This is a more positive idea than mine, which may be
sufficient for was wanted.
Note, however, an implication. It is not that -histogram-
shows bars of zero height when bins are empty; rather, it shows
no bars at all when they would be of zero height if they were
contemplated at all. It follows, I think, that cloning -histogram-
would be more difficult than I hinted, and some genetic
engineering would be needed besides.
Nick
[email protected]
David Harrison
>
> If you have no zero bins in your histogram then no new
> programming is required, just use -twoway histogram- with the
> -recast(line)- option. As Nick rightly highlights though,
> this will not deal correctly with zero bins.
Nick Cox
>
> I suspect that you want what is often called
> a frequency polygon. Frequency polygons
> are not supported by official Stata. There
> are various user-written programs in
> existence, although none I think in the public
> domain. One user got excited by the problem
> a while back, but (s)he doesn't seem to have
> followed through to completion.
>
> Programming them from scratch divides into
> the easy bit, which is easy, and the difficult
> bit (making sure that zero bins are shown properly),
> which is not so easy. I guess that a good Stata
> programmer could also look at -histogram-
> and clone it, but that would be a class act.
>
> An alternative, which I believe superior, is
> to go straight to -kdensity-. Statistical
> science spent too many decades stuck
> in bins. Climb out of the bins and get
> smoother estimates!
D.Christodoulou
> > This is probably a silly question:
> > How do I get a density line in the -hist- graph command
> > without the visual of the bins, i.e. to
> > connect the pinnacles of each bin, and get rid off the bins ?
> >
> > I search the graph manual but I couldn't find it, is it there
> > somewhere?
*
* 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/
