The Class as an Experimental Unit

Publisher Website
Google Scholar
Abstract
The use of the class, versus the student, as an experimental unit has frequently been discussed in the literature. Most analytic models assume random assignment of treat ments to units. When classes are the unit and student scores are analyzed, model assumptions are violated. This article examined the use of ANOVA under the assumption that the treatment had been randomly assigned to students; in fact, the class had been the unit. Resultant effects on Type I and Type II error, effect and sample size were examined and tabulated. Data support the idea that if one can randomly assign treatments to intact classes, serious consideration should be given to doing so.

This publication has 2 references indexed in Scilit: