Treatment Components

Anna C. Salter, Ph.D.

How to Decide

Meta-analysis of factors related to recidivism Theory Skill Set of Staff Tradition

Key Research

Hanson & Bussiere Hanson & Morton-Bourgon Hanson

1998 2005 2009

Meta-analysis Hanson & Bussiere, 1998

61 Studies N = 28,972 Correlation coefficients

Meta-analysis Hanson & Morton-Bourgnon, 2005

Studies = 82
N = 29,450 35 from 1998 meta-analysis Mean difference

Traditional Clinical Assessment

P-graph deviant arousal pattern children Deviant arousal Personality disorders, e.g., APD Empathy Denial Family problems Psychological problems Sexual abuse as a child Social skills Substance abuse

Traditional Clinical Assessment

P-graph deviant arousal pattern children Deviant Sexual Preference Personality disorders, e.g., APD Empathy Denial Family problems Psychological problems Sexual abuse as a child Social skills Substance abuse

Yes Yes Yes No No No No No No No

Traditional Clinical Assessment

Deviant arousal pattern children Deviant Sexual Preference Personality disorders, e.g., APD Empathy Denial Family problems Psychological problems Sexual abuse as a child Social skills Substance abuse

.32 .22 .16 .03 .02 .08 0 -.01 -.04 .03

Clinical Vs. Actuarial Assessment

Dixon, 1974 Epperson, Kaul and Huot, 1995 Florida Dept. of Health & Human Services, 1984 Khanna, Brown, Malcolm & Williams, 1989 (Hanson & Bussiere, 1996)

Clinical vs. Actuarial Assessment

Reddon, Studer, and Estrada, 1995 Rice, Quinsey and Harris, 1989 Ryan and Miyoshi, 1990 Schram, Milloy and Rowe, 1991 Smith & Monastersky, 1986 Sturgeon & Taylor, 1986 (Hanson & Bussiere, 1996)

Clinical vs. Actuarial Assessment

r Clinical Assessment Actuarial Assessment .10 .46

(Hanson & Bussiere, 1998)

Accuracy of Clinical and Actuarial Risk Prediction

0.35 0.3 0.25 0.2 0.15 0.1 0.05 0 -0.05 a b c d e f g h i j k l m clinical actuarial

Hanson Clinical vs. Actuarial

1.

Variability of clinical great Some clinical worse than chance None of clinical showed results better than worse of actuarials (ns)

2.

3.

Meta-analysis Hanson & Morton-Bourgnon, 2005

Studies = 82
N = 29,450 35 from 1998 meta-analysis Mean difference

Value of d is approximately 2 times the correlation coefficient from same data

 Large

> .30 >.20 .10 <.10

Moderate

Small

Not

useful

Cohens d
.80

large moderate small

.50

.2

Hanson & Morton-Bourgon 2005

Categories
Sexual

deviancy Antisocial orientation Sexual attitudes Intimacy deficits Adverse childhood environment General psychological problems Clinical presentation

Categories

Sexual deviancy Antisocial orientation Sexual attitudes Intimacy deficits Adverse childhood environment General psychological problems Clinical presentation

.30 .23 .17 .15 .09 .02 -.02

Hanson & Morton-Bourgon 2005

Sexual Deviancy

Sexual interest in children Any deviant sexual interest Sexual preoccupation

.39 .31 .21

Hanson & Morton-Bourgon 2005

Anti-social

General regulation problems PCL-R Antisocial personality disorder

.37 .29 .21

Hanson & Morton-Bourgon 2005

Intimacy Deficits

Emotional identification with children Conflicts with intimate partners Social skills deficits Loneliness

.42 .36 -.07 .03

Attitudes towards sexual crime Child molester attitudes

Yes No

Hanson & Morton-Bourgon 2005

No go Childhood neglect or abuse Sexual abuse in childhood Lack of empathy Low self-esteem Loneliness Denial of sexual crime .10 .09 -.08 .04 .03 .02

The prototypic sexual recidivist is not upset or lonely; instead, he leads an unstable, antisocial lifestyle and ruminates on sexually deviant themes. (Hanson & Morton-Bourgon, 2005, p. 1158)

Hanson & Morton-Bourgon 2005

Not Correlated
Child

molester attitudes General psychological problems Sexually abused as a child Social skill deficits Loneliness Anxiety

Hanson & Morton-Bourgon 2005

Not Correlated
Depression Low

self-esteem Victim empathy Denial Minimization Poor progress in treatment

Survey of Treatment Programs

Community Programs

% Victim empathy 94.8 Social skills training 80 Family support networks 72.8 Arousal control 63.6 Antisocial attitudes ? (McGrath et al., 2003)

Survey of Treatment Programs

Residential Programs

% Social skills training 89 Victim empathy 86.8 Arousal control 59.6 Family support networks 43.3 Antisocial attitudes ? (McGrath et al., 2003)

Arousal Control
Com % Covert sensitization 50 Odor aversion 25 Masturbatory satiation 24 Aversive behavioral rehearsal 23 Verbal satiation 16 Minimal arousal conditioning 18 Orgasmic conditioning 16 One or more 63 Res % 48 18 19 18 14 19 19 60

