Estimated Effect of Earning the More Positive Performance Label at Different Cutoffs and on Different Outcomes for Urban, Low-Income Students Scoring near the Cut Point
| Grade of Test | Needs Improvement/Warning | Proficient/Needs Improvement | Advanced/Proficient |
|---|---|---|---|
| Panel I: College Attendance | |||
| Eighth grade | 0.021~ (0.009) | 0.001 (0.028) | 0.007 (0.035) |
| h = 3 | h = 8 | h = 4 | |
| 5,801 | 6,313 | 1,248 | |
| Tenth grade | N/A | 0.008 (0.010) | 0.051* (0.020) |
| h = 6 | h = 8 | ||
| 8,280 | 4,171 | ||
| Panel II: High School Graduation (in four years) | |||
| Eighth grade | 0.028*** (0.004) | 0.019 (0.014) | −0.006 (0.023) |
| h = 5 | h = 6 | h = 10 | |
| 13,832 | 7,517 | 4,962 | |
| Tenth grade | N/A | 0.021 (0.014) | −0.002 (0.012) |
| h = 3 | h = 9 | ||
| 6,035 | 7,103 | ||
| Panel III: Tenth Grade Mathematics Scores | |||
| Eighth grade | 0.032** (0.003) | −0.006 (0.019) | 0.022~ (0.010) |
| h = 2 | h = 6 | h = 6 | |
| 7,974 | 11,695 | 5,155 | |
Notes:
↵~ p < 0.10;
↵* p < 0.05;
↵** p < 0.01;
*** p < 0.001. Cell entries include the parameter estimate, standard error (in parentheses), optimal bandwidth used, sample size, and approximate p-value. All inferences from two-tailed hypothesis tests. Estimated effects from a local linear regression-discontinuity model from Equation 1 using observations within one bandwidth on either side of the cutoff, with the following control predictors: student race, gender, whether the student was new to the state, was currently or formerly classified as limited English proficient or required special education, and the fixed effect of cohort.