Objective Personality Assessment
Fb interpretation
Compare F and Fb
Correction Scale [K]
Correction scale was developed to detect more subtle attempts to deny psychopathology and present favorably [i.e. defensiveness] Scale development: -The items were derived by comparing two groups [empirical keying]: psychiatric patients who produced normal MMPI results compared to a normal sample. -30 items - cover a range of problems areas a person can deny. -Items are more subtle than the Lie scale items.
[F] minimizing symptoms
Inpatient F<54, Outpatient F<54, nonclinical F<39. Minimizing symptoms, faking good.
[F] Real symptoms
Inpatient F=55-79, Outpatient F=55-69, Nonclinical F=55-64. Real symptoms, deviant social or political ideas.
[F] Symptom exaggeration
Inpatient F=80-99, Outpatient F=70-89, Nonclinical 65-79. Symptom exaggeration, cry for help, serious psychological symptoms.
[F] Invalidates
Inpatient F>= 100, Outpatient F>=90, nonclinical F>=80 Invalid: responded randomly, All True, or malingering
Fb invalidation Clinical
Inpatient/Outpatient -Fb >=110 and Fb-F>=30. Approach changed. Interpret scales only from the first half: F, L, K and Clinical scales.
S scale items
Items were factor analyzed to create sub scales based on item content. Fell into 5 areas.Elevations on the sub scales reveal particular content areas in which the client is denying difficulty (10 point elevation in one or two areas).
K scale average score
K = 40-64
K scale unfavorable light
K<40 depending on other scales, can suggest an attempt to present oneself in an unfavorable light.
L open or exaggeration
L < 50. Person was open and honest. Depending on other scales, can suggest symptoms exaggeration.
Other issues regarding reliability
Length of the test. Format T/F less reliable because of the 50% chance of endorsing an item in the "scored" direction vs. 20% chance with a multiple choice test. Tests with sub scales should report the reliability results for the whole test AND each of the sub scales. Acceptable levels of reliability depend on the variable being measured.
Objective Personality Assessment
MMPI-2, Wechler Adult Intelligence Scale, California Personality Inventory. Structed, unambiguous test stimuli, often self-report. Usually direct questions regarding the person's opinion of him/herself. Test results are interpreted by comparing the person's responses with a set of criterion-refernced normative data.
Validity scales
The first step in interpreting an MMPI protocol is evaluating whether or not it is valid (the respondent's test-taking attitude). Many different influences can result in an invalid protocol: Deliberate minimization of symptoms, deliberate exaggeration of symptoms, lack of comprehension or concentration.
Content Validity
The representativeness and relevance of the test to the construct being measures.
Split Half Reliability
The test is given only once, but the items are split in half, and the two halves are correlated. This give a measure of internal consistency of the items rather than their stability over time. (Problem: how to split the items. Compare odd vs. even. You need a lot of items to do this.)
What is personality
The totality of emotional and behavioral traits that characterize the person in day-to-day living under ordinary conditions; it is relatively stable and predictable.
Collect a large pool of items
They collected personality statements from case histories, textbooks, and earlier personality scales. Psychoanalytic tone.
Criterion Validity
This is also known as predictive validity and is determined by comparing test scores on the new measure with performance on an outside measure that has theoretical relation to the variable.
Format of the MMPI
True/False format.
How do we test or measure personality?
Two major categories of personality tests: Projective & Objective
Scale 1 development
criterion group of 50 hypochondriacs - characterized as having a preoccupation with the body and having fears of illness and disease. -32 items
Convergent Validity
scores correlate with variables that are theoretically similar.
Revision of the MMPI
-Concern of the original comparison sample was not representative of the US pop. Archaic language, sexist attitudes, assumed Christian beliefs. Some areas of clinical interest were not addressed by the original items, (i.e. suicide potential and drug use)
FBS scale interpretation
-Experts supporting addition of the FBS to the standard set of MMPI-2 validity scales recommended that: -Raw scores above 22 should raise concerns about the validity of self-reported symptoms. -Raw scores above 28 should raise very significant concerns about the validity of self-reported symptoms, particularly with individuals for whom relevant physical injury or medical problems have been ruled out.
