2

I need to run multiple regression with 2 IVs and 1 DV and include a control variables. The control variable is "Type of House" and has 4 categories: 1=bungalow, 2=apartment, 3=penthouse and 4=condominium.

How do I enter the control variables into the model? Is it just the way it is (i.e., nominal) or do I have to create dummy variables? If I need to create dummy variables, do I have to enter only 3 categories (i.e. k-1)? Lastly, how do I present the table and interpret?

kjetil b halvorsen
  • 63,378
  • 26
  • 142
  • 467
Muzi
  • 101
  • 1
  • 6
  • What software do you use for fitting the regression? – COOLSerdash Jun 06 '13 at 12:34
  • 2
    You need to create dummy variables. Normally, the software will do that automatically, even if you are not aware of it (I don't know how SPSS handles that). You only need $k-1$ dummies if you have $k$ categories. For the interpretation, see my answer [here](http://stats.stackexchange.com/questions/60817/significance-of-categorical-predictor-in-logistic-regression/60821#60821). The coefficients of the categories are usually the difference between the reference category and the other categories. But because it is a control variable, you are probably not interested in the significance. – COOLSerdash Jun 06 '13 at 12:47
  • 2
    A search here for [dummy coding](http://stats.stackexchange.com/search?q=dummy+coding) suggests you can find a few hundred answers and examples :-). Here's one that's particularly close to your set of questions: http://stats.stackexchange.com/questions/24242/how-to-apply-coefficient-term-for-factors-and-interactive-terms-in-a-linear-equa – whuber Jun 06 '13 at 14:22

0 Answers0