin & Jackson, Frans W. P. Van der Slik and Geert W. J. M. Driessen 107 1982). Missing data on family income have been set to the average family income of the year in which the GSS survey was administered. The potential consequences of this substitution were taken into account by including a dummy, measuring whether respondents have given valid information about income. In general, these subsamples are very similar. The following differences were significant: (1) mothers valued conformity less than fathers, an outcome which has been found repeatedly, (2) the mean age of the mothers was around 43 years, while fathers were around 47 years of age; (3) fathers have experienced a divorce more often than mothers, while mothers reported more often to be unhappily married than fathers, and (4) fathers are more likely to provide information on family income. We found no significant differences between the two subsamples in regard to the other variables (spouses’ education, employment, cohort experiences, period, and income). In short, the subsamples of mothers and fathers appeared to be very similar with respect to the variables of interest (see Table 1). TABLE 1. Characteristics of Selected Subsamples of Mothers and Fathers Mothers (N = 3,005) Fathers (N = 2,634) Mean S.D. Mean S.D. T/F p ⫺.09 .99 .10 1.01 ⫺7.35 Mother’s education 2.11 1.01 2.15 1.04 n.s. Father’s education 2.26 1.21 2.28 1.21 n.s. 3.50 n.s. –both paid work .35 .48 .36 .48 ⫺ ⫺ –only father paid work .44 .50 .40 .49 ⫺ ⫺ –only mother paid work .05 .21 .06 .23 ⫺ ⫺ –both no paid work .16 .36 .18 .39 ⫺ 43.61 14.36 47.34 15.23 ⫺9.45 < .001 < .01 Dependent variables: Conformity to parental rules < .001 Independent variables: Covariates: Spouses’ employment: Age Having been divorced ⫺ .15 .36 .18 .38 ⫺2.33 1.40 .55 1.35 .52 3.51 Birth cohort experiences 14.01 4.12 14.10 3.63 n.s. Interviewed in the 1980s .42 .49 .41 .49 n.s. 9.41 2.49 9.50 2.51 .07 .26 .04 .20 Unhappily married Family income Missing on family income < .001 n.s. 5.14 < .001 108 MARRIAGE & FAMILY REVIEW Analyses We use diagonal reference models, introduced by Sobel (1981; 1985), in order to assess the relative effects of own and spouse’s education. Diagonal reference modeling is particularly suited to estimate spouse effects when marital partners are heterogamous with respect to an independent variable. Diagonal reference models have been used in various studies analyzing spousal effects (De Graaf & Ganzeboom, 1990; De Graaf & Heath, 1992; Sorenson & Brownfield, 1991; Van Berkel, 1997; Van der Slik et al., 2002; Sobel et al., 2004). The basic assumption behind these models is that, within educationally homogamous couples, the educational level of one’s spouse cannot have an additional effect on one’s own child-rearing values. Of course, educational homogamous spouses, could influence each other on the basis of characteristics in which they differ, but that would be subject for a subsequent study. As a consequence, the child-rearing values of homogamous couples can be considered as the reference for other, heterogamous couples. Within diagonal reference modeling, the child-rearing values of an educationally heterogamous parent (say with educational level 2, while spouse’s educational level is 4), are modeled as a function of the child-rearing values of homogamous couples with educational level 2, and child-rearing values of homogamous couples with educational level 4. If the child-rearing values of this heterogamous parent resemble those of homogamous couples with educational level 2, own education has the larger effect. In contrast, if this parent’s child-rearing values come close to those of homogamous couples with educational level 4, the spouse’s education has the larger effect. Note that, in the latter case, it is assumed that this parent’s child-rearing values have changed in such a direction that spouse’s education could become predictive. Table 2 provides an overview of the diagonal models representing our hypotheses. Since the models are equivalent for fathers, we present only the models for mothers. The models for fathers are made up in similar fashion. The baseline Model A represents the “male dominance” hypothesis. Male dominance will be found if the weight factor (w), referring to the relative effect of female’s education on her child-rearing value conformity is significantly less than .5. In that particular case, the relative effect of her spouse’s education would outweigh wife’s education (1-w). It is assumed that w lies in-between the 0-1 interval. Frans W. P. Van der Slik and Geert W. J. M. Driessen 109 TABLE 2. Model Specifications, Given for Mothers Only Model A Model B Model F1 Model F2 Model F3 Model F4 Model C1 Model C2 Male dominance (baseline model) MCFijk = (w) · (meduci) + (1 - w) · (feducj) + ΣBl Xijkl + eijk Where: i = level of education of mother (1-5) j = level of education of father (1-5) w = weight (0 ⱕ w ⱕ 1) k = 1 through 3,005 (mothers) l = number of covariates e = error term Status maximization If: mother’s education < father’s education w=1⫺w Employment w = (p + pjj · jj + pnj · nj + pnn · nn) Aging w = (p + page · age) Unhappily married w = (p + punhap · unhap) Been divorced w = (p + pdivorced · divorced) Cohort w = (p + pcoh · coh) Period w = (p + pp80 · p80) ΣBl Xijkl = B1 · jj + B2 · nj + B3 · nn + B4 · age + B5 · unhap + B6 · divorced + B7 · coh + B8 · P80 + B9 · income + B10 · misinc a MCF = mother’s conformity to external authority score; meduc = mother’s educational level; feduc = fa- ther’s educational level; jj = both spouses paid job; nj = only mother paid job; nn = both spouses no paid job; age = years of age; unhap = unhappily married; divorced = having been divorced; coh = birth cohort experience; p80 = period: interviewed in the 1980s; income = family income; misinc = missing on family income Model B, representing the “status maximization” hypothesis, contains a symmetric heterogamy effect. This is modeled parsimoniously by inverting the weights (w) and (1-w) if the spouse has the higher education. Whereas Model A assumed that, independent of the level of the spouse’s education, the effect of the mother’s education always has to be weighted by (w), Model B assumed that this only applies if her education is higher than her spouse’s education. If the spouse’s education is the higher one, Model B assumes that the effect of the spouse’s education has to be weighted by (w). 110 MARRIAGE & FAMILY REVIEW In order to test the subsequent hypotheses, interaction effects have been incorporated in baseline Model A. Models F1 through F4 (representing respectively, the “spouses’ employment,” “aging,” “having been divorced,” and “unhappily married” hypotheses), and Models C1 and C2 (representing the “birth cohort,” and “period” hypotheses), incorporate interaction effects on the weight factor (w). For example, the “unhappily married” hypothesis predicts that unhappily married mothers adjust less strongly to their spouse’s values than do happily married mothers. This would imply that the relative effect of the education of an unhappily married mother on her child-rearing value conformity will be stronger than the effect of a mother’s education in case she is happily married. The weight factor (w = p + punhap ⋅ unhap) models this interaction effect (see Table 2, Model F2). When a mother is married unhappily (unhap = 3), the weight factor of mother’s education would be: (w = p + 3. punhap), whereas in cases that the mother reported to be married happily (unhap = 1) the weight factor of mother’s education would be (w = p + punhap). By incorporating interaction terms in the Models F1 through F4, C1 and C 2, these models are nested within the baseline Model A, and use additional degrees of freedom, as compared to the baseline Model A. We used the standard Likelihood-ratio test (e.g., Sobel 1981; 1985) for the comparison of these nested models with the baseline model. The nested model results in a significant improvement of the baseline model (p < .05) if the difference in the Likelihood-ratio is at least 3.84 against 1 degree of freedom. By design, for Model B the same number of parameters has to be estimated as in Model A. So, Model B uses the same number of degrees of freedom as Model A. This implies that any improvement related to Model A will be considered as significant (e.g., Hendrickx et al., 1990). We conducted the analyses by means of Non Linear Regression (NLR), which is a procedure within SPSS (1990). RESULTS Model Testing The general fit statistics for both mothers’ and fathers’ models are presented in Table 3 (the equation belonging to each hypothesis is presented in Table 2). The “male dominance” hypothesis is represented by Model A, the “status maximization” hypothesis by Model B. Models F1 Frans W. P. Van der Slik and Geert W. J. M. Driessen 111 TABLE 3. Goodness-of-Fit Statistics for Diagonal Reference Models Predicting Mother’s and Father’s Child-Rearing Value Conformity to Parental Rules Fit Statistics Mothers Model Description df Fathers 2 Diff.in ⫺2L df 2 Diff.in ⫺2L Aa Male dominance 16 ⫺ 16 ⫺ Ba Status maximization 16 3.97 16 14.71 F1b Employment 17 ⫺4.23* 17 ⫺3.16 F2b Aging 17 ⫺3.01 17 .37 F3b Married unhappily 17 ⫺4.38* 17 ⫺1.17 F4b Been divorced 17 ⫺.81 17 .03 C1b Birth cohort 17 .18 17 ⫺1.04 C2b Period 17 .70 17 1.38 Notes: Differences are calculated from Model A in each case. See Table 2 for model specifications. a Model A is the preferred model for fathers. b Models F1, and F3 are the preferred models for mothers. * p < .05 For Model A (mothers): ⫺2L2 = 604.57; for Model A (fathers): ⫺2L2 = 401.22 through F4 represent the family hypotheses, while Models C1 and C2 represent the contextual hypotheses. With five parameters for the diagonal, 1 weight parameter w, and 10 parameters for the covariates, the baseline Model A uses 16 degrees of freedom. For fathers as well as for mothers, Model B, representing the “status maximization” hypothesis, was clearly less appropriate than Model A. Instead of a decrease in the Likelihood-ratio, an increase was found, for both mothers and fathers (cf. Table 3, Diff. 2L2). These results imply that a status maximizing (or resource) interpretation was not justified by the data. Next, Models F1 through F4 incorporate interaction terms to model family effects of spouses’ employment, aging, being married unhappily, and having been divorced (see Table 2). For both mothers and fathers, the “employment” hypothesis turned out to be only partially supported. The hypothesis that mutual dependency will be found when mothers and fathers are both employed, seemed to be refuted, since the interaction terms, modeling these effects, did not result in a significant decrease of the Likelihood-ratio. In addition, the “male dominance” hypothesis in case both parents are unemployed is refuted, since the interaction terms, modeling these effects, did not result in a significant decrease of the Likelihood-ratio as well. The accompanying interaction 112 MARRIAGE & FAMILY REVIEW terms were, therefore, not considered for further analyses. Only the interaction effect modeling whether or not the mother is the sole breadwinner, resulted in a significant improvement of the model fit (F1: ⫺2L2 decreases 4.23, against 1 df, p < .05). Next, for mothers and fathers, the “aging” (F2) hypothesis was rejected, since no significant decrease in the Likelihood-ratio was found. The “unhappily married” hypothesis seemed to be supported for mothers (F3: ⫺2L2 decreases 4.38, against 1 df, p < .05). No support has been found for the “having been divorced” hypothesis. Finally, Models C1 and C2 incorporate interaction terms to model birth cohort experiences, and trends over time. No support was found for these hypotheses. In sum, for mothers Model, F1, and F3 (explained variance is 16%) were preferred, while for fathers Model A (explained variance is 15%) was favored. An important qualification is that, in line with previous studies, the explained variance is quite small. The Relative Effects of Spouses’ Education The parameter estimates of the baseline Models A, and models F1 and F3 for mothers, can be found in Table 4. The weight factors (w) and (1-w) of Model A for mothers indicated that the relative effect of the mother’s education (w) on her valuing conformity was .61 (p < .01), while the relative effect of her husband’s education (1-w) was .39 (p < .01). Within heterogamous couples, it appeared that own education is more important than spouse’s educational characteristics. This observation also holds for fathers (Table 4, Model A for fathers). The relative effect of the father’s education (w) on his conformity orientation was .76 (p < .01), while the relative effect of his spouse’s education (1-w) on his conformity scores was .24, which is by no means negligible (p < .01). Although the relative effect of the mother’s education on her husband’s conformity was not as strong as the relative effect of the husband’s education on his wife’s conformity, these results point to mutual dependency. Spouse’s education did affect one’s own child-rearing values in a substantive way, suggesting that the “male dominance” hypothesis is refuted. This finding of mutual dependency seems to imply that both father’s and mother’s child-rearing values have changed into the direction of their spouse’s child-rearing values. For mothers, the “employment” hypothesis was partially supported, while this hypothesis was rejected for fathers. In families where mothers are the sole breadwinners, the relative effect of their education is reduced to a negligible .13 (see Table 4, Model F1), implying that male Frans W. P. Van der Slik and Geert W. J. M. Driessen 113 dominance operates regarding the mother’s child-rearing value conformity. In addition, the results do not replicate Spade’s (1991) finding that, within a sample of dual earner couples, spouse’s education is irrelevant in predicting parental child-rearing values, since we found that the interaction term, modeling the effect of education for dual earner cou- TABLE 4. Parameters, and in Parentheses Their Standard Errors, for Preferred Diagonal Reference Models (F1 and F3 for Mothers), and Base Line Modelsa Mothers Weight factors of: Own education (w) Spouse’s education (1-w) Mother’s education, when mother is the sole breadwinner Mother’s education, when mother is not the sole breadwinner Mother’s education, when mother reports to be married happily Mother’s education, when mother reports to be married not too happily Level of conformity of educational homogamous parents Less than high school High school More than H.S. less than Bachelor’s Bachelor’s Graduate Covariates (unstandardized B’s): –Both spouses paid job –Both spouses no paid job –Only mother paid job Age/100 Unhappily married Having been divorced Birth cohort Period: interviewed in the 1980s Controls: Family income Missing on income Fathers Model A Model F1 Model F3 Model A Base line Employment Unhappily married Base line .61 (.05) ⫺ ⫺ .76 (.07) .39 (.05) ⫺ ⫺ .24 (.07) ⫺ .13 (.22) ⫺ ⫺ ⫺ .64 (.05) ⫺ ⫺ ⫺ ⫺ .52 (.14) ⫺ – – .94 (.09) – .97 (.11) .97 (.11) .98 (.11) .71 (.14) .46 (.11) .45 (.11) .48 (.11) .42 (.14) .23 (.12) .23 (.12) .22 (.12) .12 (.15) ⫺.10 (.13) ⫺.12 (.15) ⫺.10 (.14) ⫺.25 (.16) ⫺.22 (.16) ⫺.22 (.16) ⫺.18 (.16) ⫺.37 (.18) .03 (.04) .03 (.04) .03 (.04) ⫺.06 (.04) .04 (.06) .04 (.06) .04 (.06) ⫺.12 (.06) ⫺.07 (.08) ⫺.11 (.08) ⫺.07 (.08) .08 (.08) .02 (.14) .02 (.13) .03 (.14) .54 (.14) ⫺.07 (.03) ⫺.07 (.03) ⫺.08 (.03) ⫺.08 (.03) ⫺.01 (.05) ⫺.01 (.05) ⫺.00 (.05) ⫺.11 (.05) ⫺.01 (.00) ⫺.01 (.00) ⫺.01 (.00) .00 (.01) .14 (.04) .14 (.04) .14 (.04) .05 (.04) ⫺.05 (.01) ⫺.05 (.01) ⫺.05 (.01) ⫺.04 (.01) .06 (.06) .06 (.06) .06 (.06) ⫺.11 (.09) a Reference categories are: Only husband paid job; Interviewed in the 1970s 114 MARRIAGE & FAMILY REVIEW ples, was not statistically significant. Thus, the finding of mutual dependency, found for couples in which the father is the sole breadwinner, applied for dual earner couples as well. Contrary to our expectations, however, we also observed mutual dependency for parents who are both unemployed. The rejection of the “aging” hypothesis implies that no evidence was found for a change in the power balance within the family across the life span. Though we did observe a trend into the direction of an increasing effect of the wife’s education on her child-rearing values, the improvement of the model fit was nonsignificant (see Table 4: F2: ⫺2L2 decreases 3.01, against 1 df, n.s.). The “unhappily married” hypothesis was confirmed for mothers (see Table 4, F3). When mothers report that they are happily married, the relative effect of their education on their conformity scores is .51 (p + punhap × unhap = .31 + .21 ⋅ 1), while this effect is .94 (p + punhap ⋅ unhap = .31 + .21 ⋅ 3) when mothers reported to be married unhappily. As expected, no such pattern was found for fathers because, whether or not they are married unhappily, the relative effect of the fathers remained .74. These outcomes imply mutual dependency when mothers reported being happily married and independence when they reported being unhappily married. Having been divorced seemed to have had no significant impact on the power balance between spouses because the related hypotheses were rejected. Fathers and mothers who have experienced divorce were affected in their child-rearing values by their spouses in much the same way as mothers and fathers who have not been divorced. Such results imply that parents who had been divorced either adjust to their newly wed spouses or that their new spouses have about the same educational level as the spouses from whom they were divorced. In addition, the women’s liberation movement of the late 1960s and 1970s appeared to have had no great impact on the power balance between spouses since mutual dependency in regard to child-rearing values was as present in the 1970s as it was in the 1980s. The five diagonal reference intercepts of Model A for mothers represent mothers’ levels of conformity for educationally homogamous mothers, while the five intercepts of Model A for fathers represent levels of fathers’ conformity scores for educationally homogamous fathers. These intercepts revealed that, after taking the effects of the covariates into account, mothers and fathers with less than high school Frans W. P. Van der Slik and Geert W. J. M. Driessen 115 valued conformity most strongly, while those with the highest degrees valued conformity the least. In addition, we found significant independent effects of family income for both mothers and fathers. Mothers and fathers who gave invalid information about family income appeared to be very similar to those who did not answer this question, implying that substituting invalid answers by average year income did not affect the outcomes of the analyses presented in this study. Those having more family income valued conformity less. Older fathers valued conformity more than younger fathers did, and fathers who have been divorced valued conformity less than did fathers who have never been divorced (these effects do not hold for mothers). Finally, mothers interviewed in the 1970s valued conformity less than those interviewed in the 1980s, though fathers did not. It appears that for mothers the 1980s can be characterized as a period of restoration. This finding is not in line with the results of Alwin (1986), who made use of virtually the same GSS data sets as used in the present study. Examining trends in the 1970s and 1980s in more detail, Alwin (1986) finds no apparent trend. Alwin, however, made use of the entire adult population, and did not differentiate between fathers and mothers. Whether the different trends for mothers and fathers, found in the present study, have continued in the 1990s, and the beginning of this century, cannot be answered unambiguously because the Kohn’s items have not been included in the GSS surveys for the past fifteen years. DISCUSSION The main objective of the present study was to examine the degree to which own as well as spousal educational attainment affects the value placed on a child’s conformity to external authority. The overall pattern was one of interdependency between American spouses with regard to this value. It appeared that early studies in which the mothers’ education was not accounted for in predicting either own or husbands’ (fathers’) values, have resulted in incomplete outcomes. Specifically, the spouse’s own educational background does matter in a substantive way. In fact, a mother’s level of education was more important in predicting her conformity values than was her husband’s educational attainment. Another objective of this study was to examine whether several family conditions