Can Early Family Formation Explain the Lower Educational Attainment of U.S. Conservative Protestants?

Using nationally representative data, this article assesses whether the timing of life course transitions (i.e., marriage and childbirth) can explain the lower educational attainment of individuals raised in conservative Protestant (CP) households. A key finding is that early family formation affects educational attainment for both white and black women raised in CP households. For white women, the timing of family formation, net of controls, reduces the negative effect of childhood CP to insignificance. For black women, the timing of family formation reduces but does not eliminate the negative effect. For black men and white men, the negative effect of childhood CP on adult educational attainment is explained away by control variables. An earlier version of this article was presented at the annual meeting of the Population Association of America, Atlanta, Georgia, 2002. We thank the special issue editors, Christopher Ellison and Darren Sherkat, and the anonymous Sociological Spectrum reviewers, and Charles Brody for their comments. Notes 1To avoid possible pejorative connotations associated with the term fundamentalist, we primarily refer to these traditions as CP in this article. 2We experimented with multiple alternate coding schemes before selecting the approach used in the present analyses. The most promising alternate approach used information identifying the year education was stopped and the year of first marriage and/or first birth. This allowed us to construct variables indicating (1) respondent's age at each of these events (if they took place) and (2) whether cessation of education took place in the year after, before or concurrent to the life course events. While this approach may have provided a more precise test of our theoretical argument, the empirical model was plagued by muliticolinearity issues resulting from the necessary use of age-related information in construction of multiple timing variables. 3Our approach conceptualizes the timing of life course events as being relative to subpopulation (i.e., gender and race categories) norms rather than individually specific. 4This index taps into beliefs about appropriate gender roles and division of labor within households. Support for traditional roles and divisions of labor will likely lead to early cessation of education for women as they transition into marriage and motherhood. Inclusion of this gender traditionalism measure allows us to isolate the effect of being raised in CP households above and beyond traditional gender beliefs. In other words, after including this measure in the model we can interpret the effects of CPism as being unique to CPs and not simply a manifestation of traditional/conservative beliefs. ∗p < .05; ∗∗p < .01; ∗∗∗p < .001. ∧Controlling for age, Hispanic, father's education and occupational prestige, mother's education and occupational prestige, family disruption, urban residence, region, and number of siblings. 5Some researchers believe that differences between CPs and the rest of society have narrowed in the later part of the twentieth century (Woodberry and Smith Citation1998) This suggests that the effects of being CP will vary by generation (i.e., socioeconomic "penalties" of CPism will be greater for older respondents than younger). We model this possibility by including an interaction term between childhood CPism and age. However, "[d]ifferences by age shown by cross-sectional data may or may not be age effects, because the people (or other entities) of different ages are members of different cohorts and may have been shaped by different formative experiences and influences" (Glenn Citation2005, p. 3). Accordingly, we experimented with creating a dichotomous cohort variable "(Pre-WWII" and "Baby Boomers"). Because age and cohort are highly correlated the inclusion of both measures produces unreliable estimates for the two variables, so we ran models that substituted cohort for age. The models using cohort exhibit the same pattern as age (complete results available upon request). The interaction term for CP and cohort are statistically significant only for black women. The effect on education, net of controls, for black women Baby Boomers who were raised CP is −.128; for members of the Pre-WWII cohort the effect is −1.45. Because (1) this pattern is consistent with the CP by age interaction reported above, and (2) the use of cohort does not change the statistical significance of any of the coefficients for key variables nor improve model fit, and (3) we have no way of determining with these data whether the relationship is ultimately the product of age, cohort or period effects, we present results using age and age interactions only. ∗p < .05; ∗∗p < .01; ∗∗∗p < .001. ∧Controlling for age, Hispanic, father's education and occupational prestige, mother's education and occupational prestige, family disruption, urban residence, region, and number of siblings. ∗p < .05; ∗∗p < .01; ∗∗∗p < .001. ∧Controlling for age, Hispanic, father's education and occupational prestige, mother's education and occupational prestige, family disruption, urban residence, region, and number of siblings. ∗p < .05; ∗∗p < .01; ∗∗∗p < .001. ∧Controlling for age, father's education and occupational prestige, mother's education and occupational prestige, family disruption, urban residence, region, and number of siblings. ∗∗p < .001; ∗p < .05. Pearson Chi-Square (Asymptotic Significance, two-sided). 6A reviewer suggested that we also examine whether CPs marry more quickly than others after completing schooling—perhaps even getting married more frequently before they leave school. If so, this might provide evidence that the primary reason for school cessation is (anticipated) family formation. In response to this suggestion we calculated the time lag between age at first marriage and age at education cessation and tested whether this time lag was different for CPs than others within each of the gender and racial categories. Unfortunately, education cessation is only reported by year (as opposed to specific month or date) making our measure fairly crude. Statistically significant differences emerged only for black women and black men. The average time lag between getting married and stopping schooling for black CP women is .27 year and it is −1.52 years for black non-CP women. In other words, black CP women, on average, wait until after completing schooling to get married while the black non-CP women tend to marry before completing education. This is consistent with the interpretation that marriage and children are defining events for CPs that propel them into adult responsibilities that preclude further education. In contrast, both CP and non-CP black men tend to get married after the cessation of schooling. The average time lag for black CP men is 3.04 years and 1.20 years for black non-CP men.