Dummy variable interactions

#1
I want include an interaction term in my regression from two dummy variables. Does anyone have any experience with doing this? I can't simply multiply the values because then I get three of the four interactions with the same value (0)

i.e. if
gender (since apparently the s-word is a dirty word) (0=F, 1=M)
Race (0=white, 1=non-white)

my interactions look like this:
F*white=0*0=0
M*white=1*0=0
F*non-white=0*1=0
M*non-white=1*1=1

Not good.

Any ideas would be great
 

Dragan

Super Moderator
#2
I want include an interaction term in my regression from two dummy variables. Does anyone have any experience with doing this? I can't simply multiply the values because then I get three of the four interactions with the same value (0)

i.e. if
gender (since apparently the s-word is a dirty word) (0=F, 1=M)
Race (0=white, 1=non-white)

my interactions look like this:
F*white=0*0=0
M*white=1*0=0
F*non-white=0*1=0
M*non-white=1*1=1

Not good.

Any ideas would be great
You're regression model should appear as follows:

Yhat = b0 + b1G + b2R + b3G*R

where G*R is your interaction term (Gender crossed with Race).


Note: What you are describing is logically analogous to a 2 x 2 ANOVA. You many want to think about you're design in this context.