Use of special software (separate from what is being tested) to control the execution of tests and compare actual with predicted outcomes