MerQuR+Statistics Glossary

Statistics Glossary → Equivalence test (TOST)

Equivalence test (TOST)concept

A test used to show that the difference between two means lies within a pre-specified range regarded as negligible, i.e. that the groups are practically equivalent.

In a classical test, p > 0.05 does not mean 'no difference'; it only means there was not enough evidence to show one. In an equivalence test the hypotheses are reversed: the null hypothesis becomes 'the difference lies outside the negligible range'. The researcher first sets the lower and upper bounds regarded as practically negligible.

TOST (two one-sided tests) tests with two separate one-sided tests that the difference is above the lower bound and below the upper bound. If both are significant (p_TOST = the larger of the two p values < α), the groups are considered equivalent. This is the same as the (1 − 2α) confidence interval of the difference lying entirely within the bounds.

An everyday example

When a cheaper generic version of a drug is brought to market, it must be shown to work the same as the original. Saying 'no difference was found' is not enough; it must be shown that the difference in blood levels stays within a narrow range agreed in advance. An equivalence test does exactly this.

In MerQur

The 'Equivalence test, TOST' check box of the independent-samples t-test, with lower and upper bound fields.

Analyses where this concept appears

Independent-Samples t-Test

Related concepts

Confidence level and confidence intervalCohen's d (effect size)