The big idea
⚠️
Doubled from what?
Relative risk without a baseline is uninterpretable — and it is the format that makes effects look largest.
Statistics for Everyday Life · Lesson 3
Risk and Uncertainty
beginner⏱ 18 min · 7 cards
Try it
Convert the Risk
“Doubles a side effect occurring in 1 in 20,000 people.”
Convert to absolute numbers first, then decide.
Always ask: doubled from what?
Turn headlines into absolute numbers, then decide.
Compare
The same fact, twice
Relative
"Risk doubled"
Tap.
tap for example ↻
Absolute
1 in 20,000 → 2 in 20,000
Tap.
tap for example ↻
The same move runs in reverse in advertising: "reduces risk by 50%" from 4 in 10,000 to 2.
Key term
When the maths is wrong
Positive expected value — always take the bet?
tap to reveal ↻
Not if you cannot continue
Expected value describes an average over many repetitions. For a one-shot decision where losing ends the game, it is the wrong tool. The arithmetic does not know about ruin.
October 1995
The pill scare
A warning that certain contraceptives roughly doubled thrombosis risk was widely reported as a ratio.
The absolute figures — roughly 15 versus 25 cases per 100,000 users a year — were far less prominent. Many women stopped; unintended pregnancies rose, carrying their own risks.
Quick check
Why can a random sample of 1,000 represent millions?
You did it
Nice work
Mark this lesson complete to track your progress.