[R] masked from package:base?

[Arne Henningsen]

On Friday 02 December 2005 09:00, you wrote:
> On 12/1/05, Arne Henningsen <ahenningsen at email.uni-kiel.de> wrote:
> > On Friday 02 December 2005 02:50, August Berg wrote:
> > > I am confused by the following description in
> >
> > http://www.maths.lth.se/help/R/.R/library/systemfit/html/hausman.systemfi
> >t .
> >
> > >html
> > >
> > > what does the "Not run" mean?
> >
> > This means that the command "library( systemfit )" is not executed when
> > the
> > package is checked by "R CMD check".
> >
> > @Jeff: I think we should remove "library( systemfit )" from the example
> > section, because everybody should know this.
> >
> > > if we do not load systemfit, how can we run
> > > the following code?
> > >
> > > ## Not run: library( systemfit )
> > >
> > > data( kmenta )
> > > attach( kmenta )
> > > ...
> > >
> > > I install the package of systemfit, and run the code.
> > >
> > > I got the warning:
> > > > library( systemfit )
> > > > data( kmenta )
> > > > attach( kmenta )
> > >
> > >         The following object(s) are masked from package:base :
> > >
> > >          q
> >
> > The data set "kmenta" has a (column) name called "q"
> > R> names(kmenta)
> > [1] "q" "p" "d" "f" "a"
> >
> > If you attach these data to the search path, there are two objects with
> > the
> > name "q": the quantity from Kmenta's data set and the function to quit R.
> > Hence, attaching this data set masks the function "q" (quit) from the
> > package
> > called "base". However, you don't have to worry, because "q()" and "?q"
> > still
> > work as expected.
> >
> > @Jeff: Since using "attach" is often troublesome, I suggest that we
> > remove "attach( kmenta )" from the example section and use "systemfit(
> > ..., data =
> > kmenta )" instead.
> >
> > Best wishes,
> > Arne
>
> Thanks. But still report such an error:
> > pval <- 1 - pchisq( h, dim( fit3sls$bcov )[1] )
>
> Error in pchisq(q, df, lower.tail, log.p) :
>         Non-numeric argument to mathematical function

The object "h" is not numeric, but a list:
R> class(h)
[1] "hausman.systemfit"
R> names(h)
[1] "q"    "qVar" "m"    "df"   "pval"

You probably want this:
R> pval <- 1 - pchisq( h$m, dim( fit3sls$bcov )[1] )

Or you can just type
R> h
2.53565 (P-value: 0.924389)

This is documented in the systemfit manual, please type
R> ?hausman.systemfit

Best wishes,
Arne

-- 
Arne Henningsen
Department of Agricultural Economics
University of Kiel
Olshausenstr. 40
D-24098 Kiel (Germany)
Tel: +49-431-880 4445
Fax: +49-431-880 1397
ahenningsen at agric-econ.uni-kiel.de
http://www.uni-kiel.de/agrarpol/ahenningsen/