Personality

1
Personality

• First text published by G. Allport in the 1930s
• First theoretical models date back to William
  James and Sigmund Freud (both late 1800s)
• Considerable variability in explanations for
  personality (biological/genetic models,
  psychodynamic, self-theory, etc.)
• Common Elements
• Stability (situations and time) differentiated
  from mood
• Research on stability over the lifespan (greater
  as we age)
• Affective, cognitive and behavioral components

2
Personality assessment

• Recent survey of practicing Ph.D.s, PsyD.s, and
  Ed.s revealed that only 32 use personality tests
  and only 43 do treatment planning.
• De-emphasis in personality training occurred at
  the same time as Mischel shock in 1968, so
  clinicians trained in the late 1960s and 1970s
  did not value personality assessment
• Today, treatment planning based on assessments
  is essential from both an ethical standpoint and
  for insurance reimbursement

3
Objective assessments?

• How can personality assessment be more objective
• assess any biases and correct for them (lie,
  defensiveness)
• find a method to avoid such biases
• look for convergence with reports from others
• assess with low face valid instruments and look
  for consistent patterns (though this only really
  addresses intentional faking)
• Personality assessment is used to further
  describe the client, just as a diagnosis does
  (note that you would not say that depression is
  causing the patient's behaviors, you merely use
  the term to summarize a cluster of behaviors. The
  diagnosis itself also does not necessarily imply
  a causal mechanism nor an explanation - those
  from different perspectives would define it
  differently)
• e.g., if someone is depressed it could be
  explained biologically, cognitively,
  behaviorally, or even in psychodynamic terms

4
The structure of personality

• Personality involves stable patterns of
  behavior, affect, and cognitions. So how stable
  is stable? (states vs. traits)
• Levels of analysis
• 1. factors - groups of traits that show better
  global predictive utility (e.g., Big 5 of N, E,
  O, A, C The Big 3 of N, E, P Big 2)
• 2. traits - clusters of consistent individual
  behaviors
• 3. habits - consistent (over time) individual
  behaviors
• 4. single acts - individual behaviors
• All levels are used to predict future behavior
  with the top being the most robust
• Consider this model when recommending or
  implementing change in clients

5
Predicting behavior

• Difficult to predict specific single behaviors
  from global trends (Epstein, 1983)
• For clinical evaluations, if the context of
  interest is known, then you may want to trade off
  the generalizability and give a specific
  prediction
• e.g., Pt.s test scores indicate that he is
  generally impulsive. This may be exacerbated when
  in the company of other individuals who are also
  impulsive and when the individual is drinking, as
  alcohol minimizes any inhibition processes that
  he might have. This substantially increases the
  likelihood that he will act impulsively when...

6
Two key discussions

• Read material in advance and know your MMPI
• Scheduled discussion
• Should we use projective tests?
• Are they tests or techniques?

7
Assessing Axis I and II

• Personality addresses both AXIS I and AXIS II
  disorders.
• What are some AXIS I disorders that might be
  related to personality traits? e.g.,
• depression and NA/Neuroticism
• anxiety and NA/neuroticism
• impulse control disorders extraversion/sensatio
  n seeking
• AXIS II personality disorders explicitly link up
  with personality assessments (video DSM-IV)
• Cluster A (odd) Paranoid, Schizoid, Schizotypal
• Custer B (emotional) ASPD, Borderline,
  Histrionic, Narcissistic
• Cluster C (anxious) Avoidant, Dependent,
  Obsessive-Compulsive
• PD NOS features of several Dx,but does not
  meet criteria for any one.

8
Selecting a test battery (see Beutler, 1995)

• What is the referral question?
• Single most important determinant
• Are there any limiting factors with regard to the
  client?
• Context of the evaluation? (work, school,
  hospital, etc.)
• Follow up assessment relevant to trait findings
  (e.g., patients who show impulse control problems
  should also be assessed for potential for acting
  out violently)
• Problem focused or broad, multipurpose battery
• Nomothetic (allows for normative evaluations) or
  ipsative (allows for the evaluation of the
  individual) analysis

9
If using qualitative methods, consider

• 1. Method appropriateness are there
  quantitative methods that you could use instead?
• 2. Openness make clear the theoretical
  orientation that undergirds the qualitative
  assessment
• 3.  Theoretical sensitivity use qualitative
  methods that are based on accepted theories not
  your own theories 4.  Bracketing of expectation
  you must explicitly state where your
  conclusions depart from accepted theories 5. 
  Responsibility how were the qualitative methods
  administered and interpreted
• 6. Saturation/generalizability when assessing
  traits, sample from a large number and wide range
  of situations 7. verification of methods
  cross-validate your methods using other reports,
  other test material to see if it agrees with your
  conclusions, do findings predict outcomes, etc.

10
If using qualitative methods, consider (cont)

• 8. grounding stay close to the data when making
  interpretations (no big theoretical leaps)
• 9. coherence do all of the interpretations fit
  together to make a coherent story
• 10. believability/usefulness does the use of
  the qualitative method provide more info on the
  client, or just raise more questions? Does it
  result in a believable narrative?
• 11. Intelligibility Is the report readable and
  jargon free?

