| Recent growth in the U.S. immigrant population has had a tremendous impact on the American educational system. Despite longstanding federal requirements that English language learners (ELLS) be provided with equitable educational opportunities, these students are specifically highlighted in the No Child Left Behind Act of 2001, which mandates accurate assessment of the academic achievement of these students. While this assessment mandate engenders positive outcomes, these testing requirements represent serious challenges for students, teachers, schools, and test developers. These relate to difficulty in interpreting ELLs' achievement test results, negative consequences resulting from inappropriate uses of tests and test scores, and the shortcomings of test administration accommodations. In order to effectively serve ELLs in the K-12 context, alternative methods of assessment must be developed.;One avenue for the collection of this precise data is the use of selected-response test items that incorporate linguistic simplification and/or visual support. This pilot study sought to explore what types of selected-response items allow ELLs to best demonstrate what they know and can do in the area of science. Specifically, the study compared ELLs' performance on multiple-choice items with Standard English to their performance on multiple-choice items using visual support and/or linguistic simplification. The study further compared student performance on multiple-choice items with and without linguistic simplification and visual support to four selected-response item formats (yes/no, classification sorting, matching, and sequencing) that use both linguistic simplification and visual support. Because ELLs are not a uniform group, three student characteristics were also considered: level of English language proficiency, first language, and grade within two grade-level clusters (grades 3-5 and 6-8). Fifty-two 3rd-5th grade ELLs and 46 6 th-8th grade ELLs completed a series of four content-parallel tests that incorporated the different item formats.;Among the most significant findings, split-plot ANOVAs revealed that, of the six item types investigated, two seemed to hold the most promise for improving the accuracy of ELL assessment: yes/no and multiple-choice items, both with linguistic simplification and visual support. However, given a variety of factors, all of the item types may warrant further investigation. The study concludes with a plan for its replication. |
