Affordable Access

Bias and prevalence effects on kappa viewed in terms of sensitivity and specificity.

Authors
  • Hoehler, F K
Type
Published Article
Journal
Journal of Clinical Epidemiology
Publisher
Elsevier
Publication Date
May 01, 2000
Volume
53
Issue
5
Pages
499–503
Identifiers
PMID: 10812322
Source
Medline
License
Unknown

Abstract

Paradoxical effects of bias and prevalence on the kappa coefficient are examined using the concepts of sensitivity and specificity. Results that appear paradoxical when viewed as a 2 x 2 table of frequencies do not appear paradoxical when viewed as a pair of sensitivity and specificity measures where each observer is treated as a predictor of the other observer. An adjusted kappa value can be obtained from these sensitivity/specificity measures but simulation studies indicate that it would result in substantial overestimation of reliability when bias or prevalence effects are observed. It is suggested that investigators concentrate on obtaining populations with trait prevalence near 50% rather than searching for statistical indices to rescue or excuse inefficient experiments.

Report this publication

Statistics

Seen <100 times