11
MMPI (Hathaway McKinley, 1943)

• 10 clinical scales and 3 validity scales
• Empirical scale development with items selected
  based on their ability to differentiate normals,
  from a target group (another clinical group with
  similar symptoms was sometimes also employed)
• Clients should be 18 or older 6th grade
  education
• Generally lower face validity (breaks with
  tradition of items that clearly sample the domain
  of interest) most relevant for clinical
  population

12
MMPI development

• Item pool derived from psychological and
  psychiatric reports, textbooks, previous scales,
  etc.
• Criterion group composition
• Minnesota normals 724 relatives and visitors of
  patients at the U. of M. Hospitals, 265 recent
  high school grads, 265 administration workers,
  and 254 medical patients
• Clinical groups 221 patients representing the
  major psychiatric categories (excludes those with
  multiple diagnoses, or questionable diagnoses)
• Item analysis to identify those items
  differentiating the clinical and normal groups

13
MMPI development cont.

• The items that could differentiate were then
  cross validated with new groups of normals and
  patients
• Later developed two non-clinical scales
• M/F initially to identify male homosexuals was
  augmented with broader items
• Si derived from an introversion/extraversion
  scale and cross validated by predicting
  involvement in college activities in a second
  sample (all female college students)
• Validity scales were either derived rationally (L
  K) or from baserates in the normal group (F)

14
Utility of the MMPI

• Not considered a diagnostic inventory (as was
  originally intended)
• Ineffective at differential diagnosis (based on
  how it was originally developed)
• Numerical scale labels was intended to further
  minimize the connection with a specific
  diagnostic label

15
Some problems with MMPI

• Method of determining the criterion group
• The PIGs were not a truly random group (relatives
  and friends of those in the hospital though
  largely the medical patients) convenient
• Criterion and PIGs were largely from the midwest,
  in the late 1930s/early 1940s
• Utility of some of the scales as it matched
  diagnostic concerns of that era, dated and
  culture-specific item content, and
  representativeness of the norm group.

16
MMPI vs. MMPI-2 (1989)

• MMPI was the most widely used personality test in
  all pops (though only validated for inpatient
  adult samples)
• MMPI validation and norm samples were ones of
  convenience with limited variability on education
  (M8 years), coming from a rural background in
  the midwest
• Normative data collected in the 1930s
• Clinical cut-off now defined by t-score of 65 vs.
  70 on the MMPI

17
MMPI vs. MMPI-2

• Advantages of updating the test
• more representative norms (based on projected
  census data)
• relevance of the items
• language employed for the items (both temporally
  laden references like drop the hanky, and
  gender biases in item content)
• addition of new scales of relevance today
• Uniform T-score transformation now used so that
  T-scores reflect percentile ranks that are the
  same across all clinical scales

18
MMPI vs. MMPI-2

• Disadvantages to all updates
• over 20,000 published studies no longer apply
• MMPI-2 must revalidate all of the scales
• inability to make comparisons with adolescent
  scores (MMPI-2 vs. MMPI-A)
• Many of the new scales are very short and lack
  appropriate psychometric properties
• How often should we redevelop or renorm the scale?

19
MMPI-2 (1989) 567 items

• Norm group 2,600 community based subjects
• 1138 m 1462 f, aged 18-85 (M41, SD15.3),
  education 3 yrs - 20, 61 married median incomes
  25-35,000, 3 of m and 6 of f receiving mental
  health treatment
• 81 Caucasian, 12 A-A, 3 Hispanic, 3 Native
  American, 1 Asian-American

20
Validity scales

• Assumption that the clinical population will not
  be able to answer forthright
• Lie naive or unsophisticated lying (low SES and
  education)
• K less obvious (high SES and education)
  defensiveness is a component of all responding
• F answering questions in such a way so as to be
  different from 90 or more of the population
  (non-normative responses) See fake bad/fake good
  profiles
• F K Index can be used to indicate fake bad,
  with larger numbers making it more likely (little
  evidence to suggest that fake good can be
  detected) see p. 38

21
Clinical Scales

• 1. Hs - exaggerated concerns re physical
  illness, or tendency to report symptoms
• 2. D - Clinical dep unhappy pessimistic about
  the future
• 3. Hy - conversion reactions (substitute illness
  for emotions)
• 4. Pd - History of delinquency, antisocial
  behavior (non-conventional re moral standards)

22
Clinical scales - continued

• 5. Mf - prototypical gender identity (military
  recruits, stewardesses, homosexual males
  students)
• 6. Pa - paranoid symptoms (ideas of reference,
  persecution, grandeur)
• 7. Pt - anxious, obsessive-compulsive, guilt
  ridden, self-doubts
• 8. Sc - thought disorder, perceptual
  abnormalities (various types of Schiz.)

23
Clinical Scales - continued

• 9. Ma - exhibition of mania, elevated mood,
  excessive activity, distractibility, (possible
  manic-depression or BP II)
• 10. Si - college students scoring in the extreme
  range on introversion - extra.
• Costa McCrae (1990) suggest that the MMPI-2
  wont work in the normal pop. As people dont
  respond passively to items

24
New Validity Indexes

• Basic validity comes from L, F, K
• VRIN (variable response inconsistency)
• 47 pairs of items that should be answered
  similarly or the opposing direction. Client gets
  a point for each inconsistent response.
• A completely random response set results in T
  scores of 96 for m and 98 for f (gt80 inval.)
• acquiescent responding T 50

25
New Validity cont.

• TRIN (true response inconsistency)
• 23 pairs of items that are opposite in content
• either T/T or F/F to assess acquiescent or
  non-acquiescent responding
• larger raw scores true responding while smaller
  raw scores false responding
• raw scores should be between 6 and 12 in order to
  consider the profile valid
• Fb - back infrequency items for latter part

26
Coding the Profile

• List scale codes in order of their T-score
  elevations (from highest to lowest)
• usually only interpret 4 scale codes and order
  does not matter
• Welsh coding system involves adding symbols to
  numerical scale codes
• e.g., L F K 1 2 3 4 5 6 7 8 9
  0
• T 57 75 43 69 88 75 94 52 81 75 79 59 65
• Welsh 4268371095 FLK

27
Codes (listed to the right)

• 100-109, 90-99, 80-89, 70-79, 65-69,
  -60-64, /50-59, .40-49, 30-39
• Some coding forms use ! to denote scores of
  110-119 and !! for 120 or greater
• Underline identical T-scores (and list in
  ascending order) as well as those within one
  point of each other
• e.g., 4268371095/ FL/K.
• Code Types 2,3 and 4 point codes 5 point diff
  between lowest code T and T of highest scale not
  in the code.

