Proportion of the Correlation Between Siblings’ Behavior Explained by the Direct Effect (Based on the Joint Dynamic Probit Model)
| Estimates of the Effect of Old Sibling’s Behavior on Younger Sibling’s Behavior at the Same Age | Proportion of the Coefficient Due to Sibling Effect, No bias correction (5) | Proportion of the Coefficient Due to Sibling Effect with Bias Correction (6) | ||||
|---|---|---|---|---|---|---|
| NLSY 97 Data (1) | Simulated Data, Uncorrected Sibling Influence (2) | Simulated Data, Bias Corrected Sibling Influence (3) | Simulated Data, Zero Sibling Influence (4) | |||
| No controls | ||||||
| Smoking cigarettes | 0.226*** | 0.208 | 0.209 | 0.192 | 0.080 | 0.082 |
| (0.011) | (0.016) | (0.016) | (0.019) | 0.001 | (0.065) | |
| Drinking alcohol | 0.259*** | 0.255 | 0.254 | 0.243 | 0.046 | 0.043 |
| (0.012) | (0.012) | (0.012) | (0.013) | 0.034 | (0.034) | |
| Smoking marijuana | 0.157*** | 0.145 | 0.145 | 0.145 | 0.001 | 0.004 |
| (0.011) | (0.015) | (0.015) | (0.017) | 0.085 | (0.084) | |
| Using hard drugs | 0.092*** | 0.082 | 0.084 | 0.070 | 0.143 | 0.167 |
| (0.012) | (0.018) | (0.019) | (0.022) | (0.222) | (0.215) | |
| With controls | ||||||
| Smoking cigarettes | 0.173*** | 0.171 | 0.172 | 0.154 | 0.098 | 0.100 |
| (0.011) | (0.016) | (0.016) | (0.019) | 0.079 | (0.079) | |
| Drinking alcohol | 0.173*** | 0.168 | 0.167 | 0.153 | 0.092 | 0.084 |
| (0.012) | (0.012) | (0.012) | (0.013) | 0.051 | (0.052) | |
| Smoking marijuana | 0.135*** | 0.129 | 0.130 | 0.129 | 0.003 | 0.005 |
| (0.011) | (0.015) | (0.015) | (0.017) | 0.093 | (0.093) | |
| Using hard drugs | 0.076*** | 0.072 | 0.074 | 0.060 | 0.162 | 0.189 |
| (0.012) | (0.018) | (0.018) | (0.021) | (0.252) | (0.243) | |
Note: In Column 1, standard errors are clustered at the household level. Parametric bootstrap standard errors in Columns 2–4 based on 150 replications. The controls consist of male, black, and Hispanic dummies, younger sibling’s age dummies, highest grade completed by age 19, mother’s highest grade completed, AFQT percentile score, number of siblings, birth order dummies, whether the respondent reported that her house had been broken in by age 12, whether she reported that she had been a victim of bullying by age 12, whether she reported having witnessed a shooting by age 12, and whether she lived with both biological parents at age 12. The numbers in Column 5 are obtained by taking the difference between the numbers in Column 2 and Column 4 and dividing by Column 2. The numbers in Column 6 are obtained by taking the different between numbers in Column 3 and Column 4and dividing by Column 3.