On Fri, Sep 11, 2009 at 4:44 PM, Ivan Kassal <kassal(a)fas.harvard.edu> wrote:
We are now on Sample D.
If the averages of the scores for the identical samples are not equal up to
some statistical criterion (is it the p test that one is supposed to use?
the t test? I never learned that properly) I will conclude that the guinea
pigs don't know what they're talking about and I will order the coffee that
I prefer.
That seems a bit intellectually dishonest. You should conduct a
double-blind taste test on yourself to see if _you_ can tell the
difference between the samples to a statistically significant degree.
If this is not the case you should simply order the cheapest brand,
even if everyone thinks it tastes like used motor oil, and convince
yourself that it's just a nocebo effect you're experiencing.
Ville
--
Ville Bergholm
ville.bergholm(a)iki.fi
http://iki.fi/ville.bergholm/
US mobile +1-617-3869378
EU mobile +358-44-5722802