Topic 5: Profiling and Classification Flashcards
Why is classification important?
important for theoretical development: classifying offenders into meaningful groups assists in the identification of between-group differences
guides assessment practices: knowing something about a given offender type informs the psychologist about relevant areas of inquiry
guides treatment practices: depending on the basis of grouping, or organizing principle, there may be common treatment needs
Why would classification be problematic?
important individual differences may be obscured
knowledge base concerning group or type may be irrelevant to the patient in question
may preempt more meaningful diagnostic exploration
categorization (“labeling”) is often seen as pejorative: often used in legal defenses
What are the defenses to the idea that classification is problematic?
humans have a strong tendency to categorize and will do so anyway
better to acknowledge this and proceed systematically, than to ignore it and do so haphazardly
recognizing commonalities among group members does not equal ignoring individual differences
What are weak classification systems?
are devoid of any organizing principle
e.g., one group is defined on the basis of behavior, and another by some other construct such as mood or intelligence
individuals may belong to multiple groups
low psychometric reliability, therefore poor chances of survival
DSM-5-TR is arguably an example of this
What are some specific issues or decisions where classification is relevant?
level of risk
institutional housing (roommate) decisions
degree of supervision required versus privileges
offender needs
treatment inclusion: D/A, anger management, medication, etc.
Are between-group differences theoretically informative?
developmental trajectories/risk factors
circumstantial risk management post-release
What is the Chaiken & Chaiken (1984) classification system?
8 “dimensions” based on past crimes
yes/no: assault, robbery, burglary, drug dealing, theft, car theft, fraud, credit card fraud
yields 256 possible combinations: but 10 of these accounted for 59% of inmates
what’s wrong with this: there is no rational basis, some of the crimes are pretty close together (theft or car theft)
What characteristics describe a “good” classification system?
comprehensive: covers every offender
unambiguous: categories are clearly defined
statistically reliable assignment procedures
type distinctions are valid
sensitive to individual changes
treatment relevant
economical to apply
What is the historical perspective of classification system?
systems based on impressions (clinical or administrative) have borne little fruit
systems based on a cohesive personality/criminogenic theory fair a little better
most attention is currently focused on empirically derived systems; these are the most testable and tend to make the most accurate predictions
What are impressionistic systems?
an individual, or group of individuals, think they recognize a pattern or type among some members of a clinical/institutional sample
no theory or systematic method of measurement guides that impression
the essential features are identified after the fact
lots of problems arise concerning diagnostic reliability: prognostic validity tends to be very poor
How is the DSM-5-TR an example of an impressionistic system?
although the diagnostic validity of the specific disorders has come a very long way since the original DSM, it is fraught with problems: personality disorders with “cluster B features”, many NOS categories, systematic disagreements, easy for lawyers to tear up during cross examination
like it or not, there are a number of personality disorders that go hand-in-hand with increased risk of criminal activity: most obvious is APD, BPD isn’t far behind in the B cluster, definition is highly behavioral, therefore predictions are circular
assumes discontinuity among the personality disorders: discrete taxons, these may nevertheless enjoy some empirical support
What are theoretically derived classification systems?
typically stem from a developmental or personality theory
testing the utility of the related classification scheme really puts the whole theory on the line
relates specific attributes or points of maturation to antisocial behavior
may appeal to empathic capacity, social reasoning, intelligence, etc.
measurement instruments may draw heavily from concepts central to that theory: can introduce significant bias
What is the Hunt & Hardt (1965) classification system called the conceptual level?
based on ability to adopt other perspectives
attempts to tap reasoning process more so than outcome
assessed via an incomplete sentences task
both of these systems represent efforts to match the treatment to the offenders’ need/level (responsivity principle)
e.g., lower conceptual level = more behaviorally-oriented interventions
higher conceptual level kids paired with staff more sensitive to manipulation attempts
showed favorable outcomes over general assignment
What are empirically derived classifications?
mainly based in two statistical techniques: cluster analysis and factor analysis
What is cluster analysis?
referred to as numerical taxonomy
groups subjects according to level of similarity as measured by variables selected by the researcher
What is factor analysis?
essentially correlational
good for reducing a large number of variables to a smaller number
may help identify “underlying dimensions”
What is the Megargee & Bohn (1979) MMPI based scheme that uses cluster analysis?
