Module 11 - Risk assessment

Question 1

Risk assessment: What is it?

Answer 1

• Pre 1990s = dichotomy
• Now dimension of probability that changes over time and interaction of characteristics
• Risk management = prediction + management

Question 2

Civil setting

Answer 2

• Civil commitment: involuntary hospitalization, Mental Health Act
• Child protection
• Immigration
• School + labor regulations
• Duty to warn

Question 3

Criminal setting

Answer 3

• Disclosure of risk with considerations for client-solicitor privilege
• Long-term offenders
• Early release (parole eligibility)
• Capital mitigation
• Correctional placement
• Sentencing

Question 4

Types of prediction

Answer 4

• True + = correct and risk
• True - = correct and non-risk
• False + = incorrect and risk
• False - = incorrect and non-risk

Question 5

Limitations of threat assessment

Answer 5

• Base rate problem
• Evaluative (methodological) issues
• Judgement and error biases
• Lack of theory
• Sampling (white males)

Question 6

Base rate problem

Answer 6

• % of population that commits crime
• Either too high or too low (too low = false +)
• Variables: group/population of interest, predictions, follow-up period
• Baxstrom and Dixon study = low BR and high rate of false positives

Question 7

Methodological issues

Answer 7

• Ideal experiment not feasible due to ethical considerations
• Weaknesses: limited risk factors studied, measuring criterion variable (use of records but unreported crime = underestimate), defining criterion variable (requires severity, type, target, location, motivations)

Question 8

Judgement error

Answer 8

• Heuristics
• Illusory correlation
• Ignoring BR + relying on more salient/unique cues
• Overconfidence

Question 9

Unstructured clinical judgement

Answer 9

• Professional discretion with no guidelines
• Informal, subjective, impressionistic
• Historically more common
• Cognitive biases + adversarial allegiance
• Predictability at 33%

Question 10

Actuarial/mechanical prediction

Answer 10

• Defined risks based on empirical research
• Doesn’t allow for individualized risk appraisal
• No consideration of situational factors
• Predicts recidivism + measured on continuum
• PCL-R, LS-CMI, V-RAG, Static-99R

Question 11

Structured professional judgement

Answer 11

• Predetermined list of risk factors based on presence and severity
• HCR-20: historical (violence, relationships, employment), clinical (symptoms of major mental disorder, instability, violent ideation), and future risk factors (living situation + stability, stress/coping mechanisms, professional services + plans)
  o Best prediction with women

Question 12

Tests for risk assessment

Answer 12

• PCL-R = psychopathy checklist revised, 20 items rated 0/1/2 (e.g., grandiose sense of self, shallow affect, lack of empathy, etc.)
• LS/CMI = level of service/case management inventory, part of RNR (most widely used), based on criminogenic risks/needs and responsivity
• V-RAG = violent recidivism assessment guide, mostly static factors, high accuracy rate (75%), works across genders + cultures, 12 items
• Static-99R = risk of future dangerousness for male adult sex offender, only static risk factors, good predictor (41%)

Question 13

Ideal assessment approach

Answer 13

• Empirically valid risk factors
• Method for measuring
• Procedure for combining scores
• Violence risk estimate

Question 14

Traditional types of risk factors

Answer 14

• Static: historical and clinical
• Dynamic (criminogenic needs): dispositional and contextual, easier target for treatment
• Now = more of a continuum (static – stable dynamic – acute dynamic)

Question 15

Types of risk factors

Answer 15

• Historical (static) = past behavior (violent/nonviolent), age of onset (early = chronic), childhood history of maltreatment, social history (employment problems)
• Dispositional (dynamic) = demographics (age, males), personality (impulsiveness, psychopathy), attitudes, traits, tendencies, style
• Clinical (static) = mental disorder symptoms (substance abuse), affective disorders + schizophrenia
• Contextual (dynamic) = situational, access to victim/weapon, perceived stress, lack of social support (instrumental, emotional, appraisal, information)

Question 16

Draw Coping-Relapse model of criminal behavior

Answer 16

Notes

Question 17

Protective factors

Answer 17

• Mitigate/reduce likelihood of antisocial acts (dynamic)
• Prosocial involvement
• Strong social support, positive social orientation
• Strong attachments
• Intelligence

Question 18

Risks of risk assessment

Answer 18

• Lack of agreement
• Field reliability
• Further validation
• Terminology

Question 19

History

Answer 19

• 1st generation: unguided clinical judgment, inaccurate + high error rate
• 2nd generation: static actuarial measures, VRAG + Statis 99R, limited factors with food predictability
• 3rd generation: static + dynamic, education, vocational skills, family, employment, substance abuse
• 4th generation: case management (aka treatment plant), making recommendations

Question 20

Process

Answer 20

• Case file review: previous record (medical/criminal), childhood, goal = reveal antisocial behavior
• Interview: semi-structured, social network, irritability, impulsivity

Question 21

Mental illness

Answer 21

substance use disorder, adjustment disorder, dependent disorder, explicit in media/implicit in justice system

Question 22

Sexual offenders

Answer 22

15% reoffence (non-married with no kids), sexual deviancy, prior sexual offenses, sexual fantasies

Question 23

Female offenders

Answer 23

fraud/theft/prostitution/drug charges i.e., survival responses

Question 24

Young offenders

Answer 24

short-term = high recidivism, anti-social peers, callous/unemotional trait

Question 25

Testimony

Answer 25

• Lowest accuracy influences juries the most
• Unstructured allowed into evidence at state level

Question 26

What did Kaplan et al. find in terms of lay people that serve in juries?

Answer 26

They don’t understand coercion in interrogations as well as expert statements