Performance of intraclass correlation coefficient (ICC) as a reliability index under various distributions in scale reliability studies
-
- Shraddha Mehta
- Allergan plc, 2525 Dupont Drive Irvine CA 92612 USA
-
- Rowena F. Bastero‐Caballero
- Allergan plc, 2525 Dupont Drive Irvine CA 92612 USA
-
- Yijun Sun
- Allergan plc, 2525 Dupont Drive Irvine CA 92612 USA
-
- Ray Zhu
- Allergan plc, 2525 Dupont Drive Irvine CA 92612 USA
-
- Diane K. Murphy
- Allergan plc, 2525 Dupont Drive Irvine CA 92612 USA
-
- Bhushan Hardas
- Allergan plc, 2525 Dupont Drive Irvine CA 92612 USA
-
- Gary Koch
- University of North Carolina at Chapel Hill 135 Nottingham Drive Chapel Hill NC 27517 USA
この論文をさがす
説明
<jats:p>Many published scale validation studies determine inter‐rater reliability using the intra‐class correlation coefficient <jats:styled-content>(<jats:italic>ICC</jats:italic>)</jats:styled-content>. However, the use of this statistic must consider its advantages, limitations, and applicability. This paper evaluates how interaction of subject distribution, sample size, and levels of rater disagreement affects <jats:styled-content><jats:italic>ICC</jats:italic></jats:styled-content> and provides an approach for obtaining relevant <jats:styled-content><jats:italic>ICC</jats:italic></jats:styled-content> estimates under suboptimal conditions. Simulation results suggest that for a fixed number of subjects, <jats:styled-content><jats:italic>ICC</jats:italic></jats:styled-content> from the convex distribution is smaller than <jats:styled-content><jats:italic>ICC</jats:italic></jats:styled-content> for the uniform distribution, which in turn is smaller than <jats:styled-content><jats:italic>ICC</jats:italic></jats:styled-content> for the concave distribution. The variance component estimates also show that the dissimilarity of <jats:styled-content><jats:italic>ICC</jats:italic></jats:styled-content> among distributions is attributed to the study design (ie, distribution of subjects) component of subject variability and not the scale quality component of rater error variability. The dependency of <jats:styled-content><jats:italic>ICC</jats:italic></jats:styled-content> on the distribution of subjects makes it difficult to compare results across reliability studies. Hence, it is proposed that reliability studies should be designed using a uniform distribution of subjects because of the standardization it provides for representing objective disagreement. In the absence of uniform distribution, a sampling method is proposed to reduce the non‐uniformity. In addition, as expected, high levels of disagreement result in low <jats:styled-content><jats:italic>ICC</jats:italic></jats:styled-content>, and when the type of distribution is fixed, any increase in the number of subjects beyond a moderately large specification such as <jats:styled-content><jats:italic>n</jats:italic> = 80</jats:styled-content> does not have a major impact on <jats:styled-content><jats:italic>ICC</jats:italic></jats:styled-content>.</jats:p>
収録刊行物
-
- Statistics in Medicine
-
Statistics in Medicine 37 (18), 2734-2752, 2018-04-29
Wiley
- Tweet
詳細情報 詳細情報について
-
- CRID
- 1361137046448969216
-
- DOI
- 10.1002/sim.7679
-
- ISSN
- 10970258
- 02776715
-
- データソース種別
-
- Crossref