I have a longitudinal dataset. I wish to do some exploratory data analysis on it. However, it has the problem of missing observations/non-response.

Q: How does one do exploratory data analysis on this dataset so as not to bias the hypothesis generated?

My intention is to do multiple imputation first and then conduct exploratory data analysis. However, if missing mechanism is MNAR, this will certainly bias the hypothesis and testing which should be detected after analysis. I only want to do exploratory analysis without analysis to generate hypothesis.



Your Answer

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