Is it legitimate or valid to include a binary independent variable in sem analysis?

[Taka Sakamoto]

Dear Stata-listers:

Is it legitimate or valid to include a binary independent variable in sem analysis? I've looked it up, but so far haven't come across a solid answer. On the one hand, the Stata Manual (on gsem) has an example of a binary independent affecting a dependent variable. On the other hand, some researchers say that sem treats a binary independent variable as a continuous variable, so sem can include a binary independent variable. But in this case, I don't know if it is a valid thing from a statistical and methodological perspective.

I would be grateful if someone could point me in a correct direction.

Thank you for your help in advance.

Best,

Taka Sakamoto

[Richard Williams]

I’m not aware of anything that says sem treats a binary independent variable as continuous. I can’t think of any type of analysis where it isn’t legitimate to include a binary independent variable, but I suppose there may be something.

[Leonardo Guizzetti]

I second Richard Williams' advice. I can think of some software (MPlus) which will give a warning about including binary independent variables having potentially erroneous standard errors, but that's not typically a real concern.

[Taka Sakamoto]

Thank you so much for your kind help. I feel very hopeful and happy as I can use my binary variable in my research. Many thanks, both of you.

Best,

Taka