I am using the World Values Survey and the answers to my independent variable are as follows:

Daily Weekly Monthly Never

Answers to my dependent variable:

Always Usually Never

As you can see, the categories of the ordinal variable are not equally spaced. Would it be any possible to use this variable in my regression analysis?


For the dependent variable, the usual starting point is ordinal logistic regression. This makes some assumptions (chiefly that of proportional odds) and there are ways of dealing with violations of this. These have been discussed here before.

For the independent variable, things are a little murkier. Some people treat it as categorical, some as continuous (e.g. by coded them with numbers). One general approach is optimal scaling. I don't know if this is available in Stata.

For your particular variable, one approach that I have used is to convert it into times per month. So, daily - 30, weekly = 4, monthly = 1, never = 0. You can do a sensitivity analysis by changing these a bit (after all, people who answer "daily" might not be exactly 30 times per month).

  • $\begingroup$ Thank you. Since you are new,the usual thing to do when an answer meets your needs is to "accept" it by clicking the check mark. $\endgroup$ – Peter Flom - Reinstate Monica Dec 24 '18 at 12:20

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.