All the control variables were there because they affected the dependent variables in previous studies. However in my dataset, all are insignificant and do not affect the main effects.

Should I remove them all? Not sure what this means.....

p.s. I tried PLS and standard regression and the results are similar.


If you leave your control variables in, then you can say that they are not significant and do not affect the main effects; you can show that they are different then in previous studies and, if you are using exactly the same main effects and covariates as earlier studies, you can compare the parameters directly.

You can then ask why your results were different.

Was it sample size? Were the parameter estimates similar in size as earlier studies?

If not, what was different about your sample or the way you measured things?

| cite | improve this answer | |
  • $\begingroup$ The main effects is new and has not been studied before. Many different sample sizes have been used before on the control variables -not sure how it affects the results. Parameter estimates always vary, but I am not sure how big a variation actually mean something... $\endgroup$ – Transhumanist Jul 11 '14 at 13:28

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.