GLM Recap and Principles of Offender Risk Assessment

Recap of GLM and Introduction to Offender Risk Assessment

• Micah to finish discussing the GLM and move on to offender risk assessment.

• Questions are welcome throughout the lecture.

Principles of Effective Correctional Practices

• Recap of last week's topics and principles of effective correctional practices.

• Heavy link between correctional practices and risk assessment.

• Risk-Need-Responsivity (RNR) model:

  • Core principles are risk, need, and responsivity.

  • Full model consists of 15 principles in total.

  • Decades of research support the model across diverse offenses.

• Good Lives Model:

  • Originally proposed as an alternative, often pitted against RNR.

  • Devon Polaschek offers balanced summaries of both approaches.

Strengths and Weaknesses of the Risk-Need-Responsivity Model

Strengths

• Unifying power and external consistency.

• Based on decades of theory and research.

• Explanatory depth across the 15 principles.

• Integrates political, social, neighborhood, individual, and family factors.

• Great research support and practical utility for correctional systems.

• Fertility allows room to grow in different areas.

Weaknesses

• Lack of simplicity and parsimony due to complexity.

• Overarching principles or guidelines rather than a specific model.

• Underdeveloped areas, especially specific responsivity.

• Criminogenic needs concept underdeveloped theoretically and in research.

Dissemination Issues

• Difficult writing style and lack of clarity.

• Detached and formal academic writing style.

• Terminology makes it hard for students to learn.

Implementation Problems

• Critiques of RNR model stem from implementation issues.

• People often don't implement the model as described properly.

• Quality of implementation is a serious concern.

• Tony Ward and the Good Lives Model pushed RNR proponents to expand on things.

• Differences between the two models evaporate when applied.

Why "What Works" Doesn't Always Work

• Integrity issues in assessment and intervention principles.

• Intensive services not given to higher-risk individuals.

• The wrong people going to the wrong program.

• Programs not delivered as specified.

• Increased focus on structure and manualization of treatment.

• Emphasis on manuals and step-by-step books oversimplifying treatment, and a lot of stuff has sometimes been lost.

• Inexperienced staff delivering treatment by reading aloud from manuals.

• Accreditation processes evaluate programs by written documentation.

• Simplistic approach to monitoring treatment integrity.

Importance of Staff Characteristics

• Characteristics of staff are crucial to program effectiveness.

• Research by Bill Marshall on staff characteristics and change among offenders.

• Important staff traits: empathic, warm, rewarding, directive, and non-confrontational.

Ongoing Research Areas

• Dosage: How much treatment is enough for high-risk offenders?

• Key criminogenic needs for different subgroups of offenders.

• Linking changes in criminogenic needs to reoffending rates.

• Elucidation of the responsivity principle.

• Staff practices and management for program implementation.

Overall Comments on Science and Correctional Practices

• Science is cumulative.

• Incremental improvements over time.

• Avoid throwing out old ideas for new trends.

• Good ideas in the General Linear Model (GLM).

• Integrate good aspects of new ideas and discard problems in old ones.

• Expand, refine, and improve existing practices for greater effectiveness.

Risk Assessment

Introduction to Risk Assessment

• Overview of speaker's background in risk assessment research.

• Involved with Static-99R, Stable 2007, and Acute 2007 risk assessment tools.

• Consultation on court cases and risk assessment reports.

Ubiquity of Risk Assessment in the Criminal Justice System

• Risk assessment is a ubiquitous task at every stage.

• Police prioritizing calls depends on risk assessments.

• Decisions to release on bail or keep custody.

• Sentencing decisions and treatment program planning.

• Determining security levels, probation frequency, and parole readiness.

History of Risk Assessment

• Freudian-influenced hypotheses and untestable subconscious factors.

• Psychologists and psychiatrists claimed unique insights.

• Challenge to professional opinion: Paul Meehl's book on clinical versus statistical prediction in the 1950s.

• Key case in the US in 1960: Inmates detained in psychiatric hospital against their will.

• Supreme Court decision led to the release of many offenders.

• Follow-up study showed low rearrest rates and limited dangerousness.

• Shift from expert opinion to evidence-based risk assessment methods.

