Abstract
The selection of optimal designs for generalized linear mixed models is complicated by the fact that the Fisher information matrix, on which most optimality criteria depend, is computationally expensive to evaluate. We provide two novel approximations that reduce the computational cost of evaluating the information matrix by complete enumeration of response outcomes, or Monte Carlo approximations thereof: an asymptotic approximation that is accurate when there is strong dependence between observations in the same block; and an approximation via kriging interpolators. For logistic random intercept models, we show how interpolation can be especially effective for finding pseudo-Bayesian designs that incorporate uncertainty in the values of the model parameters. The new results are used to evaluate the efficiency, for estimating conditional models, of optimal designs from closed-form approximations to the information matrix derived from marginal models. Correcting for the marginal attenuation of parameters in binary-response models yields much improved designs, typically with very high efficiencies. However, in some experiments exhibiting strong dependence, designs for marginal models may still be inefficient for conditional modelling. Our asymptotic results provide some theoretical insights into why such inefficiencies occur.
Original language | English |
---|---|
Pages (from-to) | 677-693 |
Number of pages | 16 |
Journal | Biometrika |
Volume | 102 |
Issue number | 3 |
DOIs | |
Publication status | Published - Sep 2015 |
Keywords
- Bayesian design
- Binary response
- Blocked experiment
- Count response
- Generalized linear mixed model
- Kriging
- Outcome enumeration
- Quasilikelihood