Statistical methods

Choose a method from the response distribution, dependence structure, and scientific estimand. Each guide connects the model definition to the fitted objects, diagnostics, and worked examples that implement it.

Linear mixed models

Continuous Gaussian outcomes with grouped random intercepts or slopes. Covers ML and REML, covariance estimation, conditional and population prediction, and simulation.

Linear mixed models

Generalized mixed models

Binary and count outcomes with Gaussian random effects. Covers supported families, conditional interpretation, first-order Laplace fitting, calibration, and simulation.

Generalized linear mixed models

Repeated-measures models

Continuous longitudinal outcomes with structured within-subject residual covariance. Covers visit ordering, missing-response handling, contrasts, degrees of freedom, and covariance diagnostics.

Mixed models for repeated measures

Choose quickly

Scientific structure	Start with	Primary dependence model
Continuous response; groups or nested units	LMM	Random effects
Binary or count response; grouped observations	GLMM	Gaussian random effects on the link scale
Continuous response at scheduled visits	MMRM	Structured residual covariance

If both a random trajectory and a repeated-measures residual structure appear plausible, decide which representation matches the scientific question before selecting an engine. The analysis chooser and analysis matrix compare the supported execution paths in more detail.

• Linear mixed models
  • When to use an LMM
  • Declare a model
  • ML and REML
  • What the fit returns
  • Prediction modes
  • Variance and random-effect interpretation
  • Diagnostics
  • Reference-engine scope
  • Worked examples
  • API
• Generalized linear mixed models
  • Supported fit subset
  • Fit a Bernoulli random-intercept model
  • Laplace calculation
  • Interpret coefficients on the correct scale
  • Random-effect modes
  • Diagnostics
  • Simulation and predictions
  • Reference-engine scope
  • Worked examples
  • API
• Mixed models for repeated measures
  • When to use MMRM
  • Declare covariance and mean
  • Visit order is scientific state
  • Missing responses
  • Fixed effects and contrasts
  • Degrees of freedom
  • Covariance checks
  • Diagnostics
  • Reference-engine scope
  • Worked examples
  • API