Command for SEM standardized coefficients on y not xy?

[Ummul Kathawalla]

Hello,

I am running a path analysis with SEM on Stata and I know the stand command gives me standardized coefficients with x and y standardized. I have a categorical variable in my model (coded as dummy variables) and I would like to report the standardized coefficient where just y is standardized. If I am understanding correctly, I think that would be easier to interpret. I know that in MPLUS there is a command for stdy standardization. Is there a command for Stata?

I found listcoef command for regressions. But nothing for SEM.

Thank you for your help!

[Richard Williams]

I haven't tried it. Some ideas:

If y is observed, standardize it yourself.

If y is latent. you could change the way it gets identified, e.g instead of fixing one of its indicators at 1, fix its error variance at whatever value will makes its total variance one.

If you want to make sure you've done things right, compare the standardized solutions you get before you fidle around with y with the ones you get afterward. They should be same.

Having said all that, I am not that big of a fan of y standardization. But I imagine there is some way to do it.

[Ummul Kathawalla]

Thank you so much for your response and help! These ideas are very helpful. I think it will be the easiest for me to just standardized y.

If you do have a chance to reply, why are you not a big fan of y standardization? (Even for binary x's?)