AtlasSession 3L7: Systematic errorsp.61–64, 88–89
Selection of the reported result
- 4 earlier topics
- What is a hypothesis
- Test, null and alternate hypotheses
- Alpha error
- Tests, p-values, significance
- Selection of the reported result
Fig. 1
Rat rotator-cuff repair, PCL patch vs no patch: outcomes per study, no true difference (failure load, N)
Fig. 1. Studies that each test 10 outcomes 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
nothing drawn yet
Level 1 · First look · Lecture 7 · p.61–64, 88–89
See why reporting only the exciting results gives a false picture.
Explain why reporting only the significant outcomes or time points biases the evidence, and how a pre-published protocol prevents it.
Highlight reel (the step you are on)
A striker's video showing only goals hides all the misses.
A striker’s highlight reel shows only the goals. These 0 studies of a patch that does nothing each measured 10 things; — (no studies yet: press Play or +1) studies had at least one that looked impressive.