ABSTRACT
This chapter describes an approach to test case design using a proven statistical method known as design of experiments (DoE). Benefits and costs quantified by a study comparing effectiveness of a manual test case design approach to effectiveness of DoE-based test case selection methods are described. James Bach has taken issue with designating DoE-based and combinatorial methods as best practices. He points out, we think quite rightly, that this method is not a solution to every testing challenge. DoE-based methods can be applied at a high level of abstraction in the early stages of testing to explore the application. Efficient test plans will test for as-yet-untested two-way defects with each new test case, whereas inefficient plans will test for the same potential defect again and again. DoE-based methods optimize plans generated to maximize efficiency.
