Hi all,

I am running a hierarchical linear regression where the first block contains demographic variables including:

sex (male=1, female=0)
spouse (yes=1, no=0)

And have been trying to add an interaction term because it looks like there is a significant sex*spouse interaction. But each time I add these dummy variables:

maleSpouse (1,0)
maleNoSpouse (1,0)
femaleSpouse (1,0)

With females having no spouse as the reference group, my main effects (spouse, sex) disappear from the model -- they are just not in the coefficient list anymore. What is happening here?

Thanks,
statz_grl :wave:

Maybe I'm missing something, but shouldn't the dummy variables be coded as:

maleSpouse (1,1)
maleNoSpouse (1,0)
femaleSpouse (0,1)