This article builds on my article in the January 2002 issue of JACM on the use of run charts to analyze data variation and document process improvement. It describes a more powerful tool than the run chart for analyzing variation and measuring process improvement, namely, the control chart developed by Walter Shewhart. This article first explains the basic elements of a control chart, the tests for detecting a special cause variation, and how to choose the best control chart for the type of data being collected. It concludes with one case study using a control chart to document a successful intervention and some suggestions regarding computer software.