Using a multiple regression, do all observations have to be from the same dataset? For example, assume a survey is conducted with some likert scale questions along with other predominantly categorical data. Can pure continuous data from another dataset be included even though this data would not be from each individual respondent in the survey? The population is however the same. Would it be legitimate to combine these datasets/ their results?

  • $\begingroup$ Definitely not. $\endgroup$ – llewmills Mar 6 '19 at 3:54
  • $\begingroup$ Sorry that was a bit brief. To make inferences about the effect of one predictor on the outcome while controlling for the effect of the other predictor, each observed instance of the outcome needs to have an observation of both predictors. $\endgroup$ – llewmills Mar 6 '19 at 4:01

Your Answer

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

Browse other questions tagged or ask your own question.