and contextual conditions can clarify or moderate the existence of asymmetric patterns of influence. Two effects were found: 116 MARRIAGE & FAMILY REVIEW mutual dependency within couples with regard to their child-rearing values was moderated by mothers’ sole employment, and the degree of unhappiness with the marriage. No such effects were found for fathers and it was evident that the sole employment of mothers is associated with male dominance instead of female dominance. We believe that this finding may be interpreted as the result of male status violation. In order to preserve the husband’s masculinity, the mother–willingly or unwillingly–gives in and appears to accommodate to the husband’s viewpoints. The power balance in families in which the father is the sole breadwinner is comparable to the power balance in dual-earner families. These outcomes seem to undermine the general accepted idea that female employment will increase the wife’s power base within the family, at least in the 1970s and 1980s. Obviously, these findings appear to be in direct contradiction to the core of feminist emancipatory philosophy, and it remains to be seen as to whether similar analyses of more recent data will result in a less disturbing picture. The reported outcome, however, does not appear to be coincidental because Sobel et al. (2004), using large-scale quantitative British data, and Tichenor (1999), using small-scale qualitative American data, also report patterns of influence consistent with the male status violation thesis. Regarding the value mothers place on a child’s conformity to external authority, the evidence suggests that happily married mothers adjusted more strongly to their spouse’s position than did unhappily married mothers. In the reverse circumstances, where mothers reported being unhappily married, it appeared that independency between spouses exists regarding the child-rearing value conformity to external authority. Although these outcomes are consistent with the “marriage gradient” (Peplau, 1983) and the “status violation” thesis (Brines, 1994), the causal link between the mother’s marital unhappiness and the occurrence of independence between spouses with regard to child-rearing values is by no means clear. Nevertheless, it may be hypothesized that independency between spouses regarding childrearing values (and perhaps a variety of values and behaviors as well), and particularly when associated with an unhappy marriage, is a gateway to a potential divorce. Yet another “solution” remains possible, however, particularly in the case where intense marital discord exists. Van der Slik et al. (2002), for example, have found that, among Dutch parents who experience intense marital discord, the mother often gives in and accommodates almost completely to her spouse’s child-rearing values. Why such an Frans W. P. Van der Slik and Geert W. J. M. Driessen 117 accommodation among Dutch mothers takes place is unclear, though one possibility might be that they try to protect both their children and the marital relationship. The outcome of spousal interdependency found in this study is inconsistent with the findings of other American studies, past or present. Earlier American studies have found support for male dominance in regard to child rearing (e.g., Maccoby, 1961), while more recent work has documented patterns of independence between spouses, at least for dual-earner couples (Spade, 1991). Unlike these earlier studies, this study has made use of a large-scale longitudinal database that is representative of English-speaking American individuals who are married and English-speaking. Such a sampling advantage means that greater confidence can be placed on the validity of the results obtained in the present study than some of the previous work. In short, mutual dependency within couples as regards their child-rearing values was not moderated by the effects of age, divorce, period, or cohort experiences. Given the major social changes that American society faced during the 1970s and 1980s, the rejection of the “birth cohort,” “period,” and “divorced” hypotheses in particular, is an important finding. Of particular prominence, for example, are results indicating that recent improvements in women’s social position as well as trends toward increased individualism (Thornton & Young-DeMarco, 2001), have failed to impact on the balance of power between husband and wife in regard to child rearing values. Moreover, no effect was found for having been divorced, a counter-intuitive finding that should be explored further for replication with a more current data base, particularly because more than half of today’s U.S. children have experienced a divorce. A potential drawback of the present study might be that the outcomes are the result of partner selection before marriage (Tynes, 1990; Sutton, 1993). However, assuming that there are countless criteria for mate selection, it seems rather unlikely that researchers would find some kind of systematic link between partner’s education and their ideas about raising children, who have not yet been born (Van Berkel & De Graaf, 1995). Even in the hypothetical case that partners do choose each