28
MMPI-2 practice case M.S.

• Integrate the MMPI-2 data with the client
  information (vs. laundry list). Note profile
  valid.
• e.g., profile 3-2/2-3 should revolve around the
  discussion of depression and the manifestation of
  symptoms (physical symptoms tend to be
  substituted)
• How does this relate to M.S.?
• Recent loss, seeing her physician, isolation
• What does the 8 (or 2-3-8) tell you?
• How might psychotic symptoms relate to M.S.?
• Confusion from malnutrition, confusion as a
  result of depression, her age re dementia? All
  are possible

29
M.S. - continued

• Include discussion of (or section on) prognosis,
  recommendations, and diagnosis
• Axis I 296.24, Major depression, single episode,
  with psychotic features
• AXIS II No diagnosis (or deferred)
• AXIS III Malnutrition, dehydration, poor hygiene
  personal care
• AXIS IV Death of spouse (Severity extreme
  (acute event)
• AXIS V GAF Current, 24 highest past year, 52

30
MMPI-2 with other pops.

• MMPI was originally developed using Caucasian
  groups of patients
• Although some research has shown mean score
  differences between majority and minority groups,
  this is less relevant to the issue of whether
  there is differential predictive validity (few
  studies on this)
• Hall, Bansal, Lopez, 2000, have conducted a
  meta-analysis of 30 years research on minority
  groups and the MMPI (both versions)

31
Hall et al., 2000 - summary

• AA first note that cultural identification
  moderates all findings (cf. acculturation)
• Inconsistent findings re mean differences, with
  F, 8, 9 sometimes higher by approximately 5
  T-score points
• Many matched grouped studies of patients have
  found no differences, though Ns were small
  (meaning what?)
• Generally no differences in predictive validity
  that achieve statistical or clinical significance
  and any differences can be attributed to SES and
  age
• MMPI-2 has representative norms
• Minimal information on the supplemental scales
  and even less for the content scales

32
Hall et al., 2000 sum cont

• Hispanics likewise show few differences from
  Caucasians
• Possible differences for scales 3 and 0, with
  Hispanics scoring higher on 3 and lower on 0, but
  these effects were small with minimal clinical or
  statistical sig.
• Much stronger effect for acculturation in this
  ethnic group
• Few studies on Native Americans, but they show
  this pop. to score slightly higher on most scales
• Few studies for Asian Americans, and they show
  slight elevations for scales F, 2, 8.
• Generally valid to use for these pops given
  appropriate acculturation and understanding of
  the language

33
Other populations

• Given its original construction, there should be
  no problems using the MMPI in medical settings
• Medical problems do not necessarily result in
  higher scores (i.e., more distress)
• In substance abuse settings, no profile emerged
  to detect substance abuse, but scale 4 was a good
  predictor (see also the supplemental scales)
• We will discuss forensic applications later in
  the semester (see chapter 13)
• MMPI-2 can be used in non-clinical settings to
  screen for psychopathology, but there are some
  concerns.
• False positives are more common
• Has not been validated to predict success in
  other settings (e.g., jobs) which is true of most
  personality tests (predict interest)

34
MMPI-A (1992)

• Do we need a different inventory for adolescents?
  Why? Scales of concern?
• M/F for adolescents may be less defined
• Theoretically Pd is thought to be elevated, but
  actually it tends to be lower
• Personality is less stable overall so we need
  different norms to better interpret scores and
  relevant items for this age group
• Valid for those aged 14-18 (for 18 y.o., the
  decision is based on life circumstances e.g. at
  home? working?)
• Important to score on both adult and adolescent
  norms as there can be substantial differences
  (T-score shifts of 15 points)
• 478 items (some new some from the original
  inventory)
• written auditory forms both in English and
  Spanish

35
MMPI-A

• Includes all of the clinical, some new
  supplemental content scales. So we use
  basically the same scales but different
  descriptors (i.e., a high score on Hs will not
  mean exactly the same thing for the MMPI-A e.g.,
  Pd equates more with acting out)
• Biggest change was with the F scale since it is a
  norm defined scale (we need new norms)
• Norms 805 boys 815 girls aged 14-18 solicited
  randomly from schools in 7 states. Represents the
  U.S. for SES and ethnicity (again minimal diffs
  for ethnicity)
• Change from MMPI which had separate norms for
  different adolescent age groups (now only one)
• F scale now has 2 parts F1 1st part of test,
  F2 2nd part (Ftotal)

36
MMPI-A New scales

• New Supplemental scales
• Alcohol/drug problem proneness (PRO)
  empirically derived to assess the likelihood of
  alcohol or other drug problems. Items
  differentiate adolescents in tx from those having
  other psychological problems
• Alcohol/drug problem acknowledgement (ACK) face
  valid items that reflect the admission of
  problems
• Immaturity (IMM) reporting behaviors,
  attitudes, and perceptions that reflect
  immaturity (e.g., poor impulse control, judgment,
  and self-awareness). Items predict academic
  problems and cognitive limitations.
• Check for diagnoses such as oppositional-defiant,
  conduct disorder, and in adulthood ASPD

37
MMPI-A Psychometrics

• For the most part, the psychometric properties of
  the MMPI-A are sound. The reliability values are
  lower than the MMPI-2 values, but still within
  acceptable limits.
• Why might there be less temporal stability in the
  MMPI-A?
• General interpretative data from the MMPI-2 can
  be generalized to the MMPI-A, but this data
  should be considered in light of the clients
  position in life (i.e., consider how the scores
  relate to school life, problems with parents,
  need for independence, etc.)
• Note no K-correction for clinical scales even
  though a defensiveness score is calculated. So
  what are the clinical scale implications for a
  high K?

