American Sign Language proficiency: A study of agreement among raters

This dissertation examines several approaches to the measurement of overall American Sign Language (ASL) proficiency. It concentrates on the recent movement towards ASL proficiency testing, embodied in the American Council on the Teaching of Foreign Language/Educational Testing Service (ACTFL/ETS) oral proficiency and rating scales. The impetus for ASL proficiency comes from several studies that a native-like competency in ASL is positively correlated with elevated language proficiency levels in the written form of English among Deaf children. ASL proficiency testing for classroom teachers is in high demand so that they can provide an environment in which Deaf children will acquire ASL, upon which bilingual education programs in the United States are based in order to deliver the expected results.;The dissertation outlines the methodological background to the development of ASL proficiency interview (ASLPI) instrument. It then describes an experiment in which the ASLPI procedures were used in assessing the proficiency of a group of college program students who use ASL. Seven Deaf individuals were informally trained to rate candidates on the basis of ASLPI rating scales. They then observed the recordings of interviews made with twenty-one undergraduates with ASL minor and graduate students of the Deaf Education program at McDaniel College. The researcher, following the ACTFL/ETS procedures, conducted the interviews.;Analysis of the results of the experiment showed that raters were able to reach a high degree of concordance in their application of the ASLPI rating scale, even when their ratings were made after only ten minutes exposure to an interview. There was clear evidence that raters arrived at their judgments globally, rather than separately evaluating discrete components of language proficiency. These results provided evidence in favor of the practicality, reliability, and validity of the ASLPI format. The implications of wider use of such tests are discussed, together with some suggestions for further research and possible improvements in the ASLPI procedure.