S scale Interpretation Invalid
Clinical S>= 70 Nonclinical S>=75 Invalid Defensive approach, fake good response set.
L average scores
Clinical/Nonclinical L=50=59
L interpret cautiously
Clinical L=65-79 and Nonclinical L=70-79. Interpret cautiously if at all: Person trying to appear virtuous and well-adjusted.
Scale 4 development
-criterion group of 75 psychiatric patients and 100 federal prisoners - all with diagnosis of psychopathic personality -they were characterized as by delinquent acts such as lying, stealing, sexual promiscuity, and excessive drinking -50 items
Clinical scale development
-criterion group with members having the diagnosis of interest (and only that diagnosis) -empirical keying procedure to select items.
Clinical scale interpretation
-higher scores generally associated more severe symptoms. -Interpretive inferences (or descriptors) are suggested for high scores based on research - high scores associated with particular behaviors and/or symptoms. -low scores not interpreted, except for scales 5 and 10.
MMPI original validity scales
1) Cannot Say [?] 2) Lie [L] 3) Infrequency [F] 4) Correction [K]
Eight clinical scales assess the following major psychiatric categories.
1) Hypochondriasis 2)Depression 3) Hysteria 4) Psychopathic Deviate 5) Paranoia 6) Psychathenia 7) Schizophrenia (3 subtypes) 8) Hypomania
Clinical scales
1) Hypochondriasis [Hs] 2) Depression [D] 3) Hysteria [Hy] 4) Psychopathic Deviate [Pd] 5) Masculinity-Feminity [Mf] 6) Paranoia [Pa] 7) Psychasthenia [Pt] 8) Schizophrenia [Sc] 9) Hypomania [Ma] 0) Social Introversion [Si]
Additional validity scales- MMPI-2
1) Variable response inconsistency [VRIN] 2) True response inconsistency [TRIN] 3) Back infrequency [Fb] 4) Infrequency psychopathology [Fp] 5) Superlative self-presentation [S]
Original vs. MMPI-2
550 vs. 567, final version includes most of the items contributing to the original Validity and Clinical Scales. Normative sample was collected using the 1980 census data as a guide. 7 sites in 7 different states.
Original intention vs. actual use
Apparent that the MMPI was not succeeding in its original purpose of providing a valid specific diagnosis of new patients according to scores on the 8 original clinical scales. Failure attributed to factors such as: the overlap of items between scales, and the unreliability of the early diagnoses used to develop the original scales. Used in a way that is different from what was originally intended. Now able to interpret many personality characteristics, traits and behaviors that are associated with scores on scale or combination of scales.
L unsophisticated defensiveness
Clinical L=60-64/ Nonclinical L=50-59. Unsophisticated defensiveness.
Large scale omitted items
Can result in lower scores on various scales. The MMPI-2 manual suggests invalidating if [?] raw score is greater than 30. Cutoff: Graham suggest invalidating or proceeding with caution if [?] raw score is greater than 10.
Four scales that assess the test-takers approach
Cannot Say Scale (?); Lie Scale (L); Infrequency Scale (F); Correction Scale (K).
Lie Scale [L] interpretation invalid
Clinica/Nonclinical L>= 80. Denied negative characteristics, claimed virtues.
K scale invalid
Clinical K>=65 Nonclinical K>= 75. Defensive approach, Fake good response set.
Scale 2
Depression [D]
Scale 2 interpretation
Depression: -depressive symptoms (particularly if T>70) -feel depressed, sad, blue, and dysphoric -hopeless and pessimistic about the future -talk about suicide -feelings of self-depreciation and guile -anhedonia -lack of energy -have diagnosis of depression -report physical complaints and poor sleep -agitated, tense and fearful -have poor concentration
Symtom Validity Scale [FBS]
Detect non-credible reporting of emotional distress among personal injury claimants and was added later to the MMPI-2. Scale Development: -items for the FBS were selected rationally based on some unpublished frequency counts of supposed malingerers' MPPI responses and observations of personal injury malingerers. -does not accurately identify those who are feigning emotional distress, and also tends to classify those with real distress as over reporters. -FBS successfully differentiates between persons who gave non credible reports of cognitive deficits and those know to have such deficits.
F minus K index
Developed by tough (1950) who found that those who were trying to appear to have severe psychopathology scored considerably higher on the F scale than on the K scale. F scale (raw score) - K scale (raw score). Score >= 11 suggest a fake bad protocol.
Lie Scale [L]
Developed to detect deliberate and unsophisticated attempts to portray oneself in an unrealistically favorable light. Scale development: -The items were written specifically for the scale (rational item selection). -15 items- all deal with minor flaws or weaknesses most of use are willing to admit. -These items are obvious -All the items were written so that an answer of False adds a point (most scales have combination of T & F responses.
Infrequency Scale [F]
Developed to detect deviant or atypical ways of responding to the test items. Scale development: -The items were those answered in the scored direction by fewer than 10% of the normal sample. -MMPI had 64 items, MMPI-2 has 60. -A factor analysis found 19 varied dimensions [e.g. paranoia, antisocial attitudes, hostility, poor health, etc.] -High scores are associated with elevations on the clinical scales.
Scales assessing content-nonresponsiveness
Did the test take respond to items without consideration of their content? If so, was it: -Systematic - over falling or over truing -random -intentional -unintentional - unable to read or comprehend the questions, poor attention. 1) Cannot Say [?] 2) Variable response inconsistency [VRIN] 3) True response inconsistency [TRIN]
Scales assessing content-responsiveness (overreporting)
Did the test taker intentionally report more problems or symptoms? Motivations can vary: -Cry for help -Avoid responsibility for their actions [e.g. not guilty by reason of insanity] - Monetary gains [e.g. personal injury or disability claims] Scales: 1) Infrequency Scale [F] 2) Back Infrequency Scale [Fb} 3) Infrequency Psychopathology Scale [Fp} 4) Symptom Validity Scale [FBS]
Scales assessing content-responsiveness- underreporting
Did the test taker intentionally underreport or minimize problems or symptoms? Motivations can vary: -Avoid stigma of emotional/behavioral problems -Embarrassment -Trying to get a job -Retain or gain custody of children Lie Scale [L] Correction Scale [K] Superlative scale [S]
Back Infrequency Scale [Fb]
Fb scale was developed to detect deviant or atypical ways of responding to the test items on the second half of the test. Scale development: -the items were those answered in the scored direction by <10% of the comparison sample. -40 items (first question is item #281)
Infrequency Psychopathology Scale [Fp]
Fp scale was developed to detect malingering as a supplement to the F scale. Scale development: -the items were those score infrequently by the normal sample and a group of psychiatric inpatients. -27 items -The items are less likely to reflect true psychopathology compared to the F scale.
use of the MMPI
Grew rapidly in clinical settings. Adapted for use in personnel screenings. Military and law enforcement. Meets the Daubert Standard for admissibility of evidence. Hundreds of new scales were developed using the 550 items. Some scales were developed using the empirical keying approach, other were developed rationally.
Scale 1 Interpretation
Hypochondriasis [Hs] -Excessive bodily concerns -conversion disorder (if T>80 and scale 3 is also very high) -somatic delusions (if T>80 and scale 8 is also very high. -vague somatic complaints -chronic pains, headaches, GI discomfort -chronic weakness, fatigue, sleep problems. -preoccupation with health problems - physical see increasing along with increasing stress. -diagnosed with somatoform,pain depressive, or anxiety disorders. -described as selfish, self-centered -not good candidates for therapy- don't like psychological explanations for physical symptoms.
Scale 1
Hypochondriasis [Hs].
Scale 3 development
Hysteria -criterion group of patients having hysterical reactions to stress -60 items
Scale 3
Hysteria [Hy]
Scale 3 interpretation
Hysteria [Hy] -often feel overwhelmed -react to stress and avoid responsibility by developing physical symptoms (particularly if T>80) -report headaches, stomach discomfort, chest pains, weakness -have symptoms that appear and disappear suddenly -report feeling sad, depressed, and anxious at times. -may have diagnosis of conversion, somatoform or pain disorder -lack insight concerning possible causes of their symptoms -self-centered, narcissistic -expect a great deal of attention and affection from others (and get angry she they don't get it) -resistant to psychological interpretations and may terminate if therapist focuses on psychological causes of symptoms.
Collect 2 criterion groups
Minnesota Normals consisted of 724 relatives and visitors of patients at the hospital, 265 high school students, 265 government workers, and 254 medical patients. Almost all were white, christian, blue collar workers, and lived in the rural area near the Minneapolis hospital.
Can a measure be valid if it is not reliable?
No- a measure must have a certain level of reliability to be considered valid.
K scale moderate defensiveness
Nonclinical 65-74 Moderate defensiveness: take into account when interpreting tother scales.
[F] Average
Nonclinical F=40-54. Average score, no problems or only minor problems reported.
L defensiveness
Nonclinical L= 65-69. Interpret taking defensiveness into account: Symptoms minimized, overly positive self presentation.
Scale 4
Psychopathic Deviate [Pd]
Scale 4 interpretation
Psychopathic Deviate: -difficulty incorporating values and standard of society (T>75) -likely to engage in a variety of asocial, antisocial, and criminal behavior (T>75) -rebellious toward authority -stormy relationships with family, blame family for their difficulties -history of underachievement -impulsive, strive for immediate gratification, act without considering consequences -take risks, show poor judgment -insensitive to the needs of others
Issues in Test Construction
Reliability and Validity
Fp interpretation
Research has shown the Fp scale may be more accurate than the F scale in identifying psychiatric inpatients and outpatients who are instructed to exaggerate their symptoms and problems. Suggested interpretation: -MMPI-2 manual suggested Fp scores> 100 are likely to indicate "faking bad" [i.e. malingering]and an INVALID protocol. -Meta analytic study suggested a raw score of greater than 7 as the cut off (T greater than 94 for men, T greater than 97 for women.)
Other validity indicators
Response Bias Scale [RBS]: intended to measure over reporting of cognitive symptoms. Dissimulation Scale {ds}: developed to identify persons who are simulating or exaggerating psychopathology. Subtle-Ovious Index: meant to compare the relative endorsement of subtle vs. obvious items as a measure of under vs. over reporting response sets.
Projective Personality Assessment
Rorschach, Thematic Appreception Test, House-Tree-Person.
Average S scale
S < 70 average score not very defensive.
S scale defensive interpretation
S = 70-74 defensive: take into account when interpreting other scales.
Superlative Self-presentation scale [S]
S scale was developed to detect the tendency to present oneself as highly virtuous, responsible, free of psychiatric symptoms, no moral flaws. Scale development: -The items were derived by comparing two groups [empirical keying]: Airline pilot applicants compared to the MMPI-2 comparison sample -50 items -items are more subtle than the Lie scale items -correlates highly with the K scale.
5 scale areas
S1: Beliefs in Human Goodness S2: Serenity S3: Contentment with life S4: Patience/Denial of Irritability S5: Denial of Moral Flaws
Conversion disorder
Scale 1 (Hypochondriasis) T>80 and scale 3(Hysteria) is also very high.
Somatic delusions
Scale 1(hypochondriasis) T>80 and scale 8 (Schizophrenia) is also very high.
Categories of validity scales
Scales assessing: 1) Content-nonresponsiveness 2) Content-responsiveness: -overreporting symptoms -underreporting symptoms
Cannot Say [?]
Scoring: the number of items omitted or double endorsed. Reason test-takers omit/double endorse: carelessness, confusion, poor reading comprehension, indecisiveness, avoid admitting undesirable facts.
TRIN cutoff
T score greater than 80 (raw score of 13) indicates inconsistent responding with a true bias that invalidates the protocol. (suggests an acquiescent approach). T score greater than 80 (raw score less than or equal to 5) indicates inconsistent responding with a false bias that invalidates the protocol (suggests a non-acquiescent approach).
VRIN cutoff
T score greater than or equal to 80 (raw score of 13). Indicates inconsistent responding that invalidates. -A completely random response set proceeds a t-score on the VRIN scale of 98. -An all T or F response set produces a t score of 50. -Malingering or those honestly admitting serious pathology will typical have near average T-scores.
