Chapter Nine

CHAPTER 9:
The MMPI-A: Extending the Use of the MMPI to Adolescents

- The MMPI-A works similarly to the MMPI-2
- It was originally used primarily with juvenile delinquents
- The MMPI-A is successfully used to measure the success of intervention programs
- The first extension of the MMPI to youth is believed to have been done by Dora Capwell in the 40s
o She compared delinquent and non-delinquent girls and found differences on all scales except for Scale 3 (Hysteria) and L
- Monachesi replicated Capwell’s findings with boys. The scales on delinquent boys that most obviously exceeded controls were 4 (psychopathic deviate), 6 (paranoia), 7 (psychasthenia), 8 (schizophrenia), and 9 (hypomania), with scale 4 being most different.
- Delinquent girls seemed to have a tendency toward sensitivity and feelings that they were unduly controlled.
- A follow-up study by Capwell found that slight elevations on scale 6 (paranoia) proved to be beneficial later on for these girls.
- In 1963 a large landmark study was conducted by the University of Minnesota Graduate School and the National Institute of Mental Health (Hathaway & Monachesi, 1963)
o 15300 Grade 9 students from Minnesota
o Collected MMPI data and they collected information about each participant from the schools, law and social services.
o Used to see how the MMPI can predict delinquent and acting-out behaviours
o Used to look at how the MMPI can be used with adolescents compared to adults

Adult-Adolescent Differences on the MMPI

Item Endorsements

- differences in item endorsement likely represent actual differences between adolescents and adults, e.g. adolescents gain weight more frequently and have different family relationships than adults do.
- girls differ more from women than boys do from men
o some of these are related to gender-role changes
§ e.g. flirting, stereotypically feminine-type occupations
- Adolescents tended to express greater excitement, e.g. stirring up excitement when bored
- Adolescents tended to dislike intellectual pursuits more than adults (e.g., reading history)
- Adolescents tended to endorse emotionality more so than adults (e.g., fits of laughing and crying that I cannot control)
- Adolescents were more likely to admit to pathological things like bizarre thoughts, ideas of reference, urges to do something shocking, feelings of unreality and strange experiences.

Code Types

- A no-high point profile type (one with no T scores greater than 54) was found more in adults
- Adults, especially women, had more code types indicative of neuroses (elevations on scales 1 - hypochondrias, 2 – depression , 3 – hysteria)
- Adolescents had more profiles indicative of sociopathic and psychotic qualities (elevations on scales 4 – psychopathic deviate, 8 - schizophrenia, and 9 – hypomania; with the 4-9/9-4 code type being most frequently occurring)
- Little validity was found for the use of code types among adolescents. Only the 4-6/6-4 code type was found to have external validity (4 – psychopathic deviate, 6 – paranoia)
- Using code types with adolescents is still not encouraged because there is not enough external validity was of yet.

Scale Elevations

- normal adolescents score at least one s.d. higher than adults on the original F, Pd (psychopathic deviate), Pa (paranoia), Pt (psychasthenia), Sc (schizophrenia) and Ma (hypomania) scales when using the original adult norms
- with the new norms the differences decrease but are still substantial
- these differences were the primary motivators behind creating an MMPI just for adolescents
- instead of having a mean of 50, the average for the MMPI-A t-scores are around 55-60
MMPI-A

- This first MMPI-A manual to use the MMPI-A with kids are young as 12 was by Philip Marks et al. (1974)
- They suggested that older adolescents might still use the adult MMPI
- In clinical settings, the adolescent norms tended to produce false-negative results (i.e., kids who were mentally ill were showing up as normal)
- In the 80s two additional norm sets superficially for kids in clinical settings were created

Interpreting the MMPI-A

- Adult interpretive approach: Used the same procedures as were used when assessing adults
- Adolescent interpretive approach: Marks, Seeman, Haller (1974). Age-appropriate norms and code-type descriptors were used to derive interpretive statements.
- Mixed interpretive approach: combining adolescent norms with adult code-type descriptors
- Combined interpretive approach: Archer (1987). Combined adolescent descriptors and adult descriptors, proving narrative descriptions for 29 different code types.
- Scale Descriptor Interpretive Approach: Focus on scale descriptors instead of code type descriptors

Problems with using the original MMPI with adolescents

- Also see the chart, Table 9-2, on p. 220

- Item problems

o MMPI’s content was confusing because of out-of-date phrases, awkward wording etc.
o Some items were not appropriate for young adolescents
o The limited item pool did not have themes specific to adolescents and their problems
o Items did not directly assess important problems that occur frequently in adolescents, such as eating disorders and substance use
o The limited item pool did not measure other important areas like strengths, motivations for treatment or change potential
o There were too many items

- Scale and norm problems

o Scales developed specifically for adolescents populations were not available
o No research data were available from adolescents for the potentially useful content scales, subscales or other special scales. No adolescent norms were available for these scales (The one exception was the MacAndrew Alcoholism Scale)
o Scale norms were out-of-date
o Results of available research were unclear about which norms were most appropriate to use with adolescents
o Mean profiles of adolescents in clinical settings were in the normal range
o The K correction was not used when creating the adolescent norms which contributed to the adolescent-adult differences in scores
o Other adolescent norms were proposed in the 80s but none were from large, representative samples

- Interpretation problems

o Predictive and descriptive accuracy for adolescents was lower for adolescents than for adults
o Many interpretive statements were available only from research using adults
o Code types and interpretations would change dramatically when going from adult to adolescent norms
o Often the interpretations based on the adolescent code type descriptors differed from those based on adult code type descriptors
o Use of potentially valid content measures was not possible because of limited research with adolescents
o Williams and Butcher (1989a) was the only source of MMPI standard-scale descriptors from a large adolescent clinical sample (they are the authors of our book)
o Williams and Butcher (1989a) found only limited validity for the code-type approach in their study
- General criticism
o There was not one widely used method of interpretation
o Much less research was conducted with adolescents

Advantages of using the MMPI-A

- see also Table 9-3 p. 223

- General

o A wide variety of psychological problems is assessed with the MMPI-A
o Requires little professional time for administration
o Many adolescents who are reluctant to admit to having problems in an interview will respond less defensively using this measure
§ E.g. A boy was admitted to hospital and was very quiet and withdrawn in the admission interview. He was given the MMPI-A and his scores indicated poor impulse-control, substantial risk of suicidal gestures, and possible initial symptoms of psychosis. He was put on close supervision as a result. One day later he became extremely combative and psychotic and the staff was ready because the MMPI-A had put them on the alert.
o Most adolescents like the T-F format
o The MMPI-A can also be administered using computers and adolescents prefer that.
o The MMPI-A can be scored quickly
o The MMPI-2 can be used with parents while the MMPI-A is used with adolescents so that the adolescent does not feel singled out. Adolescents are more likely to cooperate if they know that their family members have to do it too.

- Item

o The MMPI Restandization Committee changed the item content to make it more appropriate.
o 70 items with problematic wording were rewritten for the MMPI-A booklet
o Offensive items were deleted and sexual activities questions were deleted
o Unique items with adolescent-specific themes including peer-group influences family relations and issues about school were added.
o Items were added to assess common adolescent problems like eating disorders and substance use
o MMPI-2 items that were used to develop the Negative Treatment Indicators content scale were also included in the MMPI-A

- Scale

o Continuity between the original MMPI and the MMPI-A was maintained for several validity scales, the clinical scales and MAC, A, and R
o Scales 5 (Masculinity-Femininity) and 0 (Social Introversion) were shortened
o Statistical analyses and rational procedures about adolescent development were used to refine the F scale, VRIN, TRIN, the content scales, and two scales assessing drug and alcohol problems
o Some of the MMPI-A content scales are unique to the MMPI-A

- Norm/Descriptor

o MMPI-A scale norms were based on a contemporary sample and included youth belonging to minority groups
o MMPI-A norms, like MMPI-2 norms, were based on a uniform T-score transformation that ensured percentile equivalence across the difference scale scores.
o Both the MMPI-A and the MMPI-2 norms were developed using the same target distribution so as a person ages, his or her scores can be compared to his her previous scores
o Descriptors were separated for boys and girls
o A Spanish version is available with Hispanic norms. This was developed by Butcher et al. (1998) with Spanish-speaking youth from Puerto-Rico, Florida, California, and Mexico

- Interpretation

o The MMPI-A manual presents one norms set to use with adolescents.
o The same cut off for clinical interpretations (65) is used with adolescents and adults. However, clinicians should also be on the look-out for scores above 60
o Interpretation can be based on scales and descriptors derived from studies of adolescents
o Code-type congruence between the MMPI-A and original adolescent norms is substantial, allowing for the use of the code-type descriptors created by Marks, Seeman and Haller in 1974. However, the contemporary MMPI-A samples do not provide evidence for the validity of these. Future research is needed.