VIF and multicollinearity diagnostics

In the book I use the car package to get VIF and other multicollinearity diagnostics. I’ve occasionally found this breaks down (usually through mixing different versions of R on different machines at work home or on the move). I recently saw the mctest package and thought it would be useful to use that as a backup – and also because it offers a slightly different set of diagnostics.

If you’d like to try it out I’ve written a helper function that makes it easier apply directly to a linear model. You can find the function and a simple example here.


CI for difference between independent R square coefficients

In an earlier blog post I provided R code for a CI of a difference in R square for dependent and non-dependent correlations. This was based on a paper by Zou (2007). That paper also provides a method for calculating the CI of a difference in independent R square coefficients based on the limits of the CI for a single R square coefficient. I’ve also been experimenting with knitr and unfortunately haven’t yet worked out how to merge the R markdown output with my blog template, so I’m linking to RPubs for convenience.

You can find the function and a few more details here.