I won't repeat comments I made earlier in the thread --
giving criticisms of this back-to-back design --
but will just add one more.
The back-to-back format suggests pairing, as does
your "pyramid" terminology. But the pairing is one
of rank only. In that situation, if the bars for men
went from left-to-right, not right-to-left, the intended
interpretation might come more naturally.
Nick
[email protected]
Gawrich Stefan
> Took me some time to find out, but the solution to the
> question I posted
> last week is rather simple in the end:
>
> I wanted to create a population-pyramid-like graph for
> sex-specific cancer
> frequencies (sorted by sex-specific rank) with all-positive
> barlabels):
>
> Men Women
>
> Prostate ************26******** *********32************* Breast
> Colon ********19****** ******24*********** Colon
> Lung *******14***** ****13******** Lung
> Stomach ***7** **6***
> Ovaries
> etc.
>
> (In fact, it's the same double bars but otherwise doesn't
> have too much in
> common with the pop-pyramid.)
>
>
> This graph can be done by combining two hbar graphs:
> The men's graph is "mirrored" with the "yreverse" option (so
> that bars go
> from right to left without having to use negative numbers).
> The woman's graph has a "xalternative" option, so that
> category labels are
> on the right.
>
> For both sexes, each graph shows the n most frequent disease groups in
> frequency order.
> I do this by collapsing sex-specific cases by disease
> category and computing
> a descending ranks var, by which the over-var is sorted in the graph
> command.
*
* 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/