38
MCMI-III (Millon, 1990)

• 175 item scale assessing problematic personality
  styles and classic psychiatric disorders (drawn
  from the DSM)
• In contrast to the MMPI, this scale was derived
  theoretically to match the nosology (taxonomy) of
  the DSM to facilitate diagnosis and intervention
  planning. Assumes that any assessment is theory
  driven (vs. MMPI which tried to be a theoretical)
• The theory is grounded in evolutionary principles
  assessing 4 spheres existence (from serendipity
  to an organized structure), adaptation
  (survival), replication (reproductive styles that
  maximize diversity), and abstraction (the
  emergence of competencies to foster planning).
• Scored according to a polarity model. e.g., self
  vs. other orientation (reproduction), pleasure
  vs. pain (existential, or aim of, existence)
• Illustration Schizoid is marked by deficits in
  both pleasure and pain as indicated by the lack
  of emotion and apathy

39
MCMI-III properties

• A brief inventory (175 items) that takes only 30
  minutes to complete
• 3 modifier scales that correspond to the validity
  scales
• Disclosure defensiveness
• Desirability favorable response set
• Debasement lying
• 11 clinical personality patterns schizoid,
  avoidant, depressive, dependent, histrionic,
  narcissistic, antisocial, aggressive (sadistic),
  compulsive, passive-aggressive, self-defeating
• 3 scales denoting severe personality patterns
  schizotypal, borderline, paranoid
• 7 clinical syndromes anxiety, somatoform,
  bipolar, dysthymia, alcohol dependence, drug
  dependence, PTSD
• 3 severe syndromes thought disorder, major
  depression, delusional disorder

40
MCMI-III- continued

• Scales interpreted based on base rates for each
  dx and it assumes that disorders are
  interconnected (consistent with comorbidity data)
• Initial studies had classification rates of 90,
  but follow-up studies have been much lower (50
  or less)
• Validity data has been equivocal and the
  reliability data is likewise lower than the
  MMPI-2 (these are related, and both linked to
  number of items)

41
CPI (Harrison Gough)

• Developed at the same time as the MMPI and served
  as the personality test for the normal population
  (MMPI for the clinical pop.). Drew from a similar
  item pool.
• 480 T/F questions (some overlap with MMPI and
  others are new)
• Emphasizes more positive/normal aspects of
  personality
• 3 validity scales well being (normals asked to
  fake bad), good impression (normals asked to fake
  good), communality (popular/obvious responding
  that may reflect defensiveness and conformity)
• 15 general scales assessing a wide range of
  traits such as intellectual efficiency, capacity
  for status, achievement via conformity
• Grouped into 4 quadrants (factors) Norm favoring
  vs. norm doubting and externalizing vs.
  internalizing

42
CPI - continued

• CPI was revised in 1986 with norms based on
  13,000 males females
• Most commonly used personality inventory overall
• It has been replaced by the NEO-PI as most common
  in the last 15 years.
• Psychometrically sound (reliability and validity
  coefficients are high and stable for different
  pops), but a very long instrument.
• Also some question as to the need for validity
  scales in the normal pop.
• Burisch suggests this is unnecessary provided 1)
  no reason to lie, 2) knowledge of the
  construct(s), and 3) self awareness.

43
NEO-PI (Costa McCrae, 1985, 1992)

• Based on the empirically derived 5 factor model
• Assumption that 5 factors can represent all of
  normal personality
• Evaluated this model in a variety of contexts,
  with samples from all over the world and in
  different languages
• Assumes that language is the best place to start
  examining how to describe behavior (132 Eskimo
  words for snow indicates it is a meaningful
  construct)
• Neuroticism (emotional stability), extraversion,
  openness to new experience, agreeableness
  (quality of interactions) and conscientiousness
  (dutiful, organized).
• 5 factors have been recovered from other
  inventories like the Myers-Briggs, 16PF, etc.

44
NEO-PI

• Full version is 220 items and has 6 facets for
  each of the 5 factors
• Short form (NEO-FFI) has 60 items and provides
  factor scores only
• Norms are available for adults, college students
  and adolescents (though minimal differences
  between the latter two groups)
• Strong psychometric properties including very
  stable retest coefficients, internal reliability,
  and validated with other personality scales.
• Can be used to predict job interests (though
  vocational inventories such as the Strong
  Interest Inventory are better suited for this),
  but they do not predict job success (same is true
  for interest inventories)
• Often used for intuitive purposes and not
  empirically validated purposes (e.g., assume that
  a manager should be low on N and high on C vs.
  empirically testing this assumption with current
  managers)

45
Structure of affect and other issues

• Big two (PA/NA) vs. 5 factor
• Bipolarity of affect (vs. orthogonality)
• Temporal question for what defines affect vs.
  personality
• Problem of temporal language (e.g., at this
  moment)

46
Measures of Affect

• Note The EPI (Eysenck) likewise measures
  personality (extraversion and neuroticism) in the
  normal population, and these two factors are
  usually the first two to emerge in factor
  analysis.
• These factors correspond to the Big Two affect
  constructs (PA and NA)
• Note most of these measures do not address
  validity of responding
• Nevertheless, research suggests that these scales
  tend to be fairly accurate and reflect actuarial
  rates for affective disorders (5-9 of adult
  women and 2-3 of adult men)
• BDI published in 1961 and revised in 74, 78,
  and 96.
• Among the most commonly used inventories with a
  comprehensive manuals published in 1987, 1993,
  and 1996 (BDI-II)
• Normed for adolescents and adults aged 13 and
  older. 21 items with items arranged in a Guttman
  approach (increasing order of severity)
• Suicide potential in items 2 and 9. For dx of
  Depression see neurovegetative items

47
BDI - continued

• Internally consistent and reliabilities range
  from .48 to .86 for periods ranging from several
  hours to four weeks
• Why are retest coefficients smaller?
• No way to correct for faked scores
• Validated extensively for use in clinical
  settings
• BDI-II validated on 500 outpatients drawn from
  across the country and a student sample of 120
• 1 week retest was .93 and coefficient alphas were
  .92 or higher
• Average BDI-II scores are 3 points higher than
  the original BDI
• BDI-II time frame for each item focuses on last
  two weeks to match the DSM criteria

