Table 1.
(a) Baseline characteristics difference between treatment and control groups by province. (b) Baseline characteristics difference between treatment and control groups using pooled sample.
| Panle (a) by Province | ||||||
| Gansu (n = 762) | Shaanxi (n = 1325) | |||||
| Variables | Control Group | Treatment Group | Difference |
Control
Group |
Treatment Group | Difference |
| (1) | (2) | (3) | (4) | (5) | (6) | |
| Student: | ||||||
| Grade (1 = fourth) | 4.607 | 4.637 | 0.03 | 4.604 | 4.58 | −0.024 |
| (0.489) | (0.481) | (0.041) | (0.489) | (0.494) | (0.029) | |
| Male sex (1 = yes) | 0.509 | 0.525 | 0.016 | 0.498 | 0.472 | −0.026 |
| (0.500) | (0.500) | (0.036) | (0.500) | (0.500) | (0.031) | |
| LogMAR | 0.501 | 0.489 | −0.012 | 0.519 | 0.524 | 0.006 |
| (0.257) | (0.241) | (0.025) | (0.23) | (0.210) | (0.018) | |
| Not wearing eyeglasses at baseline (1 = yes) | 0.887 | 0.859 | −0.028 | 0.834 | 0.83 | −0.004 |
| (0.318) | (0.348) | (0.022) | (0.372) | (0.376) | (0.031) | |
| Baseline standardized math score | 0.137 | 0.149 | −0.012 | 0.290 | 0.360 | −0.070 |
| (0.100) | (0.945) | (0.070) | (1.007) | (0.974) | (0.055) | |
| Family: | ||||||
| One or both parents with ≥12 years of education (1 = Yes) | 0.217 | 0.252 | 0.035 | 0.193 | 0.208 | 0.015 |
| (0.412) | (0.434) | (0.037) | (0.394) | (0.403) | (0.027) | |
| Family asset value, 10 thousand | 1.433 | 1.539 | 0.105 | 3.709 | 4.25 | 0.541 |
| (1.832) | (1.963) | (0.136) | (3.547) | (3.912) | (0.435) | |
| One or both parents out-migrated for work (1 = yes) | 0.517 | 0.53 | 0.013 | 0.438 | 0.419 | −0.019 |
| (0.500) | (0.5) | (0.049) | (0.497) | (0.494) | (0.042) | |
| Teacher & Schools: | ||||||
| Teachers finished college or higher education |
0.478 | 0.496 | 0.019 | 0.516 | 0.521 | 0.005 |
| (0.500) | (0.500) | (0.100) | (0.500) | (0.500) | (0.095) | |
| Share of teachers that held professional titles | 0.989 | 0.836 | −0.153 * | 0.917 | 0.853 | −0.064 |
| (0.429) | (0.212) | (0.073) | (0.134) | (0.192) | (0.038) | |
| Class size | 47.358 | 47.021 | −0.338 | 42.286 | 44.284 | 1.998 |
| (10.281) | (11.196) | (2.437) | (11.738) | (13.133) | (3.034) | |
| Half and above material taught on blackboard (1 = yes) | 0.879 | 0.783 | −0.095 | 0.638 | 0.501 | −0.137 |
| (0.327) | (0.413) | (0.079) | (0.481) | (0.500) | (0097) | |
| Number of observations | 379 | 383 | - | 609 | 716 | - |
| Panel (b) Pooled Sample | ||||||
| Pooled Sample (n = 2087) | ||||||
| Variables | Control Group | Treatment Group | Difference | |||
| (1) | (2) | (3) | ||||
| Student: | ||||||
| Grade (1 = fourth) | 4.605 | 4.600 | −0.005 | |||
| (0.489) | (0.49) | (0.024) | ||||
| Male sex (1 = yes) | 0.502 | 0.490 | −0.012 | |||
| (0.500) | (0.500) | (0.024) | ||||
| LogMAR | 0.512 | 0.512 | 0.000 | |||
| (0.241) | (0.222) | (0.015) | ||||
| Not wearing eyeglasses at baseline (1 = yes) | 0.854 | 0.840 | −0.014 | |||
| (0.353) | (0.367) | (0.022) | ||||
| Baseline standardized math score | 0.231 | 0.286 | 0.055 | |||
| (1.006) | (0.969) | (0.043) | ||||
| Family: | ||||||
| One or both parents with ≥12 years of education (1 = yes) | 0.202 | 0.223 | 0.021 | |||
| (0.401) | (0.415) | (0.022) | ||||
| Family asset value, 10 thousand | 2.836 | 3.305 | 0.469 | |||
| (3.204) | (3.602) | (0.378) | ||||
| One or both parents out-migrated for work (1 = yes) |
0.468 | 0.458 | −0.010 | |||
| (0.499) | (0.498) | (0.033) | ||||
| Teacher & Schools: | ||||||
| Teachers finished college or higher education | 0.501 | 0.512 | 0.011 | |||
| (0.500) | (0.500) | (0.070) | ||||
| Share of teachers that held professional titles | 0.945 | 0.847 | −0.098 ** | |||
| (0.288) | (0.199) | (0.037) | ||||
| Class size | 44.232 | 45.237 | 1.005 | |||
| (11.465) | (12.555) | (2.113) | ||||
| Half and above material taught on blackboard (1 = yes) | 0.729 | 0.611 | −0.118 | |||
| (0.444) | (0.495) | (0.042) | ||||
| Number of observations | 988 | 1099 | - | |||
(a) Notes: Columns 1 to 2 and 4 to 5 present means with standard deviations reported in brackets. Columns 3 and 6 show coefficients estimated by regressing each variable on the treatment dummy, with standard errors, in parentheses, clustered at the school level. * p < 0.10. (b) Notes: Columns 1 to 2 present means with standard deviations reported in brackets. Column 3 shows coefficients estimated by regressing each variable on the treatment dummy, with standard errors, in parentheses, clustered at the school level. ** p < 0.05.