Exploring gendered behavior in the field with experiments: Why public goods are provided by women in a Nairobi slum

Kenyan women are more likely to contribute funds to the greater public good in environments absent of gender and ethnic diversity.


Governments in poor countries often fail to provide adequate public goods, such as community wells, infrastructure, and schools.  As a result, communities in low-income countries often raise their own funds and resources to pay for and maintain municipal services and institutions. In the Nairobi slums, there is a strong tradition of community fundraising for public goods and services. “Harambee” (meaning “let’s pull together” in Swahili), for example, is a community fundraising mechanism that helps to raise money for public goods and services. Similarly, another community organization that operates in the Nairobi slums is ROSCA, a micro-lending institution formed by a core set of members, primarily women, who contribute to a communal fund so that loans may be granted to individual members. In both of these institutions, women tend to be more involved, often to the exclusion of men. In this study, given the high levels of female participation in such community institutions, the authors examine whether gender plays a role in the voluntary provision of public goods in a Nairobi slum. They administer a public goods game in which each participant is given money and placed in a group with other participants, then rated on how giving they were when asked to contribute to public goods.


Women’s public good contribution was the same as men’s when placed in a same-sex group, but lower than men’s in a mixed-sex group. In contrast, men contributed the same to the public good in both environments.

  • Men do not distinguish between same-sex or mixed-sex groups in terms of their expectations or contributions, giving 32% in same-sex groups and 34% in mixed-sex groups.
  • Women are equally likely to contribute in same-sex groups, 31%, but less likely to contribute in mix-sex environments, at 21%.
  • A strong positive association was seen between expectations and contributions. Men and women who expected a higher contribution from others gave a higher contribution themselves. In mixed-sex groups, women expected lower contributions from other group members, which accounted for their lower contributions in these groups.
  • Women tended to underestimate others’ contributions in mixed-sex groups, whereas men tended to overestimate the result. Women expected mixed-sex contributions to average 26% of the endowment, where men’s contributions in these groups actually averaged 35% and women’s averaged 21%. In these groups, men predicted a 37% contribution rate, while the actual rates were 22% for women and 33% for men. Women and men’s expectations were found to be generally accurate in same-sex groups.
  • Ethnic diversity also played a role in reducing expectations and contributions. In extremely heterogeneous groups, contributions were 27 Kenyan shillings lower (ksh) on average than in completely homogenous groups.

In short, women in a poor Nairobi neighborhood were far more likely to give higher contributions to the greater good in situations of greater homogeneity.  When women were in the presence of men or other tribes, they gave lower contributions based on their level of trust and expectations of what others would give.


A random sample of 270 participants was chosen from the Kwa Reuben neighborhood of Embakasi Slum in Nairobi, and their socioeconomic and demographic data was recorded. The majority of volunteers selected were already involved in “harambee,” or community programs and fundraising: 64 percent of participants belonged to a local group; 30 percent of participants volunteered within the community; and 70 percent of participants had contributed to a fundraising project. All experiments were run as single-blind trials. Ten game “workshops” in all were executed and observed, with two of each same-sex session and six mixed-sex sessions. To play the game, each participant was allotted Ksh 50, in the form of 10 coins, of which to contribute a portion or all. Each participant received a confidential number with which to bid, and an envelope in which to put the contribution for confidentiality. After contributions had been made, all respondents were asked to fill out a survey to illuminate the participants’ expectations of others and perceptions of trust during the course of the games. All participants received their earned portion of the allotted money in private after completing the games and the surveys.

Related GAP Studies