Asked in

How do you interpret coefficient of binary variable in linear regression?


User Avatar
Wiki User
March 31, 2008 10:04PM

This is my best shot. I've been trying to find this answer since I'm doing regressions right now.

Let's say you have a dummy variable "male" where 1 = male, 2 = female.

You regress:

toads_owned = c(1) + c(2)*male

You get the result:


Coefficient: 2

T-test: 3.1

toads_owned = c(1) + 2*male

So now, I think that means that if you are a male, you are likely to own 2 more toads on average than if you were a female.

The coefficient on a dummy variable simply says how different you are from the base group (the group that equals 0) if you equal 1.