Quality, innovations and statistics · 3 steps
Power, confidence intervals and the forest plot
Finding a significant result needs adequate power; power and narrow confidence intervals depend on sample size, and meta-analysis pools studies by weighting them for their precision.
Left: H₀ and H₁ distributions with α, β and power areas, confidence intervals from 100 repeated studies, and a forest plot with five studies added one by one. Right: power against sample size, the share of intervals containing the true value and 1, study weights and the pooled effect.
Finding a significant result needs adequate power; power and narrow confidence intervals depend on sample size, and meta-analysis pools studies by weighting them for their precision.
Source: Type I/II error, power and confidence interval from the book (29.4); the five forest-plot studies from Figure 29.7; pooled effect and I² calculated on this page by fixed-effect inverse variance. Power by normal approximation (two groups, two-sided).
How to use: a single click on the drawing pauses and resumes; ↑ ↓ change the effect size in the power step and the sample size in the confidence interval step (use the buttons below for α); use the time bar below to go to any moment; ← → change steps.
Presentation mode ↗The animation alone, sized for full screen; for projecting in lectures and talks. ← → step, space plays/pauses.
Embed
Paste this code into your own page (course site, slides, learning platform). The animation arrives working, with its source written underneath.
In the book
In the book this animation runs in place of these figures: