Overestimation of LPI Ratings for Native-Korean Speakers in the TOEIC Testing Context: Search for Explanation

Wilson, Kenneth M.
Publication Year:
Report Number:
ETS Research Report
Document Type:
Page Count:
Subject/Key Words:
Korean Speakers Language Proficiency Interview LPI Rating Multilevel Analysis Native Speakers Target Language Learning Rate Test of English for International Communication (TOEIC)


Regression-based guidelines have been developed for predicting ratings of speaking proficiency in English as a foreign language (EFL), as assessed using the Language Proficiency Interview (LPI) procedure, from scores on the TOEIC (Test of English for International Communication) Test. These guidelines, which reflect the regression of LPI rating on TOEIC scores in a combined sample composed of native speakers of Japanese, French, Spanish, and Arabic, respectively, were found to overestimate LPI rating when applied to data for native speakers of Korean. It was hypothesized that this outcome would be expected if development of EFL speaking proficiency (as assessed by the LPI procedure) tends to lag relatively more behind development of the proficiencies that are assessed by the TOEIC Test, for native-Korean-speaking EFL learners than for demographically comparable EFL learners in other national/linguistic settings (at least those represented in the study sample). An indirect assessment of this hypothesis was undertaken in a series of exploratory analyses involving section- and total-score means for native-language groups on the familiar, three-part version of the Test of English as a Foreign LanguageTM (TOEFL). Findings were interpreted as being generally, albeit indirectly, supportive of the developmental lag hypothesis, as well as of working assumptions underlying introduction and analysis of TOEFL native-language reference group means as group analytic variables. Needed lines of inquiry are suggested.

Read More