[R] results from t-test

Duncan Murdoch dmurdoch at pair.com
Sat May 1 22:03:58 CEST 2004


On Sat, 1 May 2004 13:54:26 -0500, "Robert W. Baer, Ph.D."
<rbaer at atsu.edu> wrote:

>1)  Why the null AND the alternative weren't stated (guessed the answer was
>parsimony)

In one-sided tests the p-value is the same whether the null is "the
mean is zero" or "the mean is less than or equal to zero", so which
statement should it give?

>2)  Why if only one was stated, the convention was to re-state the
>alternative rather than the null hypothesis.

The p-value depends on whether a one-sided or two-sided alternative is
used, so it's important to state that.

>3)  Since technically the hypothesis is made a priori, why is it not
>re-iterated until AFTER producing the test results on the output.

That might be preferable, but I doubt that anybody who misinterprets
it now would notice the difference.  It would probably be sufficient
to add "the":

alternative hypothesis: the true mean is not equal to 0

However, the code that prints this line also prints results of lots of
other tests, and it doesn't really seem worth the time required to
know if "the" is appropriate in all situations.  

Duncan Murdoch




More information about the R-help mailing list