부산대학교 도서관

This article examines the reliability of content analyses of state student achievement tests and state content standards. We use data from two states in three grades in mathematics and English language arts and reading to explore differences by state, content area, grade level, and document type. Using a generalizability framework, we find that reliabilities for four coders are generally greater than .80. For the two problematic reliabilities, they are partly explained by an odd rater out. We conclude that the content analysis procedures, when used with at least five raters, provide reliable information to researchers, policymakers, and practitioners about the content of assessments and standards. Reprinted by permission of the publisher.