Differential Vocabulary Performance by Dual Language Learners' English Proficiency

• In Event: Toward an Asset-Oriented Approach in Exploring Vocabulary Performance Among Dual Language Learners in the United States

Abstract

Objectives: The objective of this paper was to examine academic vocabulary assessment items that function differently for limited English proficient (LEP; those with limited English proficiency who need additional language support), initially fluent English proficient (IFEP; those who have full English proficiency by the time they enter school), redesignated fluent English proficient (RFEP; those originally classified as limited English proficient but attained sufficient English proficiency to be reclassified) students relative to English-only (EO) students in the 6-8th grades.

Theoretical framework: The percentage of dual language learners in US public schools is rapidly growing. Accurate and reliable assessment of their vocabulary knowledge is a national concern given the extensive use of tests in high-stakes decision making. Research has shown that test performance is often dissimilarly reliable and accurate for dual language learners and EO students as vocabulary learning is tightly connected to other language skills in first and second language.

Data sources: The data for this study were collected from an efficacy trial of an academic vocabulary intervention, Word Generation. We used the Word Generation Academic Vocabulary Test (WG) and the Vocabulary subtest of the Gates-MacGinitie Reading Test (GM) to assess students’ vocabulary knowledge. The sample included 13,780 students who attended thirteen middle schools in a large urban district in California during the 2010-2012 academic years (Cohort 1: 2010-2011, Cohort 2: 2011-2012).

Methods: Differential item functioning performance (DIF) between EO students and dual language learners (LEP, IFEP, and RFEP) was examined using three multi-level generalized mixed linear models. Each model included the group effect (reference versus focal group), item effect, group by item interaction, controlling for cohort differences via inclusion of an explicit measure of ability (a latent ability estimate based on the WG and GM items).

Results: Figure 3 depicts item easiness for EOs against item easiness for each of the dual language learner groups. The results suggested that most WG items (total number of unique items = 80) showed evidence of DIF favoring the EO students (n = 60), whereas a handful of items (n = 13) seemed to favor the LEP students. In contrast, most items did not show evidence of DIF when comparing EOs and IFEP students. Specifically, 24 of 80 items showed some evidence of DIF, only three of which favored EOs, and 21 of which favored the IFEPs. When comparing RFEP students to EOs, we found 11 items showing DIF that favored the RFEPs and 27 items showing DIF that favored the EOs. Additionally, it is clear that item difficulty differentially related to DIF across these subgroups. Specifically, it is the most difficult items that tended to favor LEP students, whereas easier items tended to favor the EOs.

Scholarly significance: Our results suggest that a richer conceptualization of student verbal knowledge is needed to understand the performance of students on vocabulary items across language proficiency levels. Understanding the linkage between differential vocabulary performance and test characteristics that are potentially responsible for item difficulty for dual language learners is important when developing vocabulary assessment items.

Authors

• Paulina Kulesz, University of Houston Times

• Autumn McIlraith, University of Houston

• Rebecca E Knoph, University of Oslo

• Joshua Fahey Lawrence, University of Oslo

• David J. Francis, University of Houston