Workshop 11.2a: Generalized Linear Mixed Effects Models (GLMM)

Murray Logan

07 Feb 2017

Generalized Linear Mixed Effects Models

Parameter Estimation



lm –> LME
(integrate likelihood across all unobserved levels random effects)

Parameter Estimation



lm –> LME
(integrate likelihood across all unobserved levels random effects)


glm —-………–> GLMM
Not so easy - need to approximate

Parameter Estimation



Penalized quasi-likelihood (PQL)

Iterative (re)weighting

Penalized quasi-likelihood (PQL)

Advantages

Disadvantages

Laplace approximation



Second-order Taylor series expansion - to approximate likelihood at unobserved levels of random effects

Laplace approximation



Second-order Taylor series expansion - to approximate likelihood at unobserved levels of random effects

Advantages

Laplace approximation



Second-order Taylor series expansion - to approximate likelihood at unobserved levels of random effects

Advantages

Disadvantages

Gauss-Hermite quadrature (GHQ)

Gauss-Hermite quadrature (GHQ)

Advantages

Gauss-Hermite quadrature (GHQ)

Advantages

Disadvantages

Markov Chain Monte Carlo (MCMC)

Markov Chain Monte Carlo (MCMC)

Advantages

Markov Chain Monte Carlo (MCMC)

Advantages

Disadvantages

Inference (hypothesis) testing

GLMM

Depends on:

Inference (hypothesis) testing

Approximation Characteristics Associated inference R function
Penalized Quasi-likelihood (PQL) Fast and simple, accommodates heterogeneity and dependency structures, biased for small samples Wald tests only glmmPQL (MASS)
Laplace More accurate (less biased), slower, does not accommodates heterogeneity and dependency structures LRT glmer (lme4), glmmadmb (glmmADMB)
Gauss-Hermite quadrature Even more accurate (less biased), even slower, does not accommodates heterogeneity and dependency structures, cant handle more than 1 random effect LRT glmer (lme4)?? Does not seem to work
Markov Chain Monte Carlo (MCMC) Bayesian, very flexible and accurate, yet very slow and complex Bayesian credibility intervals, Bayesian P-values Numerous (see this tutorial)

Inference (hypothesis) testing

Feature glmmPQL (MASS) glmer (lme4) glmmadmb (glmmADMB) MCMC
Variance and covariance structures Yes - not yet Yes
Overdispersed (Quasi) families Yes - - Yes
Mixture families limited limited some Yes
Complex nesting Yes Yes Yes Yes
Zero-inflation - - Yes Yes
Resid degrees of freedom Between-Within - - -
Parameter tests Wald \(t\) Wald \(z\) Wald \(z\) UI
Marginal tests (fixed effects) Wald \(F\), \(\chi^2\) Wald \(F\), \(\chi^2\) Wald \(F\), \(\chi^2\) UI
Marginal tests (Random effects) Wald \(F\),\(\chi^2\) LRT LRT UI
Information criterion - AIC AIC AIC, WAIC

Inference (hypothesis) testing


Additional assumptions



Worked Examples

Worked Examples

\[ \tiny \begin{align} log(y_{ij}) &= \gamma_{Site_i} + \beta_0 + \beta_1 Treat_{i} + \varepsilon_{ij} \hspace{1cm} \varepsilon \sim{} Pois(\lambda)\\ \text{where} \sum{\gamma}&=0 \end{align} \]