Graphical Methods.

This article was prepared for a yearbook of the Encyclopedia of Physical Science and Technology, Academic Press, Inc, and it is intended for a general audience of engineers, scientists, and college educated readers. Graphical methods that are appropriate for several different types of statistical problems are discussed and illustrated through examples. Line plots, histograms, and stem-and-leaf plots are useful for displaying and comparing values for one or two groups of objects. Quantile-quantile plots convey both general features and specific details about similarities and differences between two distributions. When there are three or more groups of objects, adjacent box plots are especially valuable for displaying the distributions and comparing the groups. For problems in which there are measurements on two or more variables, scatter plots can be used and augmented in a variety of ways. The needs from analyzing larger, multivariate problems and the ever-increasing computing power have lead to current developments in dynamic graphical methods, where the user interacts directly with the data through some input device and sees immediate changes in the plots on the computer screen. Three types of dynamic graphical methods are brushing, point cloud rotation, and alternagraphics.