Examining Validity Issues in Classifying and Reclassifying English Learners Through Dynamic Views of Languaging Practices

• In Event: Centering Multilingualism for More Equitable Education Systems in Assessment: Global Perspectives

Abstract

Objectives
The purpose of this paper is to identify issues of fairness around cut score setting practices that impact English learner (EL) classifications and reclassifications in the U.S. context. High-stakes decisions regarding the education of ELs have relied mostly on English language proficiency (ELP) assessments and the ELP-related classifications. These classifications impact decisions as to whom to designate as EL, whom to administer content assessments to, and who benefits from instructional support programs. One of the decision rules that is behind EL/non-EL classifications or reclassifications is proficiency cut scores used to determine the degree of eligibility for proficiency. These scores are typically identified through standard-setting procedures.

Theoretical Background
When cut scores are used to classify students as English language proficient or not, then the validity of these cut scores (and the standard-setting procedures used to arrive at these cut scores) is essential to supporting the validity of the ELP assessments (Buckendahl & Davis-Becker, 2012; Kane 2006) and EL classifications. Very few validity studies have focused on the use of cut scores (Florez, 2012; Wang, Niemi, & Wang, 2007). The issue has not yet been theorized through dynamic views of language positing that standards can not be set for bilinguals’ language use (Valdes, Poza, & Brooks, 2017).

Research Methods
Against this backdrop, the paper first presents a review of current EL cut score setting and classification practices followed across 50 states. In 2014, information on EL classification practices was collected from State Department websites and phone and online surveys were conducted regarding cut score determination practices used to classify students. Specifically, the following aspects of cut score determination practices were examined: (i) specific cut scores used across states, (ii) how, if any, adjustments to these cut scores are made, (iii) specific exit criteria used across states and how they are determined, and (iv) finally, the length of exemption (if any) from content assessment provided across states.

Findings/scholarly significance
Information about the pre-screening instruments was obtained for all states except Hawaii. 49 states indicated they use a home language survey/questionnaire to pre-screen students who speak a language other than English and who might potentially take the ELP tests and/or need instructional support. The lack of consensus in the use of screening instruments was further exacerbated by the diversity of cut score practices and the criteria used to adjust these cut scores. Once a student is classified as EL, the variability in exit criteria used across states also held implications for inconsistency in EL classifications across the states. Besides the variability in classification/reclassification criteria, there was also apparent variability in the length of time the reclassified ELL student is provided exemption from mainstream content testing. Given the variability in the exit criteria, the paper highlights fairness issues in blocking exemption in content testing. This paper provides a novel discussion about the classification practices in the US by taking a dynamic multilingualism perspective. It discusses what missing chains of validity evidence could help align the classification practices with the dynamic languaging practices of multilinguals.

Author

• Sultan Turkan, Queen's University - Belfast