초록 열기/닫기 버튼
The purpose of this study is to examine how consistently raters evaluate the pronunciation of native and nonnative English-speaking oral performance. Eleven native English- speaking raters participated in scoring twelve test-takers; eight of the test-takers were Koreans with various proficiency levels and one African- American, one British, one South African and one Filipino. The quantitative results showed that the raters showed variations in severity of pronunciation judgments especially for the African-American, the South African and one Korean with a high level of oral proficiency. The verbal protocol data revealed the following potential factors of the rater variation on pronunciation ratings: (1) raters’ different perceptions regarding standard English, (2) raters’ different degrees of familiarity with test-takers’ accents, (3) raters' utilization of criterion-irrelevant features, (4) testing method effects and (5) raters' application of their own scoring criteria. The notions of native accents and construct validity issues for the pronunciation ratings are discussed.
