I NO RITE?!?!?
Sorry if it appears to be getting under my bonnet unduly, but they do have a famously bad track record for extrapolating wildly from tentative studies - making mountains of medical molehills, as it were. So really, I shouldn't be surprised by this article.
(LongJohnBaldry, Wed 12 Jan 2011, 11:42, 1 reply, 15 years ago)

300 people isn't a bad sample size
it's just getting from there to the assumptions made in that article is wrong.
You could as easily say, Birmingham is the cleanest station.
(PsychoChomp, Wed 12 Jan 2011, 11:45, Reply)

Not when that 300 is spread over several stations.

(Applebite Uh! A cow!, Wed 12 Jan 2011, 11:47, Reply)

That's not really the issue.
It's more that this is a field investigation and the other variables haven't (I assume, mail never links it's sources) properly been accounted for. Loads of excellent studies have been done with smaller sample sizes.
(PsychoChomp, Wed 12 Jan 2011, 11:51, Reply)

You can only use small sample sizes
(and this is a small sample size) for a very definite single or dual-conclusion study. And even then it depends upon the statistical relevance. If you studied 300 people, 150 southern and 150 northern, and said that 57% of the northerners didn't wash their hands whereas only 43% of the southerners did, you're talking about an actual difference of 10 washers/non washers in each case. Statistically as good as useless.
( the mighty badger Aphrodite, on a bar stool, by your side, Wed 12 Jan 2011, 11:57, Reply)

From what I can tell they checked for multiple germs on each subject
that would give me a big enough data set to do some statistical analysis.

The analysis would be useless though, because they haven't chosen the subjects specifically enough or randomly enough.
(PsychoChomp, Wed 12 Jan 2011, 12:04, Reply)

I would love to see the journal/report they took this from.

(Applebite Uh! A cow!, Wed 12 Jan 2011, 11:57, Reply)

It'll be key a stage 2 science survey

(PsychoChomp, Wed 12 Jan 2011, 11:58, Reply)

Worse, it'll be a study for a soap manufacturer.
Bet you.
( the mighty badger Aphrodite, on a bar stool, by your side, Wed 12 Jan 2011, 12:01, Reply)

I think this may be the truth.

(PsychoChomp, Wed 12 Jan 2011, 12:05, Reply)

yes it is.
In this situation, it's little better than throwing some papers with random conclusions in the air and then catching one blindfold
( the mighty badger Aphrodite, on a bar stool, by your side, Wed 12 Jan 2011, 11:48, Reply)

Well, exactly.
That, and major train statons are always going to be filthy. It's not the sample size I take issue with in this case, it's - as you rightly point out - all the assumptions that have to be made to get anywhere near that conclusion. Particularly the way they take a tenuous shred of evidence to declare a brand new North/South divide.
(LongJohnBaldry, Wed 12 Jan 2011, 11:49, Reply)