[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: RE: bihistogram

From   "Austin Nichols" <[email protected]>
To   [email protected]
Subject   Re: st: RE: bihistogram
Date   Wed, 27 Aug 2008 13:02:27 -0400

Probably the other sources of help or examples cited below are a poor
substitute for
 help twoway bar##remarks3
combined with -egen- and -collapse- as necessary (though it is the
-egen- and -collapse- step that no doubt is confusing to most users
who want to create their own variant on a histogram), but I agree with
other posters that the bihistogram or population pyramid often
obscures rather than illuminates, and I tend to prefer overlaid kernel
density plots to various alternatives (but there will be issues of
bandwidth or bin selection in any solution, I think).

One can even use -kdensity- or -tw kdensity- with already aggregated
data, though it stands out in part because it does not require the
data to be already binned and collapsed (and allows weights); using a
rectangular kernel you can in fact produce a histogram. But compare
the output to an "overlaid histogram" by pasting this into your
Command window:

sysuse pop2000, clear
ren agegrp a
ren maletotal m
ren femtotal f
loc c `"bw(.5) k(rec) yti("") xti("")"'
loc c `"`c' yla(,angle(0)) ysc(r(0)) xla("'
forv i=1/17 {
 loc c `"`c' `i' "`:label (a) `i''""'
loc o `", `c',labs(tiny)) leg(lab(1 "Male") lab(2 "Female"))"'
tw kdensity a [aw=m]`o'||kdensity a [aw=f]`o' name(k, replace)
g a1=a*3-1
g a2=a*3
cap labmask a1, val(agestr)
if _rc!=0 {
ssc inst labutil
labmask a1, val(agestr)
loc o `"yti("") yla(,angle(0)) ysc(r(0)) xla(2(3)50,val labs(tiny))"'
tw bar m a1 || bar f a2,`o'

On Tue, Aug 26, 2008 at 5:30 PM, Sergiy Radyakin <[email protected]> wrote:
> Hi there!
> What is the difference between bihistogram and a population pyramid?
> Is it another way of drawing (horizontal vs vertical)? Or is it just
> another term?
> If so, this might be helpful:
> and this too:
> net search pyramid

>> On Tue, Aug 26, 2008 at 1:54 PM, Austin Nichols <[email protected]> wrote:
>> > See also
>> >
>> > but I would fix the negative labels--no reason not to make them
>> > positive in both directions.
*   For searches and help try:

© Copyright 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index