Chapter 14: Risk Assessment:

Disadvgs of Actuarial Instruments

1. Most rely almost entirely on historical risk factors and fail to include a wide variety of risk factors (implies that an indv's future risk will not change over time) 2. while they outperform Clinical, have been questioned based on their generalizability (df): how well these instruments perform outside the o.g. population on which they were created

Components of Risk Assessment

1. Prediction: Likelihood of future criminal act 2. Management: Development of strategies to manage or reduce level of risk

Approaches to assessing risk

1. Unstructured Clinical Judgment 2. Actuarial prediction/instruments

Disadvgs of Unstructured Clinical Judgment

1. relatively weak bc the lack of feedback about success or failure; mental health experts also tend to ignore base rate info in making their risk predictions (# of ppl in a partic population who actually become violent); if base rate is low (beh is infrequent), our ability to predict that beh will be v limited 2. Also, bc clinicians often unaware of how rarely an offender was likely to reoffend (base rate of re-offense), they were prone to over predicting violent beh

What % of "dangerous" mentally ill patients released from prison hospitals had been arrested or hospitalized for violent beh?

14.5%

What % of inmates were falsely classified as "dangerous"?

92.8%

Kansas v. Hendricks

> 1st person to be subjected to involuntary civil confinement under Kansas's new SVP law > ruled that SVP laws didn't violate double jeopardy & ex post facto const.al concerns bc these laws imposed civil confinement rather than criminal punishment, & const.al restrictions on punishment only applied to criminal matters

Unstructured Clinical Judgment

> Intuitive approach/subjective judgment > Focuses on the indv > No rules specify how a clinician should collect and combine information to decide future risk

Categories associated w violent beh used to assess risk level

> Memory Trigger: History proves Dynamically Risky 1. Historical Markers 2. Dynamic Markers 3. Risk Management Markers

Barefoot v Estelle

> Memory trigger: Bare risk predictions are the Estelles of the death penalty > case ex. of Future Dangerousness Standard > held that expert predictions of risk were admissible in capital sentencing despite their potential inaccuracy

HCR-20

> Not bound to weigh the factors in any equation—gives weight based on clinician descretion > doesn't suffer the generalizability problem of actuarial instruments bc its factors were not based on a specific pop, but instead are all factors freq id'ed in the research literature > allows researcher to include dynamic, risk management, or rare risk factors in its prediction > signif improvement of accuracy over intuitive clinical judgments

Dynamic Markers

> Not fixed, can be adjusted over time or through treatment. > Moods, attitudes, attributions & through processes, ex. lack of insight into ones self/others. > Psychiatric Symptomatology > Threat/Control-Override (TCO) symptoms > Impulsivity > Lack of responsiveness to treatment

Actuarial Prediction

> Relevant risk factors be systematically combined to calculate an estimate of future violence > Quantitative approach, based on research on reoffending > Based on data from large groups

Psychiatric Symptomatology

> a dynamic marker > "active" psychiatric symptoms (readily apparent in one's thoughts and behs) v. "dormant" symptoms

Threat/Control-Override (TCO) symptoms:

> a dynamic marker > beliefs (common in schizos) that other ppl or forces are controlling one one's thoughts or implanting thoughts in one's mind; this perceived threat from others overrides self control

Risk Management Markers

> ability to manage one's beh & environ post-release > Stable, supportive environment post-release is associated with lower risk > size of social network is irrelevant; quality of ppl in it seems to matter

Guided Professional Judgment Instruments

> designed to combine accuracy of actuarial with the flexibility of clinical decision-making > HCR-20

Future Dangerousness Standard:

> for a defendant to receive death penalty, sentencing jury must find beyond a reasonable doubt that "there's a probability that the defendant would commit criminal acts of violence that would constitute a continuing threat to society" > used in some states such as Texas and Oregon > a majority of states that permit the death penalty allow expert testimony on this issue

Historical Markers

> the most useful risk predictions > Static factors that cannot change, part of the indv's life story & personality. > Past violent behavior, age at first offense, early abuse of alcohol/drugs, major mental disorders, psychopathy, early maladjustment at home/school.

Chesire murders: How did the system fail?

When he was being sentenced, Komisarjevsky's 18 prior burglaries were never give to the judge and thus weren't considered

Traditional risk assessment models may result in significant rates of...

false positives

Preventative Detention:

holding someone in a jail or hospital because he or she might become violent.

Ex of risk assessment used in practical matters:

parole boards and prison release review boards must decide if an inmate is likely to commit future acts of violence if released

Ethical issue of balance in risk assessment:

the need to protect society v. the rights of indvs to be protected from harrassment by authorities.