Fig. 2.

a Availability of water in the household at 4 survey times: baseline, 12, 24, and 48 months. Educational levels are indicated by the following categories: None = none or basic, Primary = some or completed primary school, Secondary = some or completed secondary school (or beyond). Women’s educational level related to water availability at baseline (p = .004), and after 4 years (for the three educational levels: 47% vs. 23% and 19%, p < .0001). Likewise, water accessibility improved more in households where men had achieved at least some secondary education compared to households where men had less education (secondary education from 25% to 39%, primary education from 14% to 24%, no education from 12% to 15%, p < .0001). *p < .05, **p < .01, ***p < .001, ^p < .0001. b Toilet availability in the household at 4 survey times: baseline, 12, 24, and 48 months. Availability of an improved toilet increased in all three education groups for both women and men, but was highest in the best educated group at baseline (both, p < .0001). Households with the best educated women were more likely to have access to an improved toilet at each time point (baseline, p < .0001, 12 months p = .002, 24 months p = .002, 48 months p < .0001). In contrast, access men’s education only related to availability of improved toilets at baseline