From: Tom Lane on
>> But my new stumbling block is that I'm getting p-values of zero.
>> Everyone likes low p-values, but a zero p-value reeks of an error
>> someplace.
>>
>> My specific line reads p=1-fcdf(1.0041e+004,2,71)
....
> Maple> with(Statistics): evalf(1-CDF(FRatio(2,71),10041));
>
> .8234018829e-87

Sorry, Nathan, I didn't suspect this would be an issue. There is a way to
evaluate that more accurately:

>> fcdf(1/1.0041e+004,71,2)
ans =
8.2340e-088

In other words, take the reciprocal of the F statistic value and swap the
numerator and denominator degrees of freedom. Thanks Walter, for finding
another way to get this result.

-- Tom