More statistical syllogisms
Two, plucked randomly from Radio 4 today.
Tory MP on Today insisting that Whitehall produces more paperwork: “the size had gone up 300 per cent, that’s nearly three times…”
No, going up by 300 per cent is four times. In fact it’s confusing - percentages should be avoided generally, and specially if they’re over 100 per cent when it’s an increase.
Later, on a 1pm show, woman talking about IVF treatments: “Everyone has the right to three cycles of IVF treatment, and as each one has a chance of conception of 20 per cent, it would be nice if it could be five….”
Because five times 20 per cent equals 1, right, so you’d be sure to conceive on or before number 5, right? Wrong.
The chance of conceiving in five IVF cycles (assuming each has an independent 20 per cent chance of success) is 67.3 per cent. Because it’s 1 - (0.8^5) [where ^ means “to the power of” - can’t do superscript just now.]
- These posts might be related (the database thinks..):
- What's wrong (and could be right) with feed readers, aka aggregators (7 September 2004; score: 32.74%)
- And so then we come to the topic of nuclear power again (6 May 2008; score: 10.38%)
- Is it really so hard? (7 July 2004; score: 0%)



