Table 6.
Multiply imputed moderators
| Moderator | (1) MI, math | (2) MI, reading | (3) MI, RCT | (4) MI, QES |
|---|---|---|---|---|
| Proportion of girls | 0.0045 [−0.0197, 0.0288] | 0.00245 [−0.0039, 0.0088] | 0.0046 [−0.0039, 0.0130] | −0.0112 [−0.0369, 0.0145] |
| Proportion of minority students | 0.0002 [−0.0086, 0.0090] | −0.00064 [−0.0039, 0.0027] | −0.00003 [−0.0024, 0.0023] | −0.0013 [−0.0212, 0.0187] |
| Proportion of low income students | −0.0009 [−0.0120, 0.0101] | 0.00053 [−0.0042, 0.0053] | 0.00079 [−0.0035, 0.0051] | −0.0019 [−0.021, 0.017] |
| Constant | 0.333 [0.146, 0.522] | 0.177 [0.113, 0.241] | 0.167 [0.0958, 0.237] | 0.364 [0.183, 0.544] |
Note: Confidence intervals, based on robust standard errors, in brackets. The number of clusters, effect sizes, and student observations in the respective specifications are 25, 36 and 14,961 in column 1; 51, 158 and 64,230 in column 2; 49, 135 and 50,499 in column 3 and 17, 59 and 28,692 in column 4.
Abbreviations: RCT, randomised controlled trial; QES, quasiexperimental study.