Skip to content
Prev 367019 / 398506 Next

how to interpret t.test output

Hi Somayya,
When you perform a t-test on two sets of numeric values, the answer
you get tells you how likely it is that those two sets of numbers came
from the same distribution. What most people are interested in is
whether the means of those two distributions are different. Let's see,
you seem to be with NHS in Scotland, so you are probably interested in
something like whether the mean reported severity of hangover is the
same for whisky as vodka in a sample of drinkers, controlling for
number of standard drinks of course. About all you can say from a
t-test is that those drinking one drink felt worse the next day _if_
the p value is small enough. Beware of doing lots of t-tests and
looking for <0.05.

Jim


On Tue, Feb 7, 2017 at 1:46 AM, Gardee, Somayya
<Somayya.Gardee at ggc.scot.nhs.uk> wrote:
Message-ID: <CA+8X3fWO_admoNNPJgJmr9R-G_QneCAHaKV_76piLLWTmZeTTQ@mail.gmail.com>
In-Reply-To: <57658195CF796048AA3FDA2C90A0F6418D03702E94@LAPPWGGCPMB04.ggc.scot.nhs.uk>