Premium
Analysing Intraclass Correlation for Dichotomous Variables
Author(s) -
Mak Tak K.
Publication year - 1988
Publication title -
journal of the royal statistical society: series c (applied statistics)
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.205
H-Index - 72
eISSN - 1467-9876
pISSN - 0035-9254
DOI - 10.2307/2347309
Subject(s) - intraclass correlation , statistics , econometrics , correlation , mathematics , psychometrics , geometry
SUMMARY This paper deals with the analyses of intraclass correlation in littermate data with a dichotomous response variable. The common correlation model which includes many well‐known parametric models as special cases is studied. It is seen that the classical sample measure of intraclass correlation based on formulae in variance component models is a consistent estimator of p, a population parameter measuring the strength of intraclass correlation under the common correlation model. The asymptotic variance of this sample measure is also derived and the approximation to the asymptotic variance suggested by some researchers is demonstrated to be inadequate. The asymptotic theory, which does not require large litter sizes, is believed to be more relevant in some biological or toxicological applications than the uses of some existing algorithms for approximating asymptotic variances. A new kappa‐type sample measure of intraclass correlation is also proposed and its efficiency as an estimator of p is compared numerically with the variance component estimator based on the derived variance formulae. The results established are also applied to a real example.