Lecture 10: Risk Assessment

Question 1

Why should we care about risk assessment?

Answer 1

Because Risk assessment informs: Sentencing (especially Dangerous offender hearings), classification, treatment needs, treatment intensity, parole decisions, level of supervision, notification decisions, release conditions.

Question 2

What is risk assessment?

Answer 2

A concept involving two components: (1) risk prediction - assessing the risk that people will commit violence in the future, and (2) risk management - developing effective intervention strategies to manage that risk.

Question 3

What are the goals of risk assessment?

Answer 3

Improve accuracy (making correct decisions in regards to high or low risk)
Improve transparency (people like the offender, the victim, the public,  need to understand why someone is released or not released)
Improve consistency (the same decisions should be made)

Question 4

What is a true positive?

Answer 4

A correct prediction that occurs when a person who is predicted to engage in some type of behaviour (e.g., a violent act) does so

Question 5

What is a true negative?

Answer 5

a correct prediction that occurs when a person who is predicted not to engage in some type of behaviour (e.g, a violent act) does not

Question 6

What is a false positive?

Answer 6

an incorrect prediction that occurs when a person is predicted to engage in some type of behaviour (e.g., a violent act) but does not

Question 7

what is a false negative?

Answer 7

An incorrect prediction that occurs when a person is predicted not to engage in some type of behaviour (e.g., a violent act) but does

Question 8

How are the last two types (false negative and positive) dependant on each other?

Answer 8

Minimizing the number of false positive errors will result in an increase in the number of false negative errors.

Question 9

What do we consider in risk assessments?

Answer 9

Risk factors- which are variables related to recidivism

Question 10

What is risk management?

Answer 10

Develop effective intervention strategies to manage that risk

Question 11

What are static risk factors?

Answer 11

Fixed and unchanging, Most convenient, Most frequently used, Can be reliably measured and are very predictive , If its historical or previous involvement like the commission of a crime or a history of mental illness it is static, because the past cannot be changed.

Question 12

What are dynamic risk factors?

Answer 12

Change with time, Less convenient, less reliable, Less frequently used, BUT: sensitive to change. With intervention, can change level of risk

Question 13

What are some examples of static risk factors?

Answer 13

Demographic variables (Being male is a static risk factor, men are more likely to reoffend), History of criminal behaviour, History of mental disorder

Question 14

What are the 2 Main types of dynamic risk factors?

Answer 14

Stable dynamic: persistent and change slowly, if at all (e.g., impulse control, coping ability)
Acute dynamic: rapidly fluctuating (e.g., intoxication, personal support, employment)

Question 15

What are the big 4 risk factors?

Answer 15

Criminal history
Procriminal personality (impulsive, aggressive)
Procriminal attitudes
Procriminal associates

Question 16

What variables are commonly mistaken for risk factors?

Answer 16

Low Socioeconomic status (not an INDIVIDUAL risk factor for reoffending, may increase exposure to crime but is not a predictor for reoffending)
Personal distress/psychopathology (also includes low self esteem or depression
Fear of punishment
Verbal intelligence
Remorse/empathy
Offense severity

Question 17

What are the 3 approaches to risk assessment?

Answer 17

Unstructured clinical judgement, actuarial prediction, and structured professional judgement

Question 18

What is unstructured clinical judgement?

Answer 18

Decisions characterized by a substantial amount of professional discretion and lack of guidelines. Subjectively selecting what factors to look at, analyzing these factors and then interpreting the risk. This is usually a disaster
Advantages: very flexible (because you are not restricted by anything), idiographic (person specific, case by case basis)
Disadvantages: inconsistent, low accuracy

Question 19

What is actuarial prediction?

Answer 19

Actuarial tools collect pre-specified risk factors that are entered into a statistical model that weights them, combines them, and produces a probability of offenders
Advantages: consistent and high accuracy
Disadvantages: Nomothetic, Validity across different samples (it will be low). Sometimes can be quite poor in making transparent decisions, that is, making decisions that people understand

Question 20

What is an example of an actuarial tool?

