HANDLING ATTENTION POINTS◆ 189It can be useful to look at a single box-and-whisker plot of the
statistics from one stem-and-leaf. But these plots’ real value is in
allowing you to compare the variation across different sets of
data points, as shown in Figure 7.4. Here one can see at a glance:
◆ variations in the central level of the three different sets of
observation, as shown by comparing the vertical position of
the medians and of the middle boxes; and
◆ variations in the spread of their data, shown by the vertical
size of the shaded middle boxes, the vertical size of the
boxes plus whiskers, and the presence or absence of outliers.
This is a sophisticated, multi-indicator comparison, yet accom-
plished in a very intuitive and accessible way. It can greatly
assist your understanding of the data, and it can also convey
a lot of information effectively to readers.
Smoothing datais another very useful data-reduction technique
for any kind of information that is analysed using line graphs,
especially over-time movements of any kind of index. There are
many cases where we acquire a large number of observations in
a volatile data series, one that zigzags up and down a lot, such
as the movements of stock markets, or commodity markets, or
public opinion polls showing the popularity of a government.
The key difficulty here is to try and separate out the meaning-
less or temporary fluctuations from the underlying, long-run
0102030405060Variable A Variable B Variable Cmediantop scorebottom scoreupper quartilelower quartileoutliersFigure 7.4 An example of a box-and-whisker chart
comparing across variables