48
BAI (Beck Steer, 1993)

• 21 item symptomatic inventory
• Items rated on a 0-3 scale
• Validated for use for inpatient (N 1,086),
  outpatient (N 160) and college student samples
  (N65).
• Shows convergent validity with other measures of
  anxiety and some disciminant validity with
  depression measures (though they are correlated
  sharing 10-25 variance)
• Rapid self-report tool

49
CES-D (Radloff, 1977)

• Developed by NIMH for use as a screening tool in
  the general population (also in college and
  geriatric pops)
• Optimal test for this purpose in this population
• 20 likert type items focusing on the last week
• Better than the BDI-II at differentiating among
  those experiencing lower levels of depression
• Internal consistency is high (.85 in general pop.
  and .90 in patient samples).
• Retest figures tend to be low (.48) but this is
  less relevant for this construct
• A score of 16 is clinical cutoff and it assesses
  depressed affect, positive affect, somatic
  activity, and interpersonal functioning

50
MAACL-R (Zuckerman Lubin, 1985)

• Originally published in 1965 and revised in 85.
  (132 checklist type items)
• Normed on over 1500 adults, 400 adolescents
  (approx. 90 Caucasian, 10 Black)
• Scores for Anxiety, Depression, hostility, PA,
  and SS (the latter has very poor internal
  reliability)
• A rapid assessment but not as good
  psychometrically
• Can be used to evaluate states or traits and
  reliability figures are better (though not very
  high) for the latter
• Scales dont corr with social desirability and do
  converge with MMPI ratings

51
Behavioral Assessments

• Assumption behaviors can reflect cognitions and
  emotions (e.g., FACS Ekman Friesen, 1978)
• Proliferation of behavioral assessments with
  limited validity due to the assumption that
  behavior can be easily defined and that it
  represents a meaningful (typically underlying)
  construct e.g., sweating, pacing
• How to improve behavioral assessments?
• Identify the actual behavior being assessed (lip
  turned downward vs. sadness)
• Habitual behaviors may indicate underlying
  condition
• Acknowledge role of both traits and situations

52
Beh assessments cont.

• Also influenced by factors such as social
  desirability (varies depending if one is aware of
  the assessment)
• Difficult to organize and systematize behaviors
  (e.g., how does one smile equate with the absence
  of a frown re depression?)
• Very inconsistent findings regarding the
  organization of individual behaviors (even
  physical symptoms) via F.A.
• Why might self-report and behavioral assessments
  not overlap? What does this mean?
• Recall behavioral reactivity phenomenon change
  in behavior as a function of its assessment

53
Physiological measures

• Some people want to fill the world with silly
  physiological measures. And what's wrong with
  that? (McCartney et al., 1976)
• Biofeedback long history but very mixed
  findings
• Plethysmography changes in blood volume that
  may relate to emotional changes
• Pupillary responses attraction and fear?
• Polygraph arousal related to lying?

54
Cognitive testing refresher

• WAIS-III score interpretations for reports
• With regard to the index scores, which declines
  the most with age?
• Quick, its PS!
• Which show the greatest decrements secondary to
  organic dysfunction (trauma or disease)?
• PS, WM, and PO Depends on the area of the brain
  that is damaged. If diffuse, then all three. If
  temporal then WM, if more right hemisphere then
  PO.
• Which is the best indicator of premorbid
  functioning?
• VC (or subtests of vocabulary, similarities
  info.)

55
Cognitive and personality functioning

• What are meaningful ways to integrate these two
  pieces of information?
• What interpretations might one make for high IQ
  individuals relative to low IQ individuals re
  personality?
• Overlap with maturity? Less complex
  presentations?
• What PD is associated with extremist thinking
  (splitting), inability to recognize subtleties?
• Other implications?
• Ease of use for clients, alternative test format,
  wider range of responses (variability),
  alternative approach to detecting pathology,
  difficult for client to identify socially
  desirable or undesirable responding, theory based
• Defensiveness strategies (see MMPI-2)?

56
Projective test/technique

• MMPI/MMPI-2 is most frequently used test in
  inpatient settings
• Rorschach TAT are not too far behind
• Advantages of projectives?
• Disadvantages of projectives?
• Administration and scoring is generally less
  standardized so reliability and validity are
  compromised

57
Minimal criteria for a test

• Standardized administration
• Rorschach has numerous administration procedures
  (Bleck, Klopfer, Exner, etc.)
• Standardized scoring
• Rorschach has numerous scoring approaches (Bleck,
  Klopfer, Exner, etc.)
• Standard of comparison for interpretations (norm
  group)
• Minimal information with regard to representative
  norms

58
Exners scoring system

• Location part of the blot
• W, D, d, S, (WS)
• How common is the location (normative comparisons
  from manual)
• Determinant what led to response
• Form, Color, FC or CF, Movement, etc.
• Evaluate form quality (normative decision based
  on manual of responses). Low F psychosis/poor
  reality contact
• Content focus on what specifically
• Human or animal, whole or detail, nature, etc.
• Populars determines normative responding

59
Rorschach Exner

• Exners (1987) scoring system involves an attempt
  to increase validity by objectifying the scoring,
  increasing the number of responses (14), and
  standardizing the administration
• This has resulted in significant improvements in
  the tests reliability and validity
• In a meta-analysis, Hiller et al. (1999) found
  the Rorschach (using Exners scoring) to have
  larger validity coefficients than the MMPI-2 for
  studies using objective criterion variables

60
Other projective tests

• TAT (Thematic apperception test, Murray)
• Stimuli are less ambiguous than the ink blots
• Tell a story, though little standardization re
  which pictures to be used, scoring (typically a
  content analysis), etc.
• Used extensively with less literate pops like
  children (CAT), geriatric pops (GAT), non-English
  speaking individuals, etc.
• Draw-a-figure test (figure drawings)
• Person, family, house, tree, etc. all are
  interpreted as you
• Minimal standardization for scoring
• Sentence completion
• Sentence stems like Mom is, Life, etc.
  largely scored for a thematic standpoint
• Bender-Gestalt (the same test used for
  neuropsychological screens)
• Copying figures and making personality
  interpretations

61
Test or technique?

• Review articles and come up with an opinion. Come
  ready to debate/discuss.
• On Tuesday.

