EngineersofAI LogoCoursesBlogResearch LabAI LettersThe LabCode BankInteractive 3DKodrEarnestJobs
Skip to main content
Interactive 3D/Power Analysis
Effect Size δ
δ = 0.50medium
02
Sample Size n
n = 30
5200
Significance α
α = 0.05
0.010.20
Results
Power (1-β):86.3%
Type II (β):13.7%
n for 80% power:25
✓ Adequately powered. You'd detect δ=0.50 86% of the time.
Try: Set δ=0.2 (small effect) - need n~200+ for 80% power. Then set δ=1.0 (large effect) - only n≈10 needed. Most published studies are underpowered.

Power Analysis - Interactive Visualization

Statistical power = P(reject H₀ | H₁ is true) = 1 - β. Low power means experiments are underpowered - you'll often fail to detect real effects. This visualization shows H₀ and H₁ distributions overlapping, with Type I error (α, false positive), Type II error (β, false negative), and power (1-β, true positive) shaded. Increasing n (sample size) or δ (effect size) increases power.

  • See Type I error (α), Type II error (β), and power (1-β) shaded simultaneously
  • Adjust effect size δ to see distributions separate and power increase
  • Increase n to see both distributions narrow and power improve
  • Adjust α and see the tradeoff: lower α → lower power
  • Foundation for A/B test design and ML experiment planning

Part of the EngineersOfAI Interactive 3D - free interactive visualizations covering every major concept in machine learning and AI engineering. Hover any element for a plain-English explanation. No code required.