4. Our advice

4.1 Consider the context of the decision to be made

Concluding whether or not there is a significant bias is a statistical interpretation that involves two risks: admitting a difference when there is equivalence (alpha risk), or conversely, admitting equivalence when there is a difference (beta risk). Cf. bibliography by E. Morice.

What's more, it's important not to confuse the decision (a difference declared "significant") with reality (what is the real impact of this difference?). Depending on the case, the risk of error in the decision can have significant consequences that need to be taken into account. For example, water remains liquid between 10°C and 30°C, whereas it freezes from 0.1°C to -0.1°C.