Lesson 4 - Assessment and the DSM-V

Reliability and Validity

The importance of the initial assessment is great. The initial assessment provides a baseline for later comparison and makes possible checking the effectiveness of the ongoing treatment to see if adjustments are necessary. For this reason, assessment must be reliable. With reference to psychological testing, reliability relates to the degree to which a measurement is consistent. Reliability can be divided into three categories: test-retest, internal, and inter-rater. The first category, test-retest reliability, requires that the psychologist obtain similar results with the same patient at two different points in time. With internal reliability, different parts of a test must produce similar results. In the last category, inter-rater reliability, two or more psychologists (or “raters”) must come to similar conclusions about the patient’s diagnosis.

Assessment must also be valid. Validity refers to the degree to which a technique or test measures what it is designed to measure. Validity can be divided into five categories. Face validity refers to a test or technique appearing to measure what it should measure. With content validity, the test or technique used actually measures the content or areas that are most important. The test or technique used may not measure everything, but it does not have to as long as it measures the major areas of concern. The third category, concurrent validity, requires that similar tests yield the same results. For example, two different I.Q. tests should yield similar results for the same individual. As the name suggests, predictive validity refers to how accurately behaviour can be predicted. For example, different patients with the same diagnosis are predicted to respond the same way to the same treatment. The last category of validity, construct validity, requires that a test measure what it is designed to measure. For example, if a test is designed to measure musical ability, actually measures verbal reasoning, it would not be considered a valid test.