Re: st: RE: graph command: xline() with by() option

[Nick Winter <nwinter@virginia.edu>]

Ooops....read the question too quickly to notice the version 7 
reference.  So consider this an answer to Nick C's "or even in Stata 10."

--Nick W

Nick Cox wrote:

> This is a good idea for Stata 8 up, but won't work in Stata 7, which is
> what Achille has. 
> Nick 
> n.j.cox@durham.ac.uk 
> Nick Winter
>
> How about something like this?  you would need to choose the value of 
> -top- to be the right height, which might require some experimentation. 
>   And you might want to add options to the -pcspike- plots to set the 
> line color, etc.
>
> . gen zero = 0
> . gen top = 0.05
>
> . graph twoway histogram stat || pcspike zero lower_es top lower_es || 
> pcspike zero upper_es top upper_es , by(grp)
>
> Nick Cox wrote:
>
>
> > Unfortunately, it is possible to say why this doesn't work. When Stata
> sees any references to local macros in commands, it substitutes them
> immediately with their contents and then tries to execute the command.
> That applies too to references of form 
> > `= <some calculation>'. In short, that is done once only. Thus there
> is no way that what you want will be done separately according to the
> distinct values of a variable specified in a -by()- option. 
> > Nor can I think of a neat way to do this in Stata 7, or even Stata 10.
> I guess you need to produce individual graphs, and then combine them.
> -forval- or -foreach- would help. 
> > Nick 
> > n.j.cox@durham.ac.uk 
> > Achille M. Luongo
> >
> > I would like to draw a histogram with two vertical lines showing
> > upper and lower estimates of a statistic for two (or more) groups
> using
> > the graph command of Stata 7.0.
> >
> > This is a snippet of the
> > example dataset:
> > grp      
> > stat           
> > lower_es           
> > upper_es
> > 1         
> > 91           
> > 43.45387           
> > 60.79625
> > 1         
> > 60           
> > 43.45387           
> > 60.79625
> > 1         
> > 87           
> > 43.45387           
> > 60.79625
> > 1         
> > 52           
> > 43.45387           
> > 60.79625
> > [..removed..]
> > 2         
> > 100           
> > 79.70421           
> > 124.2349
> > 2         
> > 60           
> > 79.70421           
> > 124.2349
> > 2         
> > 77           
> > 79.70421           
> > 124.2349
> > 2         
> > 125           
> > 79.70421           
> > 124.2349
> > [..removed..]
> >
> > I have already read the existing
> > post at http://www.stata.com/statalist/archive/2003-08/msg00282.html
> about
> > a similar topic.
> >
> > Nevertheless, I can&rsquo;t figure out how to
> > draw upper and lower estimates with &lsquo;xline()&rsquo; option when
> the
> > &lsquo;graph&rsquo; command is combined with the &lsquo;by&rsquo;
> option
> > like that:
> >
> >     graph stat, by(grp)
> > xline(`=lower_es[_n]' `= upper_es[_n]') 
> > In fact, the graph
> > command always draw vertical lines in the same position in all the
> > histograms combined with the by option (I suppose it always draws
> > lower_es[1] and upper_es[1]).
> >
> > I considered the possibility to
> > create histograms using the &lsquo;by&rsquo; clausole like that:
> >
> >     by grp: graph stat, xline(`=lb[_n]' `=ub[_n]') xlab(0
> > 20 to 140)
> >
> > but I obtained the the same previous result.
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/statalist/faq
> *   http://www.ats.ucla.edu/stat/stata/
--
--------------------------------------------------------------
Nicholas Winter                                 434.924.6994 t
Assistant Professor                             434.924.3359 f
Department of Politics                  nwinter@virginia.edu e
University of Virginia          faculty.virginia.edu/nwinter w
PO Box 400787, 100 Cabell Hall
Charlottesville, VA 22904

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/