Please Respond to [log in to unmask] not myself
Hello Everyone
I am trying to get some figures for sample surveys in different industries.
In particular the sort that lead to claims made in advertising, e.g.,
In tests 8 out of 10 cats prefer Whiskers (I know it's now owners and who expressed a
preference)
Our tests show 90% of woman have visibly reduced fine lines after using our product.
Even with my role as chairman of the Business & Industry Section of the RSS I am unaware
of any standard level.
My feeling is people normally use whatever resource is available to them but generally wish
they had more data.
My company uses small scale for internal testing but moves to larger tests n=100-1000 when
working with the general public.
What do other companies/businesses use (only need very rough figures)? Any data received
will be treated in confidence.
As statisticians are there levels (of sample survey size) that we would deem unacceptable?
However the old statistics joke (probably politically incorrect now) about testing vaccines on
rabbits
1/3 of our sample died
1/3 of our sample survived
and the other one ran away
does illustrate the point.
Thanks in advance
Trevor
-----------------------
T J Duguid Farrant
Statistics Section Head
P&G, Gillette R&D
+44 (0) 118 9231669
-----------------------
|