Thread Subject: p value from F statistic and n

Hello everybody

I have a silly doubt that has taken too much of my afternoon. If somebody could help I would be really grateful!

I need to calculate the probabilities (p-values) associated with a huge nxm matrix of F-statistics and the companion nxm matrix with the number of observations. It seems something trivial, but I can’t find the right function to do it in matlab.

Thanks in advance!

> I need to calculate the probabilities (p-values) associated with a huge
> nxm matrix of F-statistics and the companion nxm matrix with the number of
> observations. It seems something trivial, but I can’t find the right
> function to do it in matlab.

Admittedly I don't know what F test you want to perform, how the companion
matrix relates, and what numerator and denominator degrees of freedom you
would use.

But if you have the Statistics Toolbox, the fcdf function computes the lower
tail probability Prob(F<t) as a function of t. Depending on your test, you
may be interested in the upper tail (Prob(F>t) = 1-Prob(F<t)) or in both
tails.

-- Tom

Thank you very much Tom. I wil try to be more specific in next posts! but your fcdf function does the trick. Cheers!

"Tom Lane" <tlane@mathworks.com> wrote in message <he6sif$r9v$1@fred.mathworks.com>...
> > I need to calculate the probabilities (p-values) associated with a huge
> > nxm matrix of F-statistics and the companion nxm matrix with the number of
> > observations. It seems something trivial, but I can’t find the right
> > function to do it in matlab.
>
> Admittedly I don't know what F test you want to perform, how the companion
> matrix relates, and what numerator and denominator degrees of freedom you
> would use.
>
> But if you have the Statistics Toolbox, the fcdf function computes the lower
> tail probability Prob(F<t) as a function of t. Depending on your test, you
> may be interested in the upper tail (Prob(F>t) = 1-Prob(F<t)) or in both
> tails.
>
> -- Tom
>