Assessing Writing Portfolios: issues in the Validity and Meaning of Scores

Although portfolio assessment has gained popularity in recent years, the technical base undergirding it is weak. Can portfolios be scored reliably? Is there evidence of the transfer and generalizability of scores? Are portfolio scores meaningful indicators of student achievement? This study examines a sample of student portfolios to provide preliminary data on these important measurement issues. Results show that although it is possible to score portfolios and classroom writing assignments consistently, substantial differences arise depending on whether judgments of students' performance are based on a standard writing assessment, individual samples of student work, or portfolio collections as a whole. This article concludes with implications for the design of portfolio assessments, and possible tensions between design requirements for large-scale assessment and those for teachers' classroom uses are discussed.