Robustness Checks, Using Data from Firstborn Females
| High School Completion | University Admission | |||||||
|---|---|---|---|---|---|---|---|---|
| Baseline Estimates (1) | Add Birthweight Percentile (2) | Add Gestation Length (3) | Add Birth Spacing (4) | Baseline Estimates (5) | Add Birthweight Percentile (6) | Add Gestation Length (7) | Add Birth Spacing (8) | |
| A next brother (D) | −0.0651 (0.0261)* | −0.0529 (0.0260)* | −0.0598 (0.0272)* | −0.0641 (0.0257)* | −0.0429 (0.0235) | −0.0325 (0.0235) | −0.0354 (0.0246) | −0.0420 (0.0232) |
| Morethan2 (M) | −0.0955 (0.0342)* | −0.0685 (0.0344) | −0.0963 (0.0354)* | −0.0737 (0.0333)* | −0.0669 (0.0309)* | −0.0446 (0.0311) | −0.0609 (0.0320) | −0.0652 (0.0301)* |
| A next brother (D) × Morethan2 (M) | 0.0960 (0.0401)* | 0.0829 (0.0401)* | 0.0836 (0.0419)* | 0.0748 (0.0398) | 0.0636 (0.0362) | 0.0520 (0.0363) | 0.0497 (0.0379) | 0.0624 (0.0359) |
| Mean birthweight percentile of 2nd birth | 0.0210 (0.0023)* | 0.0176 (0.0021)* | ||||||
| Gestational length of 2nd birth | −0.0004 (0.0007) | −0.0001 (0.0006) | ||||||
| Ln (birth spacing between first 2 births) | −0.0071 (0.0044) | −0.0074 (0.0040) | ||||||
| Decomposition | ||||||||
| Average total effect | 0.0019 (0.0012) | 0.0019 (0.0013) | 0.0014 (0.0013) | 0.0020 (0.0012) | 0.0023 (0.0011)* | 0.0022 (0.0011)* | 0.0018 (0.0011) | 0.0023 (0.0011)* |
| Average indirect effect | 0.0211 (0.0075)* | 0.0151 (0.0076) | 0.0212 (0.0078)* | 0.0207 (0.0073)* | 0.0148 (0.0068)* | 0.0098 (0.0069) | 0.0134 (0.0071) | 0.0144 (0.0066)* |
| Average direct effect | −0.0192 (0.0076)* | −0.0133 (0.0076) | −0.0199 (0.0079)* | −0.0187 (0.0074)* | −0.0125 (0.0069) | −0.0076 (0.0069) | −0.0116 (0.0072) | −0.0121 (0.0067) |
| Sample mean | 0.246 | 0.248 | 0.247 | 0.246 | 0.177 | 0.178 | 0.178 | 0.177 |
| Sample size | 416,315 | 410,203 | 403,775 | 415,897 | 416,315 | 410,203 | 403,775 | 415,897 |
Notes: This table reports robustness checks on the 2SLS results for the educational outcome equation and the decomposed effects of sibling gender on firstborn females. Though not reported here, we find no evidence of nonlinearity in gestation length. Since boys are heavier than girls at birth on average, we use gender-specific birthweight percentiles and calculate the mean percentile of secondborn twins. We control implicitly for the full set of dummies for urban, the subject’s age and district of birth, parents’ education and years of birth, and maternal age at the first birth. Robust standard errors are reported in parentheses. Estimates for firstborn males are reported in Appendix Table A7.