This paper finds that a randomly assigned civic education course in Mali widened the gender gap, when it increased civic activity among men while decreasing that among women. Qualitative evidence reveals mechanisms by which the intervention generated perverse consequences for women.
In a place where women are traditionally unwelcome actors in the public sphere, the course heightened the salience of civic activity, thus increasing social costs for female participators. Women report implicit and explicit threats of sanctions from male relatives and village elders. The intervention did, however, work to close the gender gap in civic and political knowledge. Together, these findings suggest that information asymmetries constrain civic participation, but information alone cannot overcome discriminatory gender norms – and may even exacerbate them.
Key findings:
- Examining the impact of an information intervention on civic engagement demonstrates that the provision of civic and political information is insufficient to improve aggregate levels of civic participation. While information deficits proved a surmountable obstacle to participation among men, women faced additional constraints to civic participation that overshadowed any informational benefit they might have received. In fact, the particular information intervention actually increased existing gender disparities in civic participation in Mali. This perverse effect is caused, in part, by the intervention making civic participation more salient and thus more costly for women.
- Using a decision-theoretic framework, the study derived two explanations for gender disparity in civic participation. First, differential costs – both material and social – impede women and other marginalized groups from participating. Second, civic and political information deficits are more likely to affect women because of reduced access to information and lower levels of education. The second of these explanations was tested with a field experiment in which random assignment of a civic education treatment allowed for rigorous identification of the effect of information on civic engagement. The experiment yielded no significant effects in the aggregate and even decreased civic activity among women. The error in prediction comes from treating the decision to engage in civic activity as a choice-theoretic problem without recognizing the strategic aspects and preferences of other actors.
- To better understand these strategic aspects, I use qualitative evidence on men and women’s responses to treatment. In the rural setting in which the information intervention was rolled out, women are typically unwelcome actors in the public sphere. Religion and custom prescribe a separate and unequal role for women in the community and civic and political activity fall outside the domain of women. The civics course made civic participation more salient and thus more costly for women. Women report both implicit and explicit threats of sanctioning in response to attempts to engage in civic activity following treatment.
- Higher social costs dampening women’s civic participation is a more general phenomenon. Determinants of gender discrimination in the Malian context help identify scope conditions and explain cross-country variation in gender disparity on the African continent. An analysis of why the four outlier villages in the qualitative study exhibit less gender discrimination than the rest of the sample generates several explanations: underdevelopment, strength of cultural traditions, and strength of Islam. Proxy measures underdevelopment and strength of Islam are strongly correlated with a greater gender differential in civic participation among the 19 countries in the Afrobarometer.
- The divergent effects of the civics course render a nuanced policy prescription. On the one hand, an information intervention increased civic participation among a higher status group, male leaders, and raised civic and political knowledge among both men and women. On the other hand, treatment perversely affected civic participation among lower status groups, in this case, women.
- Evidence of normatively divergent impacts of democracy interventions from this and other studies suggests future attempts should be more attuned to the social constraints faced by women and other marginalized groups. One immediate and low-cost remedy would be to deliver democracy promotion interventions to women only or to women and men separately – an idea advocated by men in treatment villages who said custom and religion are intolerant of the mixing of genders in public. This will not, however, mitigate the deeper issue of repressive gender norms. That a civic education course failed to instil democratic ideals such as equal opportunity casts doubt on the ability of brief external interventions to alter social norms. Such a shift will likely require a combination of grassroots mobilization and longer-term investment from outside actors.