Reading focus
Syllabus: What Can Statistical Evidence Tell Us?
Connect the justification of equivalence bounds to the units and consequences discussed in the opening pair.
Why you’re reading it. Distinguish absence of a meaningful effect from an inconclusive result.
A question to consider. Before seeing the school’s results, how would you justify the smallest worthwhile gain, and what would an interval crossing that boundary leave unresolved?
Why this reading
Lakens, Scheel and Isager ask how evidence can rule out effects large enough to matter. Their reanalyses distinguish statistical equivalence, a difference from zero and an inconclusive result.
Before you read
Equivalence bounds specify effects judged too consequential to dismiss. The two one-sided tests procedure assesses both boundaries. The bounds need justification before seeing the result; a smallest effect of interest depends on the scientific or practical question.
Read for
- Compare theoretical, practical and resource-based reasons for choosing bounds.
- Explain why an estimate can be both different from zero and equivalent within justified bounds.
Consider afterwards
- A cheap classroom intervention and an expensive one produce the same estimate and interval. Should their smallest worthwhile benefits be identical? Explain the evidence and judgments needed.
Editorial context
Equivalence is relative to the selected bounds and model assumptions; it does not prove an exactly zero effect. Choosing wide bounds after inspecting data undermines the inference. The opening example has an apparent arithmetic slip: 0.015 minus 0.06 is −0.045, rather than the printed −0.055. This is an editorial observation, not a published correction.