0

Good day! I have encountered this problem for my thesis and really hoping to get some answers.

I have 3 variables SP, SO, and OO -all positively correlated but when I used linear regression ( 3 variables as a predictor) for SE, SP and OO is not anymore significant.

Hazel
  • 1
  • 1
  • This is probably normal. If SP, SO, and OO are conveying very similar information, then in the regression model together they can cause the denominator of the standard error to be very small, causing the p-value to be big. You can read up on "collinearity" or "multicollinearity" for this phenomenon. For diagnosis, check out the statistics "Variance inflation factor (VIF)" or "Tolerance." – Penguin_Knight Feb 17 '15 at 15:27
  • See also [this](http://stats.stackexchange.com/questions/14500/how-can-a-regression-be-significant-yet-all-predictors-be-non-significant/14502#14502). – Richard Hardy Feb 17 '15 at 18:49

0 Answers0