Nobel Prize Winners Live Longer

anthemaniac writes "A new study finds those who won Nobel Prizes between 1901 and 1950 lived about 2 years longer than nominees who didn't win. The researchers conclude that the instantly conferred social status leads to health benefits. From the story: 'The research rules out the possibility that intervening prize-related money itself adds the years through improved prosperity.' If you're thinking of aiming for the prize, pick the right field. Nobel laureates in physics lived nearly a year longer than winners in chemistry."

Re:Another reason I hate science "reporting"

[gvc]

After some effort, I found the actual article. The popular press account was bad, even for the popular press, failing to give the title of the paper and giving the author's name only parenthetically.

In any event, here is the article: http://ideas.repec.org/p/wrk/warwec/785.html [repec.org]

The article contains at least one claim to "significance at the 5% level" but as far as I can see it is a working paper, not (yet) published in a refereed venue. The author appears to have other credible publications relating to the effect of windfalls on people.

So it's not significant after all? :P

[Moraelin]

The standard deviation in the life expectancy of the general population is about 10 years (meaning - 2/3 people die between 67 and 87), although IIRC it's got a lot of skew.

Anyway, the smaller of the two samples is 135 people, so the error in the estimate of that mean is roughly 10 / sqrt(134) ~= 10/12, so two sigma is about 20 months, and the life expectancy difference is 24 months, so it's significant to 5%.

Well, then you've really made the point as to why the article is bogus, eh? Yes, they make a "nearly two years" claim at the top, but if you read a bit further: "The average lifespan for the nominees (including winners) was 76 years. Winners worldwide lived 1.4 years longer on average, and winners from the same country as non-winning nominees lived another two-thirds of a year, on average."

So lemme see. If you take the whole sample, the difference was 1.4 years, or 1.4 * 12 = 16.8 months. I'm still not done with the morning coffee, so please correct me if I'm wrong, but 16.8 months is a bit lower than the 20 months you've calculated for two sigma.

I find it more interesting when they restrict it to winners from the same country, since, well, only then it's really apples to apples. (You'd expect that someone from the USA would live longer than someone from, say, India. Doubly so when there's data from the early 1900's.) Then it's only 2/3 of a year, or 8 months difference. Quite a bit lower than 20 months, I would say. Plus, it's inherently a lot of smaller samples, so even the 20 months figure would become larger.

More importantly, that difference between "winners vs nominees everywhere" and "winners vs nominees from the same country" tells me that the first one might not be entirely unbiased as samples go. If, say, more winners come from the top industrialized nations with high standards of living, while the larger nominees sample include more people from some poorer countries too, that alone could account for the the 8.8 months difference in the two figures.

I haven't properly studied the names and countries of origin for everyone, but for physics and chemistry it sounds at least like a _believable_ kind of bias: you don't see third world countries building big cyclotrons (for advanced physics research) or having advanced big pharma companies (for advanced chemistry research.) Something like, say, the prize for literature might have been a less biased sample: you don't need lab equipment and funding in the billions to write a book. And if the only cause there is that winning a prize and resulting alpha-monkey status instantly gives you some extra months, then the effect should be the same there too.

This gets funnier when you add this quote into the mix: "Oswald and Rablen found that Nobel laureates in physics lived an average of almost a year longer than laureates in chemistry."

Err... wait a minute. Let's do some maths there, then. Assuming there have been roughly as many winners in physics and chemistry, to keep the average, then the 16.8 months figure becomes something like 22.8 months for physics and 10.8 months for chemistry. It may look like now the physics number is finally signifficant, but it also means half the sample, so sigma is 120 months / sqrt(67) ~= 120 / 8 = 15 months, so two sigma is 30 months. Hmm, now even the figure for physicists is still less significant, and the figure for chemists is outright useless.

Let's apply that piece of wisdom for the "winners vs nominees from the same country", since, again, that's really the only one which doesn't have a built-in bias. To keep the 8 month average and assuming again equal numbers from the same country it becomes 14 months for the physicists and 2 months for the chemists. Frankly, living 2 months longer as a chemistry winner already starts to sound thoroughly insignifficant. But probably that 1 year difference doesn't apply here too, or is proportioanlly reduced too, so let's ignore this.

Was there some other difference between

Give the study some credit

[Ambitwistor]

Every time I see a social science study posted here on Slashdot, everyone comes out of the woodwork with "correlation doesn't equal causation", or "this study is obviously [true|false] because of so-and-so obvious effect", etc. Please give the authors some credit. They did consider various biasing effects, such as Nobel nominee age, the fact that nominees may die before being awarded the prize, they examined alternative causal factors such as the possibility that the winners' longevity was due to their increased income, and so on. Sure, correlation isn't causation and this study doesn't prove anything, but it's not as shoddy as the Slashdot armchair experts seem to think. Read the paper [columbia.edu], or a brief summary [columbia.edu] by a statistician unrelated to the study.