Gender, Coethnic Favoritism, & Cooperation in Groups: A Public Goods Experiment

• In Event: Playing Nicely with Others?: Gender, Cooperation, and Collaboration

Abstract

Do ethnic group cleavages affect the level of cooperation among women to the same extent that they do for men?

In ethnically divided societies, it is commonly believed that individuals cooperate more in same-ethnic than in mixed-ethnic group settings. A prominent explanation for this is based on a logic of coethnic favoritism that links norms of within-ethnic group cooperation to access to material and psychological rewards that are not necessarily present in mixed-ethnic interactions. I argue that this logic likely resonates with men’s experience of clientelism in ethnic political systems, but that there are reasons to expect women may behave differently from men under certain conditions.

In many places where distributive politics overlap with ethnic group membership, women are known to benefit less from access to goods and services through the kinds of clientelistic networks that tend to dominate ethnic political systems (Benstead, 2016; Wantchekon, 2003). In the absence of clear material rewards for cooperating more within ethnic groups than across them, it is unclear that women would develop the same norms of coethnic favoritism believed to evolve among men. In the absence of clear norms to guide behavior that would favor coethnics over non-coethnics, it may be that women are not less likely to cooperate in mixed-ethnic than in same-ethnic social settings.

Alternatively, it is possible that coethnic favoritism exists among women, but only for the (limited) subset of women with access to material benefits through coethnic networks or who are strongly attached to their ethnic group identity. At the same time, women’s systematic exclusion from participation in politics in ethnic political systems could create a situation where women’s shared gender group identity is more salient than it would be if gender identity did not automatically shift their experience of access to the political, social, and economic system in which they reside. Since women’s gender identity is associated with strong norms of cooperation and prosociality in social psychology (particularly relative to men) (Eagly and Wood, 1991), this suggests that not only could gender identity be more salient for women than men, but it could also be more likely to prime norms of cooperation with other women in single-gender social settings. This norm of cooperation linked to women’s gender identity could plausibly be expected to work against norms of cooperation with coethnics over non-coethnics in all-women group settings.

I evaluate these possibilities using data from a public goods experiment implemented with 120 single-gender groups in Lebanon. Individuals were randomly assigned to participate in either a homogeneous or a heterogeneous sectarian group. I then examine whether the average level of cooperation varies depending on the ethnic and/or gender composition of the group (all-men versus all-women). I find weak but suggestive evidence that men cooperate less in mixed-ethnic than in same-ethnic groups, but no evidence that women are less cooperative in mixed-ethnic compared to same-ethnic groups. I further examine the heterogeneous treatment effects for those with strong versus weak access to benefits through politicians, those with strong versus weak sectarian identity, and those with strong versus weak gender identity. In line with my expectations, I find that women and men differ in how the strength of their attachment to their sectarian identity affects their willingness to cooperate in mixed-ethnic groups compared to homogeneous ethnic groups: strong sectarian men cooperate less in mixed-groups (consistent with the literature), while strong sectarian women do not cooperate at a significantly different level in mixed-sect compared to same-sect groups. However, I also find that while strong gender identity reduces cooperation in same-sect groups for women, it is associated with higher levels of cooperation with coethnics for men. This suggests that gender-based norms may prescribe different behaviors in mixed-ethnic and same-ethnic group settings.

Most interestingly, I find that both men and women with access to benefits through politicians show more coethnic favoritism in cooperation. This is important, since it suggests that there could be key differences among women in terms of how they cooperate within ethnic political systems. Rather than being unconditionally cooperative with each other, it suggests that one potential barrier to cooperation among women could reside in the extent to which there are divisions among women in access to benefits through elite networks. I discuss the implications of these findings for policymaking and collective action across ethnic lines.

Author

• Leslie Marshall, University of Pittsburgh