Intellectual Assessment and Interpretation in Death Penalty Appeals

1
Intellectual Assessment and Interpretation in
Death Penalty Appeals

• Frank M. Gresham
• Louisiana State University

Symposium Controversies in Determination of
Mental Retardation in Death-Penalty
Appeals August 19, 2007 American Psychological
Association, San Francisco
2
Walker v. Commonwealth of VirginiaVirginias
Definition of Mental Retardation

• A disability originating before age of 18 years
  characterized by
• Significantly subaverage intellectual functioning
  as demonstrated by performance on a standardized
  measure of intellectual functioning that is at
  least 2 standard deviations below the mean
• Significant limitations in adaptive behavior as
  expressed in conceptual, social, and practical
  adaptive skills
• Assessment of intellectual functioning shall
  include administration of at least one
  standardized measure generally accepted by the
  field of psychological testing
• With respect to intellectual functioning, the
  Virginia statute provides that the Commissioner
  shall maintain an exclusive list of standardized
  measures of intellectual functioning generally
  accepted by the field of psychological testing

Tests move on and off this list quite frequently
3
RulingU.S. District Court Eastern District of
Virginia

• Petitioner has failed to show by preponderance of
  evidence that he was mentally retarded before age
  of 18
• Petitioner has shown that he sufferers from below
  average mental intelligence
• Petitioner has failed to show by preponderance of
  evidence that the scored 2 standard deviations
  below the mean on an approved measure of
  intellectual functioning
• Only scores of 70 or less are considered 2
  standard deviations below the mean
• Petitioner argues that his WISC-R score of 76 and
  WAIS-III score of 80 are both 2 standard
  deviations below the mean in light of the
  standard error of measurement, practice effect,
  and Flynn Effect.
• Court considered these explanations speculative

4
Interpretive Issues in Intellectual Assessment

• Nature of Intellectual Functioning
• Flynn Effect
• Practice Effects
• Measurement Error
• School Diagnoses

5
Nature of Intelligence

• Fluid versus Crystallized Intelligence Liquid or
  Mineral?
• Walkers IQ Data
• Crystallized Intelligence
• Woodcock-Johnson-III (33 years) Verbal78
• WISC-R (11-6 years) VIQ 70
• WISC-R (14-9 years) VIQ 75
• WAIS-III (26 years) VIQ 87
• WAIS-III (27 years) VIQ 80
• Mdn 78
• Fluid Intelligence (33 years)
• CTONI-68
• Raven- 63
• GAMA- 61
• Mdn 63
• Scoring errors admitted in Court

6
Flynn Effect

• Well-established finding of 0.3 point increase in
  mean IQ per year
• Findings hold for U.S. and worldwide
• Norms become less accurate over time
• WAIS-III particularly susceptible to increase
  (normative problems)
• Flynns Deposition in Walker case
• WAIS-R (1978) WAIS-III (1995) 5.24 points
• WISC-III (1989) WAIS-III (1995) 1.64 points
• WAIS-III (1989) SB-IV (2001) 3.16 points
• WAIS-III (1989) WISC-IV (2001) 0.76 points
• WAIS-III inflates scores compared to
  contemporarily normed tests
• WAIS-III score of 79 (obtained in 2000) should
  be interpreted as a score of 75 (7 years X
  0.32.102.34 WAIS-III adjustment.

Dates reflect when test was normed, not when it
was released
7
Flynn Effect MMR DiagnosesKanaya, Scullin,
Ceci (2003)(Based on almost 9,000 referrals to
SPED)

• Size of Flynn Effect in MMR Borderline Range of
  IQ
• WISC-R-WISC-R 79.0-80.2
• WISC-R-WISC-III 78.4-73.9
• WISC-III-WISC-III 78.5-78.1
• WISC-III-WISC-IV 78.5-74.5
• Changes in MMR Classification Because of Flynn
  Effect
• Change from WISC-R to WISC-III almost DOUBLED
  rate of MMR classification (19 to 34)
• Impact of Flynn Effect on Borderline IQ Ranges
• IQ 71-75 on WISC-R
• 3-fold increase in IQs 66-70 on when retested
  with WISC-III (14 to 40)

Not part of the Kanaya et al. (2003) study
8
Practice Effects IQ Test Scores

• Practice effects occur on IQ tests that are
  repeatedly administered (5-8 points)
• Shorter the retest interval, the greater the
  practice effects
• Low initial scores on a test tend to increase on
  retest
• Differences difficult to interpret when initial
  test retest measures are different
• Differences in results may depend on item content
  in the test
• Differences between repeated testing results may
  be due to differences in motivation, interest, or
  the stakes of testing
• Petitioner has failed to present evidence that
  an adjustment in test scores due to practice
  effects would be anything other than
  speculation.
• Experts have not made courts understand that
  practice effects are real, not speculative,
  particularly in death penalty cases where
  defendants are repeatedly tested

9
Role of Measurement Error

• What is the most appropriate estimate of error in
  death penalty cases?
• Internal consistency estimates?
• Stability estimates?
• Both?
• Consistency estimates typically yield less
  measurement error
• Consistency How reliable is this individuals
  score on an IQ test on a given day?
• Stability Will this individual will obtain a
  similar score if retested with the same test in
  the future?
• Fourth Circuit refused to use the standard error
  of measurement to lower IQ scores in Atkins cases
  due to the inherent speculation of using the
  standard error of measurement to lower an IQ
  score when it could just as likely be used to
  raise an IQ score.
• Experts have not made courts understand band of
  error concept probability of true score

States expert refers consistently in his report
to the standard error of measure (sic)
10
School Diagnoses

• Walker classified as learning disabled and
  emotionally disturbed by schools
• Walker never received a MR label from schools
• Court took this a evidence that Walker never has
  been mentally retarded
• What does the research say about schools
  assigning the MMR label?

11
IQlt75MacMillan, Gresham, Siperstein, Bocian
(1996)
MR 14 M63
SLD 44 M68
The labyrinth of IDEA School decisions for
referred students with subaverage general
intelligence. American Journal on Mental
Retardation, 101, 161-174.
12
IQlt70Kanaya et al. (2003)
SLD 48.1
MR 48.5
M66
M64
13
Whats the Take Away Message?

• Courts do not understand the nature of
  intelligence and how an individual might obtain
  different scores on different tests at different
  times
• Psychometric True Score (IQ) vs. Absolute True
  Score (DNA)
• Courts consider the Flynn Effect to be a theory
  rather than fact
• Courts consider practice effects to be
  speculative rather than fact
• Courts do not understand concept of measurement
  error regression artifacts
• Courts often take failure of schools to diagnose
  defendants as MR to indicate absence of MR
• COURTS JUST DONT UNDERSTAND THE CONCEPT OF MMR
  AND THE CLOAK OF COMPETENCE
• Psychological evidence in these cases may not be
  very convincing