Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: st: Bar graph
Nick Cox <firstname.lastname@example.org>
Re: st: Bar graph
Tue, 4 Jun 2013 16:07:54 +0100
It would perhaps be more interesting to spell out why I think this is
often a poor kind of graph.
As so often happens, the popularity of a graph bears essentially no
relationship to its utility. For example, people have been explaining
at length for a century or so that pie charts are a bad idea but it
doesn't seem to have had much effect. Perhaps it is time for satire,
not logic: see
The objection to pyramids is this. So, you want to compare values for
males or females (or whatever), but a pyramid design requires you to
pick up one bar and mentally superimpose it on another. I don't think
we are very good at this on the whole.
In demography, this is especially absurd, because very often the
numbers of males and females are similar, yet differences of the order
of 1% can be interesting and important.
Better designs are to
1. Superimpose directly, e.g. using -graph dot-.
2. Calculate the difference (or ratio) and look at that directly.
On 4 June 2013 15:43, Nick Cox <email@example.com> wrote:
> I wrote a program -slideplot- (SSC) a while back.
> I am no longer fond of this approach, so have not really developed it
> since. But the program is still lurking there, heedless of my lack of
> An overarching comment is that if -graph hbar- gets awkward, you
> should switch to -twoway rbar-.
> On 4 June 2013 15:19, Jörg Eulenberger <firstname.lastname@example.org> wrote:
>> Dear Statalisters,
>> i want to draw a graph like a Agepyramid to compare two groups, like this:
>> But i don't want on the left side the Agegroups. I need on the left
>> side different Items. Each row should show the values of one item, on
>> the one side the value of man and on the otherside the value for woman.
>> The Graph i want is on page 23:
>> Know every one a solution or a graph command?
>> Thanks a lot,
>> * For searches and help try:
>> * http://www.stata.com/help.cgi?search
>> * http://www.stata.com/support/faqs/resources/statalist-faq/
>> * http://www.ats.ucla.edu/stat/stata/
* For searches and help try: