Interaction term - dummy*continuous variables

[Patsousasilva]

Hi all,

I'm really hoping that someone can help me with this!
I am running logistic regression models. In the models, I included two independent variables: 1 is a dummy variable that accounts for a complete partisan alternation in government; and the other is a continuous variables that captures the ideological differences between two governments.

Now I wanted to include an interaction term with these two variables, as I want to know the specific impact of government alternation and a change in government ideological profiles.
The problem is
1. Is this feasible? Is it possible to run an interaction term with a dichotomous and a continuous variable?
2. I do not know how to interpret the regression coefficients signs. A negative sign indicates the effect of inexistent partisan alternation in government or does it indicate the effect of small ideological differences between governments?


Can anyone help me with this?

Thank you!

[Sarraf, Shimon Aaron]



I am currently out of the office and will return Monday, July 16. I will respond to your email upon my return. If you need immediate assistance, please call 812-856-5824.

Thank you,

 

Shimon Sarraf

Center for Postsecondary Research

Indiana University Bloomington

 

[Poes, Matthew Joseph]

There is absolutely nothing wrong with a dummy interaction with a continuous variable, or even two dummy variables.  Think of them like switches.  The interaction term is interpreted as the difference in slope coefficient for when the dummy variable is equal to 1.  You will have an effect for just the dummy variable, and this tells you what the difference in intercept is for the complete partisan government.  The Continuous variable tells you how much change there is for each 1 point change in ideology for the non-partisan government.  The Interaction tells you what the difference in slope is for the complete partisan government from the non-partisan.

So that means you have the intercept and slope for the non-partisan group, and then the difference in intercept and slope from the non-partisan group to the partisan group.  That means that in order to know the actual intercept, you must add the coefficient of the dummy variable to the intercept value.  In order to know the value of the slope, you must add the slope coefficient of the interaction to that of the coefficient of the continuous variable.  In order to calculate point estimates for each group, you must do all of this together.

In order to make interpretation easier, I strongly recommend plotting your interaction.  This can be done via online downloadable excel documents fairly easily.  I highly recommend this, as it will make it more clear for you and for anyone else looking at this.

It's important to remember, dummy variables "goof up" main effects, you can't interpret them as straight main effects anymore.  The solution to that is to use effect coding, or in some cases, people simply run the model with and without the dummy variables.

Hope this all helps.

Matthew J Poes
Research Data Specialist
Center for Prevention Research and Development
University of Illinois
510 Devonshire Dr.
Champaign, IL 61820
Phone: [phone]
email: [hidden email]


-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Patsousasilva
Sent: Monday, July 09, 2012 4:18 AM
To: [hidden email]
Subject: Interaction term - dummy*continuous variables

Hi all,

I'm really hoping that someone can help me with this!
I am running logistic regression models. In the models, I included two independent variables: 1 is a dummy variable that accounts for a complete partisan alternation in government; and the other is a continuous variables that captures the ideological differences between two governments.

Now I wanted to include an interaction term with these two variables, as I want to know the specific impact of government alternation and a change in government ideological profiles.
The problem is
1. Is this feasible? Is it possible to run an interaction term with a dichotomous and a continuous variable?
2. I do not know how to interpret the regression coefficients signs. A negative sign indicates the effect of inexistent partisan alternation in government or does it indicate the effect of small ideological differences between governments?


Can anyone help me with this?

Thank you!


--
View this message in context: http://spssx-discussion.1045642.n5.nabble.com/Interaction-term-dummy-continuous-variables-tp5714078.html
Sent from the SPSSX Discussion mailing list archive at Nabble.com.

=====================
To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD