This paper extends latent class models for the analysis of rater agreement to the case of agreement on ordered category ratings. Among the important features of this approach are that it provides a method for combining multiple ratings--which is ordinarily difficult to accomplish when ratings are made on an ordered category scale--and estimates the accuracy of individual and combined ratings. It also permits statistical analysis of the specific effects of rater bias and measurement error in determining rater agreement and disagreement. A plausible logistic function/threshold model is applied to reduce the number of parameters that require estimation. The model is equally applicable to the analysis of agreement among two, or more than two, raters. Useful special cases of the model are discussed, methods described for evaluation and comparison, and the model shown to fit actual data well. A variant of the model applicable to the case of replicate measurement is also presented.
This report is part of the RAND Corporation Paper series. The paper was a product of the RAND Corporation from 1948 to 2003 that captured speeches, memorials, and derivative research, usually prepared on authors' own time and meant to be the scholarly or scientific contribution of individual authors to their professional fields. Papers were less formal than reports and did not require rigorous peer review.
This document and trademark(s) contained herein are protected by law. This representation of RAND intellectual property is provided for noncommercial use only. Unauthorized posting of this publication online is prohibited; linking directly to this product page is encouraged. Permission is required from RAND to reproduce, or reuse in another form, any of its research documents for commercial purposes. For information on reprint and reuse permissions, please visit www.rand.org/pubs/permissions.
The RAND Corporation is a nonprofit institution that helps improve policy and decisionmaking through research and analysis. RAND's publications do not necessarily reflect the opinions of its research clients and sponsors.