the Minnesota Multiphasic Personality Inventory (MMPI)
1) First published in 1943 by Starke Hathaway, Ph.D., and Charnley McKinley, M.D.-University of Minnesota Hospitals
MMPI used emprical keying approach
1) Used the EMPIRICAL KEYING APPROACH in the construction of the various MMPI scales. This was novel at the time,
2) Responses to individual test items were treated as unknowns and empirical item analysis was utilized to identify test items that differentiated between criterion groups ( e.g., depressed patients vs. nonclinical individuals)
MMPI- Clinical Scale Development
1) Collection of a large pool of inventory items From 1000 potential test items, 504 were selected
2) normals and psychiatric patients
3) An item analysis was conducted separately for each of the clinical groups to identify the items in the pool of 504 that differentiated significantly between the specific clinical group and the normal group
Clinical Scale development- modified approach to code types (8)
4 Psychopathic Deviate
0 Social Introversion
MMPI Clinical scale development: Mf
Masculinity-Femininity Scale (Mf)-Scale was originally intended to distinguish between homosexual and heterosexual men. It was not effective and later broadened to distinguish between items endorsed by men versus women
MMPI Clinical scale development: Si
Social Introversion Scale (Si)-Scale was developed in 1946 and became one of the basic MMPI scales
MMPI- Validity Scale Development: Cannot say score (?)
1) Four scales were developed to detect deviant test-taking attitudes
2) total number of items in the MMPI that the individual either omitted or responded to as true and false
MMPI- Validity Scale Development: L scale or lie scale
Designed to detect unsophisticated and naïve attempts of the test takers to present themselves in an overly favorable way (e.g., True-I read every editorial in the daily newspaper)
MMPI- Validity Scale Development: The Infrequency (F) scale
Designed to detect individuals whose approach to test-taking is different from that intended by the test developers (elevated F scale-people who responded to the items without reading or understanding their content)
MMPI- Validity Scale Development: The Correction (K) scale
designed to detect clinical defensiveness. A high K-scale was intended to indicate defensiveness and call into question the person's responses to all of the other items.
Modified Approach to the MMPI-CODE TYPES
1) Clinicians began to refer to the profiles of individuals based on their numbers such as, "4-9" or "1-2-3", etc
2) Thus the test evolved to be used to generate descriptions and inferences about individuals ( patients and non-patients) on the basis of their scores. *** It is this behavioral description approach to the utilization of the test in everyday practice that has led to its tremendous popularity.***
Reasons for the revision of MMPI
1) The original MMPI was the most frequently used personality test in the U.S. It was even strongly endorsed by groups such as the Society for Personality Assessment (which primarily uses projective techniques) and the American Association for Behavioral Therapy.
2) The MMPI had not been revised since its publication in 1943.
3) The original standardization sample was not adequate (not generalizable)
4) There was concern regarding the item content of the original MMPI (language was archaic, some of the language was sexist, items dealt with Christian religious beliefs, etc)
5) The original MMPI had not undergone editorial review ( there were grammatical errors, double negatives, etc)
6) The original MMPI was too narrow in scope. Issues relating to suicide attempts, drug and alcohol abuse, were not addressed
Normative data collection for MMPI-2
1) 2900 participants were testes (1138 men, 1462 women)
2) Racial composition-81% Caucasian; 12% African American; 3% Hispanic; 3% Native American; 1% Asian American
3) Age range 18-85 years
MMPI-2 Additional changes
Concurrent with the development of the adult test, large normative data was collected and test items were revised for the development of the adolescent version-the MMPI-A, which was published in 1992
Final version of MMPI-2
1) Published in 1989 and consist of 567 items
2) Is very similar to the original MMPI, but more contemporary and has a better representative standardization sample
3) Alternate form of MMPI-2-the MMPI-2 Restructured Form or MMPI-2-RF. It includes the Restructured Clinical scales.
Coding the MMPI-2 profile
1) Much of the interpretation of the MMPI-2 is based on the pattern of the scores in relation to one another-CODE TYPES.
2) Most of the empirical information is based on two-point and three-point code types.
3) First determine the highest and the 2nd highest clinical scale scores (excluding 5 and 0). For example, a protocol in which the Hs (1) scale has the highest T score and the Hy (3) has the 2nd highest score would be "1-3" code type
Validity Scales-test taking attitudes
Cannot Say (?)
