Entering edit mode
3.1 years ago
RNAseqer
▴
280
If you feed step the output of lm() using the all possible covariates, and the output of lm() using all the possible covariates using both directions and AIC criteria it will search for the most informative model correct? If you extract the summary function on the step() output you can then extract pvalues, tvalues etc for your variable of interest from the summary output.
But here is my question, what model by default is step using to get the statistics you find in summary? Is it using lm() because you gave it two lm() outputs? Or does it use something else to calculate those statistics?