Offender Risk Assessment and the GLM

Recap of Last Week's Lecture

• Continued discussion on the General Linear Model (GLM).
• Clarified misunderstandings from the previous session.
• Reiterated key points regarding the Risk-Need-Responsivity (RNR) model and its 15 principles.
• Highlighted the extensive research support for the RNR model across diverse types of offenses.

Good Lives Model

• The Good Lives Model was initially proposed as an alternative to the RNR model.
• There were critiques and debates between proponents of both models.
• Noted that some academics, like Devon Polaschek, have provided balanced summaries of both approaches, recognizing their strengths and weaknesses.

Strengths and Weaknesses of Risk-Need-Responsivity Model

• Strengths:
  • Unifying power and external consistency.
  • Explanatory depth across various principles explains why some interventions work and others don't.
  • Integrates different factors from political, social, neighborhood, individual, and family levels.
  • Great research support and practical utility for correctional systems.
  • Fertility allows room for growth in different areas.
• Weaknesses:
  • Complexity can be a strength and weakness.
  • Lack of simplicity and parsimony due to numerous principles.
  • The complexity allows diverse approaches within the model.
  • Some areas are underdeveloped, especially specific responsivity.
  • The concept of criminogenic needs is underdeveloped.

Dissemination Issues

• The model suffers from difficult writing, lack of clarity, and archaic writing.
• The detached and formal academic writing style makes it challenging for students.

Implementation Problems

• Critiques of the RNR model often stem from implementation issues rather than inherent problems with the model itself.
• Many corrections administrators claim to implement the RNR model without fully understanding or adhering to its principles.
• Quality of implementation is a serious concern.

Contributions from Tony Ward and the Good Lives Model

• Tony Ward and the Good Lives Model have pushed for expanding underdeveloped aspects of the RNR model.
• Differences between the two models diminish when applied.

Addressing Why "What Works" Doesn't Always Work

• Issues with integrity in the assessment and intervention principles of a program.
• Intensive services should be given to higher-risk individuals, but many programs exclude difficult or uncooperative individuals.
• Problems include the wrong people attending the wrong programs and programs not being delivered as specified.

Trends in Research

• Increased focus on structure and manualization of treatment for research and consistency purposes.
• Emphasis on manuals can oversimplify and lead to loss of important aspects.
• Manuals enable inexperienced staff to deliver treatment by simply reading aloud, without psychological training.
• Accreditation processes may evaluate programs primarily by written documentation rather than actual practice.
• Monitoring treatment integrity becomes a simplistic checklist approach.

Importance of Staff Characteristics

• The characteristics of the staff delivering the treatment can be as important as the content of the treatment itself.
• Empathic, warm, rewarding, directive, and non-confrontational staff tend to be more effective.
• Research shows the importance of interpersonal skills in therapists.

Areas for Further Research

• Determining the appropriate dosage of treatment for high-risk offenders.
• Investigating whether key criminogenic needs differ for different subgroups of offenders.
• Linking changes in criminogenic needs to reoffending rates.
• Further elucidating the responsivity principle.
• Addressing staff practices and management support for program implementation.

Overall Comments

• Science is cumulative, with incremental improvements over time.
• Avoid discarding previous knowledge for new trends without empirical validation.
• The GLM model has great content, but lacks empirical validation.
• Integration of new ideas with existing knowledge is more effective than complete overhauls.
• The RNR model holds up well, with opportunities for improvement and expansion.

Science and Correctional Practices

• Focus should be on expanding, refining, and improving existing practices for greater effectiveness.
• Address the fundamental things that contribute to crime seem remarkably stable.

The core risk factors for crime are surprisingly consistent in large parts of the world.

Examples of Integration

• Individuals are integrating components of both the GLM and RNR models in practice.
• Evaluations of the effectiveness of integrated approaches are lacking.
• Some clinicians and researchers value aspects of both models, seeing them as compatible.

