Bias Corrected Effect of Shifting the Older Sibling’s Probability of Behavior from 0 to 1 on the Younger Sibling’s Probabilities of Behavior Relative to Baseline (Based on Dynamic Probit Model with both Contemporaneous and Lagged Sibling Effects)
 |  |  |  |  |  | |
|---|---|---|---|---|---|---|
| Panel A: Smoking Cigarettes | ||||||
| Baseline | 0.3314 | 0.3686 | 0.3797 | 0.3923 | 0.3872 | 0.3757 |
| (0.0101) | (0.0113) | (0.0134) | (0.0175) | (0.0231) | (0.0315) | |
Shock  | 0.0716 | 0.0182 | 0.0054 | 0.0017 | 0.0006 | 0.0002 |
| (0.0834) | (0.0218) | (0.0066) | (0.0022) | (0.0008) | (0.0003) | |
Shock  | 0.1976 | 0.1120 | 0.0329 | 0.0103 | 0.0035 | 0.0012 |
| (0.0934) | (0.0540) | (0.0163) | (0.0054) | (0.0019) | (0.0007) | |
| Shock both | 0.2715 | 0.1307 | 0.0385 | 0.0120 | 0.0041 | 0.0014 |
| (0.0970) | (0.0588) | (0.0177) | (0.0058) | (0.0021) | (0.0008) | |
| Panel B: Drinking Alcohol | ||||||
| Baseline | 0.4508 | 0.5012 | 0.5443 | 0.5711 | 0.5832 | 0.5789 |
| (0.0126) | (0.0126) | (0.0162) | (0.0202) | (0.0267) | (0.0374) | |
Shock  | 0.2241 | 0.0419 | 0.0084 | 0.0017 | 0.0004 | 0.0001 |
| (0.0538) | (0.0106) | (0.0025) | (0.0006) | (0.0002) | (0.0001) | |
Shock  | 0.0561 | 0.0340 | 0.0066 | 0.0014 | 0.0003 | 0.0001 |
| (0.0563) | (0.0320) | (0.0063) | (0.0013) | (0.0003) | (0.0001) | |
| Shock both | 0.2826 | 0.0763 | 0.0151 | 0.0031 | 0.0007 | 0.0002 |
| (0.0658) | (0.0334) | (0.0068) | (0.0015) | (0.0004) | (0.0001) | |
| Panel C: Smoking Marijuana | ||||||
| Baseline | 0.2087 | 0.2528 | 0.2755 | 0.2888 | 0.2953 | 0.2999 |
| (0.0106) | (0.0117) | (0.0164) | (0.0225) | (0.0296) | (0.0433) | |
Shock  | 0.1960 | 0.0357 | 0.0078 | 0.0018 | 0.0005 | 0.0001 |
| (0.1062) | (0.0197) | (0.0046) | (0.0012) | (0.0003) | (0.0001) | |
Shock  | 0.3571 | 0.0571 | 0.0124 | 0.0029 | 0.0007 | 0.0002 |
| (0.1319) | (0.0736) | (0.0155) | (0.0037) | (0.0010) | (0.0003) | |
| Shock both | 0.5748 | 0.0956 | 0.0208 | 0.0048 | 0.0012 | 0.0003 |
| (0.1646) | (0.0757) | (0.0160) | (0.0039) | (0.0010) | (0.0003) | |
Note: “Baseline” corresponds to probabilities of simulated behaviors using the dynamic probit model with both contemporaneous and lagged sibling effects. The next line corresponds to an exogenous shift of the older sibling’s probability of behavior from 0 to 1 in the period preceding the first observation of the younger sibling. The next line corresponds to an exogenous shift of the older sibling’s probability of behavior from 0 to 1 in the first period we observe the younger sibling. “Shock Both” correspond to an exogenous shift of the older siblings’ probability of behavior from 0 to 1 in both periods. In all three scenarios, we set the older sibling’s behavior in the periods following the shock(s) to its baseline values. The numbers reported in all the rows except “Baseline” refer to the average change in said probabilities due to the corresponding exogenous switches in older siblings’ behavior, divided by the baseline probability of these behaviors. Parametric bootstrap standard errors based on 150 replications in parentheses.