Milestones in Risk Assessment

• Vernon Quinsey's research on predicting reoffending.

• Comparing psychiatrists to high school English teachers.

• Psychiatrists showed little agreement, did not incorporate important criminal history information, or show expertise, which was all needed to show good assessment practices.

• Questioning the validity of professional opinion without data.

Definition of Risk Assessment

• Estimating the likelihood of a future event based on secondary indicators.

• Risk assessment guides case management decisions.

• Risk assessment methods should be:

  • Supported by research.

  • Fair and unbiased.

  • Transparent.

  • Consistent.

  • Accurate.

  • Helpful for managing and reducing risk.

Risk Factors: Building Blocks for Risk Assessment

• No single risk factor is a "secret answer."

• Good risk assessment considers multiple factors in diverse domains.

• Each risk factor has a small predictive relationship with reoffending.

• Assess multiple risk factors to achieve moderate to large predictive accuracy.

• Consider the specific outcome you are trying to predict.

Static vs. Dynamic Risk Factors

• Static Risk Factors: Criminal history-based demographic characteristics.

  • Cannot be changed.

  • Examples: prior convictions, age at first offense.

• Dynamic Risk Factors: Factors that can be changed, and therefore should be targets for treatment.

  • Changes over time and with effort.

• Stable Risk Factors: Characteristics that change slowly over time.

• Acute Risk Factors: Factors that can change rapidly.

Advantages and Disadvantages

• Static: fast to code, high reliability, predictive validity but lacks information to intervene.

• Dynamic: time-consuming, expensive but provides treatment goals and measures change.

Protective Factors

• Protective factors: conditions, circumstances, or attributes that lessen the likelihood of negative outcomes.

• Focus on the positive

• Acknowledge resiliencies of people.

Psychologically Meaningful Risk Factors

• Risk factors and protective factors are different ways of measuring the same thing.

• Risk assessment has focused on the neutral to negative and should have focused on the positive end of the scales more.

Measuring

• Measuring things in a bunch of different ways adds to the assessments.

• You tend to get incremental and improve your accuracy by finding new ways to target dynamic risk factors.

• Therapeutic relationship will increase if focus shifts to protective factors.

• Criminogenic needs: the things that must be worked on to lower the risk.

• Targeting the 8 Criminogenic Factors:

  • History of anti-social behavior (static).

  • Antisocial personality.

  • Antisocial cognition.

  • Antisocial associates.

  • Family/marital circumstances.

  • School/work.

  • Leisure/recreation.

  • Substance abuse.

Methods of Risk Assessment

Unstructured Professional Judgment

• Subjectively selecting, analysing, and interpreting factors.

• Rooted in general psychology and psychiatry.

• Focus on mental illness.

Advantages

• Convenient

• Flexible.

• Widely Used

• Case Specific

Disadvantages

• Subjective.

• Biased.

• Weak Agreement Between raters.

• Least Accurate

Actuarial Risk Assessment Tools

• Research Based scales.

• List of items.

• Detailed Coding Rules

• Research showing results of re-offending.

Advantages

• Most Objective we have.

• Gives probability estimate of recidivism.

• Highest Reliability, predictive validity

Disadvantages

• All based on data and averages.

• Doesn't allow for a unique evaluation.

Structured Professional Judgment Tools

• Assessment Tools

• List of risk factors.

• Coding Manual that helps score.

Advantages

• Adapts faster to knew research and situations.

Disadvantages

• More subjectivity.

• Biases are at a higher degree.

• This can lose some good things like consistency and objectivity

Myth Busting

History

• A lot of distortions of Andrews and Bonta.

• It does not relate to whether a tool has static or dynamic aspects in it.

Which works best?

• Can't rate them. There's a lot of differences and nuances.

When is Mechanical Good?

• Not when it relates to human behavior.

• A synthases overall:

  • Good tools and it's possible to use them effectively either way.

  • Unstructured judgment is not defensible.

Overrides

Realities

• Clinical opinion always needed for the assessment.

• Most people end up always overriding or thinking higher risk. It degrades predictive accuracy.

• Also degradation is bad because they end up overriding due to biases.

Conclusion

• Use risk-tool as a guide

• Look for research to help fill that gap

• Deference to what research tells us is paramount