On Thu, Mar 21, 2013 at 12:16 PM, Nick Cox <njcoxstata@gmail.com> wrote:
> Worth noting that John Tukey in his Exploratory data analysis (Reading, MA,
> Addison-Wesley, 1977) did include a detailed example making it clear that
> box plots could not do justice to bimodality.
And this generalizes to the level of understanding. Statistics may be taught, and is often learned, as a set of formulae - pick the right test, get the arithmetic mechanically, and that's your "answer". People think "I'm doing a t-test" not "I'm comparing two means, on the assumption that the mean is the critical parameter for these groups." In respect of box plots and any other graph, are you drawing the graph to gain understanding of the data (analysis) or to demonstrate a feature that you have already identified (presentation)?
Allan
*
* 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/
