Week 2 Flashcards
methods of self-report test construction: Rational
very straight forward; items are written to capture understanding of what a trait is
characteristics:
tend to be face valid
susceptible to response biases (easily faked)
may be internally consistent or valid
Methods of self-report test construction: Factorial
generate a bunch of items and use factor analysis to cluster each item and see which go together; items are selected on the basis of factor analysis
Characteristics:
highly internally consistent
tend to be face valid
somewhat susceptible to response bia
Methods of self-report test construction: Empirical
generate a bunch of items and give to 2 groups, one group that is “fine” and one group with something you are looking at (ex depression –> so a group of folks with depression) and see which items differentiate the groups; harder to fake; items are selected on their ability to empirically distinguish one group from another
Characteristics:
often have low internal consistency
often items are not face valid
may be less susceptible to response bias
Thinking errors
fundamental attribution bias
confirmation bias
distinction bias
just world phenomenon
Clinical judgement; Dawes Faust and Meehl
using statistical methods to derive algorithms (i.e. MMPI)
some studies use artificial tasks that failed to tap their expertise
humans do have superior observational skills but better to enter them into decision-making algorithm
Clinical Judgement; Paul Eckman
facial recognition/facial expression; no difference between different groups of professionals and the lay public in effectively knowing when someone is lying
EXCEPT: secret service
nature of some errors in clinical judgement (1998 APA reports)
- may only elicit info that confirms hypotheses and ignores questions/info that would disprove it
- may compare patient to prototype rather than systematically evaluating on specific criteria
- may be overconfident rather than appropriately tentative
- Hindsight bias; wrongly assume we could have predicted the results AFTER being told the results
- may not consider relative frequency of the event they are predicting (rare events are harder to predict than common ones)
Clinical decision making
clinical intuition is very fallible, but we tend to ignore this fact
actuarial algorithms are better than clinical judgement
6 factor model
- extraversion
- conscientiousness
- openness to experience
- emotionality (rotated version of neuroticism)
- agreeableness
- honesty-humility (morality and prosociality)
Unidimensional
widely used for quick assessment of a specific issue
multidimensional
personality test, batteries that contain multiple scales, often include validity scales
unidimensional examples
BDI
DES-II
Y-BOC
MAST
problems with self report data
impulsivity
maturity
based on one’s theory of self (could be limiting but tells us how the person views themself)
not responsive to behavior change that can be tracked
types of psychometric data
observational data (teachers/parents)
life data (context)
self-report data
performance-based data (some folks dont respond honestly but performance data helps tease that out)
informants data (collateral)
s data
self report data
o data
other report data
unidimensional report measures from collateral
Conners-3
BDI
Multidimensional report measures from collateral
Child behavior checklist (CBCL)
BASC
Shedler Weston Assessment Procedure
SWAP; clinical measure done by therapist about client; collateral source
History of MMPI
developed by Stark Hathaway and CS McKinley
1930-40s
First version of MMPI
published in 1943
scales 5 and 0 were added later
Clinical scales of MMPI
- hypochondriasis
- depression
- hysteria
- psychopathic deviate
- masculinity-femininity
- paranoia
- psychasthenia (anxiety)
- schizophrenia
- hypomania
- social introversion
which scales were added later on
masculinity-femininity
social introversion
Uniform T scores for MMPI
mean: 50 SD: 10
uniform distribution forced on the clinical scales (except 5 and 0) and content scales
significant T scores
38 = 8 percentile
50 = 55 percentile
65 = 92 percentile
General interpretive guidelines for T score
65 and above = elevated
60-65 = interpretable on validity and content scales
don’t pay as much attention to low scales but consider scales!
Administration info for MMPI
MMPI = 18 years and older
MMPI-A = for adolescents 14-18
requires an 8th grade reading level
if IQ is 80-85 > may not be appropriate test
test format
soft cover booklet
hardcover booklet (institutional setting)
audio recording
Spanish version
computerized administration (minor differences)
MMPI-2-RF
shorter and viable alternative
MMPI-3
update of the MMPI2-RF with new items and scales
MMPI shortened
test can be scored if the first 370 items are answered; if someone is having a problem during it they can end at those; counts but not best practice
ethical considerations for protecting test data
ethical responsibility to protect the integrity of the test
unethical to let someone take the MMPI or other psychological tests home with them
Administration of MMPI
- establish rapport
- let them know why they are taking it, who will get results and that they will get feedback
- check reading levels by having them read the first aloud
- advise them to answer each question as they are now (not how they were in the past)
- when done look over to make sure they answered all items
scoring methods for MMPI
handscore (inexpensive portable 20-30 minutes)
computer scoring (quicker, fewer errors, generate more scales
validity scales check for..
“test taking attitude”
distorted responding
different response styles
malingering
social desirability
claiming excessive virtue
acquiescence
non-acquiescence
extreme responding
validity scales
?
L
F
K
Fb
VRIN
TRIN
S
? scale
“cannot say”, sum of omitted and double scored items
10 items omitted interpret with caution
30 items omitted = invalid test
high ? score
suggests carelessness
uncooperative
poor reading skills
ignored specific content area, indecisive,
lack of experience,
severe disturbance
L scale
“lie scale”
only rationally developed scale > collection of unlikely virtue
detects naive deliberate unsophisticated attempts to be favorable
16 questions
high L score
claiming excessive virtue
suggests that other scores show better picture than reality
“want to look good”
“things are better than they are”
F scale
“infrequency scale”
very heterogeneous scale
detects deviant/atypical ways of responding
rare answers “not like most people”
elevations are confusing and confound with psychic disturbance and distress
elevated F scale
reading difficulties
random responding due to confusing or inattention
uncooperative
malingering or conscious exaggeration
cry for help
demands for attention
true disturbance accurately reported
VRIN scale
use VRIN to understand elevation in F scale
pairs of items that should be answered similarly
raw score >13 = inconsistent responses and probable invalidity
High F and High VRIN
reading difficulties
random responding due to confusion or inattentive
willful random responding (uncooperative)
High F and Normal/Low VRIN
malingering or exaggeration
cry for help or narcissistic demand for attention
true disturbance accurately reported
Low F and Low VRIN
obsessive, indecisive or perfectionistic
especially if scale 7 is elevated
K scale
social desirability
subtle index of attempts to deny and down play socially undesirable traits
Fb Scale
“back page infrequency scale”
similar to F scale but covering last part of the test (after 370)
T > 80 = caution with interpreting items after 370
T >120 = items after 370 invalid
if F = valid but FB = invalid the person likely quit paying attention
TRIN scale
True response inconsistency scale
detects tendency to answer true and tendency to answer false
high raw score = tendency to answer “true”; >13=invalid
low raw score = tendency to answer “false” ; <5 suggests invalid
not super useful
S scale
superlative scale
detects presentation as highly virtuous, responsible, psychologically healthy
highly correlated with K
Fp scale
infrequency psychopathology scale
useful in identifying faking bad
useful in forensic cases
Mp scale
positive malingering scale
useful in detecting attempts to present in a favorable light
FBS scale
fake bad scale
controversial
T score = 80-99 > possible inconsistent responding > check VRIN
if consistent responding = high FBS could reflect:
credible medical problems or symptom exaggeration
K correction
Paul Meehl
way to adjust for excessive effort to downplay problems
accomplished by adding a % of K scale raw to scores