other on the base of their child-rearing values, this still would mean that something like male dominance exists (Van der Slik et al., 2002). Another shortcoming, perhaps, might be that so few of our hypotheses were supported by the data. In our view, however, a major task of social sciences is to explode myths. In doing so, refuting research hypotheses is equally valuable as finding support for them. 118 MARRIAGE & FAMILY REVIEW This study has contributed to family research on patterns of influence in another way. The finding of spousal interdependency with regard to child-rearing values implies that the outcome of male dominance for voting behavior and class identification can now be considered as domain-specific (e.g., Van Berkel, 1997). Future research is needed that explores the differences among various aspects of family life more systematically. The relevant questions are now, which domains show male dominance, which are of equal concern to both spouses, and finally, which display female dominance? Diagonal reference modeling, the technique we applied in this study, proved to be a helpful tool to study such dominance patterns. NOTES 1. In more recent GSS surveys, instead of the Kohn-items, phrasings adopted from Lenski (1961) have been used to measure the parental value preferences for a child’s conformity versus autonomy. This change in operationalization severely complicates the inclusion of more recent GSS samples. 2. One might like to consider income differences between spouses as well, because resource theory would predict that money yields power within the family. A study by Spade (1991), however, using data of American dual-earner couples, reported no significant effects of spouse’s income on the value placed on a child’s conformity. Unfortunately, within the GSS surveys selected, no exact information is tractable on a respondent’s as well as his/her spouse’s income. 3. As compared to the entire population in the 1970s and 1980s, the average married American parent was more likely to be either older than 25 years or younger than 70 2 2 years (χ = 710.99; df = 13; p < .001), to be male (χ = 28.27; df = 1; p < .001), to have a 2 2 higher income (χ = 425.78; df = 11; p < .001), to have attended junior college (χ = 2 45.29; df = 4; p < .001), to have experienced divorce (χ = 377.33; df = 1; p < .001), and 2 to have been interviewed in the 1980s (χ = 23.50; df = 1; p < .001). As regards employment no significant differences were found. Given that unmarried childless persons are more likely to be either younger or older, that divorced females are less likely to remarry than divorced males, the observed differences are not unexpected. 4. A number of respondents have ranked just one of the child qualities. Lower educated persons tend to have more missings than higher educated (Krosnick & Alwin, 1987). 5. A number of people have more years of schooling than just High School; yet do not possess a higher degree. These persons, as well as those who graduated from Associate/junior College, were coded as having educational level (3). 6. No information is available about spouse’s age. Therefore, spousal differences in age could not be used as an explanation for asymmetric effects. 7. School enrollment rates of males and females (aged 18 and 19 years) have been published by the US Bureau of the Census (1998) for the years 1947 to 1996. No data are available for respondents older than 18 years of age in 1947. We imputed for these respondents the difference between enrollment rates as measured in 1947. Frans W. P. Van der Slik and Geert W. J. M. Driessen 119 REFERENCES Acker, J. (1972). Women and social stratification: A case of intellectual sexism. American Journal of Sociology, 78, 936-45. Adorno, T. W., Frenkel-Brunswik, E., Levinson, D. & Sanford, R. N. (1950). The authoritarian personality. New York: Harper & Brothers. Alwin, D. F. (1984). Trends in parental socialization values: Detroit, 1958-1983. American Journal of Sociology, 90, 359-82. Alwin, D. F. (1986). Religion and parental child-rearing orientations: Evidence of a Catholic-Protestant convergence. American Journal of Sociology, 92, 412-40. Alwin, D. F. (1989). Changes in qualities valued in children in the United States, 1964 to 1984. Social Science Research, 18, 195-236. Alwin, D. F. (1995). Parental shaping of gender in child-rearing. In J. J. Hox, B. F. Van der Meulen, J. M. A. M. Janssens, J. J. F. Ter Laak, & L. W. C. Tavecchio (Eds.) Advances in Family Research (pp. 17-32). Amsterdam: Thesis Publishers. Alwin, D. F. & Jackson, D. J. (1982). The statistical analysis of Kohn’s measures of parental values. In K. G. Jöreskog & H. Wold (Eds.) Systems under indirect Observation (pp. 197-223). Amsterdam: North-Holland Publishing Company. Alwin, D. F., & Krosnick, J. A. (1985). The measurement of values in surveys: A comparison of ratings and rankings. Public Opinion Quarterly, 49, 535-52. Baxter, J. (1994). Is husband’s class enough? Class locations and class identity in the United States, Sweden, Norway and Australia. American Sociological Review, 59, 220-35. Blood, R. O. & Wolfe, D. M. (1960). Husbands and wives: The dynamics of married living. Glencoe, Ill: Free Press of Glencoe. Brines, J. 1994. Economic dependency, gender, and the division of labor at home. American Journal of Sociology, 100, 652-88. Bronfenbrenner, U. (1958). Socialization and social class through time and space. In E. Maccoby, T. M. Newcomb, & E. L. Hartley (Eds.) Readings in social psychology (pp. 400-25). New York: Holt, Rinehart & Winston. Chodorow, N. (1974). Family structure and feminine personality. In M. Z. Rosaldo & L. Lamphere (Eds.) Women, culture, and society (pp. 43-66). Stanford, CA: Stanford University Press. Crosbie-Burnett, M. & Giles-Sims, J. (1991). Marital power in stepfather families: A test of normative-resource theory. Journal of Family Psychology, 4, 484-96. Davis, J. A. & Smith, T. W. (1989). General social surveys, 1972-1989: Cumulative codebook. Chicago: National Opinion Research Center. Davis, N. J. & Robinson, R. V. (1988). Class identification of men and women in the 1970s and 1980s. American Sociological Review, 53, 103-12. De Graaf, N. D. & Ganzeboom, H. B. (1990). Cultuurdeelname en opleiding: Een analyse van statusgroep-effecten met diagonale referentiemodellen. Mens en Maatschappij, 65, 219-44. De Graaf, N. D. & Heath, A. (1992). Husbands’ and wives’ voting behaviour in Britain: Class-dependent mutual influences of spouses. Acta Sociologica, 35, 311-22. Freeman, C. E. (2004). Trends in educational equity of girls and women: 2004 (NCES 2005-016). U.S. Department of Education, National Center for Education Statistics. U.S. Government Printing Office. 120 MARRIAGE & FAMILY REVIEW Gabennesch, H. (1972). Authoritarianism as a world View. American Journal of Sociology, 77, 857-75. Ganong, L. H. & Coleman, M. (1992). Gender differences in expectations of self and future partner. Journal of Family Issues, 13, 55-64. Gecas, V. (1979). The influence of social class on socialization. In W. R. Burr, R. Hill, F. I. Nye & I. L. Reiss (Eds.) Contemporary theories about the family. Volume I. Research-based theories. (pp. 365-440). New York: Free Press. Glenn, N. D. 1977. Cohort analysis. Beverly Hills: Sage Publications. Hendrickx, J., De Graaf, N. D., Lammers, J. & Ultee, W. C. (1990). Models for status inconsistency and mobility: A comparison of the approaches by Hope and Sobel with the mainstream square additive model. Quality and Quantity, 27, 335-52. Hill, R. (1965). Decision-making and the family life cycle. In E. Shanas & G. F. Streib (Eds.) Social structure and the family: Generational relations (pp. 113-39). Englewood Cliffs, NJ: Prentice-Hall. Hornung, C. A., McCullough, B. C. & Sugimoto, T. (1981). Status relationships in marriage: Risk factors in spouse abuse. Journal of Marriage and the Family, 43, 675-92. Hyman, H. H., & Wright, C. R. (1979). Education’s lasting influence on values. Chicago: University of Chicago Press. Kohn, M. L. (1989). Class and conformity. A study in values. With a reassessment, 1977. Chicago: University of Chicago Press. Komter, A. (1989). Hidden power in marriage. Gender and Society, 2, 187-216. Krosnick J. A. & Alwin, D. F. (1987). An evaluation of a cognitive theory of response-order effects in survey measurement. Public Opinion Quarterly, 51, 201-19. Krosnick J. A. & Alwin, D. F. (1988). A test of the form-resistant correlation hypothesis. Ratings, rankings, and the measurement of values. Public Opinion Quarterly, 52, 526-38. Lenski, G. E. (1961). The religious factor. Garden City, NY: Doubleday. Lipset, S. M. (1960). Political man: The social bases of politics. Garden City, NY: Doubleday. Maccoby, E. E. (1961). The choice of variables in the study of socialization. Sociometry, 24, 357-71. Marks, N. F. & McLanahan, S. S. (1993). Gender, family structure, and social support among parents. Journal of Marriage and the Family, 55, 481-93. Markson, E. W. (1984). Family roles and the impact of feminism on women’s mental health across the life course. Marriage and Family Review, 7, 215-32. Meijnen, G. W. (1977). Maatschappelijke achtergronden van intellektuele ontwikkelingen. Groningen: Wolters-Noordhoff. Peplau, L. A. (1983). Roles and gender. In H. H. Kelley, E. Berscheid, A. Christensen, J. Harvey, T. Huston, G. Levinger, E. McClintock, L. A. Peplau, & D. Peterson.(Eds.) Close Relationships (pp. 220-64). New York: W. H. Freeman. Rodgers, W. L. (1982). Estimable functions of age, period, and cohort effects. American Sociological Review, 47, 774-87. Schaeffer, N. C. (1982). A general survey experiment in genetic words. Public Opinion Quarterly, 46, 572-81. Frans W. P. Van der Slik and Geert W. J. M. Driessen 121 Sidanius, J., Pratto, F. & Bobo, L. (1994). Social dominance orientation and the political psychology of gender: A case of invariance? Journal of Personality and Social Psychology, 67, 998-1011. Slomczynsky, K. M., Miller, J. & Kohn, M. L. (1981). Stratification, work, and values: A Polish-United States comparison. American Sociological Review, 46, 720-44. Sobel, M. E. (1981). Diagonal mobility models: A substantively motivated class of designs for the analysis of mobility effects. American Sociological Review, 46, 893-906. Sobel, M. E. (1985). Social mobility and fertility revisited: Some new models for the analysis of the mobility effects hypothesis. American Sociological Review, 50, 699-712. Sobel, M. E., De Graaf, N. D. Heath, A. & Zou, Y. (2004). Men matter more: The social class identity of married British women, 1985-1991. Journal of the Royal Statistical Society A, 167, 37-52. Sorenson, A. M. & Brownfield, D. (1991). The measurement of parental influence. Assessing the relative influence of father and mother. Sociological Methods and Research, 19, 511-35. Sørensen, A. (1994). Women, family and class. Annual Review of Sociology, 20, 27-47. Spade, J. Z. (1991). Occupational structure and men’s and women’s parental values. Journal of Family Issues, 12, 343-60. SPSS. (1990). SPSS advanced statistics user’s guide. Chicago, Ill: SPSS Inc. Sutton, G. S. (1993). Do men grow to resemble their wives, or vice versa? Journal of Biosocial Science, 25, 25-9. Szinovacz, M. & Harpster, P. (1993). Employment status, gender role attitudes, and marital dependence in later life. Journal of Marriage and the Family, 55, 927-40. Thornton, A. & Young-DeMarco, L. (2001). Four decades of trends in attitudes toward family issues in the United States: The 1960s through the 1990s. Journal of Marriage and the Family, 63, 1009-1037. Tichenor, V. J. (1999). Status and income as gendered resources: The case of marital power. Journal of Marriage and the Family, 61, 638-650. Twenge, J. M., Campbell, W. K. & Foster, C. A. (2003). Cohabitation and marriage-parenthood and marital satisfaction: A meta-analytic review. Journal of Marriage and the Family, 65, 574-83. Tynes, S. R. (1990). Educational heterogamy and marital satisfaction between spouses. Social Science Research, 19, 153-74. US Bureau of the Census. (1998). Table A-2. Percentage of the population 3 to 34 years old enrolled in school, by age, gender, race, and Hispanic origin: October 1947 to 1966. US Bureau of the Census: Internet Release date, August 14, 1998. US Bureau of the Census. (2002). The big payoff: Educational attainment and the synthetic estimates of work-life earnings. Current Population Reports, Issued July 2002. Van Berkel, M. (1997). Who dominates when? Asymmetrical patterns of influence among Dutch husbands and wives. PhD Thesis, University of Nijmegen, Nijmegen, the Netherlands. 122 MARRIAGE & FAMILY REVIEW Van Berkel, M. & De Graaf, N. D. (1995). Husband’s and wife’s culture participation and their levels of education: A case of male dominance? Acta Sociologica, 38, 131-49. Van der Slik, F. W. P., De Graaf, N. D. & Gerris, J. R. M. (2002). Conformity to parental rules. Asymmetric influences of father’s and mother’s levels of education. European Sociological Review, 18, 489-502. Vogt, W. P. (1997). Tolerance and education, learning to live with diversity and difference. Thousand Oaks, CA: Sage Publications. West, C. & Zimmerman, D. H. (1987). Doing gender. Gender and Society, 1, 125-51. Williams, C. L. (1989). Gender differences at work: Women and men in nontraditional occupations. Berkeley/ London: University of California Press. Wink, P. & Helson, R. (1993). Personality change in women and their partners. Journal of Personality and Social Psychology, 65, 597-605. Wright, J. D. & Wright, S. R. (1976). Social class and parental values for children: A partial replication and extension of the Kohn thesis. American Sociological Review, 41, 527-537. For FACULTY/PROFESSIONALS with journal subscription recommendation authority for their institutional library . . . If you have read a reprint or photocopy of this article, would you like to make sure that your library also subscribes to this journal? If you have the authority to recommend subscriptions to your library, we will send you a free complete (print edition) sample copy for review with your librarian. 1. Fill out the form below and make sure that you type or write out clearly both the name of the journal and your own name and address. Or send your request via e-mail to docdelivery@haworthpress.com including in the subject line “Sample Copy Request” and the title of this journal. 2. Make sure to include your name and complete postal mailing address as well as your institutional/agency library name in the text of your e-mail. [Please note: we cannot mail specific journal samples, such as the issue in which a specific article appears. Sample issues are provided with the hope that you might review a possible subscription/e-subscription with your institution's librarian. There is no charge for an institution/campus-wide electronic subscription concurrent with the archival print edition subscription.] Please send me a complimentary sample of this journal: (please write complete journal title here–do not leave blank) I will show this journal to our institutional or agency library for a possible subscription. Institution/Agency Library: ______________________________________________ Name: _____________________________________________________________ Institution: __________________________________________________________ Address: ___________________________________________________________ City: ____________________ State: __________ Zip: ____________________ Return to: Sample Copy Department, The Haworth Press, Inc., 10 Alice Street, Binghamton, NY 13904-1580