Last week we had the prank about the association between people’s IQs and the internet browsers they use. Another story doing the rounds, following the death of Amy Winehouse, is about the ’27 Club’ – that musicians have a peculiar propensity to die at this age.

Matt Parker from Queen Mary University was interviewed on the BBC Radio 4 programme, More or Less, at the weekend. He did some research which shows that a famous musician has a 2.3% chance that they will die within five years of them becoming famous (i.e. 23 famous musicians for evey 1,000). (The average age of becoming a famous musician is 25.) If you look at 27 year olds across the whole population, he says, they have only a 0.3% chance of dying in the next five years.

Matt Parker therefore concludes that being a musician greatly increases the chance of him or her dying in the next five years.

Now, this makes for a nice headline, and this might even be the case. But these figures alone are not sufficient to draw this conclusion.Why should we treat this analysis with caution?

To say that becoming a successful musician leads to a greater likelihood of a person dying earlier than they would have done had they not become a successful musician, one should compare successful musicians with people in the general population who share the same characteristics (beyond just age). My hunch is that people who become famous musicians have an already existing propensity to die early i.e. it is less about being a famous musician and more about preceding factors. They are more likely to die early irrespective of whether they become a famous musician or not.

One thing this teaches us is that in our our desire to find explanations of people's behaviours, we have to remain rigorous in our analysis of data. An approach to doing research is to start off by creating hypotheses, and then proceed to seek to disprove them. Making sure we do this will significantly help us to avoid the trap of spurious findings.

Anyway, enough of this. I’m off to change my internet browser...

Steve is Head of Thought Leadership at Starcom MediaVest, London.