[R] calculating power function

mb vera.kuehnl at uni-ulm.de
Wed Dec 18 13:30:15 CET 2013


Hello,

i know there are several functions on r to calculate the power of a test.
f.e the functions of pwr packet. In this packet the use the power function
(1-\beta) to calculate. My question is how we get this function for anova
(pwr.anova.test)?
I know that 
1-\beta=P(H1|H1) = P(F>F_[1-\alpha,k-1,N-k]) = 1-P(F<F_[1-\alpha,k-1,N-k]) =
1-P(MQE/MQR<F_[1-\alpha,k-1,N-k]) = ??....?? = 
1-P(F[k-1,N-k]<F_[1-\alpha,k-1,N-k]-k*n*f^2) = 
p.body = pf(qf(\alpha,k-1,(n-1)k,low=F),k-1,(n-1)k,k*n*f^2,low=F)
With f^2 the effect
and k*n*f^2 ncp
but how does the function which has a F[k-1,N-k] distribution look like? It
has to be the (variance between groups/df)/(variance within
groups/df)-k*n*f^2, but how can I write these so that I see the
dirstribution.
I have never found the exact way how to get to this formula. So may you can
help.

Thanks a lot.



--
View this message in context: http://r.789695.n4.nabble.com/calculating-power-function-tp4682396.html
Sent from the R help mailing list archive at Nabble.com.



More information about the R-help mailing list