Statistical Methods for Quality Improvement, Third Edition

CHAPTER 11

Graphical Methods

The cliché “a picture is worth a thousand words” could perhaps be modified to “a picture is often better than several numerical analyses” when adapted to the field of statistics. In this chapter we present some standard graphical tools for displaying data, which have been used for decades, in addition to some newer and, in many instances, more powerful methods, which have been introduced and have gained acceptance within the past 20 years.

The book How to Lie with Statistics (Huff, 1954) is replete with examples of misleading graphical displays which can serve to remind us of the much-quoted statement of Disraeli concerning different types of lies and statistics. The problem, of course, is not with statistics but rather in the way in which they are used. Football experts have been prone to say that statistics are for losers, but all of the game statistics generally indicate why a team was the loser. One issue that must be faced when using graphical displays is that there is generally not going to be one right way to proceed in terms of choosing appropriate graphical methods or in terms of how a particular method is used. There are multiple bad ways of graphically displaying data, however, and Tufte (1990, 1997, 2001) has revolutionized the graphical presentation of data through his books. Many of Tufte's main points have been summarized by Wainer (1984); see also the comments in Gunter (1988, 1995).

The graphical tools discussed in this chapter all ...