American Statistical Association
In medical and social studies, it is often desirable to assess the correlation between characteristics of interest that are not directly observable. In such cases, repeated measures are often available, but the correlation between the repeated measures is not the same as that between the true characteristics that are confounded with the measurement errors. The latter is called the hidden correlation. Previously, the problem has been treated by assuming prior knowledge about the measurement errors, or by using relatively complex models, such as the mixed effects models, with no closed-form expression for the estimated hidden correlation. We propose a simple estimator of the hidden correlation that is very much like the Pearson correlation coefficient, with a closed-form expression, under assumptions much weaker than the mixed effects model. Simulation results show that the proposed simple estimator performs similarly as the restricted maximum likelihood (REML) estimator in mixed models, but is computationally much more efficient than REML. Simulation comparison with the Pearson correlation is also made. A real data example is considered.
Jiming Jiang is Professor of Statistics, University of California, Davis. Ph.D. in Statistics, 1995, Advisor: Peter Bickel; Fellow of ASA and IMS; author of two Springer books, "Linear and Generalized Linear Mixed Models and Their Applications" and "Large Sample Techniques for Statistics"; Associate Editor of the Annals of Statistics; Co-winner of "Outstanding Statistical Application Award" (ASA 1998; joint with A. Gelman and F. Bois); Multiple grant awards of NSF and NIH R01; Research interests: mixed effects models, model selection, longitudinal data, small area estimation.
|Date:||Thursday, March 29, 2012|
|Time:||4:00 - 5:00 P.M.|
Mailman School of Public Health
Department of Biostatistics
722 West 168th Street
New York, New York