It is important that we keep in mind, however, that most of us who regularly use
exploratory methods of data analysis have strong backgrounds in inferential
methods. When we approach data exploration with fewer assumptions, we often set
aside, for the moment, much of the power of the mathematical models of statistics.
But to play data detective, we have a host of tools and experiences to draw on, many
of which stem from our knowledge of the mathematical models of statistics. As
Cleveland (1993) observes, “Tools matter (p. 1).” The tools that he was referring to
were methods of displaying data. We would add that underlying the skillful use of
such graphical tools is the skillful use of conceptual ones, which matter even more.