Introduction to Offender Risk Assessment

• Risk assessment is a ubiquitous task in the criminal justice system.
• Decisions at every stage, from policing to parole, are based on risk assessment.
• Risk assessment should be done well, given its importance.
• Decisions about whether someone can be realised on bail or have to stay in custody until their trial.
• Sentencing decisions about what treatment programs they need to attend, what level of security they should be in, if they’re incarcerated.

Risk Need Responsivity Model and Risk Assessment

• The core principles of risk need and responsivity hinge on using a good risk assessment method.

• Delivering treatment proportionate to risk level requires a good way of assessing risk level.

• Targeting criminogenic needs requires assessment of those needs.

• Good risk assessment is the huge pillar of the risk need responsivity model.

History of Risk Assessment

• Early approaches were influenced by Freudian ideas and untestable hypotheses.
• Psychologists and psychiatrists believed they had unique insights into predicting behavior.
• Paul Neal challenged this with his book on clinical versus statistical prediction in the 1950s.
• Key case in the US in the 1960s involved patients detained in psychiatric hospitals based on psychiatrists' opinions.
• The Supreme Court decision led to the release of many patients deemed too dangerous, who did not re-offend at high rates.
• This case questioned the ability to predict who is dangerous and paved the way for evidence-based risk assessment methods.
• The supreme court case that challenged the holding of mentally ill offenders by forensic psychologists was the Backstrom case.

Development of Evidence-Based Risk Assessment

• The 1980s saw the development of more evidence-based risk assessment methods.
• Vern Quincy conducted research on how to better predict who is likely to re-offend.
• He compared forensic psychiatrists to high school English teachers to evaluate their ability to predict dangerousness. Psychiatrists are usually better at predicting dangerousness than English Teachers.
• the Psychiatrists were no better than English teachers at predicting who was most dangerous.
• Psychologists are pretty good with self-reports.

Definition of Risk Assessment

• Risk assessment involves estimating the probability of a future event based on secondary indicator variables.
• It relies on synthesizing individual pieces of information to form an overall risk assessment.
• The term likelihood is preferred over probability.
• Hansen uses for risk assessment. And that it involves estimating the probability of a future event based on secondary indicator variables.

Key Components of Risk Assessment

• Risk assessment methods should ideally:
  • Be supported by research.
  • Be as fair and unbiased as possible.
  • Be transparent.
  • Be consistent.
  • Be accurate.
  • Provide insights into why someone is high risk and how to manage or reduce that risk.

General Principles of Risk Assessment

• Good risk assessment needs to consider multiple risk factors in diverse domains.
• Each risk factor tends to have a small predictive relationship with re-offending.
• Combining multiple risk factors together can achieve moderate to large predictive accuracy effects.

Types of Risk Factors

• Distinction between static and dynamic risk factors.
  • Static: Criminal history, demographic characteristics (unchangeable).
  • Dynamic: Factors that can be changed.
• Dynamic risk factors subdivided into stable and acute.
  • Stable: Characteristics that change slowly.
  • Acute: Factors that can change rapidly.
• A comprehensive risk assessment looks at static, stable, and acute types of risk factors.

Advantages and Disadvantages of Risk Factors

• Static risk factors:
  • Most reliably documented, quick, and easy to score.
  • Have the strongest predictive accuracy on their own.
  • Insufficient for case management.
• Dynamic factors:
  • More time-intensive to assess.
  • Provide treatment targets.
  • Allow measurement of changes over time.
• Implement a static assessment first, then Dynamic later to triage individuals.

Importance of Protective Factors

• Focus on positive attributes, strengths, and resiliencies.
• Should be implemented by the assessor.
• Protective factors are not just the opposite of risk factors, there's something completely different.
• Strenghts have more strengths and are doing well in life.

Psychologically Meaningful Risk Factors

• The importance of psychologists and how they are important to the risk-assessment process.
• They do a better job of distinguishing as well as measuring.
• Instead of focusing on some distinction of static, dynamic and as protective factors different, is that they might all be different ways of measuring the same thing.
• Any risk factor, any item on a risk assessment tool is probably measuring some kind of underlying psychologically meaningful propensity or characteristic related to crime.

