|
Activity Number:
|
117
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Monday, August 4, 2003 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Social Statistics Section
|
| Abstract - #300512 |
|
Title:
|
Assessing Interobserver and Intraobserver Agreement
|
|
Author(s):
|
Huiman X. Barnhart*+ and Michael Haber and Jingli Song
|
|
Companies:
|
Emory University and Emory University and Emory University
|
|
Address:
|
Rollins School of Public Health, Atlanta, GA, 30322-0001,
|
|
Keywords:
|
interobserver agreement ; intraobserver agreement ; reproducibility ; repeatability ; reliability
|
|
Abstract:
|
In evaluation of a continuous scale for practical use, one often needs to assess both inter and intra observer agreement. Often, an agreement study is designed to have multiple observers taking single reading for each subject or to have only one observer taking replicated readings for each subject. Although traditional agreement index is often reported as interobserver agreement in the first case, it is in fact a total agreement that contains both inter- and intraobserver agreement. In the second case, only intraobserver agreement can be evaluated. We advocate an agreement study design that involves multiple observers who take replicated readings for each subject so that one can simultaneously assess inter-, intraobserver agreement as well as total agreement. We propose inter-OCCC (overall concordance correlation coefficient) based on "true'' readings of the observers to assess interobserver agreement and use intra-ICC (intraclass correlation coefficent) to assess intraobserver agreement. We investigate the relationship of total agreement with inter-OCCC and intra-ICC. Simulation studies and applications to data will be presented.
|
