INVESTIGATING THE VALIDITY OF TOEFL: A FEASIBILITY STUDY USING CONTENT AND CRITERION-RELATED STRATEGIES

Abstract

The purpose of this study was to investigate the feasibility of two complementary approaches to assessing the validity of the TOEFL examination. One approach used evidence based on test content. In the context described in this report, evidence based on test content refers to the degree to which the items on the TOEFL examination are representative of the knowledge and skills required to demonstrate English proficiency in undergraduate and graduate programs throughout the United States and Canada. The content-oriented approach used in this pilot study involved item rating procedures that were designed to evaluate and document the relationship between the language tasks or behaviors previously identified as important for academic success and the test items used to measure them. The second approach used a criterion-related validation strategy. In this aspect of the study, experimental rating scales were developed for use by faculty to evaluate students' current levels of English language proficiency. These scales were designed to sample the domain of behaviors previously identified as important.