Sunday 24 July 2011

Is The 27 Club Statistically Significant?

Shortly after the sadly premature death of the undoubtably talented Amy Winehouse last night, I tweeted that yet another rock star has tragically died at the curious age of 27.

The prestigious ranks of the infamous 27 club include Kurt Cobain (Nirvana), Brian Jones (The Rolling Stones), Jim Morrison (The Doors) as well as Jimi Hendix and Janis Joplin.

I was however taken to task for the non sceptical nature of my tweet, suggesting that I was perhaps implying some supernatural driver to this interesting coincidence.

I thought I was merely commenting on an interesting coincidence, but this got me thinking. But before we even consider if there is a reason, we must first determine if its is even statistically significant enough to warrant needing a reason at all.

So, ignoring age, I listed whom I considered to be the 100 most well known dead "rock" stars, and then I looked up their ages and shoved them into an excel spreadsheet. (Not completeley blinded alas as I had a fair idea of roughly how old they all were). Anyway, I've highlighted the "27 club" in red:

Of course there are many other rock stars in the 27 Club, but I selected whom I considered the most well known across the board rather than cherry picking (or indeed ignoring) any particular age. However, I accept it is only an arbitrary sample.

I then plotted the total number of dead rock stars by age, and came up with the following graph:

27 is indeed the most common age of death for the most well known rock stars, but the red line doesn't seem as if it is any longer than mere chance might produce for any random age. Indeed if I widened my sample of rock stars to a 1,000 I'd be unsurprised to see the red line even less statically significant.

But that's not to say that there are not a number of complex reasons why 27 turns out to be the mode in my graph. Perhaps those with a constitution unsuitable for the rock star life make them vulnerable after a certain amount of exposure to drugs and alcohol which commonly coincides with this age.

I think taking an interest in the 27 Club is worth while merely to revisit the unfortunate talent that has been admitted. I'm pretty sure we can do this without invoking any supernatural or paranormal false justification.

Of Course regardless of whether this unfortunate guild has coalesced around a truly random number or a number resulting from a complex array of natural and social phenomenon, the fact remains that 27 is now the magic number that has entered our culture. Like all superstitions we are now free to willfully ignore the vast majority of Rock's tragedies that do not invoke the 27 club, and ensure that we are all over it like a tramp on chips at the inevitable next initiation ceremony.

Oops Entered Bon Scott twice, cant be bothered to update my graph though, you get the general idea. In fact why doesn't someone do a more controlled study?

Thanks to everyone who pointed out that as rock and pop has its roots in the 1950’s and was predominately an interest of the young, it was fairly inevitable that any early casualties s would be young. Indeed many of the 27’s clubs biggest stars acquired their membership in the early 70’s. If I conducted my crude analysis on rock stars who’ve died since say 1980, then the queer peak at 27 would disappear. I’m sure there’s many other logical reasons that also contribute to the fascinating but relatively insignificant peak of the 27 club.

Paul Clarke’s comment below is worth mentioning again above the fold. In my words: Does death at a young age and perhaps at the height of musical success help secure a rock star mega status that may well have faded if not inconveniently preserved by a premature death.


slakr73 said...

Very nice! I wanted to do this, but I figured the list should be longer, and I didn't want my own preferences to lead the way. (Maybe you could have added Robert Johnson?)

I always thought that 27 was statistically significant because it is most likely a product of fame more than anything. Once a young budding musician hits the big time, how long until he/she destroys him/herself? 27 seems to be the magic number. There are too many variables though. I always want to add people like John Belushi and Jean-Michel Basquiat as well. But my skeptic brain thinks that there's no statistical significance.

Jeff Pickthall said...

Funny that, I was completely unaware that Louis "Lois" Armstrong had mad any rock albums.

Tom Schaffer said...

Of course it is nothing supernatural. 94 out of 100 dead rock stars did NOT die at the age of 27.

And think of all thousands of rockstars, that are still alive and already older than 27.

Add to that the fact, that rock music isn't THAT old, so of course there are not THAT many rock stars that already died from natural causes at an average age.

And then of course there is the question: "What makes someone a 'rock'star and why is e.g. Amy Winehouse on this list?". Rock isn't even a genre that is listed on Amy Winehouse's wikipedia article. If you add other prominent figures of jazz- or soul-music to the list, it might look completely different.

Paul Clarke said...

Another distorting factor that leaps to mind is the amplifying effect of the early death per se.

Take someone, who not that notable now (let's pick on poor old Jon Bon Jovi) would perhaps, if they had expired in 1989 at the magic age, have achieved the sort of notability required to feature in this list. If he died at 47 or 57 he probably wouldn't have been in consideration even for the type of age-spread sample used above.

A complex, and somewhat circular problem, I know, but definitely an interesting one.

ABS said...

I was actually thinking of doing the same exercise. Thanxs for the trouble! :)

rollo said...

You also seem to have come up with a Brand New Test: is something "statically important"?

Paul Clarke said...

"Nothing in his life
Became him like the leaving it."


Yes, that's the line I had buzzing in the back of my head when I wrote my comment above.

Bit late now, of course.

ivan said...

Suppose that we would expect rock star deaths between the ages of 21 and 40 to be roughly evenly distributed. You have about 50 deaths between those ages in your table. (I used that range because beyond there the deaths look to thin out by year). I calculate the probability of 6 of those 50 occurring at a specific age to be about 0.026, which is roughly 1 in 40. So it happening is no great statistical oddity.

ht said...

There is a more straightforward explanation, and I am surprised that it has not been given here.

For most of the member of this 27 club, death can be at least partially explained by some sort of self destruction.

Once a rockstar reach the age of 27, he knows he is a potential member of the club. If he is already going wrong, this could trigger its death.

Bra Hunt said...

What about Brad Nowell, the lead singer of Sublime, who died from a heroin overdoes in 1996 at the age of 28. If he would have died a few months earlier would he now be a member of the 27 club?

Robert said...

Great reading! Love it and will share it. BTW, what constitutes a "premature" death?

Before THEY wanted to go?
Before WE wanted them to go?
Or simply "before old age"?

Robert in Arizona

Anonymous said...



Kimpatsu said...

A new member of the 27 Club:

Anonymous said...

One might notice though that a significant portion dies from age 50 and younger. That would actually be useful data to suggest that the rock-star life style is not conducive to good health and long life. The graph probably matches the same numbers for smokers, alcoholics, and drug users. But then it might be interesting if we could get some rock stars who actually lived healthy lifestyles and see what their longevity looks like. Maybe then it can be shown that rock is actually good for you?