True Response Inconsistency Scale [TRIN]
TRIN scale was developed to identify a tendency to respond inconsistently to the MMPI items by: -giving true or false responses indiscriminately. *Use in conjunction with the F scales & VRIN. Scoring: 20 items response pairs with opposite content. One raw-score point is added for each inconsistent pair. Depending on the question pairing, responses of T-T or F-F can contribute a point.
Four basic methods of measuring reliability
Test-retest, alternative forms, split half reliability, inter scorer or interrater reliability.
Scale 2 development
criterion group of 50 depressed patients- most bipolar in depressed episode. 57 items
Development of the MMPI
Was developed in the late 1930's by Starke hathaway and Charnley McKinley who were working in the University of Minnesota hospitals. They set out to develop a test to be useful for routine psychiatric assessments. They wanted to develop a paper and pencil personality inventory that could be administered to a group of patients and then easily scored. Their goal was a construct a more efficient and reliable way of arriving at a diagnosis.
Can a measure be reliable if it is not valid?
Yes- a measure can be reliable and not valid. It might be reliably measuring a different construct than intended.
Reliability
a measure's stability, consistency, and accuracy. There will always be a certain degree of error or noise. The goal in test construction is to reduce, as much as possible, the amount of measurement error. Two main sources of error: (1) natural variation in human performance; (2) imprecision of psychological testing methods, particularly measures of personality vs. other measures.
Test-retest
administer the test and then repeat on a second occasion. The reliability coefficient is calculated by correlating the scores across the two occasions. (The Problem: the practice effect).
Alternative Forms
create parallel forms of the test, you can then retest with less of the issue of memory or practice effect. (Problem: There is still some carry-over effect of the process of having been tested before. Also, are the alternate forms really equivalent?)
[?] greater than 10
determine on which scales the omissions occur. Are omission after item #370? Is there a theme to the omitted items? Interpret the unaffected scales.
Social Introversion Scale [Si]
developed by Drake (1946) and was added to the basic MMPI scales.
Construct Validity
developed to correct the inadequacies of content an criterion validity. The approach is to assess the extent to which the test measures a theoretical construct.
Variable Response Inconsistency Scale [VRIN]
developed to provide an indication of a tendency to respond inconsistently to the MMPI items. Use VRIN in conjunction with the F scales. Scoring: -67 item response pairs with either similar or opposite content. -One raw-score point is added for each inconsistent pair. -Depending on the question pairing, responses of T-T, F-F, T-F, or F-F contribute a point.
Defining high clinical scale scores
high score is generally considered >65. -Graham says some scales have interpretations at specific high scores.
Developmental Approach
items are chosen mathematically, ignoring their content. One empirical approach, called "empirical Keying," you give hundred of items to two pre-selected groups. The items that statistically differentiate between the groups (e.g. a group of depressed patients vs. a comparison sample) are used to create the scale. From about 1000 statements, they chose 504 that were reasonably independent of each other.
Divergent validity
low or negative correlations with variables that are dissimilar.
Validity
most crucial issue in test construction, tells us if a test is measuring what it INTENDS to measure. Establishing the validity of a psychological measure can be difficult due to the abstract nature of many of the constructs. Many psychological constructs have no tangible reality and must be inferred through indirect means.
MMPI-2
published in 1989 and is the version most commonly used. Intended for the use of persons 18 years and older. A shorter 370 item version (MMPI-2 RF) has been published and is becoming more popular.
Interscorer or Interrate reliability
the test is given by two different raters, or the same results are scored and by two different raters.
The Masculinity-Feminity Scale [Mf]
was constructed in an attempt to produce a scale that would distinguish between homosexual and heterosexual men. The item analysis did not produce enough items to differentiate these groups, so they broadened their approach to include items that were differentially endorsed by men and women. Items were added to the measure from the Terman and Miles Attitude-Interest Test to strengthen the Mf scale.
Theoretical assumption behind projective personality assessment
when confronted with a vague stimulus and required to respond in some manner, people cannot help but reveal information about themselves. Interpretation is based on a theory of human behavior and personality and it is assumed that a person's characteristics, defenses, etc., will become apparent through the testing process.