Please I am exploring for multicollinearity in my data between socio-economic characteristics. I ran a collinearity diagnostic test and I have a conditional index of 6.235 which is less than 10. I believe this shows that multicollinearity is not a serious problem. On the other hand, when i check the variance proportions for the dimensions, I find that 3 of my predictor variables had greater than .60 loading on the same dimension. Does this suggest multicollinearity? All tolerance values were greater than .2 and VIF values were less than 5.

Therefore, what I am really interested in is thus: If I have VIF values < 5 and tolerance values > .2 but I have more than 1 variable with a loading of .50 and above on 1 of the dimensions, should I be concerned with the multicollinearity or should I ignore the variance loading since all other conditions were met?


closed as unclear what you're asking by kjetil b halvorsen, mkt, Michael Chernick, mdewey, Ferdi Sep 12 '18 at 19:43

Please clarify your specific problem or add additional details to highlight exactly what you need. As it's currently written, it’s hard to tell exactly what you're asking. See the How to Ask page for help clarifying this question. If this question can be reworded to fit the rules in the help center, please edit the question.

  • 2
    $\begingroup$ What conditional index? VIF? What variance proportion? Correlation? $\endgroup$ – user2974951 Sep 12 '18 at 10:04

Welcome to CV.

Collinearity isn't a yes/no thing, it's a question of degree. It's a question of how much the collinearity is affecting your results. One way to check this out, if you have R, is the pertrub package.

However, in general, Belsley notes that troublesome collinearity occurs when there is a condition index over 10 and high variance shared on that condition index.


Not the answer you're looking for? Browse other questions tagged or ask your own question.