Stata commands for random effects adjustment: the reffadjust package

The reffadjust package provides postestimation commands to perform adjustment of random effects estimates.

The reffadjust commands are:

• reffadjustsim: simulating from the distribution of random effect variances and covariances
• reffadjust4nlcom: regression coefficient formula to pass to nlcom.

Since multilevel models including random effects, such as those implemented in MLwiN and mixed/xtmixed, return estimates of the variances and covariances of the random effects and the corresponding variances and covariances of these estimates we can use these to estimate adjusted coefficients.

The approach is described in more detail in Macdonald-Wallis et al. (2012) and Palmer et al. (2014).

Installation

To install, issue in Stata (in versions 13 and above) using the github package, first install the github package:

net install github, from("https://haghish.github.io/github/")

then run

github install remlapmot/reffadjust

Alternatively use net install directly as follows

net install reffadjust, from("https://raw.github.com/remlapmot/reffadjust/master/")

Installation of MLwiN

This is available from https://www.bristol.ac.uk/cmm/software/mlwin/.

The MLwiN libraries are now available natively for macOS (both Intel and Apple Silicon Macs) and a range of Linux ditributions.

On macOS the mlnscript and libmln.dylib files must be granted permission to run in System Settings | Privacy and Security | Scroll down to Security section.

Acknowledgements

Thanks to Gemma Clayton for discovering that reffadjustsim had stopped working under a recent version of Stata. And thanks to Stata Corp. for providing me with a fix for this issue.

Notes

• 2024-10-12 Tested under MLwiN 3.13 and StataNow 18.5
• 2024-08-12 Tested under MLwiN 3.12 and StataNow 18.5
• 2024-06-15 Tested under MLwiN 3.11 and StataNow 18.5
• 2024-03-19 Tested under MLwiN 3.10 and Stata 18.0
• 2024-01-30 Tested under MLwiN 3.09 (and Stata 18)
• 2024-01-29 Tested on Stata 18 and macOS

References

• Macdonald-Wallis C, Lawlor DA, Palmer TM, Tilling K. 2012. Multivariate multilevel spline models for parallel growth processes: application to weight and mean arterial pressure in pregnancy. Statistics in Medicine, 31, 26, 3147-3164. https://doi.org/10.1002/sim.5385
• Palmer TM, Macdonald-Wallis CM, Lawlor DA, Tilling K. Estimating adjusted associations between random effects from multilevel models: the reffadjust package. Stata Journal, 14, 1, 119-140. https://www.stata-journal.com/article.html?article=st0327