Page 221 - Six Sigma Demystified

Part 3 s i x s i g m a to o l s 201

Protruding up and down from each box are “whiskers,” the lengths of which
are defined by the following formula:

Lower limit = Q1 – 1.5 × (Q3 – Q1)

Upper limit = Q3 + 1.5 × (Q3 – Q1)

Note that quartiles typically are used because they are nonparametric (not
dependent on the distribution of the data). If normality can be assumed, the
box-whisker plot instead may use the mean and standard deviation (SD) to
define the box and whisker lengths: the edges of the box defined at ±1 sample
SDs with the whiskers extending to ±3 sample SDs.
Extreme values also may be shown, usually as dots beyond the whiskers.

Box-Whisker Chart

Minitab

Use Stat\EDA\Boxplot (use the Simple option for the result as shown in the
figure).

Excel

Using Green Belt XL Add-On
Use New Chart\Box Whisker Chart (use Each Column will create a Box-Whisker
option for the ANOVA data shown in Table T.1).

Interpretation

The analysis shown in Figure F.3 might be useful to understand the differences
in process variation observed when each of four control strategies are employed
for a given process. Categorizing the data in this way sometimes can be a good
starting point for understanding the process dynamics and converging on suit-
able metrics.
Since this is an enumerative statistical tool, care should be taken in interpre-
tation. Unless statistical control of the underlying process is established, the
statistics presented on the box-whisker chart may not be reflective of the
expected process outcome.