Understanding the Relationship Between English Language Proficiency and Academic Performance of English Learners

• In Event: Validity and Utility of English Language Proficiency Assessments: New Evidence in the Common Core Era

Abstract

Purposes & Significance

Since the authorization of the Every Student Succeeds Act (ESSA) in 2015, states’ English language proficiency (ELP) standards and assessments have undergone changes in their content (e.g., the ways academic language proficiency is measured) to correspond to the language demands of new academic content-area standards (e.g., Common Core State Standards in English language arts and mathematics). However, empirical investigations about how English learner (EL) students perform on the new assessments based on new standards are limited.
In this study, we aimed to fill this gap. Specifically, we address the following research questions:
1. What is the relationship between the content and ELP assessment performances of EL students in different states using different ELP assessments? Does the relationship exhibit a common pattern regardless of the type of ELP assessment?
2. Does that relationship vary based on students’ home language?
The findings will provide useful insight into the role of ELP assessments in understanding EL students’ content performance given the increased language demands of the new standards.

Perspective

A handful of studies have examined the relationship between EL students’ ELP and content assessment performance using assessment data in the pre-ESSA period (Cook et al., 2012; Crane et al., 2011; Kong et al., 2012; Parker et al., 2009; Wolf & Faulkner-Bond, 2016). For example, Parker et al. (2009) found that written language scores (reading and writing) of three states’ ELP assessment were significant predictors of fifth- and eighth-grade EL students’ performance on statewide content assessments. This line of research has provided important validity evidence for ELP assessment use, particularly in understanding the impact of ELP on EL students’ demonstration of content knowledge and skills. The findings also have implications in high-level EL instruction planning and in making EL reclassification (i.e., EL status exit) decisions.

Method

We obtained two states’ Grade 5 datasets from the 2018-2019 school year, including students’ home language background and scores from content and ELP assessments. We also obtained access to the assessment items. Table 1 summarizes each state’s assessments in English language arts (ELA), mathematics, and ELP. To answer the research questions, we employed structural equation modeling (SEM) and correlational analyses. Additionally, we developed and applied a linguistic-content-analysis protocol (see Table 2) to quantify the language complexity of the assessment items. The results of the linguistic content analysis were used in interpreting the results from the SEM and correlational analyses.

Results

We found that the ELP factor strongly predicted the ELA and mathematics factors in both states’ data. The reading domain had the strongest correlation with ELA and mathematics scores (see Figures 1 and 2 and Table 4; see Table 3 for the model fit statistics). This pattern held true regardless of home language group. The linguistic ratings indicated that ELA assessments were highly complex, particularly in the reading and writing domains while the two ELP assessments had varied linguistic complexity. In this presentation, we will share the detailed results and insights based on the language complexity rating of each assessment’s subdomains.

Authors

• Mikyung Kim Wolf, Educational Testing Service

• Hanwook Yoo, Ascend Learning

• Laura Ballard, Educational Testing Service

• Florencia Tolentino, Educational Testing Service

• Alison L. Bailey, University of California - Los Angeles