Re: st: box and whisker plot with proportional box size?

[Ronan Conroy <rconroy@rcsi.ie>]

On 2 MFómh 2010, at 14:54, Jeph Herrin wrote:

> A colleague has shown me a couple of box and whisker plots where
> the width of the box (that is, the dimension that is perpindicular
> to the whiskers) is proportionate to the number of observations the
> box represents. I think it's a useful modification, and have been
> trying to create this myself. However there seems to be no option
> in -graph box- to control the width of the box.

While this seems like a useful option, it can rapidly result in  
illegible graphs and crowded axis labels in the presence of groups of  
widely differing sizes. JMP does it, and I keep the option turned off  
by default.

If you want to give an idea of the amount of data in each group, then  
superimposing this as text on the graph might be more efficient.

What always seemed to me to be a better idea was giving some sort of  
display of the actual data points with a boxplot superimposed, or a  
little to one side, as you can do with Nick Cox's -stripplot-. This  
shows more and, of course, gives a good impression of the density of  
data behind each boxplot.


Ronan Conroy
=================================

rconroy@rcsi.ie
Royal College of Surgeons in Ireland
Epidemiology Department,
Beaux Lane House, Dublin 2, Ireland
+353 (0)1 402 2431
+353 (0)87 799 97 95
+353 (0)1 402 2764 (Fax - remember them?)
http://rcsi.academia.edu/RonanConroy

P    Before printing, think about the environment





*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/