8

Since today Amy Winehouse died, she joined the 27 club. The Club 27 is based around the statement that the death at the age of 27 for famous musicians is more common than in any other group of people. Is there any other similar group of people that their age of death has been compared to and showed this as a relevant fact? Is there any explanation to this? (Let's say, something in the lines of: "The abuse of drugs, or reckless living style, is more common among musicians than among actors")

Oddthinking
  • 140,378
  • 46
  • 548
  • 638
Jader Dias
  • 583
  • 5
  • 14
  • 4
    Looking at your comments to the answer below, I think you need to specify your question to be clearer in what you want. The question above seems to ask if *27* (and not 26 or 28) shows a particular incidence of death among "famous musicians" (however you may define it). In your comments below, you seem more interested in the contrast between the general population's mortality curve and the specific sub-population of famous musicians. The answer to that one is simple statistics: a distinct sub-population has no reason to follow the same distribution as the more general pop. it belongs to... – Dave Jul 24 '11 at 07:44
  • @Dave your comment is incomplete – Jader Dias Jul 24 '11 at 18:53
  • 1
    it isn't. But having to abbreviate 'population' did make it look so. Anyway, just to reiterate the point: when taking a specific subset of a population, there is absolutely no particular reason that it should follow the same distribution as the parent population. A very obvious example would be separating your population along genders (you will obtain two non-identifcal curves with peaks at different ages). When taking a subset of people with a "high-risk"/"unhealthy" occupation, a different age-of-death distribution *is* to be expected. – Dave Jul 24 '11 at 18:57
  • @Dave now I see. But the general population is still a reference population to which you can compare its subsets. – Jader Dias Jul 24 '11 at 20:20
  • Tim Harford says 27 is a special age for famous musician stars. Here is the podcast http://www.bbc.co.uk/iplayer/console/b012x12m (hopefully, it's the right one, I know that I listened to it a couple of days ago). – Alexandru Aug 07 '11 at 09:18
  • This question was examined by [Mr Science Show](http://www.mrscienceshow.com/2011/07/27-club.html) – Oddthinking Aug 27 '11 at 12:24

4 Answers4

13

I think the biggest thing going on here is confirmation bias as opposed to any other phenomenon. There are thousands and thousands of musicians, and the "27 Club" has only 45 members (as defined by wikipedia, and going back to 1908). I think this particular phenomenon also has some hallmarks of the Birthday Paradox in that people will always have a specific age when they die. Naturally, some will have the same age. (This is a weak analogy, I am thinking of a better one, and am open to suggestions.)

I think what really drives people to the confirmation bias inherent in this is that 27 is a rather young age for anyone to die. To have musicians be very famous by that age, makes them stand out even more.

As for other people dying at 27, I think that you can list many other groups that have a much higher chance of dying. For instance, in the commercial sector, being a fisherman is the deadliest demographic as far as profession goes.

according to the most recent Bureau of Labor Statistics (BLS) data, in 2006 it became in fact the deadliest job in the country with a fatality rate of 141.7. Fishers didn’t have the highest number of fatalities over all (51 reported fatalities in 2006) but due to the relatively small amount employed, it had the highest rate.

In terms of overall numbers, drivers/sales workers suffered by far the most on-the-job fatalities – 940 in 2006. The strong majority of these fatalities occurred among heavy truck drivers. Because of the large labor force of drivers, the fatality rate (27.2) was lower than for fishers.

Although, soldiers aren't mentioned. And of course, these are deaths due to accidents. Suicides, drug overdoses, and the like attributed to the "27 Club" are tracked under different methods (such as the CDC). In a cursory look at that data, it doesn't seem like there is anything particularly different about musicians.

As Jader mentions in the comments, the average death rate for people in the 25-34 age group is about 1.7%. Now look at the "27 Club" that is listed on Wikipedia. That's 45 of them in over 100 years! According to the labor bureau, there are currently 186,000 PROFESSIONAL musicians in the US for 2008, with a 10 year delta growth of just over 14,000. If anything, the 27 club seems to be UNDER represented on the wiki page if it can only find 45 in 100 years!

Also, consider that the http://www.the-eggman.com/writings/death_stats.html link gives deaths for the 25-34 age category under accidents at a higher 12% rate! I think it could be argued that most of those deaths also involved some level of "accident" such as overdose, suicide, etc. I refer you to the answer by Sklivvz.

Community
  • 1
Larian LeQuella
  • 44,977
  • 18
  • 187
  • 208
  • 3
    +1 for confirmation bias. Maybe someone should start "The Club Not 27" for musicians that don't die at 27... – Thomas O Jul 23 '11 at 20:56
  • 3
    Agreed on confirmation bias (and the common sense argument that 27 year-old artist deaths will bring more attention to them). *However*: Birthday Paradox is a statistical result with no connection whatsoever to the issue at hand... And I am not sure that statistics on the wider population at large (or any subset other than rock stars) are relevant to the question. – Dave Jul 23 '11 at 21:38
  • 2
    @Dave, fair comment. I just used the Birthday Paradox as an example of statistics (i.e. if you die, you have an age. Given a large enough group, many will have the same age). Not that it's the same mathematical formula since birth-dates are limited to 365 (unless born on leap-day), but human ages upon death will have some of characteristics that can probably be figured out (not that I am going to do that math though). – Larian LeQuella Jul 23 '11 at 21:41
  • 2
    @Larian: I understand that you were only drawing an analogy, but it still thin, in that the Birthday Paradox relies essentially on the power of combinatorial (the fact that you are looking at any pair of dates), whereas the current case sets the year first, then looks at occurrences. In birthday analogy, it would be akin to finding out that a significant number of people out of a group, are born on April 12th (or October 23rd etc). – Dave Jul 23 '11 at 22:13
  • 2
    @Dave, any suggestions on a better analogy? I was sort of drawing a blank, so that's the one I went with. – Larian LeQuella Jul 23 '11 at 22:17
  • Your answer is relevant and valid, but isn't exactly what I want. This site http://www.the-eggman.com/writings/death_stats.html says that 1.7% of the dead Americans died with an age between 25 and 34. Even if there is thousands of famous musicians, and 45 died with 27 years, maybe still this proves that they die younger than the rest of the population – Jader Dias Jul 24 '11 at 00:23
  • 1
    @Jader Dias, That's 45 of them in over 100 years! According to the labor bureau, there are currently 186,000 PROFESSIONAL musicians in the US for 2008, with a 10 year delta just over 14,000. If anything, the 27 club seems to be UNDER represented! – Larian LeQuella Jul 24 '11 at 00:35
  • Wikipedia won't list all professional musicians in the 27 club. It only cites the ones some researchers knew about, or the most famous. I think I'll make a web spider to verify the veracity of this myth http://stackoverflow.com/questions/6804211/how-to-extract-statistics-from-wikipedia – Jader Dias Jul 24 '11 at 00:38
  • 1
    @Jader Dias, I think your word choice is rather accurate. :) (Look at the last word in your comment before the link.) – Larian LeQuella Jul 24 '11 at 00:40
  • I didn't mean it though – Jader Dias Jul 24 '11 at 00:43
  • If there is 45 musicians in the 27 club, it requires 26k dead musicians to fit the average percentage. You informed us that there are 186k ALIVE professional musicians in the US alone. But I can't figure out how many DEAD musicians we have. – Jader Dias Jul 24 '11 at 00:50
  • 1
    @JaderDias let us [continue this discussion in chat](http://chat.stackexchange.com/rooms/904/discussion-between-larian-lequella-and-jader-dias) – Larian LeQuella Jul 24 '11 at 01:36
  • Sorry to revive a thread from 10 years ago, but I think birthday paradox is indeed relevant, and here is why. Let's assume that rockstars die at a uniformly random age between 20 and 30, so you would expect a spike in some number between. (This is like a balls in bins problem, and is a generalization of birthday paradox albeit not the same). So, it is possible that the 27 club was founded based on birthday paradox. (Of course, once the number was set, one can now look at this statistically provided you are willing to throw away the past data) – E-A Jun 15 '20 at 12:52
12

It's easy to be seen that this is a prime example of confirmation bias:

  1. Take a list of famous musicians with birth and death dates.
    I've used this one which lists famous composers, but any other better list can be used if found.

  2. Calculate the age at death (or lifespan)

    Lifespan calculation

  3. Plot on a graph.

    Composer lifespan Age on the x-axis, number of occurrences on the y-axis.

  4. Does the age 27 look particularly significant? No.

On the other hand we could have a "club 67"... ;-)

Sklivvz
  • 78,578
  • 29
  • 321
  • 428
  • 9
    The "club 27" should be better defined to give a statistically sound answer. For instance, if you take [this list](http://www.av1611.org/rockdead.html) of prematurely dead rock stars instead of a list of classical music composers, you get two main peaks at 27 and 38... Of course it does not mean that the Club 27 story is real (I strongly doubt it is) just that choosing the right dataset is very important. PS: forgive me for linking to a page full of absurdities, it's just the first list I found on Google and it serves the purpose. – nico Jul 24 '11 at 14:22
  • 1
    @nic, I've taken a look at the list and it doesn't seem to be reliable for two reasons. Firstly, it doesn't really show much correlation between lifespan and age as it should--see [here](http://imgur.com/TWfik). Secondly, according to that list rock musicians don't get to live over 65, which is absurd--or not absurd, since rock hasn't existed for long enough to have rockstars die of old age... :-) – Sklivvz Jul 24 '11 at 14:43
  • 1
    maybe I did not explain myself well. I expressively posted a link to a skewed dataset (**prematurely** dead rock stars). You're basing your answer on another list (of classical music composer), which is skewed by other factors (different lifestile, most of those people lived in a different century etc.). That is why I am saying "Club 27" is not well defined: does it include only rockstars? All the musicians? What about singers? And composers? How you define one musician is "famous"? And so on... – nico Jul 24 '11 at 14:49
  • @nic, I understood your point. My point on the other hand is that a pretty unbiased source (of 800+ classical composers) does not show this effect. However a biased source, like your example, does show some random stuff. In conclusion, one *can* create a whimsical club 27 by carefully choosing the sources, for example by only looking at rockstars who died young, which is why I call this a "prime example of confirmation bias". If you look at unfiltered data the age 27 doesn't look suspicious. – Sklivvz Jul 24 '11 at 14:56
  • 2
    I wouldn't say a list of classical composers is unbiased. You should include ALL types of musicians. Furthermore, that list does not include modern musicians (say, born after the 1960s-70s) when the habits of famous people clearly started to change... Again, I agree with you that the "27 argument" is probably spurious, I am just saying that you will never be able to prove it until the problem is better defined (who is included **exactly** in the club 27). – nico Jul 24 '11 at 15:03
  • No: if there is no good definition for club 27 then club 27 does not exist by definition :-) Surely "something" exists, for some definition of "something". However, I've proven that there is no correlation between being a good musician and dying at 27, which is what this question claims. – Sklivvz Jul 24 '11 at 15:18
  • I think your last graph partly confirms the Club 27, but it also creates a Club 38... – Jader Dias Jul 24 '11 at 18:52
  • 1
    @Sklivvz, I agree that your data meets the definition given by the OP. I don't think that definition matches well with what Club 27 normally means (which seems to be more focussed on contemporary rock/pop stars). I think that definition of Club 27 is not at all clear - but that doesn't mean the concept doesn't exist - just that it is impossible to meaningfully study it until you have a provisional definition. The OP provided one, so it is appropriate right that you went with that for now. – Oddthinking Jul 25 '11 at 02:57
  • 2
    @odd: well, so, I think my data disproves they hypothesis that musicians in general tend to die at 27. if we want to restrict our club to rockstars that's fine. I can't find a good data set though. Note that shifiting definitions are a bad sign... "club of the gaps"? – Sklivvz Jul 25 '11 at 05:58
  • -1: -- No original research: http://meta.skeptics.stackexchange.com/a/522/34 – Russell Steen Jul 01 '12 at 21:11
  • @RussellSteen maybe you would like to explain how this is actually research? It's just an exposition of available data. – Sklivvz Jul 01 '12 at 21:21
  • 1
    @Sklivvz -- I gave you a reason for my downvote. I believe that to be constructive. Since you find the other comment offensive I have removed it. Your answer is original research. – Russell Steen Jul 01 '12 at 21:32
  • @RussellSteen not everything that has a self made graph qualifies as original research. I haven't "researched" the musician ages, I've merely taken a list. If one asks "is it true that the Empire state building is the tallest skyscraper?" and I answer with a list of the world's tallest buildings and a graph, that's not original research. It's an exposition of facts. – Sklivvz Jul 01 '12 at 21:40
  • @Sklivvz -- By virtue of creating your own sample, I would conclude otherwise. My comment was purely to *be* constructive by saying why I applied the -1, not with a desire to argue or debate the viewpoint. My apologies for starting one and I will refrain from such comments in the future. We disagree on the point at which it constitutes "research". Intuitively I think you are correct in your result, but the answer is not valid in my view due to the method of sampling. You chose the sampling method which in my view, makes this your research. I'll refrain from such comments in the future. – Russell Steen Jul 01 '12 at 21:42
  • @RussellSteen I haven't "sampled" anything. This is merely a (full) list I found. If there is any better list, I'd rather use that, but this is what was available at the time of the answer. I do not believe the list to be particularly biased, if you look at musicians in general, any list would have many more classical musicians than modern anyways. – Sklivvz Jul 01 '12 at 21:47
1

Lenon, Elvis, Zappa, Jackson - none of them died at an age of 27. And Mozart, Beethoven, Wagner, Paganini, Liszt, Caruso, ...

'Musician' is a wide field, but what's about rock stars? Caruso wasn't a rock star? Well - Amy Winehouse isn't exactly Rock, is she? It's Soul. Is Aretha Franklin still alive? Tina Turner? James Brown was or is older than 27. Why are so damned much musicians older than 27 still alive?

Okay, back to the standards: Here are some diagrams, showing 100 dead, most well known rock musicians (without explaining, how 'well known' was measured).

But I link it, because of the skeptic questions raised in the text and in the comments:

  • does dying early make you more famous?
  • is there a high risk of drug usage in the early adult phase, which some musicians survive (Cliff Richard, Joe Cocker, ...), and after a critical age, not exactly the age of 27, but around that, their chances of getting old rise, because they learned how to handle the risk?
  • is the scope of musicians (rock, soul) adapted to dying stars with 27, after an initial group of musicians in the 70ies (Jones, Morrison, Hendrix, Joplin)
  • the age of rock music
  • 94% did not die at the age of 27
user unknown
  • 3,968
  • 1
  • 26
  • 31
-2

I think the other thing going on here is that 27 may be a common age by which an entertainer is widely known (if they are ever going to be widely known.) Let's face it, if you're an entertainer making a mark, you're probabably not too young (very few 5 year old rock and roll singers) and not too old (if you're 40 and a musician, you usually hit your peak earlier, or no one has heard of you.)

The idea that there is something mystical about musicians dying at 27 is like finding something mystical in the average age of first marriage for women being 26. Most women aren't getting married for the first time at 7 or 57.

Sklivvz
  • 78,578
  • 29
  • 321
  • 428
Rachel Sanders
  • 21
  • 1
  • 1
    Welcome to Skeptics, Rachel. Speculative answers are not allowed on this site. Please provide facts to support your thesis. http://meta.skeptics.stackexchange.com/q/5 – Sklivvz Jul 24 '11 at 12:14