Hoorcolleges Risicotaxatie
Tilburg University 2019-2020
Femke van Leth
Inhoudsopgave
Hoorcollege 1 (29-1)....................................................................................................................................... 2
Introduction to risk assessment............................................................................................................................2
Additional information based on the article for this lecture...........................................................................5
Hoorcollege 2 (30-1)....................................................................................................................................... 5
Issues surrounding risk assessment......................................................................................................................5
General principles of risk assessment.............................................................................................................6
Additional information based on the article for this lecture...........................................................................8
Hoorcollege 3 (5-2)........................................................................................................................................ 8
Measurement issues.............................................................................................................................................8
Signal detection theory..................................................................................................................................10
Hoorcollege 4 (6-2)....................................................................................................................................... 11
Risk assessment in juveniles...............................................................................................................................11
Additional information based on the compulsory reading for this lecture..................................................14
Hoorcollege 5 (12-2)..................................................................................................................................... 15
Youth and sexual offenders................................................................................................................................15
Additional information based on the compulsory reading for this lecture..................................................17
Hoorcollege 6 (13-2)..................................................................................................................................... 19
Adult sex offenders.............................................................................................................................................19
Violent risk assessment......................................................................................................................................21
Additional information based on the compulsory reading for this lecture..................................................24
Hoorcollege 7 (19-2)..................................................................................................................................... 25
Institutional violence..........................................................................................................................................25
Hoorcollege 8 (04-03)................................................................................................................................... 28
Clinical case........................................................................................................................................................28
Hoorcollege 9 (5-2)....................................................................................................................................... 31
Clinical and actuarial judgements......................................................................................................................31
Additional information based on the compulsory reading for this lecture..................................................32
Hoorcollege 10 (11-3)................................................................................................................................... 33
Dangers and pitfalls with risk assessment..........................................................................................................33
,Hoorcollege 1 (29-1)
Introduction to risk assessment
- Risk (Criminal Justice): uncertain predictions about future behaviour, with a possible
chance that the future outcome of the behaviour will be harmful or negative
o Risk assessment: probability calculation that a harmful behaviour or event will
occur -> involves assessment about the frequency of the behaviour (or event),
about its likely impact and about who it will affect -> risk assessment must
specify 3 things: behaviour (violent), potential damage or harm caused by the
behaviour + probability that it will occur and under what circumstances
We will never get it 100% right, but we can try to do it the best we can
Important to remember: Risk assessment is not ‘one size fits all’ ->
there are many possible risk assessment tools you can use for different
kinds of offenders, target populations and target behaviours
- In risk assessment you look at two types of factors: risk factors and protective factors
o Risk factors increase the likelihood of re-offence -> many kinds of risk factors
Protective factors reduce the likelihood of re-offence (positive factors)
Both factors are dependent of the offender you’re dealing with
A risk factor that can be triggered in a specific context = a mechanism
o Static risk factors: age, gender, ethnicity, first offender, type of offence
Static risk factors are not changeable or treatable + are historical ->
they provide important information but can’t be used in therapy
These risk factors are therefore not used in forensic psychiatry
Offender Groups Reconviction Scale (OGRS) -> translated in Dutch:
StatRec -> validated tool that assesses risk based on static factors
The predictive validity is good Area Under the Curve = .80
Static-99R: static risk assessment for sexual offenders used worldwide
Problem: it’s a categorical decision -> when do you grant
someone freedom, and when do you keep someone in prison?
o Dynamic risk factors: they are the children of risk prediction, because they are
treatable and changeable -> interventions aim at the changing of these
dynamic risk factors, because they are an important cause of re-offending
The relation between dynamic risk factors and risk of re-offending is
moderated by static risk factors (which you cannot control/change)
That is why it is a good thing to focus on both risk factors
Central eight criminogenic needs (Andrews, Bonta & Wormith):
antisocial attitudes, antisocial peers, antisocial personality, antisocial
behaviour patterns, absence of pro-social leisure/recreation activities,
dysfunctional family, employment issues and substance abuse
There are many treatment targets based on these needs
, o Focus on needs in therapy: the RNR-model
The big-Four (the first four needs) are the most predictive
o But: some of these factors are very hard to change
Non-criminogenic needs: tend to be needs of the person that are
predictive of re-offending, but that are not directly related to
criminality: personal distress, mental disorder, low self-esteem, low
physical activity, poor physical living conditions and insufficient fear
Stable dynamic risk factors: personal skill deficits and learned
behaviours that correlate with recidivism but that can be changed
through a process of effortful treatment -> change happens slowly
Sexual self-regulation, general self-regulation, intimacy deficits,
compliance and understanding for the need of treatment and
control, existence of supportive significant others and/or
distorted attitudes or attitudes tolerant to sexual violence
o These are the stable dynamic risk factors of sexual
violence that are assessed using the STABLE-2007
o Compliance and understanding for the need of
treatment and control is very important -> you have to
have a form of insight in your problem and you have to
be motivated to change your behaviour (very hard!)
