Lessons in statistics

This week brought two really fascinating insights into the world of statistics. The first was from a most unusual source: The Daily Mail (not my usual read – the link was posted to the Bad Science forum). They had run with the story Cracked it! Woman finds six double yolk eggs in one box beating trillion-to-one odds, which was then pretty rigorously debunked by the Mail’s own Michael Hanlon.

In Eggs-actly what ARE the chances of a double-yolker? Hanlon points out that young hens tend to produce more double yolks than older hens, and that flocks tend to be of the same age, so six double-yolkers is not an unusual occurrence for a young flock. Further more, double-yolkers are heavier than single yolked eggs, so when the eggs are sorted by weight they will tend to wind up in the same box. So really, a box of six double-yolkers isn’t that much of a surprise.

The second was from WNYC’s RadioLab, a great radio show and podcast from NPR in the States which has now become a must-listen for the gym. I love RadioLab – they cover science stories in an engaging, entertaining and though provoking way. Their programme from Sept 9 last year was called Stochasticity, “a wonderfully slippery and smarty-pants word for randomness”. The first two sections should be compulsory listening for every journalist:

A Very Lucky Wind
Laura Buxton, an English girl just shy of ten years old, didn’t realize the strange course her life would take after her red balloon was swept away into the sky. It drifted south over England, bearing a small label that said, “Please send back to Laura Buxton.” What happened next is something you just couldn’t make up – well, you could, but you’d be accused of being absolutely, completely, appallingly unrealistic.

On a journey to find out how we should think about Laura’s story, and luck and chance more generally, Jad and Robert join Deborah Nolan to perform a simple coin-toss experiment. And Jay Koehler, an expert in the role of probability and statistics in law and business, demystifies some of Jad and Robert’s miraculous misconceptions.

And then the first half especially of:

Seeking Patterns
Fine. Randomness may govern the world around us, but does it guide US?? Jonah Lehrer joins us to examine one of the most skilled basketball teams ever, the ’82 – ’83 ’76ers, and wonders whether or not the mythical “hot hand” actually exists.

Then we meet Ann Klinestiver of West Virginia, an English teacher who was diagnosed with Parkinson’s in 1991. When she began to take a drug to treat her disease, her life changed completely after one fateful day at the casino. Jonah discusses the neurotransmitter dopamine and the work of Wolfram Schultz, whose experiments with monkeys in the 1970s shed light on Ann’s strange addiction and the deep desire for patterns inside us all.

Statistics is something that you constantly see journalists getting wrong. The Bad Science forums are rife with examples of statistics abuse. It’s not surprising, because it’s actually very easy to get statistics wrong: Probability in particular can be very counter-intuitive and assumptions that seem to be common sense are frequently just our brains playing silly buggers with us. Personally, I think that all journalists should have to study statistics, even the freelances, because it’s so easy to get it wrong and so useful when you get it right. But, in the meantime, I’d settle for more people listening to shows like RadioLab and reading blogs like Good Math, Bad Math, Bad Science, or Junk Charts.

One thought on “Lessons in statistics

Comments are closed.