AtlasSession 2L6: Hypothesis Testing and Statistical Testsp.55–58
One- vs two-sided; equivalence
- 4 earlier topics
- What is a hypothesis
- Test, null and alternate hypotheses
- Alpha error
- Tests, p-values, significance
- One- vs two-sided; equivalence
Fig. 1
Serum-free vs FBS medium for MSC expansion: difference in cumulative population doublings, serum-free minus FBS (doublings)
Fig. 1. One hump: the results you would get if both media gave the same growth. Your result, −0.35 doublings, is marked. Results as far out, or further turn up about 7 in 100 times when nothing is going on.
- This far out, if nothing is going on
- about 7 in 100
- Rarer than 5 in 100?
- no
Two-Curve Test Visualiser · First look
Level 1 · First look · Lecture 6 · p.55–58
Decide whether you care about a change in both directions, in one direction, or about being 'the same'.
Decide whether a question needs a two-sided, one-sided or equivalence analysis, and explain why an equivalence margin must be set in advance.
Which way matters (the step you are on)
A new route to work: 'is it different?', 'is it no slower?' or 'is it basically the same?'
A new route to work: ‘is it different?’, ‘is it no slower?’ or ‘is it basically the same?’ Each question needs its own check: pick one with the switch and the figure changes to match.