In article <Xns9A89A214EF93goddardbenetscapenet@64.209.0.94>,
Bart Goddard <goddardbe@netscape.net> wrote:
>Miss Elaine Eos <Misc@your-pants.PlayNaked.com> wrote in news:Misc-
>E0A07B.18410922042008@news.sf.sbcglobal.net:
>
>> In article <1208913874.381162@irys.nyx.net>,
>> anon3c67@nyx.nyx.net (Bruce Watson) wrote:
>>
>>> Our sample consists of 157 centenarians living in Rome, 39 males
>>> and 118 females (ratio m/f =1:3), mean age being 101.59 +/- 1.8 years
>>> (+/-SD), 83.8% of the centenarians have never smoked, 13.5 % are
>>> former smokers, and 2.7% are active smokers.
>>
>> How do those numbers (% smokers, former & non-) compare to those who
>> died at age 50?
>>
>> Also, while it's interesting, the sample size is so small as to make it
>> statistically insignificant.
>
>No, the sample size is large enough. The real problem is
>the confounding variable of gender. Women live longer than
>men and women are less likely to smoke than men. They
>could chosen anything men are more likely to do than
>women and they would have gotten similar results. E.g.,
>what percentage of centenarians were golfers?
>
>It's funny, because confounding variables are usually sneaky
>and hard to unearth, but this one is so obvious.
Denial.