62
Assessment of malingering

• What is malingering? What must it include?
• Intentional? Awareness? Personal gain?
• Very complex phenomenon that may change over time
• e.g., A lie (or lies) that become real/true for
  the individual over time, or a truthful statement
  that becomes a lie.
• Most statements cant be categorized as one or
  the other, and typically involve aspects of both
• Berry et al (1995) suggest that faking good and
  faking bad are distinct constructs (not opposite
  ends of the same continuum)
• Harder to detect specific faking vs. general
  faking
• Content nonresponsivity (CNR) random
  responding, all true or all false
• Content response faking (CRF) fake good or bad
  research suggests that these may be independent
  dimensions (client may fake good on some parts
  and fake bad on others)
• Should always be considered (in some form) when
  there are contingencies for the patient

63
Classifications of Misrepresentation

• Are symptoms under conscious control? Are
  physical/psychological symptoms motivated by
  internal or external gains?
• Factitious Disorders intentional production of
  symptoms (feigning) that are motivated by
  internal gains
• Motivation is to assume the sick role as there
  are no external incentives for the behavior
  (e.g., economic gain, avoiding legal
  responsibility, etc.)
• Somatoform disorder unintentional (i.e.,
  unconscious) production of symptoms for internal
  gains
• Malingering intentional production or
  exaggeration of symptoms (i.e., conscious)
  motivated by external incentives
• Lack of cooperation during the evaluation,
  presence of ASPD, discrepancy between
  self-reported data and objective findings,
  medicolegal context for referral (e.g., attorney,
  police, etc.)
• Note Exaggeration rather than fabrication makes
  differential very difficult

64
Pros and Cons of Malingering Dx

• What are the costs of labeling someone a
  malingerer
• Questions all present and future clinical
  presentations
• What are the limits of our measures to make this
  differential?
• After weighing the strength of any claim of
  malingering (relatively weak given the limits of
  our measures) and the costs of making an
  erroneous judgment, we need to act very carefully
• Use converging, independent evidence to make any
  determinations
• e.g., objective inventories like the MMPI-2,
  strong contextual factors (i.e., to provide the
  motive and baserates), interview, low probability
  baserates for responding (e.g., incorrect on all
  options when this would be well below chance
  responding), and response to the evaluators
  feedback (e.g., Actually, youre doing quite
  well followed by decrements in performance)

65
Mind of a murderer the Bianchi tapes

• Identify the circumstances that could be seen as
  contingencies for malingering (reinforcers for
  malingering)
• Why would that particular malingering behavior be
  manifested?
• How could client have obtained the information
  necessary to provide the malingering profile? Any
  evidence that this information was obtained?
• Any indications of malingering in his
  presentation? (Be objective)
• What are some reasons why he might not be
  malingering?
• Predict response sets in advance of testing (vs.
  scoring in hindsight)
• What pattern of responses do you predict for the
  Rorschach?
• What pattern of responses would you predict for
  the MMPI-2?
• Whats your call?

66
Measures of malingering Berry et al

• The pasta strainer and photo copy machine
  incident
• MMPI-2 F, F-K (note these two indices are not
  independent), VRIN (random), TRIN (all true or
  all false), and Fb
• Also look for discrepancies between some of your
  subtle and obvious supplemental scales (though
  this can also just assess sophistication in
  malingering)
• The D scale has also been used with some success,
  as the items appear to reflect a less
  sophisticated (popular) view of mental illness
• MCMI evaluates random responding, low frequency
  responding, willingness to disclose information,
  debasement (willingness to endorse psychological
  problems), and desirability (unwilling to endorse
  psychological problems). Also as with the D scale
  of the MMPI, the well-being scale can likewise
  assess psychopathology

67
Measures of malingering 2 continued

• CPI (Cough, 1957) intended to assess
  personality in the normal population
• Has 3 validity scales good impression (faking
  good), communality (items with either very high
  or very low endorsement frequency that assesses
  random responding), well-being (assesses fake
  bad)
• Basic personality inventory (BPI Jackson, 1989)
  contains 12 scales each with 20 T/F items.
  Research is limited on its utility for this.
• Deviation scale is comparable to the MMPI-2 F
  scale
• Personality assessment inventory (PAI Morey,
  1991) is a 344 items
• 4 validity scales Inconsistency, infrequency,
  negative impression management and positive
  impression management
• NEO-PI-R (Costa McCrae, 1991) no effective
  validity index, so should not be used in this
  context
• 16 PF also lacks adequate validity measures and
  should not be used

68
Measures to specifically detect malingering

• These measures should be administered when the
  referral question specifically implicates
  malingering and/or when there are substantial
  contingencies to suggest that malingering is
  likely
• Structured Interview of reported symptoms (SIRS)
• Has shown some promise, though it is susceptible
  to acquiescence and false positives (claiming
  malingering when it is not)
• The M test is a 33 item T/F test with three
  scales genuine symptoms of schizophrenia,
  atypical attitudes not characteristic of mental
  illness, and bizarre and unusual symptoms rarely
  found in mental illness
• Showed some ability to differentiate patients
  from directed malingerers and from suspected
  malingerers (Note The problem with using the
  latter criterion group as there is no definitive
  knowledge about those individuals)

69
Measures to specifically detect malinger. - 2

• Test battery approach including WAIS-III and the
  MMPI-2 the more tests administered, the harder
  it is to present a consistent profile
• This approach should use baserates for incorrect
  responses as the primary means of classifying
  (see also TOMM)
• Provide response options (typically no more than
  two) such that a chance correct criterion can be
  calculated (e.g., 50 for a two item version)
  this should be no lower than 30 to avoid floor
  effects
• Track responses over at least 30 trials (the more
  the better as this minimizes chance outcomes).
• Calculate the probabilities for deviations from
  .50 correct and apply it to clients correct
  response rate (i.e., what are the odds that they
  would have missed as many as they did if they
  were truly guessing)
• Evaluate responsiveness to your feedback (e.g.,
  Youre actually not doing that bad vs. Most
  people with your type of injury do better)
• If less sophisticated malingering there will be
  an immediate and relatively large response to
  your comments

