Levene's Test calculator

Verify if multiple groups have equal variances. Use our online Levene's test calculator before running ANOVA or t-tests.

What it does

Tests whether groups have equal variances, by running an ANOVA on each value's absolute deviation from its group centre.

When to use it

As a diagnostic before a test that assumes equal spread — though for two groups the better answer is usually to use Welch's and not test at all.

Cautions

• Treating this as a gate is a two-stage procedure, and it distorts the error rate of whatever test follows.
• Like all such tests, it mostly measures sample size: large samples flag trivial differences, small ones miss real ones.
• The median-centred (Brown-Forsythe) variant is more robust for skewed data than the mean-centred original.

Alternatives

• Welch's t-test / Welch's ANOVA — you would switch on a significant result anyway — then just use them from the start

How to read the output

What it tests

Whether the spread of the outcome differs between groups, using distances from each group centre so it is not itself derailed by non-normality. Levene on the mean is more powerful; on the MEDIAN (Brown-Forsythe) it is far more robust to skew, and is the safer default.

Usually you should not need it

Welch's t-test and Welch's ANOVA do not assume equal variances at all. Testing variances first and then choosing between the classic and Welch versions is a two-stage procedure that damages the final p-value — and it buys nothing, because Welch costs almost nothing when variances ARE equal. Use Welch by default and report Levene as a description if a reviewer asks.

A significant result is not a verdict

Unequal spread can be the finding rather than a nuisance. Two treatments with the same mean and very different variability are genuinely different treatments, and that is a result worth reporting rather than correcting away.

How this calculator is validated · Which statistical test should I use?