AtlasSession 4L10: Sample size calculationsp.8–9
Minimal important difference
- 8 earlier topics
- CIs and p-values
- Beta error and power
- Why sample size matters
- Continuous outcomes
- Minimal important difference
Fig. 1
Cartilage repair trial on a 0–100 knee score (KOOS, points): which difference should it be sized for?
Fig. 1: decision flowchart “Which difference should the trial be sized for?”. Case: Pilot saw 12 KOOS points. Matrix-assisted autologous chondrocyte implantation (MACI) vs microfracture, KOOS (0–100): literature MID 10 points, SD 15 (illustrative). A pilot observed a 12-point difference. Sized on 12: 25 per group (ext). Question: Has an earlier study or pilot already seen a difference? Answers: 1 Yes, 2 No.
- Case
- Pilot saw 12 KOOS points
- Question
- observed difference?
- Your answers
- none yet
- How it’s going
- — (answer the first question to see how it’s going)
- Score
- 0/4 (no cases finished yet)
Decision Flowchart Walker · First look
Level 1 · First look · Lecture 10 · p.8–9
Explain the difference between a change you can notice and a change that matters enough to act on.
Define the minimal important difference, separate it from a just-noticeable difference, and apply the p.9 rule for designing around an observed or a desired difference.
Noticing (the step you are on)
A room less than 0.25 °C warmer feels no different; even a noticed change may not change what you wear.
A room less than 0.25 °C warmer feels no different, and even a change you notice may not change what you wear. A knee-score gain is the same: noticing it is not enough.