Bayes Factor Reference
How to use this document: Each test has a Bayes factor — a round number you can do in your head. After a positive result, multiply the prior by the Bayes factor. After a negative result, divide. When a test is much stronger in one direction, separate positive and negative Bayes factors are reported.
Pr[Disease | Positive test] ≈ BF × Prior
All Bayes factors are rounded to mental-math-friendly values: 1.5, 2, 3, 5, 10, 20, 50, or 100.
Because BF × Prior is a shortcut, the result can exceed 100%. Don't worry about it. Approach is still informative. When BF × Prior ≈ 100%, interpret it as "more likely than not to have the disease." When BF × Prior > 200%, interpret it as "very likely to have the disease." In these cases, the exact calculation is given below the table.
Probability anchors — what these numbers feel like:
| Probability | Feels like... |
|---|---|
| 0.01% (1 in 10,000) | Being in a fatal car crash this year |
| 0.1% (1 in 1,000) | Developing appendicitis this year |
| 0.5% (1 in 200) | Being an identical twin |
| 1% (1 in 100) | Having a shellfish allergy |
| 2% (1 in 50) | Having naturally red hair |
| 5% (1 in 20) | Being a vegetarian |
| 10% (1 in 10) | Being left-handed |
| 20% (1 in 5) | Drawing a face card from a shuffled deck |
| 50% (1 in 2) | A coin flip |
Sources: NHTSA 2024 (39,345 US traffic fatalities / 335M population); StatPearls appendicitis (1.1 per 1,000/year); CDC twin birth data (3–4 per 1,000 births monozygotic); FARE shellfish allergy (~1.3%); global red hair prevalence (~2%); Gallup US vegetarian poll (~5%); handedness meta-analyses (~10%); card probability (12 face cards / 52 = 23%, ≈ 20%); coin symmetry.