70
Who is your client?

• Why is this question important in addressing the
  malingering issue?
• If the suspected malingerer is your client who is
  undergoing therapy with you (or someone else) to
  whom is your obligation and what are the
  costs/benefits of undertaking an evaluation of
  malingering?
• Does it help the therapeutic process? Focus on
  why one might be deceptive to better understand
  clients behavior
• If the client is the court, then to whom is
  your obligation and what are the costs/benefits
  of undertaking an evaluation of malingering?
• Question now is to determine if client is being
  deceptive/evasive.

71
Assessing psychopathic personality

• Psychopathic personality behavior characterized
  by remorseful and callous disregard for others
  and a chronic antisocial lifestyle. Thus, most
  ASPDs are not necessarily psychopathic.
• Drawing data from various sources (at least
  three)
• In person interview
• Testing
• Independent historical information (anything that
  is not self report it is important to note that
  other official records are not necessarily based
  on anything other than self-report)
• Although all three of the above are important in
  order to provide converging evidence, the test
  data will be the strongest tool in court (due to
  its psychometric strengths)

72
Assessment (Meloy Gacono, 1995)

• The Psychopathy checklist revised (Hare, 1991)
  20 item test with a 4-point Likert scale
  response format. Largely intended for males
  (little data on females)
• To be completed by the clinician after a clinical
  interview and review of historical data (includes
  descriptors falling under a single dimension of
  psychopathy) e.g., impulsive, irresponsible,
  shallow emotions, etc.
• Items must be scored in a particular sequence,
  with more structured items first, followed by the
  least structured items (with the former
  contributing to the latter)
• Cutoff score of 30 or greater to define
  psychopathy, with higher scores denoting more
  extreme presentations
• Adequate reliability and validity, though note
  the overlap between some of the validity criteria
  and the info used to determine the score (e.g.,
  extent of criminal record is used for both)

73
Assessment (Meloy Gacono, 1995) p. 2

• The Rorschach should still pursue the minimum
  number of responses (14 or more) as suggested by
  Exner (1986)
• Include an assessment of defenses and object
  relations (both of which appear to have modest
  reliability) that suggest more narcissism
  (self-references), violations of boundaries, etc.
  in the psychopathic personality (specific ratios
  from Exners scoring system are described)
• MMPI-2 primary focus is on scale 4 (also
  content subscales drawn from 4 be cautious with
  the latter)
• If administering scale 4 alone, note that you
  will not have the benefit of the k correction.
  Thus, scores will be suppressed.
• L and F will also predict psychopathy (tendency
  to be untruthful)
• Cognitive abilities (e.g., WAIS-III) are
  unrelated to the presence of psychopathy, but may
  be informative as to the nature of the
  presentation (e.g., level of sophistication,
  concordance with traditional/normative concepts
  of intelligence, etc.)

74
Integrity testing

• Evaluating integrity as a trait, whereas such
  behavior may be situation specific (e.g., someone
  who would not lie in interpersonal settings might
  not hesitate to cheat on their taxes).
• Characterological view of integrity downplays
  situational factors
• Integrity is a very broad concept that can
  include diverse responses (e.g., passive vs.
  active lying, cheating vs. theft, etc.)
• Early paper and pencil tests were validated with
  the polygraph
• Employed in low end entry jobs when people have
  to interact with money (retail, financial
  services, etc.)
• Today, such tests attempt to predict a wide range
  of behaviors including violations of work rules,
  fraud, absenteeism, etc.

75
Integrity testing p. 2

• Overt integrity tests evaluate beliefs about
  the incidence of theft and other
  counterproductive behaviors, punitive attitudes
  towards theft, endorsement of common
  rationalizations for theft, and direct questions
  about ones own involvement in such activities.
• Personality oriented measures much broader than
  integrity tests and tend to have lower face
  validity (e.g., high conscientiousness on the
  NEO)
• Clinical measures like the MMPI validity scales
• All are difficult to validate because the
  behavior we are trying to predict goes largely
  undetected. So if a test score does not predict
  it could just mean that this is a false positive
  or someone who was not caught

76
The polygraph test

• Measures physiological arousal that is presumed
  to be associated with lying. e.g., perspiration
  as indicated by galvanic skin response, brain
  activity suggesting arousal, etc. to the question
  (not answer)
• Is this assumption reasonable?
• Confounds?
• Under what circumstances can lying not be
  associated with arousal?
• Habituation effect from repeated lying?
• Lack of awareness of the lying? (issue of
  conscious vs. unconscious)
• What is the best way to quantify arousal? Should
  we evaluate this normatively or ipsatively?
• Control Question Test (CQT) compares relevant
  questions to control questions which are intended
  to elicit a strong physiological response from
  innocent subjects (e.g., Prior to 1993, did you
  ever do anything that was illegal or dishonest?)
• While innocent people know they didnt commit the
  crime, they are either uncertain or lying about
  the CQ. Guilty persons should not respond as much
  to the CQ

77
The polygraph test p. 2

• Criticisms of the CQT
• Difficult to develop good control questions that
  will produce similar responses relative to
  relevant questions for innocent people. This
  results in many false positives (Note Bias for
  positive outcome is why most of these tests have
  artificially high success rates in forensic
  settings most are guilty)
• CQ are designed for each individual, so
  standardization is compromised
• Direct Lie Control Test (DLCT) if person
  answers truthfully to a question they are asked
  the question again and told to lie about it when
  asked again (a known lie for comparison)
• Can be standardized and the power of the DLCT is
  from the instruction (which is standardized) not
  the content of the question
• Can reduce the rate of false positives and
  generally does better than the CQT
• Initially employed absolute standards for arousal
  lying and this was not at all effective

