Interaction effects are used in logistic regression models just as they are in any other linear model. See KUTNER ET AL APPLIED LINEAR STATISTICAL MODELS 5 ed for full details

This book is available in the z-library for download Best wishes David Booth

With all two-way interactions:

Price_Deviation = B0 + B1 (prestegious_investor) + B2 (default_score) +Bk (controls) + B3 (prestegious_investor)(default_score) + B4 (default_score)(controls) + B5 (default_score)(controls)

and if you also want to include the three-way interaction you need to add the term B6 (prestegious_investor)(default_score)(controls).

When using nonlinear models such as logistic regression models interaction terms can't be interpreted in the same way as in linear models (OLS-regression). Berry et al. (2012) show that here you have to distinguish between interaction terms and interaction effects.

I would recommend to calculate and plot marginal effects. How this can be done in principle is shown in Mize (2019) (how to do this in Stata is shown in lecture material by Williams here: https://www3.nd.edu/~rwilliam/stats3/ (see "Interpreting results: Adjusted Predictions and Marginal effects").

Helpful literature explaining what is different in nonlinear models (and hence in logistic regression models) and what to do to interpret results appropriately can be found here (in alphabetical order):

- Berry, W. D., Golder, M., & Milton, D. (2012). Improving tests of theories positing interaction. The Journal of Politics, 74, 653--671.

- Mize, T. D. (2019). Best practices for estimating, interpreting, and presenting nonlinear interaction effects. Sociological Science, 6, 81-117. Article Best Practices for Estimating, Interpreting, and Presenting ...

[Strongly recommended!]

- Mood, C. (2010). Why we cannot do what we think we can do and what to do about it. European Sociological Review, 26, 67--82.

[A classic to explain the general problem]

Additionally, when you are ilnterested in mediation effects, as a starter you should have a look at:

- Kohler, U., Karlson, K. B., & Holm, A. (2011). Comparing coefficients of nested nonlinear probability models. The Stata Journal, 11, 420--438. Article Comparing Coefficients of Nested Nonlinear Probability Models

In my previous post I erroneously cited a paper by Berry et al. (2012) that is not relevant here (although interesting for those who want to interpret interaction effects). The paper that I should have cited instead is:

- Berry, W. D., DeMeritt, J. H. R., & Esarey, J. (2010). Testing for interaction in binary logit and probit models: Is a product term essential? American Journal of Political Science, 54, 248--266.

By the way: Berry et al. (2010) quote Norton et al. (2004) who describe a procedure different from investigating marginal effects as described by Mize (2019) (see my previous post). See:

- Norton, E. C., Wang, H., & Ai, C. (2004). Computing interaction effects and standard errors in logit and probit models. The Stata Journal, 4, 154--167.

Article Computing Interaction Effects and Standard Errors in Logit a...

Many thanks Dirk Enzmann !

I had heard about Norton's paper but you gave me a complete reading list. That's what I was looking for. I'm grateful for the favor.

Also thanks to David Eugene Booth and Jochen Wilhelm for their views.

The only problem with Dirk Enzmann post is that he is completely wrong. Download a copy of KUTNER ET AL APPLIED LINEAR STATISTICAL MODELS from the z-library and read about logistic regression and interaction. Remember what the Romans used to say was CAVEAT EMPTOR. Best wishes David Booth

Dear Prof. David Eugene Booth

I don't exactly know, how is he (Dirk Enzmann) wrong. If you please specify the point of contradiction, it will make more sense to me.

However, I'll study both the resources (your's and Enzmann's). I hope, I'll be benefitted from these.

Thanks and regards,

Dear Prof. David Eugene Booth

I don't exactly know, how is he (Dirk Enzmann) wrong. If you please specify the point of contradiction, it will make more sense to me.

However, I'll study both the resources (your's and Enzmann's). I hope, I'll be benefitted from these.

Thanks and regards,

Dear Prof. David Eugene Booth

I don't exactly know, how is he (Dirk Enzmann) wrong. If you please specify the point of contradiction, it will make more sense to me.

However, I'll study both the resources (your's and Enzmann's). I hope, I'll be benefitted from these.

Thanks and regards,

Dear Prof. David Eugene Booth

I don't exactly know, how is he (Dirk Enzmann) wrong. If you please specify the point of contradiction, it will make more sense to me.

However, I'll study both the resources (your's and Enzmann's). I hope, I'll be benefitted from these.

Thanks and regards,

Dear Prof. David Eugene Booth

I don't exactly know, how is he (Dirk Enzmann) wrong. If you please specify the point of contradiction, it will make more sense to me.

However, I'll study both the resources (your's and Enzmann's). I hope, I'll be benefitted from these.

Thanks and Regards,

Jochen Wilhelm was correct when he described logistic models. Please reread his post. Also get Kutner et al and read it. YOU ARE RESPONSIBLE FOR UNDERSTANDING YOUR OWN RESEARCH. DAVID BOOTH

David Eugene Booth Thankyou! I'm studying these resources (& posts), trying to understand.

Here's an application of a logistic regression that's a linear model. It's LINEAR in the regression coefficients. As Dr Wilhelm mentioned it could have been nonlinear coefficients in a different application. Best wishes David Booth

I can't reply to a bold statement such as that what I wrote would be completely wrong if it is not substantiated by any argument -- David Eugene Booth didn't even specify in which way this should be wrong. Hence, my suggestion is to read the material and see if it makes sense to you.

Logistic regression is non-linear in terms of probability (i.e. the probability P(Y=1) as a function of a predictor X is non-linear) or odds, but linear in terms of logits (or log odds). Perhaps this may help to clarify things: https://stats.stackexchange.com/questions/365391/logistic-regression-is-a-nonlinear-regression-problem.

I had a look at the book recommended by David Eugen Booth [Kutner, M. H., Nachtsheim, C. J., Neter, J., & Li, W. (2005). Applied Linear Statistical Models. New York, NY (5th ed.): McGraw-Hill]. On page 512 (in Part III "Nonlinear Regression" !) they write: "Another important nonlinear regression model is the logistic regression model."

David should try to be more carefully (especially when reading the material he himself is recommending) before jumping to such confusing and bold statements as "he is completely wrong".