Answer 20

The violence risk appraisal guide (VRAG). Consists of 12 weighted static risk factors. Added together to give overall probability of risk. Code presence of risk factors, Total the scores, Assign individual to 1 of 9 bins, Estimate probability of violence

Question 21

What is structured professional Judgement?

Answer 21

Arising from the limitations associated with unstructured clinical judgement and concern that the actuarial method did not allow for individualized risk level, a new approach to risk assessment emerged. decisions are guided by a predetermined list of risk factors that have been selected from the research and professional literature (case specific details are added in as well). Judgement of risk level is based on the evaluators professional judgement.
Advantages: flexible, nomothetic-idiographic
Disadvantages: moderate accuracy (clinical judgement),less consistent than actuarial

Question 22

What is an example of a structure professional judgement tool?

Answer 22

the HCR-20, in this approach the evaluator conducts a systematic risk assessment and refers to a list of risk factors, each having specific coding criteria and a demonstrated relationship with violent recidivism based on the existing professional and empirical literature. The HCR-20 stands for the list of 20 items organized into three main scales that align risk factors into past (historical), present (clinical) and future (risk management).
Low risk: monitor and intervene with low priority and intensity
Mid risk: monitor and intervene with some priority and intensity
High Risk: monitor and intervene with high priority and intensity

Question 23

Why do we use tools to evaluate risk assessment tools?

Answer 23

Most important: predictive accuracy, Does it predict recidivism? Do the high risk offenders reoffend more than the low risk offenders?
Also: ease of use, is it easy to train people on, how much time does it take, construct validity, consistent (inter rater reliability),

Question 24

What is the receiver operating characteristic (ROC)?

Answer 24

In this context it is a technique for measuring the accuracy of risk assessments by examining false positive and true positives across decision threshold
A true positive is a “hit”. A false positive is a false alarm. False negative or a miss (predict something and then something else happens. True negative or correct rejection. Decision thresholds Or cut-off scores: scores range varies from 0-40 (e.g., 30 out of 40 is too high risk you wont be released). Compare prediction column to reality column (violent)

Question 25

Desribe the ROC graph

Answer 25

For each possible cut-off value (score), plot false positive rate (X-axis) as function of true positive rate (y axis). Connect the dot and get a curve. We can then measure the area under that curve to get an overall measure of predictive accuracy. Ideally you want the hit rate to false alarm rate ratio to be very high

Question 26

According to the ROC how accurate are the unstructured clinical judgement, the Actuarial tool and the structured professional judgement?

Answer 26

Clinical judgements: AUC=0.55
Actuarial tools: AUC=.68 to .80
Structured professional judgement: AUC= .62to .75

Question 27

What is base rate?

Answer 27

represents the percentage of people within a given population who commit a criminal or violent act.

Question 28

What is the base rate problem?

Answer 28

To predict violence we must determine base rates. But it is difficult to make accurate predictions when the base rate is too high or to low. A problem that emerges when attempting to predict events that have a low base rate is that many false positives will occur (e.g., the past decade has seen several high profile shootings, although they generate a lot of media coverage they are infrequent , any attempt to predict which youth might engage in school shooting would result in many people being wrongly classified). The general rule is that it is easier to predict frequent events than infrequent events.

Question 29

ROC analysis

Answer 29

Stands for receiver operating characteristics analysis. Used as a procedure for measuring the accuracy of risk predictions. The height of the ROC curve, as measured
by the area under the curve (AUC), indicates accuracy. It is calculated by plotting the false positive rate as a function of the true positive rate for each possible cut-off score
on the scale. An AUC=1.0 indicates perfect accuracy and an AUC=0.50 indicates chance accuracy. The advantage of using the AUC as a measure of decision making
accuracy is that it is not threshold dependent (i.e., it does not depend on the cut-off score that was chosen by the decision maker to define when an individual will be
violent).