This article examines trends in divorce attitudes of young adult women in the United States by educational attainment from 1974 to 2002. Women with 4-year college degrees, who previously had the most permissive attitudes toward divorce, have become more restrictive in their attitudes toward divorce than high school graduates and women with some college education, whereas women with no high school diplomas have increasingly permissive attitudes toward divorce. We examine this educational crossover in divorce attitudes in the context of variables correlated with women's educational attainment, including family attitudes and religion, income and occupational prestige, and family structure. We conclude that the educational crossover in divorce attitudes is associated most strongly with work and family structure variables.
Key Words: attitude change, divorce, educational attainment, family demography, family inequality.
The spread of nontraditional family values in the United States has slowed since the 1970s, and some observers now argue that a shift back to conservative attitudes is under way, portending significant change in family behaviors and outcomes. Blankenhom (2002) summarizes this trend by asserting that:
(O)n the core social question of whether family fragmentation is a bad thing or a not-so-bad thing, a steady shift in popular and (especially) elite opinion took place over the course of the 1990s. Denial and happy talk about the consequences of nuclear family decline became decidedly less widespread; concern and even alarm became much more common. As a society we changed our minds, and as a result we changed some of our laws. And now, it seems, we are beginning to change some of our personal behavior. This is very encouraging news.
Blankenhorn's argument is particularly salient in light of growing evidence of class differences in family fragmentation. Two-parent families remain the norm among highly educated couples but are increasingly uncommon among less educated couples (McLanahan, 2004). Since the 1970s, marital and union dissolution rates (Raley & Bumpass, 2003) have diverged by women's educational attainment, as has the prevalence of divorce among mothers (Elwood & Jencks, 2004). If attitudes toward divorce diverged across women's educational levels at the same time as divorce behaviors were diverging, then such a pattern might help researchers and policymakers understand divergence in family outcomes.
No previous studies have looked for growing class differences in attitudes toward divorce. Overall, divorce attitudes have stabilized in recent decades after a period of liberalization (Axinn & Thornton, 2000; Thornton & Young-DeMarco, 2001), but no studies have studied trends for separate educational groups. Also, the causal relationships between divorce attitudes and divorce behavior are not clear. There is evidence that divorce attitudes can affect marital behavior as Blankenhorn suggested (Amato & Rogers, 1999), but the causal relationships between family attitudes and behavior can be complex and endogenous (Rindfuss, Brewster, & Kavee, 1996).
This analysis has two parts. In the first part, we determine whether women's attitudes toward divorce are indeed diverging by educational attainment. In the second part, we use a multivariate analysis to identify variables for social values, women's work, and family structure that correlate with educational trends in attitudes toward divorce. We avoid drawing conclusive causal inferences from the multivariate analysis, but we offer arguments about the extent to which attitudes toward divorce might be independent of variables linked to divorce behavior.
Trends in Education and Attitudes Toward Divorce
An analysis of trends in education and attitudes toward divorce faces some initial difficulties. The overall level of educational attainment has changed in recent decades, as have divorce rates and laws governing divorce. …