[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: OLS three-way interaction

From   Antonio Silva <>
To   Stata list <>
Subject   st: OLS three-way interaction
Date   Wed, 4 Nov 2009 08:07:50 -0500

Stata listers:
Here is my situation. I am interested in estimated an OLS regression model with three independent variables of interest, and some controls. Let us call the three variables w, x, and z. W is a continuous variable, x is a dummy (let?s call the two levels x0 and x1), and z is categorical with 3 levels (let?s call them z0, z1, and z2). I wish to estimate a model with a three-way interaction between w, x, and z. As I understand it, the model I estimate would look like this:
Y = w + x1 + z + wx1 + wz + x1z+ wx1 + x1z1 + x1z2 + wz1 + wz2 + wx1z1 + wx1z2
My first question is this: Is this correct? I think I have accounted for all the constitutive terms here, which I know is the best way to go. Second, what do you think the best way to do this is in stata? The estimation of the model seems straightforward, but interpreting the results seems much less so to me. Do you think I should rely upon lincom to come up significance levels for various values of the independent variables? Or does the Golder add-on hold the most promise. Finally, due to the difficulty in interpreting three-way interactions?especially when they involve one continuous variable, one dummy, and one categorical variable?should I take a different approach? That approach would be to estimate three separate models instead of just one?one each for people in groups z0, z1, and z2. 
I would appreciate any and all feedback.  		 	   		  
Windows 7: Unclutter your desktop.

*   For searches and help try:

© Copyright 1996–2023 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index