Making Every EL Count: Cautions and Opportunities of Administrative Data on English Learners

• In Event: 64.081 - Reenvisioning English Learner Research Using State and District Data

Abstract

Unlike academics studying the effects of an instructional intervention or a phenomenon observable in a laboratory, researchers studying trends in English Learner (EL) education rely heavily on administratively-collected data. This data on student characteristics and outcomes and school context is made available to policymakers and the public through online data dashboards or may be obtained through data-use agreements. The 2015 Every Student Succeeds Act pushed data use forward in three ways: by emphasizing evidence-based decision making by policymakers, requiring the input of families and communities in such decision-making, and greatly increasing the volume of data that states must publish annually. Particularly in the case of EL education, there is an urgency to use data to find answers to research questions with important practical applications to a growing and diverse student population, and to identify schools and districts that are not doing their part to ensure equitable educational opportunity for all.

Therefore, it is incumbent upon researchers and the public to understand the opportunities and limitations of growing data resources. This paper outlines types of data available and how general methodological constraints like comparability across populations and datasets play out with EL data. It also points to ways researchers can (1) extend the usefulness of their analyses by demonstrating how they may be replicated with public data by state or local stakeholders, (2) help system administrators provide useful data with clear descriptions of data collection methodologies and limitations, and (3) alert readers of both scholarly and non-scholarly writing to key complexities that may impact what conclusions may be drawn from data.

The publication on which this paper is based walks the reader through the types of data collected on students that would be of interest to EL researchers and what information is available to the public in state and national data sources. Additionally, it looks in depth at several critical issues in using EL data, such as the following:
•Unlike other subgroups, the EL subgroup is fluid and constantly loses its highest-performing members (Saunders & Marcelletti, 2013)
•Also unlike other subgroups, each state has different criteria to define EL status, so a student might be an EL in one state and not in another (Linquanti, Cook, Bailey, & MacDonald, 2016)
•Most states have significantly revised state standardized tests and English language proficiency tests within the last five years, making comparisons over time challenging
•State and local education agencies may not collect, or collect but not publish, key data such as the numbers of students with interrupted formal education, recent immigrants, long-term English learners (identified for six or more years), and dual-identified ELs with disabilities
•School-based data collection may inaccurately represent multilingual and multicultural populations; for example, underestimating linguistic diversity by only collecting one home language per family

With student data taking a central role in decision-making—whether mediated by scholarly research or obtained directly by community advocates—bringing these complexities to light is a critical step in ensuring data on EL education serves to advance educational equity.

Author

• Julie Sugarman, Migration Policy Institute