Measurement Models in Psychology

• Many constructs in psychology cannot be directly measured.
• There are various indicator variables that tap into the construct.
• Each indicator is not a perfect measure.
• Is it the measurement of shyness or measure a specific type of risk

Ruth Mann and Colleagues Model (2010)

• The idea of looking at latent constructs and a risk factor model.
• Kind of static dynamic and protective is kind of thinking about different types of indicators or pieces of information that have their pros and cons and how we measure that underlying psychological kind of propensity.
• If the person has blacked out has been drunk before then that person is a risk factor
• Someone might have the same thing and they are measuring the same time. Static, dynamic, predictive

Psychologically Meaingful Risk Factors

• In terms of self regulation, you could have items related to the length and diversity of someone's criminal record as a static risk factor.
• If kind of any indications of they've successfully completed education, founded a construction company and are doing very well with that. All of those things indicate kind of ability to set and achieve goals and problem solve and balance competing priorities.
• You can't find them and the person who is drunk as all this time

Combination Techniques

• What are the three main methods of risk assessment or the people that have to assess risk?
  • clinical judgemnt
  • Actuarial risk assessment tool and this is used frequently around the world. Used to calculate car scores etc.
  • Structured profrssional judgement tools also called SPJ

Actuarial Risk Assesments

• Generally like a it's a scale. It has a list of items.
• Detailed, prespecified items rooted in research, and there's some kind of rule for how you score that item. It might be very simple, like yes or no, risk factor present or absent.
• Often used by insurance company and the the decisions are objective.

SPJ - Structured professional judgment tools

• looks like an assessment tool. They're kind of developed and they're researched, and it has a list of risk factors like an actuarial tool and often a coding manual that tells you how to measure and kind of gives guidance on how score each item.
• But then there is no total score. They usually explicitly say it is wrong to calculate a total score.
• SPJ is no different than the generation. You might dress it up and look like a list of items, but at the end of the day, it is still the professional's judgment. It is just a fancy dressed up version of generation one.
• The reason to apply judgement when risk assessments goes against certain models. It should not be used in most cases, although sometimes can be very useful.
• A number of professional will have the ability to override results

Method of Risk Assessments

• is there ways of merging the benefits of actuarial and SPJ. AND where do you lie
• the importance of using professional discretion, including being able to override risk assessment results based on that professional expertise.
• It merges the strength. So we were advocating kind of this approach for about a decade. And then all the studies were starting to come in of you use the actuarial tool, this is the result it gives you, you allow that staff can override it, and they were able to compare the predictive accuracy in those.
• It has been proven that most experts would increase the risk if they have the ability to override. Almost Never Lowering it.

Good Advice For an Clinician or Assessor

• humble thyself before the data. Because I think what really this comes down to in assessment report after assessment report is all this research support of these factors that predict reoffending and then so many clinicians just, yeah, I know the research says that, but I've been doing this for years and I think that, and I think one of the the biggest challenges in the field in a lot of assessment reports is insufficient humility and deference to what the data tell us.
• Clinical should take into account that you can not just go into a case like a robot and that everything must be taken into account. But also not thinking to high about one's opinion. You shouldn't automatically increase or decrease the risk. So balance is key

What we Still Can Expand on

• What is a protective factor model.
• There still is research support with the new data tell us. That is important because new models are being developed all the time.
• There is not a lot research and how to assess if someone is a dangerous risk.
• Not put pin all the responsibility to the individual. Don't only assess just the individual data but also other environmental factors that can influence the test results.
• Environmental Factors for those high risk. What about if the environment that the person lives in increase the risk. ( Bad neighborhood).
• If the assessment tool test is good in the current time and situation.
• How to act according to external sources to get the best outcome. What to do with external information to make the best possible decision.
• Does the crime the offender committed show any pattern in the offender risk.
• Don't only assess just the individual data but also other environmental factors that can influence the test results.