The relationship between criterion-based levels of oral proficiency and norm-referenced scores of general proficiency in English as a second language

Abstract

This paper presents data from a validation study of the American Council on the Teaching of Foreign Languages (ACTFL) Oral Proficiency Interview (OPI). The “distinctness” of OPI ratings from scores generated by Test of English as a Foreign Language (TOEFL) and one other examination was assessed on the basis of comparisons among 70 sets of scores from students of English as a second language who took all three test. Results of correlational analyses provide evidence that the OPI measures an ability (or collection of abilities) distinct from the skills evaluated by the other testing instruments, and thus are supportive of the “construct validity” of this oral interview (cf. Bachman and Palmer, 1981; Dandonoli and Henning, 1990). Correlations involving the OPI were consistently lower than correlations among subscores from the other tests. Even the correlation between regression-based, “predicted” OPI ratings and actual OPI ratings was not significantly higher than the correlation between actual OPI ratings and the single best predictor among the raw test scores. However, employing one particular “cut-off” procedure rendered OPI ratings completely predictable from TOEFL scores, a finding which underlines the relevance of potential test uses to the validation process.