Psychological Testing
Psychological tests measure psychological variables: Intelligence and cognition, personality and psychopathology, behavior and emotion. Together with observation and history, psychological tests enable accurate description and diagnosis.
Types of Tests
Psychological tests can be divided into two broad types: Self-report measures and performance based measures. Self-report measures simply ask direct questions about symptoms, emotions, behaviors and history. For instance, a self-report measure for depression, such as the Beck Depression Inventory, might ask about sleep, appetite and mood, all symptoms of depression. These tests often accurately reflect the construct of interest. Nevertheless, they are subject to willful misrepresentation on one hand and poor personal awareness on the other hand. In other words, self-report measures only work insofar as the respondent wants to be truthful and knows himself thoroughly. In contrast, performance based measures do not directly ask anything; rather, they take a sample of behavior. Performance based measures are predicated on the fact that certain persons and personalities perform in certain ways. Likewise, when a person has a particular disorder, that person performs accordingly on such performance based measures. The anxious person, for example, will perform like an anxious person on such a test. In essence, performance based measures take a structured snapshot of behavior. Instead of asking direct questions, these tests engage examinees in a task and study their behavior throughout.
Accuracy: False Negatives and False Positives
Like all conscientiously constructed tests, psychological tests are designed to ensure accuracy and limit error. A test that cannot yield accurate results is not helpful and can even be harmful. Medical tests like mammograms are generally helpful in the early detection of breast cancer; they have enough accuracy to be significantly better than chance in detecting cancerous and precancerous growths. Nevertheless, mammograms do not work significantly better than chance on women with dense breast tissue. In such cases, the radiographic image cannot reliably differentiate between the healthy and the cancerous mass. Therefore, for such women results can be accurate; but just as often they can lead a healthy woman to think that she has cancer and a woman with cancer to think that she is healthy. In essence, a mammogram of a breast composed of dense tissue will contain an unacceptable number of Type I errors or false positives as well as an unacceptable number of Type II errors or false negatives. Whether the condition of interest is cancer or depression, physical or mental illness, it is essential to, as much as possible, avoid false negatives and false positives.