| The present research addresses three main questions: (1) Is there a significant interaction between raters' English language backgrounds and their attitudes toward WEs for language testing in the holistic and analytic ratings of Korean students' speech samples on the TSE picture-description task? (2) Is there a significant difference among different groups of raters (native speakers in the U.S., native speakers in Korea, non-native speakers in Hong Kong, and non-native speakers in Korea) in the rating of Korean students' speech samples on the TSE picture-description task using holistic and analytic scores as measures? and (3) Is there a significant difference among different groups of raters based on their attitudes (negative, neutral, and positive) toward WEs for language testing in the rating of Korean students' speech samples on the TSE picture-description task using holistic and analytic scores as measures?; The data for this study were collected from 133 English speakers (30 native speakers in the U.S., 35 native speakers in Korea, 30 non-native speakers in Hong Kong, and 38 non-native speakers in Korea), who have taught the English language for more than two years. The study employed two types of scales: holistic and six analytic rating scales (grammatical accuracy, vocabulary, pronunciation/accents, rate of speech, organization, and task fulfillment). Results of this study showed that raters' English language backgrounds significantly affected the rating of speech samples on grammar and organization using analytic scales. Results also revealed that raters' attitudes toward WEs in language testing significantly affected the rating on grammar, rate of speech, and task fulfillment using analytic scales.; The results help inform L2 test developers and researchers about the extent to which rater variables such as English language backgrounds and attitudes toward WEs in language testing affect test scores of non-native speakers' English language oral proficiency. Based upon the findings in this study, the implications related to the assessment of non-native speakers' English language oral proficiency are also discussed. |
