Random Effects Ordinal Time Models for Grouped Toxicological Data from a Biological Control Assay
- 2.6k Downloads
Discrete survival times can be viewed as ordered multicategorical data. Here a continuation-ratio model is considered, which is particularly appropriate when the ordered categories represent a progression through different stages, such as survival through various times. This particular model has the advantage of being a simple decomposition of a multinomial distribution into a succession of hierarchical binomial models. In a clustered data context, random effects are incorporated into the linear predictor of the model to account for uncontrolled experimental variation. Assuming a normal distribution for the random effects, an EM algorithm with adaptive Gaussian quadrature is used to estimate the model parameters. This approach is applied to the analysis of grouped toxicological data obtained from a biological control assay. In this assay, different isolates of the fungus Beauveria bassiana were used as a microbial control for the Heterotermes tenuis termite, which causes considerable damage to sugarcane fields in Brazil. The aim is to study the pathogenicity and the virulence of the fungus in order to determine effective isolates for the control of this pest population.
KeywordsAdaptive quadrature Clustered data Multicategorical data Ordinal regression Random effects
This work was part-funded under Science Foundation Ireland’s Research Frontiers Programme (07/RFP-MATF448) and Science Foundation Ireland’s BIO-SI project, grant number 07/MI/012.
- Anderson, D., & Hinde, J. (1988). Random effects in generalized linear models and the EM algorithm. Communications in Statistics—Theory and Methods, 17, 3847–3856.Google Scholar
- Aitkin, M. (1996). Empirical Bayes shrinkage using posterior random effect means from nonparametric maximum likelihood estimation in general random effect models. In Proceedings of the 11th International Workshop on Statistical Modelling (pp. 87–94). Orvieto: Statistical Modelling Society.Google Scholar
- Brillinger, D., & Preisler, M. (1983). Maximum likelihood estimation in a latent variable problem. In S. Karlin, T. Amemiya, & L. Goodman (Eds.), Studies in econometrics, time series and multivariate statistics (pp. 31–35). New York: Academic.Google Scholar
- De Freitas, S. (2001). Modelos para Proporções com Superdispersão proveniente de Ensaios Toxicológicos no Tempo (Ph.D. thesis). ESALQ, Universidade de São Paulo.Google Scholar
- Hinde, J. (1982). Compound Poisson regression models. In R. Gilchrist (Ed.), GLIM 82. New York: Springer.Google Scholar
- Hinde, J. (1997). Contribution to discussion of The EM algorithm - An old folk-song to a fast new tune by Meng, X-L. and Van Dyk, D. Journal of the Royal Statistical Society, Series B, 59, 549–550.Google Scholar
- McCullagh, P., & Nelder, J. (1989). Generalized linear models. London: Chapman & Hall.Google Scholar
- Nelder, J. (1985). Quasi-likelihood and GLIM. In R. Gilchrist, B. Francis, & J. Whittaker (Eds.), Generalized linear models: Proceedings of the GLIM 85 conference. New York: Springer.Google Scholar
- Rabe-Hesketh, S., Skrondal, A., & Pickles, A. (2002). Reliable estimation of generalized linear mixed models using adaptive quadrature. The Stata Journal, 2, 1–21.Google Scholar