Estimating the Variance of Design Parameters

Y

8

Randomized Controlled Trials; Hierarchical Linear Modeling; Correlation; Computation; Sampling; Models; Research Design; Statistical Analysis

Randomized experiments are generally considered to provide the strongest basis for causal inferences about cause and effect. Consequently randomized field trials have been increasingly used to evaluate the effects of education interventions, products, and services. Populations of interest in education are often hierarchically structured (such as students nested within classrooms, nested within schools, nested within school districts). The sampling designs used in educational experiments often exploit this hierarchical structure by sampling entire intact units (such as schools or classrooms). The two most frequently used designs in education field experiments are variants of two designs: The hierarchical design or the (generalized) randomized block design. The purpose of this paper is to consider estimators of the treatment effect heterogeneity parameters and their sampling distributions. In addition, the estimation and sampling distributions of intraclass correlations are also reviewed. This research should be of interest to investigators who wish to report empirical estimates of treatment effect heterogeneity parameters (and their uncertainties) from individual experiments or larger datasets.

ERIC

14

2015

ED562158