Misspecifying the likelihood for clustered binary data

被引:16
作者
Molenberghs, G [1 ]
Declerck, L [1 ]
Aerts, M [1 ]
机构
[1] Limburgs Univ Ctr, B-3590 Diepenbeek, Belgium
关键词
clustered data; dose-response models; likelihood estimation; litter effect; reproductive toxicology;
D O I
10.1016/S0167-9473(97)00037-6
中图分类号
TP39 [计算机的应用];
学科分类号
081203 ; 0835 ;
摘要
The effect of misspecifying the parametric response model for a clustered binary outcome from a toxicological study on the assessment of dose effect is investigated. A marginal, random effects, and conditional model are contrasted, with the emphasis on likelihood based estimation. The methods are compared through asymptotic calculations, by means of small sample simulations, and on real developmental toxicity data. It is found that the beta-binomial and conditional models exhibit satisfactory behavior in terms of testing the null hypothesis of no dose effect. Whereas the conditional model has clear computational advantages, parameters in the beta-binomial model have a straightforward marginal interpretation. (C) 1998 Elsevier Science B.V. All rights reserved.
引用
收藏
页码:327 / 349
页数:23
相关论文
共 32 条
[1]  
AERTS M, 1997, IN PRESS ENVIRONMETR
[2]  
Altham P.M. E., 1978, Applied Statistics, V27, P162, DOI DOI 10.2307/2346943
[3]  
BAHADUR RR, 1961, STANFORD MATH STUDIE, V6
[4]   BIVARIATE LATENT VARIABLE MODELS FOR CLUSTERED DISCRETE AND CONTINUOUS OUTCOMES [J].
CATALANO, PJ ;
RYAN, LM .
JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION, 1992, 87 (419) :651-658
[5]   QUANTITATIVE RISK ASSESSMENT FOR TERATOLOGICAL EFFECTS [J].
CHEN, JJ ;
KODELL, RL .
JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION, 1989, 84 (408) :966-971
[6]  
CHEN JJ, 1994, STAT SINICA, V4, P265
[7]  
COX DR, 1972, J ROY STAT SOC C-APP, V21, P113, DOI 10.2307/2346482
[8]  
COX DR, 1994, BIOMETRIKA, V81, P403, DOI 10.2307/2336971
[9]  
DECLERCK L, 1997, UNPUB BEHAV LIKELIHO
[10]  
Diggle P., 2002, Analysis of Longitudinal Data