AtlasSession 2L6: Hypothesis Testing and Statistical Testsp.34–40
p-value misconceptions
- 4 earlier topics
- What is a hypothesis
- Test, null and alternate hypotheses
- Alpha error
- Tests, p-values, significance
- p-value misconceptions
Fig. 1
Stem-cell vs cell-free scaffolds: 20 cytokines with no true difference, and cell viability (%) in small and huge studies
Fig. 1. Studies that each test 20 cytokines with no real difference. None run yet.
- Studies so far
- 0
- Latest study
- — (no studies yet: press Play or +1)
- Studies with an impressive-looking result
- — (no studies yet: press Play or +1)
Sampling Simulator · First look
Level 1 · First look · Lecture 6 · p.34–40
Know what a p-value can't tell you.
Spot and correct the ASA p-value misconceptions, including 'p > 0.05 means no difference', p-hacking and confusing significance with importance.
No result isn't proof of no effect (the step you are on)
A quick glance that misses your keys doesn't mean they're gone; small studies miss real effects.
A quick glance that misses your keys doesn’t mean they’re gone. The stem-cell scaffold here really adds 2 points of viability, yet — of — studies with 10 wells per group missed it: finding nothing isn’t proof of nothing.