Table 1.
Average marginal effects (AME) of coverage and compliance for friendship degree Heckman model.
A) First stage dependent variable: coverage |
B) Second stage dependent variable: compliance |
|||||||||
---|---|---|---|---|---|---|---|---|---|---|
Predictors | AME | Clustered Std. Err. | p-value | 95% CI | AME | Coverage=1a | Clustered Std. Err. | p-value | 95% CI | ||
LN(Friendship degree +1), [0, 3.526] | 0.067 | 0.016 | <0.001 | 0.036 | 0.099 | |||||
LN(Friendship degree +1), [>3.526] | 0.303 | 0.232 | 0.193 | −0.153 | 0.758 | |||||
Friendship betweenness centralityb | −1.513 | 1.073 | 0.158 | −3.616 | 0.590 | |||||
Health betweenness centrality [0, 0.268] | 0.420 | 0.350 | 0.231 | −0.267 | 1.107 | |||||
Health betweenness centrality [>0.268] | 0.520 | 0.769 | 0.499 | −0.988 | 2.028 | |||||
Close friends with CMD | −0.004 | 0.020 | 0.846 | −0.042 | 0.035 | 0.021 | 0.017 | 0.224 | −0.013 | 0.054 |
Trusts CMD for health advice | 0.072 | 0.016 | <0.001 | 0.040 | 0.104 | 0.048 | 0.015 | 0.001 | 0.019 | 0.077 |
Age in years | 0.001 | <0.001 | 0.003 | <0.001 | 0.002 | <0.001 | <0.001 | 0.546 | <-0.001 | 0.001 |
Female | 0.011 | 0.008 | 0.179 | −0.005 | 0.027 | 0.011 | 0.008 | 0.143 | −0.004 | 0.026 |
Social status: was or is a religious, tribe, or clan leader or on the local village council | 0.074 | 0.026 | 0.005 | 0.022 | 0.125 | 0.018 | 0.025 | 0.476 | −0.031 | 0.066 |
Fisherman or fishmonger | −0.018 | 0.019 | 0.338 | −0.055 | 0.019 | −0.002 | 0.016 | 0.907 | −0.034 | 0.030 |
Other income-earning occupation | 0.019 | 0.013 | 0.159 | −0.007 | 0.045 | 0.027 | 0.013 | 0.029 | 0.003 | 0.052 |
Highest level of education attained | −0.009 | 0.002 | <0.001 | −0.012 | −0.006 | −0.006 | 0.002 | <0.001 | −0.009 | −0.003 |
Muslim household head | −0.049 | 0.017 | 0.005 | −0.082 | −0.015 | −0.021 | 0.015 | 0.149 | −0.050 | 0.008 |
Household head belongs to village majority tribe | 0.029 | 0.016 | 0.069 | −0.002 | 0.060 | 0.038 | 0.014 | 0.008 | 0.010 | 0.066 |
No home latrine | −0.081 | 0.028 | 0.004 | −0.136 | −0.026 | 0.001 | 0.024 | 0.974 | −0.047 | 0.049 |
Home owned, not rented | 0.008 | 0.024 | 0.737 | −0.039 | 0.055 | 0.002 | 0.023 | 0.934 | −0.043 | 0.047 |
Home quality score | 0.001 | 0.002 | 0.559 | −0.003 | 0.006 | −0.001 | 0.002 | 0.709 | −0.005 | 0.004 |
Village centre greater than 0.50 km from Lake Victoria | −0.160 | 0.017 | <0.001 | −0.193 | −0.127 | |||||
Total homes in village |
−0.001 |
<0.001 |
<0.001 |
−0.001 |
−0.001 |
|||||
Obs. 14431 | ||||||||||
Households 3387 | ||||||||||
Inverse hyperbolic tangent of | −1.348 | 0.241 | <0.001 | −1.821 | −0.875 | |||||
−0.874 | 0.057 | −0.949 | −0.704 | |||||||
Wald test of independent equations, Chi2 | 31.100 | <0.001 |
Average marginal probability of compliance = 1 conditional on coverage (being offered treatment).
Entered in the Heckman selection model as grand-mean centered and a quadratic, i.e. also including the squared form.