0
$\begingroup$

I have a data having 2 continuous and 4 categorical variables. Each categorical variable has 3 levels. I want to know how to include the variables in the model. I am using SPSS

Variables:

  1. Sales - Dependent Variable
  2. Retail Price
  3. Location : TX, CA, PY
  4. Display Location in Store : DL1, DL2, DL3
  5. Price Point : 2.5, 4.5, 1.0

I have created 9 dummy variables. 3 for Location, 3 for Display Location in Store and 3 for Price Point.

I am confused on how to use these independent variables. How to add the independent variables into the models. Should I add all the nine dummy variables or two dummy variables of each categorical variables along with Retail Price.

I searched on net but , did not get any solution. I would appreciate any kind of help.

Thanks in advance.

$\endgroup$
1
$\begingroup$

It has been a while since I last used SPSS, but I am pretty sure that it can handle the creation and inclusion of dummy variables automatically (or mostly automatically). You should not need to create the dummy variables yourself, just try including the categorical variables as predictors when specifying the model and see what happens. You do need to make sure that SPSS recognizes that they are categorical variables, this is not a problem if the data is stored as "TX", "CA", "PY", but with the Price Point variable you may need to somehow specify it as a categorical variable/factor.

If you want to do it by hand, then you can only include 2 dummy variables for each variable (the one that you do not include will become the baseline level).

$\endgroup$
  • $\begingroup$ I have already created the dummy variables. So, that means we should use two dummy variables from each Categorical Variables, right? What would happen, if I include all the dummy variables in my model? Sorry for asking many questions, but this things, really confuses me. $\endgroup$ – Arushi Jul 22 '14 at 18:04
  • $\begingroup$ @Arushi, if you include all the dummy variables then you will either see an error (something like a singular x-matrix) or the program will drop the extra dummy variables (and may or may not tell you). I don't remember which SPSS does. $\endgroup$ – Greg Snow Jul 22 '14 at 18:45
0
$\begingroup$

If you start with the original categorical variables, the result depends on which procedure you use. REGRESSION treats all variables as continuous, so it would not give reasonable results here. You would need to use the dummy variables.

If you use GLM, you classify the regressors as scale or factor, and it will do the right thing. Similarly with GENLIN and several other procedures related to linear models.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.