1

I have data in which I observe the following pattern: Variable A correlates with variable C Variable B correlates with variable C BUT variable A and B do not correlate with each other, even though they are theoretically very similar.

What could cause this? Variable C causing the other 2 is very unlikely theoretically.

Thank you!

Peter Flom
  • 94,055
  • 35
  • 143
  • 276
  • 3
    I see nothing surprising in the fact that the orthogonal vectors $A=(0,1)$ and $B=(1,0)$ both have positive scalar product with $C=(1,1)$. In layman terms, I see nothing strange in the fact that both meeting a homicide maniac and cancer are heavily correlated with death at early age but hardly with each other. Voting to close. –  Sep 04 '13 at 11:01
  • All variables correlate with each other, the question is how strongly. Please tell us all three correlations and also what the variables are. – Peter Flom Sep 04 '13 at 12:50
  • How are you defining "correlates with"? My guess is that you mean the corresponding P value is less than an arbitrary cutoff. As Peter said, this question really can't be answered without knowing what the variable are and the actual value of the correlation coefficients. – Harvey Motulsky Sep 04 '13 at 17:53

0 Answers0