Personality Tests

Self-Report Personality Tests

Personality is most commonly assessed by asking respondents to endorse, rank, or otherwise acknowledge that they possess certain characteristics, hold certain attitudes and beliefs, or exhibit certain behaviors by administering one or more self-report personality inventories. Self-report measures provide specific stimulus materials, usually in terms of questions, statements, and descriptor lists. For example, one of the earliest self-report personality tests is the Adjective Checklist. This test consists of a list of more than 300 adjectives or personality descriptors in which the respondent is asked to “check” the ones that are generally accurate descriptions of themselves (examples of adjectives are happy, sad, anxious, depressed, friendly, etc.). Over the years, self-report inventories expanded beyond simple descriptors to include questions and statements about beliefs, attitudes, and behaviors, including broad measures of personality (e.g., the Minnesota Multiphasic Personality Inventory, the 16 Personality Factors Questionnaire, California Personality Inventory, etc.) as well as more narrowly focused personality measures (e.g., Beck Depression Inventory, State-Trait Anxiety Inventory, State-Trait Anger Expression Inventory).

Development of self-report measures generally involves some combination of rational (or theoretical) empiricism. Items for self-report tests are generated [Page 768]because they fit what one rationally (or theoretically) believes represents the domain of interest. For example, if one is interested in developing a self-report measure of extraversion, one would begin by constructing items that rationally describe what it means to be extraverted (e.g., I enjoy interacting with a large group of people; I am a very sociable person). Other instruments are designed to measure a specific theory of personality. For example, items on the Millon Clinical Multiaxial Inventory reflect Theodore Millon's theory of personality. Once items are conceptualized (rationally or theoretically), they are then empirically examined within the population. Items are included on the final version of the test because they discriminate consistently and significantly between people who possess or represent the domain of interest (e.g., extraverts) and those who do not possess or represent the domain of interest (e.g., introverts). Thus, self-report personality tests rely heavily on psychometric properties (i.e., reliability, content and construct validity).

One of the most critical issues regarding self-report personality testing is unfair test bias. Test bias is defined as the statistical difference found on test results among groups (e.g., difference found between women and men, difference found among racial or ethnic groups). Bias becomes unfair when the differences in test results among groups are not due to actual difference in the domain of interest, but rather due to some extraneous factor. For example, if the domain of interest is competitiveness, and the test item states, “I like to race in regattas,” only people who have knowledge of and had opportunities to engage in yacht racing would endorse this item. Respondents without such knowledge of and opportunities to participate in regattas would not endorse this item, but very well may be as competitive as those who do endorse this item. In this case, the bias may be unfair on the basis of socioeconomic status, and possibly race, ethnicity, gender, and geographic origin. That is not to say that the item does not measure the domain of interest, but rather that the item has differential validity based on various social, cultural, and demographic variables. Thus, when selecting a self-report personality test, it is crucial to inquire about the standardization sample on which the test was normed, as well as which populations the test has been found to appropriately represent. In addition, self-report inventories are not always appropriate for groups of people who may have difficulty understanding or responding to test material (e.g., children, the elderly, or people with below-average cognitive functioning). In these cases, alternative means of assessment may include interviews, behavioral assessments, or other-report inventories (from teacher, parents, partner, peer).

...