CLME stands for Constrained Linear Mixed Effects. I wrote this R package (CRAN link) during my postdoctoral work at NIEHS.
The fundamental idea is similar to the Jonckheere–Terpstra or any other test for ordered alternatives: If the treatment groups are ordinal, then a trend of some sort may be of interest. If a researcher has such a hypothesis, they can not only test for the ordered alternative, but they can constrain the estimation to respect the order from the alternative hypothesis. This results in getting more power than a comparable test that does not impose constraints or test for an order (e.g., ANOVA).
Feel free to check out the github repository. There are a variety of improvements I’d like to make, including cleaning up my code, removing dependancies, and adding some features. Just need time to get to them. If you’d like to join, give me a shout!
Disclaimer: This is really just a post to get something into the “Project” space of my website. I’ll probably modify this post later to add clarity.
