Dynamic testing has been popularized as an alternative procedure to measuring ability. Traditionally, ability is measured from performance on problem-solving tasks that depend heavily on prior knowledge or developed skills. In contrast, ability measured from dynamic tests depends on the responsiveness of an examinee to varying testing conditions. If the varying conditions are cues or instruction, it is often maintained that the dynamic test provides a direct measure of learning ability. This article focused on computerized dynamic testing, in which cues and items are presented according to objective algorithms. Both appropriate designs and psychometric models for dynamic testing are elaborated. Further, two studies are presented that support the psychometric properties of a test that measures the susceptibility of reasoning to stressors.