Moreover, most tests restrict the repertory of permitted of answers. "True" or "false" are the only allowed reactions to the questions in the Minnesota Multiphasic Personality Inventory II (MMPI-II), for instance. Scoring or keying the results is also an automatic process wherein all "true" responses get one or more points on one or more scales and all "false" responses get none.
This limits the involvement of the diagnostician to the interpretation of the test results (the scale scores). Admittedly, interpretation is arguably more important than data gathering. Thus, inevitably biased human input cannot and is not avoided in the process of personality assessment and evaluation. But its pernicious effect is somewhat reined in by the systematic and impartial nature of the underlying instruments (tests).
Still, rather than rely on one questionnaire and its interpretation, most practitioners administer to the same subject a battery of tests and structured interviews. These often vary in important aspects: their response formats, stimuli, procedures of administration, and scoring methodology. Moreover, in order to establish a test's reliability, many diagnosticians administer it repeatedly over time to the same client. If the interpreted results are more or less the same, the test is said to be reliable.
The outcomes of various tests must fit in with each other. Put together, they must provide a consistent and coherent picture. If one test yields readings that are constantly at odds with the conclusions of other questionnaires or interviews, it may not be valid. In other words, it may not be measuring what it claims to be measuring.
Thus, a test quantifying one's grandiosity must conform to the scores of tests which measure reluctance to admit failings or propensity to present a socially desirable and inflated facade ("False Self"). If a grandiosity test is positively related to irrelevant, conceptually independent traits, such as intelligence or depression, it does not render it valid.
Most tests are either objective or projective. The psychologist George Kelly offered this tongue-in-cheek definition of both in a 1958 article titled "Man's construction of his alternatives" (included in the book "The Assessment of Human Motives", edited by G.Lindzey):
"When the subject is asked to guess what the examiner is thinking, we call it an objective test; when the examiner tries to guess what the subject is thinking, we call it a projective device."
The scoring of objective tests is computerized (no human input). Examples of such standardized instruments include the MMPI-II, the California Psychological Inventory (CPI), and the Millon Clinical Multiaxial Inventory II. Of course, a human finally gleans the meaning of the data gathered by these questionnaires. Interpretation ultimately depends on the knowledge, training, experience, skills, and natural gifts of the therapist or diagnostician.
Projective tests are far less structured and thus a lot more ambiguous. As L. K.Frank observed in a 1939 article titled "Projective methods for the study of personality":
"(The patient's responses to such tests are projections of his) way of seeing life, his meanings, signficances, patterns, and especially his feelings."
In projective tests, the responses are not constrained and scoring is done exclusively by humans and involves judgment (and, thus, a modicum of bias). Clinicians rarely agree on the same interpretation and often use competing methods of scoring, yielding disparate results. The diagnostician's personality comes into prominent play. The best known of these "tests" is the Rorschach set of inkblots.
