[S] Pie Charts and Type III Sums of Squares

Frank E Harrell Jr (fharrell@virginia.edu)
Fri, 13 Nov 1998 10:53:40 -0500

My feeling is that if you take the example where pie charts are
best (e.g., Charlie Roosen's "parts of a whole" example), pie charts
are still terrible. Cleveland and others have shown why: the
perception of the relative areas is dependent on the rotation of
the pie with the horizontal. It's difficult for some rotations to even
pick the single "largest" slice when differences are not great.

Regarding whether pie charts should have been included in S-PLUS,
my feeling is that they are worth the same as Type III sums of squares.
Maybe Type III sums of squares should always be depicted using
pie charts.

-Frank Harrell

This message was distributed by s-news@wubios.wustl.edu. To unsubscribe
send e-mail to s-news-request@wubios.wustl.edu with the BODY of the
message: unsubscribe s-news