Dimitris,
Thanks for the test. But you are using a normal approximation to the
poisson distribution. This is not applicable in my case. I was more
looking for an exact test. Probably based on Garwood's (1936) confidence
interval. If there is already an R implementation available, than this
would be helpful for me.
Jan
-----Original Message-----
From: Dimitris Rizopoulos [mailto:dimitris.rizopoulos at med.kuleuven.be]
Sent: maandag 12 september 2005 11:33
To: Jan Wijffels
Cc: r-help at stat.math.ethz.ch
Subject: Re: [R] poisson mean hypothesis
you could use something like the following (in case of two-group
comparisons make the proper adjustements):
pois.test <- function(x, alternative = c("two.sided", "less",
"greater"), mu){
alternative <- match.arg(alternative)
if (missing(mu) || (length(mu) != 1 || is.na(mu)))
stop("'mu' must be a single number")
nx <- length(x)
mu.x <- mean(x)
stat <- (mu.x - mu) / sqrt(mu.x/nx)
p.value <- switch(alternative,
"two.sided" = 2 * pnorm(-abs(stat)),
"less" = pnorm(stat),
"greater" = pnorm(stat, lower = FALSE))
list("sample mean" = mu.x, "null mean" = mu, "alternative" =
alternative,
statistic = stat, p.value = p.value)
}
################
y <- rpois(50, 5)
pois.test(y, alt = "g", mu = 4)
y <- rpois(30, 15)
pois.test(y, alt = "l", mu = 16)
I hope it helps.
Best,
Dimitris
----
Dimitris Rizopoulos
Ph.D. Student
Biostatistical Centre
School of Public Health
Catholic University of Leuven
Address: Kapucijnenvoer 35, Leuven, Belgium
Tel: +32/16/336899
Fax: +32/16/337015
Web: http://www.med.kuleuven.be/biostat/http://www.student.kuleuven.be/~m0390867/dimitris.htm
----- Original Message -----
From: "Jan Wijffels" <Jan.Wijffels at ucs.kuleuven.be>
To: <r-help at stat.math.ethz.ch>
Sent: Monday, September 12, 2005 10:01 AM
Subject: [R] poisson mean hypothesis