Comparison of the Efficiency of Blocking Factors for Randomized Control Studies in Education

• In Event: Poster Session 4
  In Poster Session: Division H Graduate Student Poster Session: Using Data to Drive Educational Success

Abstract

This study examines four blocking factors and compares their efficiency in reducing variability in experimental design in school-wide intervention studies. The four blocking factors are past performance, past performance weighted by student demographics, growth trajectory, and predicted performance. The author used each of the factors to create equal quintiles with 214 schools randomly sampled from California public school system and calculated the relative efficiency. Initial findings indicate the blocking by past performance and by weighted past performance generate similar blocks; so do blocking based on the growth trajectory of school-level prior achievement and predicted score. Based on this finding, the author compares the relative efficiency of two blocking factors, i.e., weighted prior achievement and predicted achievement scores.

Author

• Ying Zhang, IMPAQ International, LLC