Class Notes (839,394)
Canada (511,324)
Psychology (2,716)
PSYC 3402 (135)
Lecture 4

Lecture 4.docx

21 Pages

Course Code
PSYC 3402
Julie Blais

This preview shows pages 1,2,3,4. Sign up to view the full 21 pages of the document.
Lecture 4 Summary - We’ve defined our outcome measure (DV): criminal behaviour - We’ve taken a look at crime in Canada - We’ve learned about the historical theories that attempted to explain criminal behaviour - We’ve seen how these historical theories helped shape our current understanding using the PIC-R theory of criminal behaviour - Synthesis of what has been done in a particular area of study - PIC-R emphasizes the importance of theoryAND empirical evidence - PIC-R aims to identify empirically defensible risk factors that are also practically and clinically relevant (Central Eight or Big 8) - PIC-R next aims to use these factors to predict risk for the correct classification, treatment, and management of offender groups Where are we headed? - How well can criminal behavior be predicted? - What can we do with that knowledge in order to reduce the chances of criminal acts occurring? - To what extent do the Central Eight risk/need factors apply to different offender samples? Risk assessment - Integral part of the Criminal Justice System (CJS) - Affects community safety, prevention, treatment, ethics, and justice. - Guides police officers, judges, prison officials, and parole boards - Occurs at every level of the CJS - We all share an interest in accurate prediction (citizen, offender, victim, professional) - Hanson (2009): - “Risk assessment involves estimating the probability of a future event based on secondary, indicator variables” (Hanson, 2009, p. 172). - Risk assessment is not a perfect science; there is always error involved. Predictive accuracy - Major concern: PREDICTIVE VALIDITY • The degree of success of any approach in foretelling the likelihood of involvement in further offending - Statistical significance: outcome that occurs beyond chance level (affected by sample size and alpha level selected) - Effect size: standardized measure of the magnitude of a given effect (beyond significance) - More people = more significant result - Effect size: .05 = I accept the fact that I may make a mistake 5% of the time Predictions and outcomes (EXAM QUESTION ON THIS!!!!) - Outcome: • Offender is released and he offends • Offender does not offend - Results: • True Positive- High risk, and is violent/offends • False Positive- High risk but not violent/offend • False Negative- Low risk but goes and recommits crimes, violent • True Negative - Low risk and doesn't offend (non violent) - Implications: - Errors have different consequences • High False Positive – said there would be a lot more high risk (to protect society). Only affects offender • High False Negative – said everyone is low risk. Society and victim safety jeopardized - Critics have highlighted high false positive rates implying imprecision - Rebuttal indicates risk predictions exceed “chance” Measuring accuracy - Selection ratio: • Selection ratio is the cut-off or cut-point for determining high and low risk cases • The choice of the selection ratio influences the type of decision error that is likely committed  Lower cut-off yields higher false positive rate  Higher cut-off yields higher false negative rate - 40+ score to be high risk - 30+ has traditionally been used to consider a protypical psychopath Measuring accuracy - Receiver Operating Characteristics curve (ROC) • Allows us to calculate theArea Under the Curve (AUC)  AUC ranges 0.5 to 1.0 • Rock curve is plotting hits (true positive rate) on the Y-axis, of your false positive rate - Survival curve analysis ROC - The ROC curve • p(Hit) on the y-axis and p(FalseAlarm) on the x-axis • Connects points to create a curve • Measure the area under the curve (AUC) to get an overall measure of predictive accuracy • AUC ranges from 0.50 (indicating chance accuracy) to 1.00 (indicating perfect accuracy) ROC interpretation - Compares hit rates to false alarms - In other words, the probability that a randomly selected recidivist will have a higher risk score than a randomly selected non-recidivist - Also, can be directly converted to a Cohen’s d effect size (no math required! Just look it up in a table!!) ROC: strengths and limitations - The only procedure that allows researchers to establish accuracy scores that are not biased by decision thresholds (i.e., scores on an assessment tool) or impacted by base rates (like the correlation) - However, it is affected by variability in the predictor AUC - What is a good AUC? • .90 – 1.0 = excellent • .80 - .90 = good • .70 - .80 = fair • .60 - .70 = poor • .50 - .60 = fail Survival analysis - Survival curve analysis • Examines the “survival rate” of a group of offenders after they have been released into the community • Can compare survival curves of different subgroups • Offenders with lower scores on the risk assessment should “survive” in the community for a longer period of time Risk assessment standards - Instruments need to be tested prior to implementation • Objective • Reliable (internal & inter-rater) • Meaningful (items make sense) • Predictive validity (predict relevant outcomes) • Dynamic validity (changes predict outcome)  Change in person should be in scale (becoming lower risk) • Sample representativeness  Socially unbiased (ethnicity & gender)  Generalization (applies to other groups and settings beyond construction sample) - Improve accuracy • Reduce the errors we make as much as possible - Improve transparency • Clarity in the findings, better communication of the results - Improve consistency • Same standards for every case First-generation risk assessment - Unstructured professional judgment - “What the studies, taken in totality, actually show very clearly is that you have to detain a much larger number of people than those who are actually dangerous in order to reach the dangerous (Mathieson, 1998, p. 461)” - “It is clear from the research literature that we cannot, and will never be able to, predict with reasonable certainty future violence (Meloy, 1992, p. 949).” 1 generation: UCJ - Unstructured clinical judgment (1930- ) • Mental health professional predicted chance of re-offence from interview and clinical impressions • Based on practitioners’clinical experience - Advantages: • Flexible and specific to individuals - Disadvantages: • Inconsistent and inaccurate - Evidence: • Steadman & Cocozza (1974) • Baxstrom Case (1966, U.S. Supreme Court) • 98 patients released against hospital advice • Only 20 of the 98 reoffended at follow-up Case study - In groups of 3-4 people, read the case study provided on cuLearn - Place the offender in one of the following risk categories (violent recidivism) • Low • Low-moderate • Moderate • Moderate-high • High - Rated: Moderate-high • Separation from parents – No = -2 • Elementary school maladjustment – Yes= 2 • history of alcohol abuse – Serious alcohol problem • Marital status – Ever married= -2 • Criminal history score – More than 3 = +3 • Failure on prior conditional release = Yes = +3 • Age at index offence= 34 through 38= -2 • Victim injury at index offence- hospitalized • Any females victim – yes = -1 • Personality disorder- yes= +3 • Final Score = + 12 • 7 of 9 risk bins nd 2 generation: actuarial - Actuarial tools (1975- ) (uses only static risk factors) • Systematic measurement of a set of factors derived from a research database • Statistically analyze relationship between factors and outcome (weight accordingly) • Derive formulas that provide risk scores based on the empirical relationships between (static) risk variables and the criterion variable (i.e., violence) • Provides risk estimates (absolute recidivism rates) 1 vs. 2 generation - Actuarial outperforms clinical judgement all the time!! (gen. 2 better than gen.1) Examples of actuarial scales - Burgess (1928) • Examined over 3,000 parolees; identified 21 factors that differentiated parole successes from parole failures (max. associated with 76% recidivism) - Statistical Information on Recidivism (SIR; Nuffield, 1982) • CSC risk scale; used in classification • Static, historical factors • Developed and used by decision-makers • Not guided by behavioural theory Violence risk appraisal guide - VRAG (Quinsey et al., 1998) - Developed on a sample (n = 618) of offenders from Penetanguishene - Consists of 12 weighted static risk factors (-26 to +38) - Sum score and place offender within 1 of 9 risk bins (absolute recidivism rates) - Predicts violent recidivism - VRAG items: • Historical items:  Separated from parents prior to 16 (+)  Never married (+) • Criminal history items:  Failure on prior release (+)  Nonviolent offence history (+)  Age at index offence (-)  Victim injury (-)  Female victim (-) • Psychological assessment items:  Elementary school adjustment (+)  Alcohol problems (+)  Personality disorder (+)  Schizophrenia (-)  Psychopathy (+) VRAG risk bins Score the case study: VRAG - 1) Separation from parents prior to 16 • -2 No • +3 Yes - 2) Elementary school maladjustment: problems between the ages of 4-13 (truant, disruptive, etc.) • 0 if no • 2 if yes - 3) History of alcohol abuse • -1 if No • 0 if slight problem • +2 if serious alcohol problem - 4) Marital status • -2 Ever married (6 months common-law) • +1 Never married - 5) Criminal history score • -2 no criminal history • 0 if between 2-3 prior convictions • +3 if more than 3 prior convictions - 6) Failure on prior conditional release • 0 if No • +3 if Yes - 7) Age at index offence • -5 if 39 or over • -2 if 34 through 38 • -1 if 28 through 33 • 0 if 27 • +2 if 26 or younger - 8) Victim injury at index offence • -2 Death • 0 Hospitalized • +1
More Less
Unlock Document

Only pages 1,2,3,4 are available for preview. Some parts have been intentionally blurred.

Unlock Document
You're Reading a Preview

Unlock to view full version

Unlock Document

Log In


Join OneClass

Access over 10 million pages of study
documents for 1.3 million courses.

Sign up

Join to view


By registering, I agree to the Terms and Privacy Policies
Already have an account?
Just a few more details

So we can recommend you notes for your school.

Reset Password

Please enter below the email address you registered with and we will send you a link to reset your password.

Add your courses

Get notes from the top students in your class.