On 10/17/2010 11:25 AM, Jan private wrote:
> So here is the next version.
>
> Why does the intercept needs lower.tail=TRUE to give the same result as
> summary() for value=0?
It doesn't.
Or rather: For all coefficients, lower/upper tail depends on the _sign_
of t, so presumably you have a negative intercept in this particular
data set.
The more traditional way is 2*pt(-abs(t), n).
>
> # See Verzani, simpleR (pdf), p. 80
> coeff.test <- function(lm.result, idx, value) {
> # idx = 1 is the intercept, idx>1 the other coefficients
> # null hypothesis: coeff = value
> # alternative hypothesis: coeff != value
> coeff <- coefficients(lm.result)[idx]
> SE <- coefficients(summary(lm.result))[idx,"Std. Error"]
> n <- df.residual(lm.result)
> t <- (coeff - value )/SE
> if (idx == 1) {
> 2 * pt(t,n,lower.tail=TRUE) # times two because problem is two-sided
> } else {
> 2 * pt(t,n,lower.tail=FALSE)
> }
> }
>
>
--
Peter Dalgaard
Center for Statistics, Copenhagen Business School
Phone: (+45)38153501
Email: pd....@cbs.dk Priv: pda...@gmail.com
______________________________________________
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.
