A quick answer to your questions:
1. Since nobody knows the "true" delta. I prefer to calculate the
power for a range of deltas. Most of the time for a range spanning - 2 *
expected delta up to 2 * expected delta. This gives an idea on how the power
changes if delta changes.
2. ?power.t.test explains how to calculate n for a given power, delta, sd and
sig.level. A quote from ?power.t.test: "Exactly one of the parameters
'n', 'delta', 'power', 'sd', and
'sig.level' must be passed as NULL, and that parameter is determined
from the others."
Cheers,
Thierry
----------------------------------------------------------------------------
ir. Thierry Onkelinx
Instituut voor natuur- en bosonderzoek / Reseach Institute for Nature and Forest
Cel biometrie, methodologie en kwaliteitszorg / Section biometrics, methodology
and quality assurance
Gaverstraat 4
9500 Geraardsbergen
Belgium
tel. + 32 54/436 185
Thierry.Onkelinx op inbo.be
www.inbo.be
Do not put your faith in what statistics say until you have carefully considered
what they do not say. ~William W. Watt
A statistical analysis, properly conducted, is a delicate dissection of
uncertainties, a surgery of suppositions. ~M.J.Moroney
-----Oorspronkelijk bericht-----
Van: r-help-bounces op stat.math.ethz.ch [mailto:r-help-bounces op
stat.math.ethz.ch] Namens Ethan Johnsons
Verzonden: vrijdag 27 oktober 2006 16:59
Aan: Peter Dalgaard
CC: r-help op stat.math.ethz.ch
Onderwerp: Re: [R] Power of test
Thank you so mcuh for the explanation, Chuck & Peter.
Two quick questions,please.
It states that delta = True difference in means. When the true diff
is unkown, can you use the expected diff for delta.
If you want to know the n (number of observations) off of power.t.test
to have i.e. 80% power, how do you calculate? Is there a way to do it
in R, or use algebra?
power.t.test(n = NULL, delta = NULL, sd = 1, sig.level = 0.05,
power = NULL,
type = c("two.sample", "one.sample",
"paired"),
alternative = c("two.sided", "one.sided"),
strict = FALSE)
Thank you,
ej
On 27 Oct 2006 16:37:08 +0200, Peter Dalgaard <p.dalgaard op
biostat.ku.dk> wrote:> "Ethan Johnsons" <ethan.johnsons op gmail.com> writes:
>
> > What would be the R formulae for a two-sided test?
> >
> > I have a formula for a one-sided test:
> >
> > powertest <- function(a,m0,m1,n,s){
> > t1 = -qnorm(1-a)
> > num = abs(m0-m1) * sqrt(n)
> > t2 = num/s
> > pow = pnorm(t1 + t2)
> > }
> >
> > Would you pls let me know if you know of?
>
> (Notice that power.t.test does this more accurately)
>
> For practical purposes, just halve a. Perfectionists may want you to
> add pnorm(t1 - t2), so that the total power becomes a when t2 == 0.
>
> BTW: -qnorm(1-a)==qnorm(a)
>
> --
> O__ ---- Peter Dalgaard ?ster Farimagsgade 5, Entr.B
> c/ /'_ --- Dept. of Biostatistics PO Box 2099, 1014 Cph. K
> (*) \(*) -- University of Copenhagen Denmark Ph: (+45)
35327918
> ~~~~~~~~~~ - (p.dalgaard op biostat.ku.dk) FAX: (+45)
35327907
>
______________________________________________
R-help op stat.math.ethz.ch mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.