This becomes an issue with narcissism
Change within these risk factors takes time -> practice
Acute dynamic risk factors: highly transient conditions that only last
hours or days -> rapidly changing because of environmental triggers or
intrapersonal stresses related to re-offence -> the ACUTE-2007 (tool)
Acute dynamic risk factors related to sexual recidivism: victim
access, hostility, sexual preoccupation, rejection of supervision,
emotional collapse, no social support and/or substance abuse
o Both risk factors (static and dynamic) can be distal and proximal risk factors:
Distal risk factors: effecting re-offending way down the line (they
happened a while ago, but they are still persisting in the present)
Examples: genetics, childhood abuse (stress), temperament
Proximal risk factors: these are more acute and obvious risk factors
Examples: substance abuse, ongoing abuse, stressful events
These are immediate vulnerabilities for a particular condition
o Protective factors: ‘treatment is not only fixing what is broken, it is also the
preservation and strengthening of what is good’ -> we can use protective
factors in treatment to reduce the likelihood of re-offending -> effective!
A protective factor is a characteristic at the biological, psychological,
family or community level that is associated with a lower likelihood of
problem outcomes or that reduces the negative impact of a risk factor
It can also be seen as the opposite of a risk factor (positive)
Examples: social support, prosocial attitudes, prosocial identity, school
competence and motivation & positive attitudes toward intervention
These factors help the treatment along + reduce re-offending
- The cause-effect hypothesis: dynamic risk factors have a dual status: they are
potential causes of recidivism from an explanatory view AND they are predictors
, o There is evidence for an association between risk factors and re-offending
This is a correlation -> in this sense, the risk factor is a predictor (!)
o If we want to conclude that the risk factor is the cause of recidivism, there
must be a causal relationship in the right direction (the cause must precede
the effect) + the cause has to be different from the consequence (important)
Other very important question you have to ask with a correlation: are
there any other intercorrelated factors for the observed association?
o When there is a cause between risk factor A and outcome B, the mechanism
must be explained -> why does risk factor A cause re-offending (outcome B)?
Important question here: can dynamic risk factors be seen as causes?
They are composite constructs -> many related constructs
o Each factor consists of several subfactors and single
factors -> that means the risk factors are very broad
o Some risk factors are social, others are biological…
They lack specificity -> many factors aren’t explained
o There are multiple causes for re-offending -> you have
to make sure your risk factor is really causing criminality
Other causes for re-offending (confounds) must
be considered and ruled out with research!
The ‘grain’ problem: dynamic risk factors are often presented
at multiple levels ranging from a general domain to a more
concrete level -> how do you decide on the appropriate level of
abstraction when using factors in an explanatory context?
o What level of specificity is needed for your question?
