Answers are grounded in this paper and cite their sources.
Chance agreement may account for a sizeable proportion of the specificity of a screening test when the disease prevalence is low. Conversely, the observed sensitivity may be largely accounted for by chance agreement when the prevalence of disease is high. We derive descriptive statistics that are analogous to sensitivity and specificity and corrected for the agreement expected by chance. These coefficients of validity are shown to be dependent on the true prevalence of disease, as well as sensitivity and specificity.
No takes yet. Share an insight, caveat, or question.
Coughlin et al. (1992) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: