On 10/03/2016 5:05 PM, Texler, Michael wrote:
It is always difficult to compare accidents rates for 'rare' events due to the
wide 95% confidence intervals.
http://www.evanmiller.org/ab-testing/poisson-means.html
As mentioned by others, there often needs to be an order of magnitude
difference (i.e. a 10 tenfold increase or decrease of an accident rate) to
demonstrate statistical significance at the 95% level (this also means that
there is a 5% chance of accepting a chance variation as being significant).
Not actually true; the degree of difference between
groups/cases/whatever that you'll need to to get a statistically
significant result (be it for p=.05 or p=.01 or whatever) will depend on
the sample size, and on the characteristics of the sample and the
population you're drawing the sample from. There is in fact a whole
sub-topic of stats that is about working out what size sample you need
for a given situation in order to be able to plausibly see any real
differences between groups, should there be a real difference to be found.
It is not lies and damned statistics, but a 5% chance that the result is in
error (using commonly accepted practice).
See:
https://en.wikipedia.org/wiki/Poisson_distribution
Comments from statisticians welcome.
I'm an experimental psychologist - not an actual full-time statistician,
but I do play one on TV (if you know what I mean).
Teal
_______________________________________________
Aus-soaring mailing list
[email protected]
http://lists.base64.com.au/listinfo/aus-soaring
