AN INTRODUCTION TO GENERALIZABILITY THEORY IN SECOND LANGUAGE RESEARCH1

Publisher Website
Google Scholar
Abstract
During the last decade researchers in the area of second language testing have adopted increasingly more sophisticated evaluation techniques in an attempt to better understand the process of learning a second language. Analysis of variance, regression analysis, and factor analysis are among the statistical procedures that have allowed for insightful evaluation of language testing data. This paper explores the feasibility of applying another procedure, generalizability theory analysis, to second language testing research. An earlier paper (Hinofotis, Bailey, and Stern 1981) reported on the development of a rating scale for assessing the oral proficiency of nonnative‐speaking teaching assistants and the actual use of the scale by a panel of raters in evaluating pilot data. The data from that study are reanalyzed here using generalizability theory, which provides simultaneous estimates of consistency along several dimensions of the research, e.g., raters, subjects, and occasions. The paper includes a comparison of generalizability theory and classical test theory, an analysis of the oral proficiency data, and a brief discussion involving the interpretation of generalizability study results.

This publication has 0 references indexed in Scilit: