DG-00000946 Abstract

ÐÜèÊÓÆµ

Ting Zhang
University of Maryland - College Park



The Role of Reading Comprehension in Large-Scale Subject-Matter Assessments



As high-stakes testing has become increasingly associated with large-scale educational assessment, ensuring accurate and valid measures of student learning in content areas has become a critical issue. Existing limited research indicates that lack of reading proficiency can adversely affect students scores in subject matter tests. As a result, the interpretations of scores may not accurately reflect the psychological constructs of content knowledge and skills that the tests are intended to measure. By incorporating reading comprehension theories into a validity framework based on evidence centered design, this project develops improved methods for partialling out construct-irrelevant variance that is associated with reading comprehension in subject areas such as mathematics, science, and civic-related social studies. In addition, this study investigates the extent to which poor reading comprehension interferes with measures of targeted proficiencies, and how students backgrounds are associated with their cognitive processing of test items. This study uses U. S. data from the IEA TIMSS mathematics study and science study of 1999, and IEA CIVED civic education study of 1999. A three-step procedure is employed. The first step is to identify the degree of reading comprehension demand for each item in each assessment. At the second step, an advanced multidimensional item response model will be fit to each test data. This approach explores the association between item characteristics and individual differences while taking domain proficiencies and reading comprehension into account. Finally, multiple regression will examine the association between students backgrounds and their domain proficiencies viewed in the light of reading comprehension.


Back to Funded Dissertation Grants Page