Balance of Covariates: Observed Properties Uncorrelated with Program Timing
| Independent Variable | TSC Latrines Per Capita | TSC Started | ||
|---|---|---|---|---|
| Mother literate | −0.46 | (0.38) | −0.01 | (0.01) |
| Mother went to school | −0.30 | (0.44) | 0.00 | (0.01) |
| Father went to school | 0.12 | (0.43) | 0.00 | (0.01) |
| Child receives extra tuition | 0.26 | (0.46) | 0.00 | (0.01) |
| School enrollment | −0.24 | (0.32) | 0.00 | (0.00) |
| Formal house | −0.45 | (0.50) | −0.01 | (0.01) |
| Electrification | 0.20 | (0.50) | 0.00 | (0.01) |
| Village has road | 0.18 | (0.61) | −0.03† | (0.02) |
| Village has ASHA | 0.18 | (0.64) | −0.01 | (0.02) |
| Village has school | −0.12 | (0.61) | 0.00 | (0.01) |
| Village has ration shop | −0.25 | (0.69) | −0.02 | (0.02) |
Notes: Observations are the same sample of six-year-olds used in our main analysis. Each estimate presents a coefficient and a clustered standard error in parentheses from a separate regression (of the form of Equation 1) of the variables listed as the dependent variable, on either latrines per capita or an indicator for a positive number of latrines per capita (“TSC started”), with district and year fixed effects. Two-sided p-value:
↵† p < 0.1.