Home › Statistics & Probability Calculators › Variance Inflation Factor (VIF) Calculator
Variance Inflation Factor (VIF) Calculator
The variance inflation factor (VIF) measures how much the variance of a regression coefficient is inflated because that predictor is correlated with the others. Paste your predictor columns, one per line, and the calculator regresses each predictor on all the others, records the R², and reports VIF = 1 ÷ (1 − R²) and tolerance = 1 − R² for every variable.
A VIF of 1 means a predictor is uncorrelated with the rest. Common rules of thumb treat VIF above 5 as high and above 10 as severe multicollinearity; the square root of the VIF is the factor by which that coefficient’s standard error is larger than it would be with uncorrelated predictors. High VIFs make individual coefficients unstable and their p-values unreliable, even when the model as a whole predicts well.
Enter only the explanatory variables, not the response. The results match statsmodels’ variance_inflation_factor (with a constant included) and the VIF column in SPSS and R’s car::vif for models without factors. The correlation matrix shown below helps identify which pairs drive a high VIF; dropping or combining those variables, or centring variables before creating interaction or squared terms, usually brings VIFs down. For a single predictor, see the linear regression calculator.
At least 2 predictors with the same number of observations. Do not include the response variable.
Result and step-by-step solution
Severe multicollinearity: coefficient standard errors are inflated more than √10 ≈ 3.2 times.
- Auxiliary regressionsRegress each predictor on all the other predictors (with an intercept) and record R²j. n = 10 observations, 3 predictors.
- VIF and toleranceVIFj = 1 ÷ (1 − R²j), tolerance = 1 − R²j
Predictor R²j Tolerance VIF Collinearity Age 0.99497 0.00503 198.9595 severe Experience 0.99562 0.00438 228.1573 severe Income 0.98713 0.01287 77.7223 severe - Pairwise correlations
Age Experience Income Age 1 0.997 0.992 Experience 0.997 1 0.993 Income 0.992 0.993 1
Frequently asked questions
- What is a good VIF value?
- Below 5 is usually acceptable. Above 5 indicates high multicollinearity, and above 10 is generally considered serious. Some fields use a stricter cut-off of 2.5.
- How is VIF calculated?
- Regress predictor j on all the other predictors, take that regression’s R²j, and compute VIFj = 1 ÷ (1 − R²j).
- What is tolerance?
- Tolerance = 1 − R²j = 1 ÷ VIF. Values below 0.1 (VIF above 10) signal a problem.
- Should I include the dependent variable?
- No. VIF is a property of the predictors only, so enter just the independent variables.
Last reviewed: October 7, 2026. Calculations run in your browser and were validated against SciPy.