# Logistic regression and collinearity

#### mtranos

##### New Member
Hi everybody!
I am working on a logistic regression model using spss. After performing a linear regression in order to use the collinearity statistics and diagnostics, I found that two variables have a large proportion of their variance (>.85) related with the dimension with the lowest eigenvalue. However, the collinearity statistics seem to be ok (tolerance > .1 and VIF< 10). What is more I have already excluded the highly correlated variables (spearman > 0.9). Could I ignore the variance proportion problem or does it mean that there is a multi-collinearity problem?
Many thanks