Placement testing: One test, two tests, three tests? How many tests are sufficient?. (July 2020)

• Record Type: Journal Article
• Authors: Hille, Kathryn; Cho, Yeonsuk
• Is Part Of: Language testing. Volume 37:Number 3(2020)
• Page Start: 453
• Page End: 471
• Publication Date: 2020-07
• Subjects: ESL -- intensive English program -- Michigan EPT -- multiple regression -- ordinal regression -- placement accuracy -- placement testing -- predictive validity -- TOEFL ITP; Language and languages -- Ability testing -- Periodicals; Language and languages -- Examinations -- Periodicals; 407.6
• DOI: 10.1177/0265532220912412
• Languages: English
• ISSNs: 0265-5322

Abstract

Accurate placement within levels of an ESL program is crucial for optimal teaching and learning. Commercially available tests are commonly used for placement, but their effectiveness has been found to vary. This study uses data from the Ohio Program of Intensive English (OPIE) at Ohio University to examine the value of two commercially available tests (the TOEFL ITP and the Michigan EPT) and a locally developed writing test for accurate placement decisions. Placement accuracy was measured in terms of the relationship between test scores and (1) appropriate placement levels for individual students according to their teachers, and (2) student performance in the classes. Findings support the continued use of multiple measures for more accurate placement decisions in the study context. However, the relationship between test scores and student performance, measured by students' grades in the actual course levels and their success in advancing to a higher course level as additional indicators of the extent to which placement tests provide an accurate indication, was weak when analyzed through multiple regression and cross-tabulation, suggesting that factors other than initial proficiency are primarily determinative of student success when students have been accurately placed.