A question about bimodality.

First question - if my Y variable has a bimodal distribution, is it at all possible to run a regression on this variable?

Second - if my Y variable is not normally distrubuted in GENERAL is it possible to run a regression?

Thanks for any advice :)


New Member
You can apply a log transformation on your Y variable data so you can make it appropriate for multiple regression. It corrects non-normal data.