Hello Juan,

As long as the factors are recast as dummy variates (or similar; k - 1 variates for a k-level factor), then there's no reason that the dummy variates, combined with the continuous score variables, couldn't be used for a bootstrap analysis.

Good luck with your work.

In R, you can use the boot package with the boot.ci function to construct a confidence interval by bootstrap of pretty much any result from a function. The tricky part is to figure out how to extract the specific parameter estimate you want from the function output.

I have an example here, https://rcompanion.org/handbook/G_10.html , in the section Optional analyses: Confidence intervals for R-squared values. It's a little complex, but in that example, we write a function Function, that runs the linear model and extracts the R-squared value. Once that function is written, and it successfully extracts the desired parameter estimate, we can pass it to the boot function.

So what boot is going to do is run the linear model many times (5000 in my example), each time with a random bootstrapped sample of the data. From there, boot.ci can estimate a confidence interval of the output of Function by a few different methods (percentile, BCa, etc.).