I think there's no easy way around it through _metan_ options (I trust
yo have the latest that uses stata 9+ graphics). _metan_ folks could
correct me ignorance. Short of modifying the command code, the fastest
way is to play with the postscript file of the graph. Easier than it
sounds.
Export to .eps and editing the .eps with a text editor.
[I presume you have a way to view .eps files (Mac views them natively
- ghostscript is a perfect free viewer for other platforms)]
1. usually the x axis labels are in the end of the .eps file. Edit it
by adding in the leading zeros.
[in your example I presume that searching for "-.3" will do the trick].
2. search for "p=0.000" - mind any spaces around "=" if applicable.
change it to "p<0.001"
Note that sometimes the same postscript command appears twice in the
file - I'm not clear why. If so you would have the edit "p=0.000" in 2
places (again use _find_ in your text editor) - otherwise you'll see
overlayed letters
tom
On Dec 18, 2007 3:45 PM, Richard Forshee <[email protected]> wrote:
> I should have specified that I am still using Stata 9.2.
>
> Richard A. Forshee
>
>
>
> -----Original Message-----
> From: [email protected]
> [mailto:[email protected]] On Behalf Of Anders
> Alexandersson
> Sent: Tuesday, December 18, 2007 3:32 PM
> To: [email protected]
> Subject: Re: st: Formatting questions for metan command
>
> Did you try the graph editor in Stata 10?
>
> Anders Alexandersson
> [email protected]
>
> Richard Forshee <[email protected]> wrote:
>
> > I have a quantitative meta-analysis manuscript that is very nearly
> accepted
> > by a journal. The editors have requested some technical changes in the
> > forest plot that I generated using the terrific "metan" command that is
> > available on SSC. Specifically, they request that leading zeroes be
> > included on the x-axis and that the p-value on the I-V overall line be
> > listed as p=0.001 instead of the p=0.000 that is generated by the package.
> >
> > My command is:
> >
> > metan coef se1, label(namevar=study) second(random) texts(120) ///
> > xtitle("Effect Size and 95% Confidence Interval") xlabel(-0.3, 0, 0.3,
> 0.6)
> >
> > In a standard twoway graph, I believe it is possible to force the leading
> > zero using a format option, e.g., xlabel(, format(%02.1f)). This does not
> > seem to be allowed with metan because when I try it I get an error:
> > "Unknown function format()".
> >
> > Any suggestions that would help me to tweak the graph to meet the
> > specifications of the editors 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/
>
>
> *
> * 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/
>
*
* 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/
