Answers are grounded in this paper and cite their sources.
Portfolios are the centerpiece of Vermont's statewide assessment program in mathematics. During the first 2 years of this program, there was only moderate agreement among readers regarding a portfolio's quality. This quality often varied greatly across the five to seven pieces in a student's portfolio. Consequently, scores were not reliable enough to permit reporting student-level results. Increasing the number of readers per piece or the number of pieces per portfolio are not operationally feasible solutions to the score reliability problem.
No takes yet. Share an insight, caveat, or question.
Klein et al. (1995) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: