Accommodations for English Language Learner Students: The Effect of Linguistic Modification of Math Test Item Sets (AELLS)

This study examined the effect of linguistic modification on middle school students’ ability to show what they know and can do on math assessments. To do so, two item sets with 25 multiple-choice items each were developed, one containing original math items and one containing these items with linguistic modifications. Items were selected from two content strands: (1) measurement and (2) number sense and operations. Efforts were made to ensure that both sets of math test items met stringent guidelines for grade and population appropriateness, content rigor, and standardized administration. In developing the two item sets, researchers solicited input from experts and collected data through cognitive interviews and pilot testing. The two sets of math items (original and linguistically modified) were administered to three subgroups of students in grades 7 and 8 who differed in their English language proficiency (ELP) and ELA skills: EL students, non-ELA-proficient non-EL (NEP) students, and ELA- proficient non-EL (EP) students. Item sets were assigned to students within each class at random, with approximately half of each subgroup receiving the original set of items and half receiving the linguistically modified set. Participating districts were asked to provide archived data about each tested student, including students’ most recent test scores in ELA and math and EL students’ level of ELP. The ELA scores were used to distinguish EP students from NEP students. The math scores were used in correlational analyses as a proxy for math ability for EP students. The ELP scores were used to identify sampled students as EL students or non-EL students and to determine each EL student’s level of English proficiency.