Thanks for the tip about dotplots, Nick. I can see that they will be useful
for exploring data sets, but the display looks too cluttered for summarizing
or presenting the information.
-----Original Message-----
From: Nick Cox [mailto:[email protected]]
Wallace, John
> >
> > The built-in "box" parameter in the graph command comes close, in
> > that the whiskers encompass 1.5*IQR and further outliers are
> > plotted individually - but you can't use a
> > by(groupvar) to get an array of related plots.
> > The hbox addition I found in the STB files allows the by(groupvar)
> > arrangement, but the whiskers encompass the entire range of
> > data, rather than a function of the IQR.
> > This allows true outliers to dominate the
> > graph in an undesirable way, IMO
Make sure you are also aware of what -dotplot-
can do. With the addition of bars for median
and quartiles, you can have a hybrid display
which increasingly I offer prefer to boxplots.
Any broad patterns and any fine structure
are often clear simultaneously, the best of both
worlds.
Nick
[email protected]
---
Outgoing mail is certified Virus Free.
Checked by AVG anti-virus system (http://www.grisoft.com).
Version: 6.0.423 / Virus Database: 238 - Release Date: 11/25/2002
*
* 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/
