Dead rock stars

Chuck Berry: fading away, not burning out

As covered in the Daily Mail and elsewhere, a study has come out of the University of Sydney which concludes that rock stars tend to die younger than the general population.

Well, duh, you might say. And rightly so. A cursory knowledge of music history tells us that rock stars are more likely to die from drug overdoses, plane crashes, suicide etc. and this is confirmed in the study.

But I was startled to find that the author has also attempted to compare rock stars' longevity to the average.

Longevity was determined by calculating the average age of death for each musician by sex and decade of death. These averages were then compared with population averages by sex and decade for the US population (per 100,000).

And she has come up with this graph...

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.

I can't find the study, so I apologise to the author if I am misrepresenting her work, but somebody raised this question with her in the comments and her reply was not very convincing.

There is a small amount of survivor bias at work here, ie. the rock stars have to live to be in their twenties (generally) before they become rock stars. That probably isn't a big issue when so few people die before the age of 20. The real problem here is a sort of reverse survivor bias, sort of like immortal time bias, but not quite either of those.

So, my question is: does this fallacy have a name?