Minnesota Multiphasic Personality Inventory
most commonly used clinical personality measure
has 10 basic clinical scales and 3 basic validity scales
clinical interpretation is based on individual scale configurations as well as profile analysis
Megargee’s cluster analysis of over 1200 convict profiles yielded 10 types
included 116 other variables to assist in profile description: demographic, educational, familial, developmental, criminal, prison behavior, clinical (from interview)
computer software was able to classify almost all profiles (96%); results have been replicated a number of times
predictive accuracy is in question; problem is compounded by relatively poor correspondence between MMPI and MMPI-2 profiles (about 60%)
other problems: high scale overlap, are 10 groups really needed? a smaller number might be just as discriminating
the profile deviation was atheoretical (as is the MMPI-2); therefore the results may be unique to the instrument; they are not theoretically informative on their own
How is the Hewitt & Jenkins (1946) classification system an example of factor analysis?
behavioral data from 500 delinquent children (many variables)
settled on a 4-factor solution:
unsocialized aggressive behavior (UA): poor prognosis
socialized delinquency (SD): better prognosis
over-inhibited (OI): more likely to be mentally ill
physical deficiency: less interpretable
What is profiling?
actually a collection of related techniques: psychological profiling, criminal personality profiling, criminal behavior profiling, investigative profiling, criminal investigative analysis
goal is to describe perpetrator(s) in order to narrow the investigative field
What are the sources of information in profiling?
autopsy results
victim verbal reports
examination of crime scene(s)/photos
in the case of psychological profiling only, there are test and interview data
What is criminal profiling?
not even close to being the advanced, precise science shown on TV: currently, relies heavily on intuition
idea is probably very old, but first formal attempts were made by US intelligence during WWII: wanted to establish a means of predicting the military behavior of known enemy officers
resurrected by the FBI in the 1970’s
best use is to direct attention of investigators at most likely suspects
in a sense, the process is akin to interpreting a personality profile: e.g., previous perpetrators that did ____ had ____ in common
as databases increase in content, the process continues to refine and improve
mainly practiced and used by large law enforcement agencies at present
seldom points to a specific individual
What are the common general assumptions in criminal profiling?
“the murderer was probably a white male aged 35-45…”
multiple stab wounds suggest extreme anger
facial lacerations are deliberately inflicted by individuals that feel physically unattractive and have unsuccessful dating histories
age, gender, race, level of education, marital status, and degree of future risk are typical assertions made by profilers
What are computer models of criminal profiling?
computer models are in development, and are beginning to demonstrate practical utility
this is an area where considerable advancement is likely
based on geographic location rather than offender’s identity
What areas has profiling been applied to?
murderers
rapists
internet sex offenders
computer network intruders
thieves
arsonists
corporate and security fraudsters
various forms of terrorists
Overall, how good is criminal profiling?
70% of police psychologists express discomfort with the idea
more an art than a science
Holmes (1989) < 50% of profiles cases are solved: may partly reflect a tendency to reserve profiling resources to most challenging cases
What was the Pinizotto & Finkel (1990) study that compared the accuracy of profilers?
compared the accuracy of: trained FBI profilers, police detectives, non-forensic psychologists (not trained in profiling), university undergraduates (not trained as profilers)
results were unimpressive: trained profilers did not slightly better with sex offenders, but no better with violent offenders
What is reconstructive psychological evaluation (RPE) - psychological autopsy?
post mortem descriptions based on a known offender
intended to add to profiling database
diverse backgrounds make identification of common features difficult
attempt to address dynamic, not just static features
What is psychological profiling?
begins with a known perpetrator
correctly though of as an assessment: probably done as part of a court-ordered report
findings are added to a database
test selection is critical
Murphy & Peters (1992) MMPI results failed to discriminate between sexual offenders and non sex offenders; PPG test did a far better job
moral: one relevant test is of greater value than many irrelevant tests
What are some examples of geographic profiling software?