78
The polygraph test p. 3

• The guilty knowledge test (GKT) not designed to
  detect deception, rather it tries to
  differentiate between those who have knowledge
  about a particular event (crime) and those who do
  not (the innocent)
• The concealed information test (CIT) is similar
  to the above approach and likewise tries to
  assess familiarity with specific information as
  opposed to lying
• Both of these approaches have the advantage of
  asking the exact same questions of all
  individuals and comparing responses both within
  and between subjects
• Minimal data on these approaches, as the bulk of
  the research is on the CQT

79
Does it work?

• Honts (1994) reviewed the literature on the
  effectiveness of the polygraph and found that it
  does about as well as chance in experimental
  settings. Most of the reviewed research uses the
  DLCT
• In real life and experimental settings, the
  majority of errors are false negatives (saying
  someone is innocent when they are guilty)
• Most deceptive individuals (up to 95) are
  misclassified
• Because the cost of a false positive (saying
  someone is guilty when really they are innocent)
  is deemed to be higher in our legal system.
  Therefore, the cutoff scores (criteria) have been
  altered so as to make false negatives more likely
• Why does it fail?
• If high arousal to control questions, then more
  difficult to discriminate
• Idiosyncratic responses to lying

80
Admissibility of the polygraph (Saxe
Ben-Shakhar, 1999)

• Courts have almost universally rejected the
  polygraph, though this question has been and
  continues to be litigated extensively
• Courts are increasingly being made responsible
  for evaluating the merits of test data, despite
  lacking the expertise to do so.
• Note The literature has become increasingly
  discrepant in its view on the polygraph
  (disagreement on its validity even in the
  scientific community)
• What criteria should be used to evaluate this
  information and what should we tell the courts?
• History
• Marston (1917) used a blood pressure cuff to
  determine truthfulness (arousal) in a defendant
  (Frye), based on the assumption that while truth
  required little or no energy, lies do rejected
  by the courts

81
History of the Polygraph

• Note the courts use of the term experimental as
  not well established evidence
• The Frye ruling adequately reflects the courts
  treatment of the polygraph even today, though now
  based on the Federal Rules of Evidence (FRE)
  which require that the evidence (polygraph or
  otherwise) be relevant and that it aid the jury
  (i.e., be valid).
• Daubert (1993) was based on the FRE and
  highlights 4 considerations when ruling on
  evidence
• Testability or falsifiability (see Popper and the
  method of science)
• Error rate
• Peer review and publication
• General acceptance
• This basically requires juries judges to
  evaluate scientific issues

82
History of the Polygraph p. 2

• In trials like Daubert, scientists with opposing
  views on the polygraph present their views and
  the jury must decide on the merits of their
  arguments
• Generally there has been no legal distinction
  between the concepts of reliability and validity
  (you can see where this is go, since, from a
  scientific standpoint, reliability limits
  validity)
• An additional problem with these concepts is that
  the data is collected as a series of discrepancy
  scores and these are then summed to reflect a
  qualitative assessment of truthful, deceptive,
  and inconclusive. Thus, very different
  discrepancy readings might still result in
  similar qualitative assessments.
• Two accepted approaches for reliability are
• Test the same person twice on the same issue
  using the same polygraph technique with 2
  different testers
• Test the person once, but have the chart scored
  by two different people

83
History of the Polygraph p. 3

• The latter approach deals on with the error
  involved in chart scoring and ignores (or
  equates) administration error
• The real issue is whether the procedure as a
  whole is reliable (e.g., the creation and
  administration of control questions), thereby
  getting at internal reliability (do different
  parts of the test agree), test retest reliability
  (different administrations of the test agree),
  inter-rater reliability (different test
  administrators agree as to the outcome)
• Note There are practical limitations to how
  often the same test could be given to the same
  individual
• What little data exists on reliability focuses
  only on the between examiners approach
  (inter-rater reliability), though this
  reliability is reasonable (not high). Thus, this
  remains an unevaluated component of the polygraph
  (major limitation)

84
History of the Polygraph p. 4

• Because the courts do not distinguish between
  reliability and validity, the minimal reliability
  that does exist carries far more weight than it
  should.
• Modern views of validity highlight the
  integrative component of validity (recall
  Messick, 1995), though to evaluate it, it is
  necessary to consider different aspects
  separately
• Different types of validity are more relevant
  depending on the question at hand
• e.g., predictive validity for integrity testing
  in job placement/hiring, vs. criterion validity
  being more relevant for determining truth/lying
• Construct validity gets at the theoretical issue
  of what is a lie. Is it a situational phenomenon
  or a trait? Can it be represented by
  physiological responding? Etc.
• No theory to explain why a stronger response
  should occur for lies vs. truth

85
History of the Polygraph p. 5

• Similar physiological responses to lying appear
  to occur for experiences such as surprise/novelty
• Note For the CQT, questions about the crime are
  expected to be well rehearsed for the criminal
• Thus, they have questionable construct validity
  (not necessarily measuring what they propose to
  measure)
• Under-represents the construct of interest and
  over-represents irrelevant constructs (surprise,
  stress, etc.)
• What criterion can be used?
• Outcome of a trial? If the case is dismissed?
• Do either of these assure that we know the
  clients status re lying?
• Note also that a true evaluation of the polygraph
  would mean that the examiner only has access to
  the polygraph data (that s never the case).

86
History of the Polygraph p. 6

• The criterion and predictor are rarely
  independent.
• e.g., if the polygraph is used to get a
  confession and the confession helps get a
  conviction, then by definition, the polygraph is
  part of the criterion (polygraphs are frequently
  used to get confessions)
• Experimental criteria for the polygraph generally
  lack external validity (is lying in an experiment
  to lying in a crime involving yourself? That
  is, are all types of deception equal?), while
  real life evaluations of the polygr