Table 4.
Robustness Checks: Dealing with Outliers of BMI
| Model | (1) Ordered Probit | (2) Ordered Probit | (3) Ordered Probit | (4) Ordered Probit | (5) Ordered Probit |
|---|---|---|---|---|---|
| Variable | Perceived depression | Perceived depression | Perceived depression | Perceived depression | Perceived depression |
| BMI_win_03 | −0.133*** (0.023) | ||||
| BMI_win_03_squared | 0.003*** (0.000) | ||||
| BMI_win_05 | −0.148*** (0.029) | ||||
| BMI_win_05 _squared | 0.003*** (0.001) | ||||
| BMI_del_03 | −0.079** (0.031) | ||||
| BMI_del_03 _squared | 0.001** (0.001) | ||||
| BMI_del_05 | −0.080** (0.040) | ||||
| BMI_del_05 _squared | 0.002* (0.001) | ||||
| Ln_BMI | −5.217*** (0.840) | ||||
| Ln_BMI _squared | 0.817*** (0.135) | ||||
| All control variables | Yes | Yes | Yes | Yes | Yes |
| Observations | 23,515 | 23,515 | 22,104 | 21,163 | 23,515 |
| Pseudo R2 | 0.030 | 0.030 | 0.029 | 0.028 | 0.030 |
Notes: ***, **, and * Indicate significance at the levels of 1%, 5%, and 10%, respectively. The values in parentheses are standard errors robust to heteroskedasticity. Yes means the corresponding variables are controlled in the regression. Pseudo R2 is reported for nonlinear models to assess goodness of fit.