Rossmo (1995) “GeoProfile” software: based on analysis and prediction of movement patterns, highly statistical
Canter et al. (2000) “DragNet” software
CrimStat
use Distance Decay Functions based on Journey to Crime Model
What is the distance decay function?
will choose a location that is close to home but not too close
What is geographic profiling software?
this approach to spatial behavior analysis yields a jeopardy surface superimposed on a topographic map: prediction are similar in accuracy
probability is usually based on minimizing the mean distance from all known crime sites attributed to that offender
probability of offender residence is usually color coded and/or shown as an elevation in the surface
map is rasterized to calculate total distance from each known crime site to every possible point on the map; Rigel uses a grid of about 562,500 cells, Dragnet uses about 13,300
the actual formula for doing this incorporates the distance decay function and is algebraically sophisticated
What are the two minimal criteria for profilers suggested by Rossmo (200)?
a minimum of three years experience investigating interpersonal crime
“superior” investigative skill
What is the circle hypothesis?
a serial offender’s home location is most likely to be located within a circle defined (diametrically) by the two most distant crime scenes
What is the Snook et al. (2000) study?
2 groups of undergrads, naive to profiling: instructed (experimental) and non-instructed (control)
DragNet printouts
What were the findings of the Snook et al. (2000) study?
DragNet did significantly better than either group in phase 1
experimental group did significantly better that controls in phase 2
DragNet and experimental groups did not differ in accuracy in phase 2
What were the conclusions of the Snook et al. (2000) study?
neither extensive experience or a computer appear to be necessary
biggest factor is knowledge of two heuristics
subsequently replicated in a much larger study
What is psychological crime scene investigation?
three classes of crime scene features: Modus Operandi (MO), signature (calling card, personation), staging
crime scenes are classified as organized, mixed, or disorganized by investigators
What is Modus Operandi (MO)?
the way the violent act was committed; operational characteristics
time, place, weapon, victim characteristics, type of crime, location, method of entry
tends to be refined over time and with practice; becomes stable eventually: conversely can de-evolve into sloppy, less patterned behavior
What is signature (calling card, personation)?
features extra to the crime itself
may become highly ritualized
pattern of mutilation or property destruction, taking “trophies”, leaving something behind, notes/writing, positioning of body, burning body or building to obscure evidence
may provide more information than an unstable MO
What is staging?
alteration made to the crime scene after the crime was committed: may not have been done by assailant, perhaps to protect dignity of victim
e.g., trying to make a suicide look like a murder
offender may attempt to psychologically undo, or reverse that act
What are organized crime scenes?
evidence of planning: reduces chances of detection, allows greater specification of victim, may become quite cocky over time
abduction, murder, disposal usually not in same locations
What are disorganized crime scenes?
reflects the work of more impulsive, perhaps less intelligent criminals
crimes committed in angry, or highly aroused state
no movement of evidence (including body) from crime scene
victims chosen more randomly
What are mixed crime scenes?
features of both
usually means a poorly planned crime fell apart
What features are considered in psychological crime scene investigation?
point of contact
primary scene
secondary screen: intermediate scene, dumpsite/disposal scene
What are the standard questions in psychological crime scene investigation identified by Turvey (2002)?
who normally goes there?
how did they/could they have arrived?
what is there, and what is absent that should be there?
what activities normally take place there?
what other criminal features are intrinsic to that area?
why was that place chosen? did it require prior familiarity?
What are the method of approach identified by Ressler et al. (1986)?
refers primarily to rapists
may reveal information about perpetrator’s relationship to victim
blitz, surprise, con
What is the blitz method of approach?
limits opportunity to work out a defense
What is the surprise method of approach?
laying in wait
knowledge of schedule
may turn 2nd degree into 1st degree murder
What is the con method of approach?
employs a ruse
may involve repeated contacts over time
What are hunting strategies identified by Rossmo (1997, 2000)?
hunter: goes out and gets a victim fairly close to home, e.g., in the same city
poacher: goes into another town or city, doesn’t operate from home
troller: crimes of opportunity; victims are encountered while other, usually routine, tasks are being carried out
trapper: lure/trick victim to a location where they have control (e.g., home)
raptor: attack is immediate upon location of victim, characteristic of most serial predators
stalker: watches, follows, waits to catch victim in a vulnerable, off-guard state
ambusher: always works close to home or another prepared and familiar location
What are motives?
can be difficult to determine
revenge: usually a prior association between victim and criminal
financial gain: may or may not have prior association, robbing on the street vs. fraud
jealousy: romantic, business, academic
power: sexual offenders, lose power in life, want to restore it, not necessarily pain or injury
non-sadistic sexual gratification: wants sex, but not pain and injury is goal
insanity
self-defense: can sometimes go too far, never legitimate to use gun to defend themself
