| Table 7.1. |
Hypothetical Growth Trajectories of Test Score Impacts, by the Number of Months Since the Start of School (Design I) |
| Table 7.2. |
Maximum Values of β1 / θ0 for Which the ANCOVA and DID Estimators Would be Preferred to the Posttest-Only and UANCOVA Estimators (Design I) |
| Table 7.3. |
Variance, Bias, and MSE Estimates for the ANCOVA and UANCOVA Estimators, For Various Values of β1 / θ0 (Design I) |
| Table 7.4. |
School Sample Sizes Needed to Equate MDE Values for the ANCOVA Estimator, by the Size of the Early Treatment Effect (Design I) |
| Table 7.5. |
Maximum Values of β1 / θ0 for Which the ANCOVA Estimator Would be Preferred to the Posttest-Only Estimator (Designs II and III) |
Appendix A: Proof of Asymptotic Results for the ANCOVA Estimator (174 KB)