Mixed effect models and SEMs

I’m increasingly convinced of the value of structural equation models (SEMs)in ecology – particularly if random effects and phylogenetic relationships can be incorporated. That is exactly what has been achieved in the new R package ‘Piecewise SEM’ by Jonathan Lefcheck recently published Methods in Ecology and Evolution.  Before this if your model variables were not normally distributed or independent this was a big problem for SEM models. This limited the use of this type of modelling in ecology as ecological data, for the most part, violates these assumptions readily. This package extends SEMs to include mixed effect models, all sorts of other data distributions (Poisson etc) and phylogentic generalized least-squares (PGLS). This seems like a really nice extension of the idea and the package looks relatively easy to implement. There are limitations, of course, such as not being able to account for bidirectional relationships, but  as Jon suggests, future work will rectify these. Nonetheless this is a useful extension of the SEM idea which I’m sure will lead to greater uptake of this useful modelling approach.

Here is the link: http://onlinelibrary.wiley.com/doi/10.1111/2041-210X.12512/full


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s