A test’s accuracy can be measured according to several factors. First, do the items measure what they are intended to measure? You would probably not include a question about a person’s favorite radio program in a test for OCD. Second, how do the items (questions, activities, and requirements) in a test relate to each other? If they do not relate to each other, it will be difficult to tell how they might group together, which is necessary in order to establish a trend in answers that will help the practitioner come up with a diagnosis. Returning to the previous example, a question about a person’s favorite radio program would not be related to a question about a person’s hand-washing habits. However, looking at hand-washing habits and sink-cleaning habits could help determine whether there is an obsession there. Third, does the test help distinguish between individuals who have an established diagnosis of OCD and those who do not? If there are no differences in scores between people who have OCD and people free from a mental disorder or people with a mental disorder that is not OCD, then the test is useless.
This is, of course, a very basic overview of checking for test accuracy, but it can give you an idea of what mental health professionals are looking for. All of the tests listed in the answer to the question preceding this one are considered reliable and valid measures of OCD because they have been developed by experts in the field, and evaluation in clinical trials has proven them accurate predictors of who does or does not have OCD or how severe a person’s OCD is.