The current spectrum of available software quality assurance techniques is wide and includes, e.g., testing, static analysis, and software model checking. For each of these techniques, a wide variety of specializations, optimizations, and tools have been developed. Determining the relative strengths and weaknesses of all of these different approaches and how they can usefully be combined is very difficult. This talk will present a highly general and customizable framework supporting the experimental, comparative assessment of different software quality assurance tools. The use of the framework will be illustrated by comparing ConTest, an IBM tool for randomized testing of concurrent Java code, and Java PathFinder, an open-source software model checker developed at NASA Ames. Joint work with Jeremy Bradbury and Jim Cordy.