Premium
In Reply:
Author(s) -
Gill Carson,
Arnold Brendan,
Nugent Sean,
Rajwani Alykhan,
Xu Michael,
Black Tyler,
Doan Quynh
Publication year - 2018
Publication title -
academic emergency medicine
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.221
H-Index - 124
eISSN - 1553-2712
pISSN - 1069-6563
DOI - 10.1111/acem.13515
Subject(s) - kappa , medicine , cohen's kappa , agreement , documentation , psychosocial , value (mathematics) , statistic , statistics , psychiatry , mathematics , computer science , philosophy , linguistics , geometry , programming language
We thank the authors for their interest in our work and highlighting the limitations of the kappa value. Cohen's kappa statistic is calculated as a ratio of observed and expected (chance) agreement, and we concur that this value is dependent on the number of categories and prevalence in each. Indeed, prevalence approaching 0 or 100% results in high chance agreement, which typically reduces or handicaps the kappa value as the authors correctly identified in the example provided. Thus, such an argument is often raised for justifying a low kappa value in the face of high observed agreement. This article is protected by copyright. All rights reserved.