Dynamic risk factors are seen as normative concepts: the
starting point is ‘normal or appropriate behaviour’
o Other problem: categorizing people as ‘low’ or ‘high’
risk is hard, because dynamic risk factors keep changing
In sum: it is difficult to predict and explain WHEN a dynamic risk factor
(DRF) will become active and lead to a crime, and WHY this happens
- There have been multiple generations of risk assessment tools throughout the years:
o First generation: unstructured professional judgement -> the ‘professional’
who is doing the risk assessment makes a random judgement, with only a few
interviews to guide him -> there was no underlying assessment instrument
Advantage: not expensive, flexible, convenient, accepted, no training
Disadvantage: lack of consistency and transparency, bias, accuracy: .50
We needed measures that were more objective and predictive
o Second generation: actuarial instruments: empirically based measures with a
focus on static factors -> the items needed to predict re-offending, but the
content didn’t matter that much -> re-offenders needed to react ‘differently’
Advantage: transparent/objective, good reliability + predictive validity
Disadvantage: a-theoretical, unchangeable, limited room for therapy
We needed measures that focus on reducing re-offending
o Third generation: structured professional judgement -> dynamic risk factors
Advantage: transparent, sensitive to change over time, good reliability
+ predictive validity and broad identification of treatment goals
Disadvantage: time consuming to examine changes + not so practical
Tilburg University 2019-2020
Femke van Leth
Inhoudsopgave
Hoorcollege 1 (29-1)....................................................................................................................................... 2
Introduction to risk assessment............................................................................................................................2
Additional information based on the article for this lecture...........................................................................5
Hoorcollege 2 (30-1)....................................................................................................................................... 5
Issues surrounding risk assessment......................................................................................................................5
General principles of risk assessment.............................................................................................................6
Additional information based on the article for this lecture...........................................................................8
Hoorcollege 3 (5-2)........................................................................................................................................ 8
Measurement issues.............................................................................................................................................8
Signal detection theory..................................................................................................................................10
Hoorcollege 4 (6-2)....................................................................................................................................... 11
Risk assessment in juveniles...............................................................................................................................11
Additional information based on the compulsory reading for this lecture..................................................14
Hoorcollege 5 (12-2)..................................................................................................................................... 15
Youth and sexual offenders................................................................................................................................15
Additional information based on the compulsory reading for this lecture..................................................17
Hoorcollege 6 (13-2)..................................................................................................................................... 19
Adult sex offenders.............................................................................................................................................19
Violent risk assessment......................................................................................................................................21
Additional information based on the compulsory reading for this lecture..................................................24
Hoorcollege 7 (19-2)..................................................................................................................................... 25
Institutional violence..........................................................................................................................................25
Hoorcollege 8 (04-03)................................................................................................................................... 28
Clinical case........................................................................................................................................................28
Hoorcollege 9 (5-2)....................................................................................................................................... 31
Clinical and actuarial judgements......................................................................................................................31
Additional information based on the compulsory reading for this lecture..................................................32
Hoorcollege 10 (11-3)................................................................................................................................... 33
Dangers and pitfalls with risk assessment..........................................................................................................33
,Hoorcollege 1 (29-1)
Introduction to risk assessment
- Risk (Criminal Justice): uncertain predictions about future behaviour, with a possible
chance that the future outcome of the behaviour will be harmful or negative
o Risk assessment: probability calculation that a harmful behaviour or event will
occur -> involves assessment about the frequency of the behaviour (or event),
about its likely impact and about who it will affect -> risk assessment must
specify 3 things: behaviour (violent), potential damage or harm caused by the
behaviour + probability that it will occur and under what circumstances
We will never get it 100% right, but we can try to do it the best we can
Important to remember: Risk assessment is not ‘one size fits all’ ->
there are many possible risk assessment tools you can use for different
kinds of offenders, target populations and target behaviours
- In risk assessment you look at two types of factors: risk factors and protective factors
o Risk factors increase the likelihood of re-offence -> many kinds of risk factors
Protective factors reduce the likelihood of re-offence (positive factors)
Both factors are dependent of the offender you’re dealing with
A risk factor that can be triggered in a specific context = a mechanism
o Static risk factors: age, gender, ethnicity, first offender, type of offence
Static risk factors are not changeable or treatable + are historical ->
they provide important information but can’t be used in therapy
These risk factors are therefore not used in forensic psychiatry
Offender Groups Reconviction Scale (OGRS) -> translated in Dutch:
StatRec -> validated tool that assesses risk based on static factors
The predictive validity is good Area Under the Curve = .80
Static-99R: static risk assessment for sexual offenders used worldwide
Problem: it’s a categorical decision -> when do you grant
someone freedom, and when do you keep someone in prison?
o Dynamic risk factors: they are the children of risk prediction, because they are
treatable and changeable -> interventions aim at the changing of these
dynamic risk factors, because they are an important cause of re-offending
The relation between dynamic risk factors and risk of re-offending is
moderated by static risk factors (which you cannot control/change)
That is why it is a good thing to focus on both risk factors
Central eight criminogenic needs (Andrews, Bonta & Wormith):
antisocial attitudes, antisocial peers, antisocial personality, antisocial
behaviour patterns, absence of pro-social leisure/recreation activities,
dysfunctional family, employment issues and substance abuse
There are many treatment targets based on these needs
, o Focus on needs in therapy: the RNR-model
The big-Four (the first four needs) are the most predictive
o But: some of these factors are very hard to change
Non-criminogenic needs: tend to be needs of the person that are
predictive of re-offending, but that are not directly related to
criminality: personal distress, mental disorder, low self-esteem, low
physical activity, poor physical living conditions and insufficient fear
Stable dynamic risk factors: personal skill deficits and learned
behaviours that correlate with recidivism but that can be changed
through a process of effortful treatment -> change happens slowly
Sexual self-regulation, general self-regulation, intimacy deficits,
compliance and understanding for the need of treatment and
control, existence of supportive significant others and/or
distorted attitudes or attitudes tolerant to sexual violence
o These are the stable dynamic risk factors of sexual
violence that are assessed using the STABLE-2007
o Compliance and understanding for the need of
treatment and control is very important -> you have to
have a form of insight in your problem and you have to
be motivated to change your behaviour (very hard!)
