On modelling mean-covariance structures in longitudinal studies

by: Jianxin Pan, Gilbert Mackenzie

Biometrika, Vol. 90, No. 1. (1 March 2003), pp. 239-244.

View FullText article

DOI, HighWire

Reviews [Write a review of this article]

There are no reviews of this article

Find related articles from these CiteULike users

abrentnall

Find related articles with these CiteULike tags

covariance, longitudinal, modelselection, statistics

Abstract

We exploit a reparameterisation of the marginal covariance matrix arising in longitudinal studies (Pourahmadi, 1999, 2000) to model, jointly, the mean and covariance structures in terms of three polynomial functions of time.By reanalysing Kenward's (1987) cattle data, we compare model selection procedures based on regressogram estimation with these based on a global search of the model space. Using a BIC-based model selection criterion to identify the optimum degree triple of the three polynomials, we show that the use of a saturated mean model is not optimal and explain why regressogram-based model estimation may be misleading. We also suggest a new computational method for finding the global optimum based on a criterion involving three pairwise saturated profile likelihoods. 10.1093/biomet/90.1.239

@article{citeulike:3033066, abstract = {We exploit a reparameterisation of the marginal covariance matrix arising in longitudinal studies (Pourahmadi, 1999, 2000) to model, jointly, the mean and covariance structures in terms of three polynomial functions of time.By reanalysing Kenward's (1987) cattle data, we compare model selection procedures based on regressogram estimation with these based on a global search of the model space. Using a BIC-based model selection criterion to identify the optimum degree triple of the three polynomials, we show that the use of a saturated mean model is not optimal and explain why regressogram-based model estimation may be misleading. We also suggest a new computational method for finding the global optimum based on a criterion involving three pairwise saturated profile likelihoods. 10.1093/biomet/90.1.239}, author = {Pan, Jianxin and Mackenzie, Gilbert }, citeulike-article-id = {3033066}, doi = {10.1093/biomet/90.1.239}, journal = {Biometrika}, keywords = {covariance, longitudinal, modelselection, statistics}, month = {March}, number = {1}, pages = {239--244}, posted-at = {2008-07-22 17:04:00}, priority = {5}, title = {On modelling mean-covariance structures in longitudinal studies}, url = {http://dx.doi.org/10.1093/biomet/90.1.239}, volume = {90}, year = {2003} }

RIS record

TY - JOUR ID - citeulike:3033066 L3 - citeulike-article-id:3033066 TI - On modelling mean-covariance structures in longitudinal studies JF - Biometrika VL - 90 IS - 1 SP - 239 EP - 244 N2 - We exploit a reparameterisation of the marginal covariance matrix arising in longitudinal studies (Pourahmadi, 1999, 2000) to model, jointly, the mean and covariance structures in terms of three polynomial functions of time.By reanalysing Kenward's (1987) cattle data, we compare model selection procedures based on regressogram estimation with these based on a global search of the model space. Using a BIC-based model selection criterion to identify the optimum degree triple of the three polynomials, we show that the use of a saturated mean model is not optimal and explain why regressogram-based model estimation may be misleading. We also suggest a new computational method for finding the global optimum based on a criterion involving three pairwise saturated profile likelihoods. 10.1093/biomet/90.1.239 KW - covariance KW - longitudinal KW - modelselection KW - statistics AU - Pan, Jianxin AU - Mackenzie, Gilbert PY - 2003/03/01/ UR - http://dx.doi.org/10.1093/biomet/90.1.239 ER -

RIS BibTeX