Funded by the U.S. Department of Education, STELLA is a research-based, online decision-making tool which is designed to guide and support educators in selecting appropriate large-scale and classroom accommodations for English learners. WIDA's ONPAR Researcher Rebecca Kopriva was the lead developer of this project. In order to maximize the accessibility of assessments and the accuracy of test results, the STELLA system uses a detailed set of algorithms for determining the specific research-based accommodations individual students should receive. The application also recommends particular pretest support activities useful for students with different needs and challenges.

To recommend accommodations based on students' particular background characteristics and needs, STELLA collects and consolidates data from three sources: school records, teachers, and parents/guardians or students. This information includes formal and informal English language proficiency (ELP) data, primary language (L1) proficiency data, students' familiarity with standardized academic assessments and classroom evaluation practices, similarity of schooling they attended in the native country to what is experienced in U.S. schools, and time and consistency attending school in the United States.

The STELLA Output and state, district and school-level summary reports are tailored to meet the accommodation requirements of individual states and districts. Besides these useful features, the web-based application includes start-up information, a comprehensive User's Manual, a detailed "Help" resource, and question by question guide to assist educators in completing forms and interpreting the Output. The system has been thoroughly vetted by groups of teachers, parents, and local, state, and academic ELL and testing experts. Two validation studies have been conducted. The first found that students who received STELLA recommended accommodations received significantly higher scores; the other found that this method was significantly more defensible and replicable than other methods currently in use. To access the application and for more information, please visit the STELLA website.