[R] Should I use T-test or prop.test

[Dieter Menne]

Az Ha wrote:
> 
> I have a set of data where two groups of animals walk around in an
> environment and I have calculated the percentages of the area covered by
> them. Each group covers different area, for example the arithmetic mean is
> 35% and 23% with n=40 and 29 respectively.
> ...
> Should i use unpaired T-test or prop.test?
> ..
> 

Distribution questions aside, you should also consider what your story is.
>From the animal's viewpoint, a doubling from 1% to 2% is as good as one from
30% to 60%, so the proportion is the measure of success.

If you are an ecologist, a change in population from 30% to 60% might be
close to disaster for the environment, so percentage points (60-30)  or
(2-1) is the right choice.

Dieter

-- 
View this message in context: http://r.789695.n4.nabble.com/Should-I-use-T-test-or-prop-test-tp3171191p3171644.html
Sent from the R help mailing list archive at Nabble.com.