The construct of academic English in tests of academic achievement and its effect on student performance: A confirmatory factor analytic study

This study investigated the extent to which Academic English Language (AEL) ability affects scores of ELL and non-ELL students on standardized tests of mathematics and science used for accountability purposes in US schools. The study addressed a perceived lack of empirical support for some assertions concerning ELL test performance made in the research literature. The research questions for the study were: (1) To what extent can AEL ability be identified as a factor influencing test scores of mathematics and science? (2) To what extent can unique grammatical and lexical features of AEL be identified as components of content test score variance? (3) To what extent and in what ways is the strength and composition of the effect of AEL ability on content area test scores different for students at different levels of language ability?;These questions were examined through the methodology of confirmatory factor analysis (CFA). A series of theoretically plausible factor models which included both content knowledge and AEL factors were tested on four groups: low- and high-reading ability ELLs and low- and high-reading non-ELLs. Data consisted of item level scores from 8th and 5th grade mathematics and science tests administered by one state, as well as relevant measures of English language ability. Actual test items were also provided. Test items were organized and item bundles created based on the presence of specific features of AEL, using an instrument designed for this purpose.;The study found no convincing evidence of an AEL ability factor for any of the groups, even the low-reading ELL groups, in either grade. Furthermore, the best-fitting model, one which did not include AEL ability factors, was found to be invariant across all four groups, thus suggesting that the tests measured the same abilities for all groups. Results were consistent in both grades. These results cast serious doubt on the appropriateness of conducting linguistic modifications on content area tests for ELLs. Finally, the results of the study suggest that the aspects of academic English investigated here should be considered a legitimate pact of content area test score interpretations.