Here's a silly example of how to use -plot()-.
. ltable time censor if treat == 0,
graph plot(function y = 1 - exp(-1), ra(time))
The test is that the argument of -plot- could fit within
. twoway <argument of plot>
e.g.
. twoway function y = 1 - exp(-1), ra(t)
i.e. it could stand on its own -twoway- feet.
You want something like
. ltable time censor if treat == 0,
graph plot((rspike <whatever>)(connected <morestuff>))
where the <whatever> and <morestuff> are for the case
if treat == 1
However, the (temporary) variables in <whatever> etc. don't even
exist when you call -ltable-, quite apart from some
other difficulties, so I doubt that this will work.
-plot()- does offer a handle for those who know how
to use it; but here it's offering rather less of a handle than
implied (or inferred), or so I guess.
Another handle which would be useful is for -ltable-
to generate as variables the quantities it tabulates;
then you could knit your own graph.
Nick
n.j.cox@durham.ac.uk
Marcello Pagano
> Has anyone had experience using the -plot- function with -ltable-?
> To be specific:
> . ltable time censor, graph by(treat)
> works fine and it gives me two graphs side by side, each
> with a single
> line--treat takes on two values: 0 & 1.
> What I want is the two lines on a *single* graph.
> Allegedly the command
> -plot- is supposed to do just this, according to the
> documentation. But the
> documented explanation is rather poor, and no example is provided.
>
> I know a workaround, but I am curious to see how -plot-
> should work here.
> Any help would be appreciated.
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/