Breaking news: Something may or may not happen
It is highly likely that you, esteemed reader of my blog, have a better than average understanding of probability. One of my pet peeves is that journalists usually have a terrible understanding of probability.
Take this headline from BBC News, the world’s 6th most popular news website, that imparts no useful information whatsoever:
Well sure they may. There’s always a chance home values will rise. There’s also always a chance they won’t. This headline summarises what we all know: we can’t be certain about what happens in the future. Something may happen and it may not.
The article’s opening paragraph doesn’t get better: we learn a building society claims there is a “reasonable chance” that home values in the UK will rise. As opposed to an unreasonable chance that they won’t?
How much is that rise that may happen? There’s a big difference in a rise of 10% and a rise of 0.1%. On a house worth 200,000 pound, that’s 20,000 pound increase compared to 200 pound. A rise of 0.1% is not different from a fall of 0.1%, statistically speaking, if one takes a margin of error in measuring into account. It’s a loss in real terms if inflation is running above 0.1%.
I’d like to hear the building society in question state the probability of a rise in home values by a certain amount. Then I’d like to see the chief economist of the building society put his money where his mouth is, and place a bet on a prediction market, such as Intrade.
Then we could get a meaningful headline, such as: “Nationwide says 35% chance of 5% rise in home values in 2009.”