Variable Response Inconsistency (VRIN)
True Response Inconsistency (TRIN)
Back F (FB)
Infrequency Psychopathology (Fp)
Superlative Self-Presentation (S)
Symptom Validity (FBS)
? Scale ( Cannot Say; Cs)
1)The ? Scale is not actually a formal scale but merely represents the number of items left unanswered on the profile sheet. To minimize the number of "cannot say" responses, the client should be encouraged to answer all questions.
2) High number of ? (30+)
- Could indicate difficulties with reading, indecision, confusion or extreme defensiveness
VRIN (Variable Response Inconsistency Scale)
1) Comprises pairs of selected questions that would be expected to be answered in a consistent manner, but they are not.
2) High VRIN(MMPI-2 T=80; MMPIA T-80)
3) Could indicate indiscriminate responding; profile should be considered invalid and should not be interpreted, especially if F is also high
TRIN (True Response Inconsistency Scale)
1) Similar to VRIN in that is comprises pairs of items, but in this case, the responses are supposed to be opposites.
2) High (MMPI-2 T=80; MMPI-A T=80)
F scale (Infrequency)
`) Measures the extent to which a person answers in an atypical and deviant manner. The F scale items were selected based on their endorsement by less than 10% of the population
2) High F (approximating T=100; fake bad cutoff for inpatients = 100, cutoff for outpatient =90, cutoff for nonclinical settings=80; cut for MMPI-A=79
3) Indicates invalid profile due to random responding, false claims
FB (F back) Scale (MMPI-2; F1 and F2-MMPI-A
1) The 40-item FB was designed to identify a "fake bad" for the last 197 items because the F scale was developed for the first 370 items.
2) High FB (and F1 & F2; T=90 for nonclinical settings and 110 for clinical settings)
3) Indicates possible exaggeration of psychopathology
Fp (Infrequency-Psychopathology ) Scale
1) 27 item scale that reflects items infrequently answered by psychiatric patients
2) High Fp (T=94 for men; T=97 for women)
3) Indicates faking psychopathology among psychiatric patients
Fake Bad Scale (FBS)
1) The Fake Bad Scale (FBS) as developed to detect personal injury claimants who were exaggerating their difficulties. Research has been equivocal regarding its ability to detect.
2) High FBS (moderately indicative if raw score=22, more strongly indicated by raw score =28
3) Indicates faking bad or malingering
L (lie) scale
1) L scale consists of 15 items that indicate the extent to which a client is attempting to describe himself or herself in an unrealistically positive manner.
2) High L (T=65)
3) Describing self overly favorable due to conscious deception, or unrealistic view of self; poor insight due to denial of flaws; low tolerance to stress
K (correction) scale
1) Designed to detect clients who are describing themselves in overly positive terms, similar to L. However the K scale is more subtle and effective. Some with high K scores is unlikely to have significant elevations in L scale. K is more a measure of defensiveness. Because defensiveness may suppress clinical scales, a K correction is added to 5 of the clinical scales (1, 4, 7, 8, 9) to compensate for this defensiveness. Correction is not added to MMPI-A
2) H K ( T=65 or 70)
3) Describes self in an overly favorable light, denying difficulties
S (Superlative )Scale
1) High -T >70 in clinical setting; T>75 in nonclinical setting
2) Because the K and L scales have been found to be only moderately effective in differentiating persons who fake good, the S scale was developed to more accurately identify persons attempting to appear overly virtuous. The 50-item scale was developed by noting the differences in item endorsement between persons in an employment situation who were likely to be presenting themselves in an extremely favorable light (airline pilots applying for a job)
MMPI-2 Clinical Scales reliability
1) In general, reliability for personality tests are lower than those for tests of cognitive functioning.
2) The test-retest reliability within a one-week period of time for the MMPI-2 ranged from
.56 for scale 6 to .93 for scale 0
3) Internal consistency reliability is relatively lower for the various clinical scales, ranging from .36 for scale 6 to .87 for scale7
MMPI-2 Clinical scales validity
There have been at least 8,000 studies investigating profile patterns. These studies provide extensive evidence of the MMPI's construct validity
Another approach to Validity of MMPI-2
1) Another approach to validity is the assessment of the accuracy of inferences based on the MMPI
2) For example, the accuracy of neurologists' diagnoses was found to increase when they added an MMPI to their patient data.
3) Garb (1998) concluded that the MMPI was more accurate than social history alone, was superior to projectives, and that the highest incremental validity was obtained when the MMPI was combined with social history.
MMPI-2 Diverse groups
1) Conclusion: It would be premature to develop separate norms for African Americans.
2) Low elevation scores on these scales should be interpreted cautiously and not ascribe clinical significance to them.
3) In general it would be premature to develop new norms for any ethnic groups ( SES and age explains most of the variance)