Tag

predictive_validity

2 cards

EducationRCT
Classroom Observation Standalone Predictive Validity
FFT classroom observation scores used as a standalone predictor yielded an IV coefficient of 0.807 (p < 0.01, SE = 0.293); while causally valid and not statistically different from 1, standalone observations exhibited roughly double the estimation error of test-based value-added measures (SE = 0.149).
EducationRCT
Student Perception Survey Standalone Predictive Validity
Standalone Tripod student perception survey ratings predicted post-randomization student test achievement with an IV coefficient of 0.940 (p < 0.01, SE = 0.311). While causally valid, standalone survey ratings exhibited roughly double the estimation error of test-based value-added measures (SE = 0.149).