182 sYMBOLizaTiOn
implied, as with some state boundaries, it can be oriented to the meridians. The legend
should indicate whether the circles are range-graded or truly proportional and also
whether the wedges represent percentages of the total or fractions of the total.
chernoff faces
Chernoff faces (Figure 9.6; Plate 9.1) were first introduced by Herman Chernoff in
1973. These are a way to show multivariate data and complex relationships in an
easy-to-grasp, eye-catching manner. With this symbol, one can show as many as 15
variables by using variations of eyes, nose, mouth, shape of head, size and placement
of ears, and color of face. The award-winning map by Eugene Turner (Plate 9.1) shows
four variables using eyes, mouth, head shape, and color. A number of computer pro-
grams are now available,
including ArcGIS, to aid in creating Chernoff faces.
Problems with Chernoff Faces
Chernoff faces are eye-catching and can be effective. However, they have two poten-
tial problems. One is that of perceived stereotyping or bias when color is used on the
faces. If black or white is used to signify poverty or crime, critics complain that races
are being stereotyped; the counterargument is that green or purple faces, which have
been used, are jarring and perhaps give the impression of extraterrestrials. Therefore,
color should be used with care. Another problem that can arise is that of too much
information. If faces with multiple variables are used on a map with dozens of enu-
meration areas, the map becomes hard to comprehend. No studies have been done
to determine the optimum number of variables, but intuitively, a smaller number of
variables would seem to be more understandable than a greater number.
