A percentage means nothing without its base.
"Crime rose 100%" — from two cases to four. "This drug halves your risk" — from one in ten thousand to one in twenty thousand. The figure is accurate and the impression it leaves is not, which is why the first question is always: a percentage of what?
"Crime rose 100%!" Sounds terrible! But from 2 cases to 4? "Disease affects 1 in 10,000 - NEW drug cuts risk 50%!" Now 1 in 20,000. Still rare! Percentages can mislead without context. Let's learn to see through the numbers!
ALWAYS ask: percentage of WHAT base number? "50% increase" from 2 = 3 (tiny!). "50% increase" from 1,000,000 = 1,500,000 (huge!). Same percentage, wildly different reality! The BASE RATE matters enormously!
Ignoring how COMMON something is to begin with! "Test is 95% accurate at detecting rare disease!" Sounds great! But if disease affects 1 in 10,000, most "positive" tests are FALSE positives! Context = base rate = how common is this anyway?
DIFFERENT THINGS! From 10% to 20% = "increased 10 percentage POINTS" OR "increased 100 PERCENT" (doubled!). Politicians love this trick - use whichever sounds better! 10% → 11% = 1 percentage point but 10% increase!
Always convert to ACTUAL NUMBERS! "200% increase in shark attacks!" From 1 to 3 attacks nationwide? Not scary. "Crime up 5%" in city of 1 million? That's 50,000 more crimes - very scary! Percentages hide scale!
Percentages without context can be deeply misleading - always ask for actual numbers!
Key concepts:
• Base rate: Starting number matters! 100% increase from 2 vs from 2,000,000
• Percentage points ≠ percent: 10%→20% = 10 points OR 100% increase
• Relative vs absolute: "50% reduction" could be huge or tiny depending on base
Common manipulation:
"Drug reduces heart attack risk by 50%!"
Reality: Risk drops from 2% to 1% (1 percentage point = 50% relative reduction). True but misleading!
Base rate fallacy example:
Disease affects 1 in 1,000 people. Test is 95% accurate.
You test positive - do you have disease?
Most people think "95% chance!" WRONG!
Out of 1,000 people: 1 has disease (test catches it), 50 don't have it (false positives from 5% error). So 1 in 51 positive tests are real = ~2% chance!
Defense questions:
1. What's the ACTUAL number?
2. Percentage of WHAT base?
3. How common is this to begin with?
4. Percentage points or percent?
Remember: Numbers don't lie, but presentations do!
🌱 Story Seed
"Our class improved 200% on the test!" "Wait - from what score?" "From 10 to 30 out of 100..." "So 200% sounds huge, but you went from failing badly to still failing. What's the ACTUAL goal?"