Does anyone have any advice to give on this question?
Using a 3-level input factor alternative so that a function(below) can compute
both a two-sided and one-sided p-values. Making the two-sided test the default.
(I think ifelse would work best??) And output information about which
alternative was tested
function(yvec,trtvec,alpha=0.05,header="") {
#################################################
# A function to compute a two-sample t-test and confidence # interval
(equal-variance, independent samples). yvec is # a numeric vector containing
both samples' data. trtvec # is a vector, same length as yvec, of treatment #
identifiers for the data in yvec. A boxplot comparing # the treatments' data
is constructed. Output is a one-row # data frame reporting the results of the
test and # confidence interval
##################################################
trtvec=as.factor(trtvec)
boxplot(split(yvec,trtvec))
title(header)
ybar=tapply(yvec,trtvec,mean)
varvec=tapply(yvec,trtvec,var)
nvec=table(trtvec)
error.df=nvec[1]+nvec[2]-2
pooled.var=((nvec[1]-1)*varvec[1]+(nvec[2]-1)*varvec[2])/error.df
diff12estimate=ybar[1]-ybar[2]
stderr=sqrt(pooled.var*((1/nvec[1])+(1/nvec[2])))
tratio=diff12estimate/stderr
twosidedP=2*(1-pt(abs(tratio),error.df))
tcrit=qt(1-alpha/2,error.df)
lower=diff12estimate-tcrit*stderr
upper=diff12estimate+tcrit*stderr
calpha=1-alpha
out=data.frame(diff12estimate,stderr,tratio,twosidedP,lower,upper,alpha)
names(out)=c("Estimator","SE","T","P-value","Lower CI","Upper CI","Confidence")
out }
_________________________________________________________________
Gear up for Halo® 3 with free downloads and an exclusive offer. Its our way of
saying thanks for using Windows Live.
[[alternative HTML version deleted]]
______________________________________________
R-help@r-project.org 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.
