Skip to content

tslumley/svylme

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

svylme

Mixed models for complex surveys

This package fits linear mixed models to data from complex surveys, by maximising a weighted pairwise likelihood

remotes::install_github("tslumley/svylme")

Advantages

It works (gives consistent estimates of the regression coefficients and variance components) for any linear mixed model and any design, without any restrictions on the sampling units and model clusters being related. For example, you could sample on home address but fit a model clustering on school.

The implementation allows for correlated random effects such as you get in quantiative genetics

Disadvantages

Linear models only

Some loss of efficiency compared to just fitting a design-based linear model (if you don't care about the variance components)

There isn't (yet) an analog of the BLUPs of random effects, eg for small-area estimation

If your sampling units and model clusters are the same, and your design isn't too strongly informative, you can likely get more precise estimates of the variance components with stagewise pseudolikelihood as implemented in Stata or Mplus.

About

Mixed models for complex surveys

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages