Establishing placement test fit and performance: Serving local needs

Abstract

Accountability in language education is often associated with top-down national policies unresponsive-or even hostile to-local needs; however, when accountability is driven by local stakeholders seeking to better understand and enhance their programs, it can foster productive cycles of action research and curriculum development. This paper reports on one such internally-motivated accountability effort, in which program insiders sought to determine the efficacy of a reading test being administered to a new population of students at one Colombian university. Descriptive statistics, reliability estimates, and item facility and discrimination measures were used to determine whether this test was sufficiently reliable and appropriately matched to test takers' ability in order to warrant its use as part of a high-stakes English-language placement exam. A detailed analysis of this test is used not only to propose specific recommendations for revision but also to illustrate a useful set of statistical tools appropriate for test analysis in other language programs. Moreover, we conclude that the involvement of local instructors as part of an iterative, self-reflective, test development process provides opportunities for professional development and deeper engagement in accountability projects.