Testing accommodations for ELL students on an achievement test battery

How well students perform on standardized tests can affect their educational paths and the rest of their lives. In addition, students' performances on state assessments will affect their schools due to the No Child Left Behind Act. For English language learners (ELLs), the success on tests may be diminished due to their inability to completely understand what they are reading on a test. Because ELLs are a growing proportion of the population and have greater risk of not performing well in school, dropping out, and not moving on to have a job, it is very important to address their educational performance. To alleviate the difficulty of not being able to understand English competently, various testing accommodations can be given. The purpose of this study was to describe how different types of accommodations are being distributed, particularly among ELL students, on an achievement test battery. Several variables were examined to assess whether they related to which accommodation would be assigned to an individual student.

This study used data from a recent 2010 national standardization of an achievement test battery, which sampled 33,226 students from grades 3-8 across the U. S. In addition to the tests, students completed a survey which asked for such information as gender, home language, and ethnicity, and test administrators reported on that same survey which testing accommodations students were given, whether students were migrants, whether students participated in a free or reduced-price lunch program, and whether students were ELL students. These variables, along with students' Reading, Math, and Science scores were used to describe the groups of students given each accommodation.

Five testing accommodations were reported by the students - giving students extended time to take the test, allowing students to use a word-to-word dictionary, reading parts of the test aloud to the students, repeating instructions, and having the test administered by an ELL teacher. Of all these, the group that was most similar to the entire sample on the test results for the Reading, Math, and Science was the group given dictionaries, a group predominantly comprised of Asian students. The other testing accommodation groups had much lower percentile ranks on average. ELL students were largely Hispanic, spoke Spanish at home, and had lower percentile ranks on the Reading, Math, and Science tests than non-ELL students. Although the majority of ELL students in this sample was not free or reduced-price lunch eligible, there was a higher percentage than was found in the sample as a whole.

It was found that several variables were potentially important in how testing accommodations were distributed among students, including grade level, ethnicity, home language, and socioeconomic status (SES). Variables which did not seem important were gender and school. Home language and SES were also important in testing performance, with low SES students performing much worse on average than the sample as a whole. Students who had Spanish as their home language did not perform as well on the tests as students whose home language was English or another language. Lastly, a gap analysis using effect sizes showed some evidence for the gap between ELL and non-ELL students being larger in higher grade levels.