0
$\begingroup$

I would like to perform hierarchical regression in which all variables are based on previous research/theory. But when I perform Pearson correlations, I found that some variables did not correlate to the DV. Should I still include those non-significant variables in the hierarchical regression as suggested by previous research even though they were not significant in Pearson correlations?

Thank you very much for your help.

$\endgroup$
0
$\begingroup$

Pearson correlation only tells you about the degree of "linear relationship" between variables. But there can actually be strong "non-linear" relationship between two variables too! Depending on what type of variable you have, you can use different correlation coefficients. So I think you should check possible non-linear relationships and associations. If the independent variables are by no means correlated to the dependent variable, then they will contribute little to a predictive model. But you can always include ALL the variables in your model and test the significance. So that option is also always open to you. Your research experience may also help you presume the presence of possible non-linear relationship with the DV.

$\endgroup$
  • $\begingroup$ Thank you very much Blain Waan. I really appreciate it. Very helpful! $\endgroup$ – Simon Nov 11 '13 at 2:47
  • $\begingroup$ You are most welcome @Simon $\endgroup$ – Blain Waan Nov 11 '13 at 9:41

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.