0

In R-, how can I get the change in the adjusted R squared and the p-value attached to this change (which would tell me whether this change is statistically significant) from one model to another ? In the case this is relevant, I am trying to compare felm models from the lfe package.

I tried a couple of functions, namely lm.deltaR2(model1, model2) and modelCompare(model1, model2) but this only gives me the change in R squared and not in the adjusted R squared. I thought about computing this change manually by subtracting the adjusted R squared from one model to another, which is an option, but then I don't have the statistical significance of this change.

What I would like to find would look something like this (Using fictive functions here)

model.compare (felm1, felm2) output : change in the adjusted R-squared = 0.02 ; p-value = 0.001

Any suggestion? I guess two of my questions are:

  • Is the statistical significance of the change in R squared the same than the statistical significance of the change in the adjusted R squared (would they show the same p-value?)
  • Is there a way to test whether the two adjusted R squared values are significantly different from one another?

Thank you so much in advance!

Jeanne Roche
  • 1
  • 1
  • It's easier to help you if you include a simple [reproducible example](https://stackoverflow.com/questions/5963269/how-to-make-a-great-r-reproducible-example) with sample input and desired output that can be used to test and verify possible solutions. But these seem like they might be statistical questions which would be a better fit for [stats.se] in the first place. – MrFlick Apr 25 '20 at 19:54
  • not very familiar with the felm package.. and seems like anova doesn't work on it. you might have to write to the authors to ask about testing.. it's a somewhat different from linear model, isolated statistical method... – StupidWolf Apr 25 '20 at 23:09

0 Answers0