Meanwhile, we also hypothesized that gender gaps should emerge—across all subgroups of African Americans regardless of linked fate—on issues that invoke a national, American identity, rather than the African-American identity more relevant to issues of domestic politics. In issues of international affairs, black men can identify, as Americans, with a group that benefits from existing inequalities in the global arena. This shift in relevant identities should bring out gender gaps among African Americans.

As can be seen in Table 3, gender gaps were found among both whites and blacks in attitudes toward military-related issues. Greater male support for military spending and military involvement occurred among both white and black Americans across a range of time periods and different issues. In those rare cases when a gap in attitudes was either minimal or not evident, we find that this holds for both white and black respondents.

This inspection of issue-based gaps for black and white respondents also undercuts the argument that the gender gap comes from the Republican emphasis on and association with war. Gender gaps are fairly consistent across blacks and whites on military issues, while the partisan gender gap is, of course, predominantly found among whites.

While our alternative explanation for the gender gap garners support from our analysis, there is clearly still much to be done here. In particular, we need to be able to address change over time: Why did the gender gap emerge only when it did—in 1964 for vote choice and 1980 for partisan identification? We have so far developed only a speculative response to this question, though we strongly suspect that it involves changes in the parties’ positions on civil rights in the 1960s, as well as declines in union membership and other social bases of male Democratic identification from the days of the New Deal coalition.

We hope, however, that we have steered the conversation toward explaining the political behavior of white men versus that of women, and that men’s egalitarianism and social and political identities are fruitful at least as directions of further inquiry into the causes of the gender gap. And, while much remains to be done in hashing out the right combination of identity and inegalitarianism, we would argue that our findings of gender gaps among African Americans who are low in linked fate, and among all African Americans on issues of war and defense, suggest there is something in this notion that is worth further study.

On less speculative ground, we conclude that explanations of the gender gap should move away from the compassion and self-interest theories. A compassion-based gender gap cannot account for African-American male opinion or changing issue attitudes over time; a self-interest or state dependency explanation cannot explain that the gender gap is preeminently an upper-class phenomenon. Future research should, as we have demonstrated, incorporate data on African Americans and other sets of respondents and issues that do not exhibit a gender gap, and attempt to use this additional variation as leverage toward a better explanation of the phenomenon. Failing to do so has helped lead explorations of the gender gap down less profitable paths of inquiry.