Re: st: RE: RE: labels on bar graphs

[Richard Goldstein <richgold@ix.netcom.com>]

I agree that Nick's suggestion would solve my, and other people's, 
problem completely

I also did not understand Frank's comment

Rich

Nick Cox wrote:

> A percent format (with notional syntax) might look something like this: 
> With format %2.0p     .15   would display as 15 
>             %3.1p     .154                   15.4 
>             %2.0ps    .15                    15%
> 		%2.0pbs   .15                    15 %
>             %3,1pbs   .151                   15,1 % 
> So that p signals percent, s an explicit sign and b an explicit blank
> before the sign. 
> I believe that this would solve Rich's problem completely. I don't
> understand Frank's comments to the contrary. 
> Of course, that is all a "would". Only StataCorp can do that, or
> something similar. 
> Meanwhile, I intend to update -mylabels- on SSC to allow user-specified
> prefix and suffix text. That is one thing that can be done as a step in
> the same direction. 
> Incidentally, I also disagree with Frank, very mildly, on the broadening
> of this thread. Martin's comments seemed to me a fair expression of a
> defensible point of view that was entirely consistent with Rich's
> original question. 
> Nick
> n.j.cox@durham.ac.uk 
> Richard Goldstein
>
> I think that if Stata had an automatic way to turn proportions (mean of 
> a 0/1 variable) to percents, with the percent sign (as an option I 
> guess), that would have completely solved my problem
>
> Since my client is insisting on the presence of the % sign, the simple 
> solution (multiply by 100) sent by Amanda does not work for me
>
> Martin Weiss 
>
> > Well, obviously I hit a sore spot with my "straw man", given that it
> > attracted a few replies... Maybe I should have turned that into a new
> thread
> > though, to keep Rich`s other concerns alive. Sorry for that.
> Frank de Libero
>  
> > Rich originally asked for help not only to display "%" but much more.
> That's
> > how I read his message (below). Then Martin (see below Rich's message)
> > introduces what seems to be a straw man:
> >
> > [Why should users be forced] go to such lengths for a mundane task
> like
> > labeling with percentages...
> >
> > And the thread then went adrift given the general frustration
> (including
> > mine) about not having an enhanced -format- that includes "%". But as
> I read
> > Rich's original post, he was asking for a method that helped automate
> the
> > placement and value of the label, and then attaching a "%". The
> "clever
> > solutions" were not so much about displaying a "%" as they were about
> where
> > to display it and automatically display the value. A simple change in
> the
> > -format- command or equivalent would not solve all that. I think the
> > subsequent aspersions on Stata are unfair and stray from the original
> > question.
> Martin Weiss
>
> > All these clever solutions have a funny tendency to remind me of one
> thing:
> > a software that takes prides in its graph capabilities should not
> force
> > users to go to such lengths for a mundane task like labeling with
> > percentages...
> Richard Goldstein
>  
> > I need to produce a sizable number of bar graphs for a client.  The 
> > client wants the y-axis to be labeled as, e.g., "15%" rather than as, 
> > e.g., .15.  The range of proportions for these graphs varies quite a
> bit 
> > (from a max of less than .05 to a max of over .2) so I don't want to 
> > force them all to be the same.
> >
> > After seeing a graph, it is easy enough to use the ylabel option to
> get 
> > what I want.  Given that I am generally willing to live with Stata's 
> > automatic choice of tick marks, is there a way to automatically get 
> > labels in the form I want (e.g., "15%") rather than the way Stata
> wants 
> > (e.g., .15)?
*
*   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/