I'm running an OLS regression in SPSS and have a question about models that feature both Scale and Nominal/Ordinal variables. I say nominal/ordinal because the variables I'm looking at range from 0-2, with 0 representing a restricted worker's rights condition, 1 a somewhat restricted condition, and 2 an unrestricted condition. So I feel since it has a logical progression from least to most freedom, that it could well be ordinal. Does such a variable require recoding if I included it in a model that also features scale variables? My DV is also scale.

Thank you!


1 Answer 1


Categorical variables, whether nominal or ordinal, should be recoded to dummies for use in ordinary regression. (The STATS CREATE DUMMIIES or Data > Create Dummy Variables extension command can do this for you conveniently, although with only three known value, COMPUTE would be adequate.) Or you could use GLM, which understands factors.

It does not matter for OLS whether these are considered nominal or ordinal. However, if you have the Categories option, Analyze > Regression > Optimal Scaling can find an optimal scale for your categorical variables in a regression.

  • $\begingroup$ Hi, thanks so much for the answer! I have ran the models with dummy variables but I can't make much sense out of what the answers mean. I created two dummies for the categorical variables, one that represents restricted rights and one that represents unrestricted rights, but I don't quite know how to interpret that in terms of the other case yet. I've only started working with regression analysis a few weeks ago so this is all still quite confusing. $\endgroup$
    – akask
    Jan 9, 2016 at 2:51
  • $\begingroup$ Think about what hypotheses you want to test. $\endgroup$
    – JKP
    Jan 10, 2016 at 4:43

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

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