Students’ and Parents’ Pre-Treatment Characteristics
| Obs. | Treatment Mean (μT) | Control Mean (μC) | p-Value of Adj. Diff. | |
|---|---|---|---|---|
| (1) | (2) | (3) | (4) | |
| Panel A: Administrative Records | ||||
| Female | 1,066 | 0.45 | 0.47 | 0.57 |
| Age | 1,066 | 9.81 | 9.79 | 0.41 |
| New student | 1,066 | 0.08 | 0.07 | 0.42 |
| Language grade | 976 | 5.10 | 5.07 | 0.85 |
| Math grade | 976 | 5.14 | 5.19 | 0.37 |
| Final avg. grade | 976 | 5.57 | 5.59 | 0.47 |
| Attendance rate | 976 | 0.89 | 0.89 | 0.53 |
| Passed grade | 1,018 | 0.95 | 0.96 | 0.57 |
| At-risk index (standardized) | 1,066 | 0.05 | 0.00 | 0.35 |
| Missing grades/attendance/pass data | 1,066 | 0.09 | 0.08 | 0.41 |
| Multiple hypotheses Wald test | 0.72 | |||
| Panel B: Parents’ Survey Data | ||||
| Standardized scales (μC = 0, σC = 1) | ||||
| Study habits | 704 | –0.07 | 0.00 | 0.51 |
| Academic efficiency | 730 | –0.09 | 0.00 | 0.16 |
| Family support | 739 | –0.12 | 0.00 | 0.06 |
| Low family supervision | 709 | –0.06 | 0.00 | 0.72 |
| Parent school involvement | 716 | –0.01 | 0.00 | 0.66 |
| Positive reinforcement | 738 | –0.06 | 0.00 | 0.31 |
| Parent scales index | 773 | –0.06 | 0.00 | 0.21 |
| Mother completed high school | 774 | 0.53 | 0.49 | 0.78 |
| Missing baseline survey | 1,066 | 0.26 | 0.27 | 0.59 |
| Multiple hypotheses Wald test | 0.39 | |||
| Panel C: Students’ Survey Data | ||||
| Standardized scales (μC = 0, σC = 1) | ||||
| Study habits | 909 | –0.19 | 0.00 | 0.10 |
| Academic efficiency | 915 | –0.14 | 0.00 | 0.15 |
| Family support | 864 | –0.15 | 0.00 | 0.12 |
| Low family supervision | 859 | 0.05 | 0.00 | 0.60 |
| Parent school involvement | 858 | –0.12 | 0.00 | 0.59 |
| Positive reinforcement | 868 | –0.04 | 0.00 | 0.90 |
| Student scales index | 962 | –0.17 | 0.00 | 0.15 |
| Missing baseline survey | 1,066 | 0.08 | 0.09 | 0.84 |
| Multiple hypotheses Wald test | 0.11 | |||
Notes: Column 1 shows the number of observations with nonmissing data, Columns 2 and 3 show the mean value of each baseline characteristic observations.in the treated and control group, respectively. Column 4 reports the p-value on the treatment coefficient in a regression using each baseline characteristic as the dependent variable. All tests adjust for classroom fixed effects, and robust standard errors are clustered at this level. Parent and student scales index are simple scales’ averages that were standardized using the control mean and standard deviation so that standardized scales for the control group have a mean μC = 0 and a standard deviation σC = 1. Observable variables in Panel A correspond to 2013 except for the new student variable that refers to 2014. The rows “Multiple hypotheses Wald test” reports the p-value of a joint test of the null that all the differences in means of the variables reported in each panel (of treated and control students) are zero. We exclude from this test the variable that reports the proportion of missing observations.