This becomes an issue with narcissism
Change within these risk factors takes time -> practice
Acute dynamic risk factors: highly transient conditions that only last
hours or days -> rapidly changing because of environmental triggers or
intrapersonal stresses related to re-offence -> the ACUTE-2007 (tool)
Acute dynamic risk factors related to sexual recidivism: victim
access, hostility, sexual preoccupation, rejection of supervision,
emotional collapse, no social support and/or substance abuse
o Both risk factors (static and dynamic) can be distal and proximal risk factors:
Distal risk factors: effecting re-offending way down the line (they
happened a while ago, but they are still persisting in the present)
Examples: genetics, childhood abuse (stress), temperament
Proximal risk factors: these are more acute and obvious risk factors
Examples: substance abuse, ongoing abuse, stressful events
These are immediate vulnerabilities for a particular condition
o Protective factors: ‘treatment is not only fixing what is broken, it is also the
preservation and strengthening of what is good’ -> we can use protective
factors in treatment to reduce the likelihood of re-offending -> effective!
A protective factor is a characteristic at the biological, psychological,
family or community level that is associated with a lower likelihood of
problem outcomes or that reduces the negative impact of a risk factor
It can also be seen as the opposite of a risk factor (positive)
Examples: social support, prosocial attitudes, prosocial identity, school
competence and motivation & positive attitudes toward intervention
These factors help the treatment along + reduce re-offending
- The cause-effect hypothesis: dynamic risk factors have a dual status: they are
potential causes of recidivism from an explanatory view AND they are predictors
, o There is evidence for an association between risk factors and re-offending
This is a correlation -> in this sense, the risk factor is a predictor (!)
o If we want to conclude that the risk factor is the cause of recidivism, there
must be a causal relationship in the right direction (the cause must precede
the effect) + the cause has to be different from the consequence (important)
Other very important question you have to ask with a correlation: are
there any other intercorrelated factors for the observed association?
o When there is a cause between risk factor A and outcome B, the mechanism
must be explained -> why does risk factor A cause re-offending (outcome B)?
Important question here: can dynamic risk factors be seen as causes?
They are composite constructs -> many related constructs
o Each factor consists of several subfactors and single
factors -> that means the risk factors are very broad
o Some risk factors are social, others are biological…
They lack specificity -> many factors aren’t explained
o There are multiple causes for re-offending -> you have
to make sure your risk factor is really causing criminality
Other causes for re-offending (confounds) must
be considered and ruled out with research!
The ‘grain’ problem: dynamic risk factors are often presented
at multiple levels ranging from a general domain to a more
concrete level -> how do you decide on the appropriate level of
abstraction when using factors in an explanatory context?
o What level of specificity is needed for your question?
Dynamic risk factors are seen as normative concepts: the
starting point is ‘normal or appropriate behaviour’
o Other problem: categorizing people as ‘low’ or ‘high’
risk is hard, because dynamic risk factors keep changing
In sum: it is difficult to predict and explain WHEN a dynamic risk factor
(DRF) will become active and lead to a crime, and WHY this happens
- There have been multiple generations of risk assessment tools throughout the years:
o First generation: unstructured professional judgement -> the ‘professional’
who is doing the risk assessment makes a random judgement, with only a few
interviews to guide him -> there was no underlying assessment instrument
Advantage: not expensive, flexible, convenient, accepted, no training
Disadvantage: lack of consistency and transparency, bias, accuracy: .50
We needed measures that were more objective and predictive
o Second generation: actuarial instruments: empirically based measures with a
focus on static factors -> the items needed to predict re-offending, but the
content didn’t matter that much -> re-offenders needed to react ‘differently’
Advantage: transparent/objective, good reliability + predictive validity
Disadvantage: a-theoretical, unchangeable, limited room for therapy
We needed measures that focus on reducing re-offending
o Third generation: structured professional judgement -> dynamic risk factors
Advantage: transparent, sensitive to change over time, good reliability
+ predictive validity and broad identification of treatment goals
Disadvantage: time consuming to examine changes + not so practical
