|
Abstract:
|
Shoukri and Donner (2001) developed a model to estimate inter-rater and intra-rater agreement when each of two observers has the opportunity to obtain a pair of replicate binary measurements on each subject. A practical limitation of this model is that it accommodates only a single binary trait. In many studies, however, more than one trait may be of interest. For example, Becker et al. (2002) reported on a study of depression and anxiety as two outcome measures obtained on 173 patients evaluated by two methods, the first method using a patient health questionnaire and the second employing a structured clinical interview. In this case, although inferences on each trait separately are usually of main interest, a question also arises as to whether the level of inter-rater agreement is stable across traits.
|
