Standardized achievement tests and English language learners

Abedi, J. (2002). Standardized achievement tests and English language learners: Psychometrics issues. Educational Assessment, 8(3), 231-257.

Background

Research had indicated that the language proficiency of English language learners (ELLs) impacted their performance on achievement tests, and the effect was found to be greater in reading, Science and Mathematics where the content areas required higher language proficiency. ELLs are students of a non-English speaking background entering classrooms where the majority of students are expert English speakers and where the teachers have no understanding of language teaching or learning issues.

Aims

The study examined the impact of students’ language background on their achievement test results.

Sample

This study analysed the standardised achievement test data and background information of the participating students from four sites across the United States. The four locations were used as cross-validated samples.

Method

The author compared the level of performance between the ELL and non-ELL students. Comparisons were also made between students with different background variables, such as family income and parent education. Other background variables included gender, ethnicity, student ELL status, and students with disabilities.

Findings

The results suggested that the ELLs performed significantly lower than the non-ELLs. The performance gap between ELL and non-ELL students was smaller in the lower grades. However, the performance difference between them was not consistent across the content areas. In content areas that required a higher level of language proficiency, such as reading and writing, the performance difference was the highest, whereas in content areas less demanding in terms of language, such as Mathematics, and Science, the performance difference was significantly smaller. Results demonstrated that among the various background variables such as family income measured by free/reduced lunch participation, and parent education, language factors had a greater effect on assessment, much larger than family income or parent education.

Implications

The author highlighted that, where there was a high level of language complexity in the assessment tool, there would be a greater performance difference between ELLs and non-ELLs. The linguistically complex structures of the test items in content-based assessments for non-native speakers of English might weaken the reliability and validity of the tests. The author called for attention to be paid to the linguistic complexity of test items unrelated to the content that was being assessed in an effort to reduce the confounding effects of language difficulty.

Original article retrieved from Google Scholar.