Citation

St. Laurent, Roy T. (1998).   Evaluating agreement with a gold standard in method comparison studies.   Biometrics 54: 537-545.

Abstract

We develop a statistical model for method comparison studies where a gold standard is present, and propose a measure of agreement. This measure may be interpreted as a population correlation coefficient in a constrained bivariate model. An estimator of this coefficient is proposed and its statistical properties explored. Applications of the new methodology to data from the medical literature are presented.

Keywords

Agreement; correlation; gold standard; intraclass correlation coefficient; method comparison study; random effects model.