Discrete DV and Multiple Interactions

[taehyun-lim]

Hello!

Thank you very much for the interflex package! It is very useful!

I had two questions using the package.

First, my DV is discrete variable, so I was wondering if I can estimate a logit or probit model with interflex package.

Second, I was wondering if I can use the package to estimate a model with three sets of interaction variables, that is, for example, y = a + b + ab + c + ac + d + a*d.

Thank you very much!

[xuyiqing]

Thank you. On your first question, yes, we're trying to extend it to generalized linear models. But it may take a few months. Right now, we can simply use linear approximation as a tentative solution. On the second question, multi-way interactions are automatically included in the model but we only focus on (flexible) marginal effect over one particular moderator. In your case, for example, you may be focusing on a*b, but a*c and a*d are already in the model during estimation when full.moderate = TRUE

…

On Mon, Feb 24, 2020 at 4:14 PM limth85 ***@***.***> wrote: Hello! Thank you very much for the interflex package! It is very useful! I had two questions using the package. First, my DV is discrete variable, so I was wondering if I can estimate a logit or probit model with interflex package. Second, I was wondering if I can use the package to estimate a model with three sets of interaction variables, that is, for example, y = a + b + a*b + c + a*c + d + a*d. Thank you very much! — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#6?email_source=notifications&email_token=AB2PKGFVOLRR3AVCQADAMOTRERPHFA5CNFSM4K2WXWA2YY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4IP466EQ>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AB2PKGADYALNAQIRU4H45P3RERPHFANCNFSM4K2WXWAQ> .

-- Yiqing Xu Assistant Professor Department of Political Science Stanford University http://yiqingxu.org/

[taehyun-lim]

Thank you!

I am excited that you are working on a generalized linear models!

On the second question, I am guessing full.moderate = TRUE would be for R. Is there a way I could use the option in stata?

[xuyiqing]

Oh, I'm sorry. The Stata version hasn't caught up with the R version yet (esp that we're constantly developing new functionalities). Apologies. We'll update the State version when the R version is stabilized.

…

On Mon, Feb 24, 2020 at 4:25 PM limth85 ***@***.***> wrote: Thank you! I am excited that you are working on a generalized linear models! On the second question, I am guessing full.moderate = TRUE would be for R. Is there a way I could use the option in stata? — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#6?email_source=notifications&email_token=AB2PKGF2VJIDTO3YDN57JGDRERQOBA5CNFSM4K2WXWA2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEM2BWEI#issuecomment-590617361>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AB2PKGDWUQATJBWERV2Q5B3RERQOBANCNFSM4K2WXWAQ> .

-- Yiqing Xu Assistant Professor Department of Political Science Stanford University http://yiqingxu.org/

[taehyun-lim]

Aha, thank you!

I had one last question. I was wondering how it will be different from your binning method if I make three dummies from my continuous moderator variable -- low, middle, and high -- and use it as a set of dummies..