Variance inflation factor: what it means and how to read it

In the report: VIF — How much the variance of B grows because the predictor is correlated with the other predictors. Empty when lambda is above 0.

How to read it

1 — no overlap with the other predictors. Above 5 (some say 10) the predictors overlap so much that their separate B are unreliable: drop one of them or set lambda above 0.

In plain words

A student tried to guess the rent of a flat from the number of rooms and from the size of the flat. But big flats nearly always have more rooms, so the two overlap. She asked for VIF. A VIF of 1 means no overlap, and above 5 the two overlap so much that the separate B numbers for each of them become shaky. Rooms and size both came out at 7. She dropped the rooms and kept the size, and the VIF fell to 1, so the B for size can be trusted more. Setting lambda above 0 would have been another way out.

How it is computed

VIF = 1 / (1 − R²ⱼ), R²ⱼ — of the predictor regressed on the other predictors

Example

The cell with VIF is selected in each report; point at a label to see its note.

— the sample data (50 rows) the examples use
— =STATS_LINEAR(Data!A1:T51, "y", "x1,x2", 0)
— old syntax: regression steps, one-sample test against 15 =STATS_ANALYZE(Data!A1:T51, "linear(y|x1|x2) oneSample(x1|15)")

Where you see it

STATS_LINEAR — Multiple linear regression (OLS) with hierarchical steps separated by | and interactions a*b.
STATS_ANALYZE — Runs several analyses on one range and returns one report.

Covariance — How the two columns vary together, in the product of their units: positive — they rise together, negative — one rises as the other falls.
Regression slope — The least-squares slope of this column on the column xCol, on the rows where both are numbers; empty with fewer than 2 rows or without spread of X.
Resampling method — pairs — whole rows are drawn with replacement; residual — the fitted values plus residuals drawn with replacement.
Slope — The least-squares coefficient B of the last predictor of the step, the other predictors of the model held fixed — computed on the data as they are.
Ridge penalty — The ridge penalty: 0 — ordinary logistic regression; above 0 the slopes are shrunk toward 0.

Get it in your sheet

One formula in a cell gives the whole report, Variance inflation factor included. ALS Stats is a free add-on for Google Sheets™.

Install ALS Stats from Google Workspace Marketplace (opens in a new tab)