From the Annals of Bad Research: rock stars die younger

Around here we're all culturally savvy enough to have heard of the 27 Club: the list of those rocks stars who have died or drink, drugs, suicide etc at the age of 27. We've always taken this to be a rather cheery finding: that if you give some 18 year old all the money, booze, drugs, success and sex they could possibly want then it still takes them 9 years to kill themselves through overindulgence. Rather puts into perspective the prodnoses complaining about our having a second glass of sherry before dinner. However, we've just had the release of a report indicating that popular musicians do indeed die younger, on average, than the general population. And thi8s really should be included in our compendious volume, The Annals of Bad Research. For the contention is that the average age at death of rock and roll, rock and pop, stars is lower than that of the general population. But as Chris Snowden points out, we cannot actually know that:

You see the problem here, I expect. Rock stars didn't exist until the 1950s and since many of them are still alive, we don't know what their average age of death is. It wouldn't be at all surprising if they die earlier on average, but the graph above tells us very little about whether this is so. When Chuck Berry (aged 88), Jerry Lee Lewis (aged 79) and Little Richard (aged 81) pop their clogs, the average is going to go up, especially if they keep breathing for another twenty years.

And, who knows? They might. Perhaps the higher risk when young is counter-balanced by the boost to longevity of having lots of money and the best healthcare in old age?

Be that as it may, you clearly can't work out the average lifespan of a rock star until at least the first generation of rock stars are dead.

Quite: you can only work out the average age of death of any particular cohort when all of that particular cohort are dead. If you try to do it before that has happened then you'll be counting all of those who die young but not all of those who don't: meaning that what you've actually calculated is the average age at death of those who die young. And, you know, people who die younger die younger isn't really all that amazing of a research finding.