Recent analysis by Ingelhart and Norris (2003) suggests that the observed gender gap between men and women in Western societies is shifting, from women being more conservative than men in ideology, electoral preference, and political attitudes (the “traditional gender gap”) to being more liberal (the “modern gender gap”). But the same analysis challenges whether this model of the links between gender and political preferences applies well in non-Western developing societies. The present analysis draws on public opinion data gathered from face-to-face interviews with over 20,000 respondents in Round 2 of the Afrobarometer (2002-2003) to explore differences in attitudes, values and behaviors between men and women in 15 African countries. We look for gender gaps in any of four key areas: 1) regime preferences (preference for democracy vs. authoritarian systems); 2) policy preferences (pro-statist vs. pro-market); 3) performance evaluations; and 4) political behavior. We find that women differ relatively little from men with regard to their preferences for political and economic regimes, and in their performance evaluations. And in most cases, where they do differ, it is not because women stake out a fundamentally different position from men, but rather, because they are more ambivalent; they are consistently more inclined to offer a “don’t know” or “it doesn’t matter” response. The main exception: women are less convinced of the need for multiparty competition within a democracy, expressing greater concerns about the potential divisiveness of party competition, and a greater tolerance for one-party systems. With regard to political behavior, however, we find that, like their counterparts elsewhere, women in these African countries appear to be more passive than men, falling well behind them in most categories of political behavior. Using multivariate analysis, we find that part of this gender gap can be explained by structural, cognitive, cultural and especially agency factors, as well as country effects, but a gender gap in political participation remains even after controlling for these other factors.