Design Parameters for Planning Multilevel Experiments Evaluating Effects of Teacher Empowerment Programs on Student Outcomes

• In Event: AERA Poster Session 5
  In Poster Session: Evaluation in Schools Poster Session (San Diego)

Abstract

Teacher empowerment programs, such as teacher professional development and innovative curriculum initiatives, represent a core strategy to improve student educational outcomes and to close student achievement gaps. Using the ECLS-K data, this study compiles empirical design parameters for experimental evaluations of teacher empowerment programs. We estimate empirical design parameter values, including intraclass correlations and proportions of variance explained by covariates, in the setting of students nested within teachers nested within schools. The results show that intraclass correlations at the school level are relatively stable and those at the teacher level increase along grade level. We illustrate the utility of the results by demonstrating how to leverage this information and costs of sampling to design more efficient experimental studies.

Authors

• Huibin Zhang, University of Florida

• Zuchao Shen, Arizona State University

• Chris Curran, University of Florida

• You You, Peking University

• Joni W. Splett, University of Florida