Effect estimates
| Model . | Exposure . | Outcome . | Effect Estimate (95% confidence interval) . | Wild Bootstrap p Value . |
|---|---|---|---|---|
| Model 1: Correlation | Years of schooling | Overweight or obese | 0.03 | <.001 |
| (0.02 to 0.03) | ||||
| Model 2: First Stage | UPE eligibilitya | Years of schooling | 1.40 | .01 |
| (0.51 to 2.30) | ||||
| Model 3: First Stage | UPE × new schoolsb | Years of schooling | 1.05 | .004 |
| (0.12 to 1.93) | ||||
| Model 4: ITT | UPE eligibility | Overweight or obese | 0.02 | .76 |
| (−0.09 to 0.12) | ||||
| Model 5: 2SLS, UPE Eligibility | Years of schooling | Overweight or obese | 0.04 | .13 |
| (−0.01 to 0.09) | ||||
| Model 6: 2SLS, UPE × New Schools | Years of schooling | Overweight or obese | 0.06 | .60 |
| (−0.04 to 0.13) |
| Model . | Exposure . | Outcome . | Effect Estimate (95% confidence interval) . | Wild Bootstrap p Value . |
|---|---|---|---|---|
| Model 1: Correlation | Years of schooling | Overweight or obese | 0.03 | <.001 |
| (0.02 to 0.03) | ||||
| Model 2: First Stage | UPE eligibilitya | Years of schooling | 1.40 | .01 |
| (0.51 to 2.30) | ||||
| Model 3: First Stage | UPE × new schoolsb | Years of schooling | 1.05 | .004 |
| (0.12 to 1.93) | ||||
| Model 4: ITT | UPE eligibility | Overweight or obese | 0.02 | .76 |
| (−0.09 to 0.12) | ||||
| Model 5: 2SLS, UPE Eligibility | Years of schooling | Overweight or obese | 0.04 | .13 |
| (−0.01 to 0.09) | ||||
| Model 6: 2SLS, UPE × New Schools | Years of schooling | Overweight or obese | 0.06 | .60 |
| (−0.04 to 0.13) |
Notes: All models are estimated using DHS sample weights. All models control for age, birth-year fixed effects, and state fixed effects. Models 1 is estimated per Eq. (1). Model 2 is estimated per Eq. (2). Model 3 is per Eq. (2) but substitutes UPE eligibility with UPE × new schools. Model 4 is per Eq. (1) but substitutes schooling with UPE eligibility status. Models 5 and 6 are per Eq. (3) but with different instruments in the first-stage Eq. (2). All models control for a continuous measure of age; the rate of female primary school enrollment in 1970; and survey, state, and birth-year fixed effects. Standard errors used to calculate 95% confidence intervals are clustered at the state level. Wild cluster bootstrap p values are calculated using procedures described in Cameron et al. (2008). Residuals are repeatedly resampled by cluster to form a pseudo-dependent variable, and the model is estimated for each resampled data set. The p values are the proportion of bootstrapped t statistics that are at least as large as the t value from the original model.
Source: Demographic Health Surveys Nigeria (2003, 2008, 2013).
UPE eligibility is based on birth year.
The coefficient of UPE eligibility is interacted with new schools per capita in the respondent's state.