I am testing the reliability of my formative measurement model and I am using Variance Inflation Factor (VIF) and Condition Index (CI) (see this earlier question asking whether to and how to do this). I am using SPSS.
All my scores for VIF are less than 5.00 and all except one for CI is less than 10.00 (i.e. it is 22.379).
- Is VIF and CI appropriate in this case?
- How do I interpret this result?
- Is there any generally accepted thresholds for VIF and CI scores?
As far as I understand, multicollinearity is NOT good for a formative measurement model (i.e. where the arrow goes from the indicators to the construct).
Also, any measure of internal consistency such as Cronbach Alpha or Split Half Reliability are not appropriate in this case.
I am told that a VIF score of 10 or less and CI score of 30 or less are acceptable.