Growing Consensus
Not correlated with general personality deficits But with Certain specific problems Sexual deviancy Antisocial attitudes Certain intimacy deficits

Motors & Brakes

Motor Faulty brakes

Sexual attraction to children Antisocial attitudes

Sexual Deviancy
Unrelated to personality traits Narcissistic or self-effacing Outgoing or introverted Depressed, anxious or hysterical Low self-esteem or high

Correlated Personality Traits

Recklessness Impulsivity Poor problem solving General regulation problems

Research has never found measures of general psychological adjustment, such as self-esteem, depression, or social competence, to be related to sexual offense recidivism (Hanson & Bussiere, 1998). Furthermore, treatment programs that improve general psychological adjustment do not result in reduced recidivism rates.(Hanson, Steffy, & Gauthier, 1993; Nicholaichuk, 1996).

The distressed offenders are at no greater recidivism risk than the happy offenders, but both types of offenders are at increased risk when their mood deteriorates. These results suggest that therapy should focus on weakening the association between negative affect and sex offending rather than on generally improving the offenders psychological adjustment. (Hanson, 2000, p. 34-35)

Research Suggests
Treatment programs should address Sexual deviancy Antisocial attitudes & beliefs Certain intimacy deficits: Emotional identification with children, Conflicts with partners

Research Suggests
Not including
Empathy Social

skills Substance abuse Personal distress variables Anxiety Depression Low self-esteem

Meta-analysis: Risk, Need & Responsivity

Hanson, Bourgon, Helmus & Hodgson, 2009

Risk-Needs-Responsivity

Focus on high risk offenders Target criminogenic needs Use cognitive behavioral methods tailored to individual learning style

Targeting Criminogenic Needs

Criminogenic Needs
Criminogenic Antisocial Attitudes Antisocial Friends Substance Abuse Impulsivity Non Criminogenic Self-Esteem Anxiety Depression

Targeting Criminogenic Needs

0.35 0.3 0.25 0.2 0.15 0.1 0.05 0 -0.05 Targets 1 - 3 Noncriminogen ic Needs Targets 4 - 6 Criminogenic Needs

Gendreau, French & Taylor, 2002

Self Esteem Vs. Criminogenic Needs

0.4 0.35 0.3 0.25 0.2 0.15 0.1 0.05 0 -0.05 -0.1

Self Esteem Criminogenic Needs

Selection of Studies
Total = 130 Accepted 23 Accepted 18 weak 5 good

Adherence to RNR Principles

Risk: Little or no service to low risk (Tx participants of higher than average risk)

Needs: Sexual deviancy, antisocial attitudes, sexual attitudes, intimacy deficits Noncriminogenic: denial, empathy, social skills
Responsivity: Cognitive behavioral with firmbut-fair therapists

Hanson et al., 2009

Recidivism Sexual Any Violent Treated % 10.9 31.8 22.9 NS Untreated % 19.2% 48.3 32 NS

Of the three RNR principles, attention to the Need principle would motivate the largest changes in the interventions currently given to sex offenders. . . An empirical association with recidivism is a minimum criterion for a factor to be considered a potential criminogenic need. . . Many of the factors targeted in contemporary treatment programs do not meet this test. Offense responsibility, social skills training, and victim empathy are targets in 80% of sexual offender treatment programs . . . Yet none of these have been found to predict sexual recidivism. (Hanson et al., 2009, p. 25)

Tx Effectiveness & RNR Model

Number of Principles Adhered To Odds Ratio

None
1 2 3

1.17 .64 .63 .21

Any Recidivism

Responsivity Number (Fixed effects)

No effect

Sexual & Violent

If there is anything to be learned from the broad debate over the effectiveness of correctional rehabilitation, it is that not all interventions reduce recidivism. (Hanson et al., 2009)

Attack on Risks/Needs/Responsivity
Lack of unifying power and external consistency Lack of fertility with respect to treatment guidance Lack of explanatory depth Incoherency; lack of scope; incomplete rehabilitation theory (Ward et al., 2006)

Focus of Treatment Reduction of Recidivism?

We believe that treating sexual offenders also involves taking into consideration human welfare issues, as well as recidivism issues.
(Ward et al., 2006, p.269)

Individuals who are assessed as low risk may exhibit a number of significant problems that adversely impact on their functioning, for example, low mood or relationship conflict. While such problems may not be criminogenic needs, individuals could still benefit from therapeutic attention. (Ward et al., 2006, p. 269)

Risk needs model has resulted in the development of a suite of empirically derived and effective treatment for a range of crimes, including sexual offending. (Ward & Hudson, 1997)

The difficulty is that in the absence of a theoretical analysis we do not know why. (Ward et al., 2006, p. 270)

Reply from Bonta and Andrews

We welcome such a debate on the relevance of various models to explain criminal behaviour . However, that debate should be structured by respect for evidence.

Theoreticism is the acceptance or rejection of knowledge in accordance with ones personal view and not in accordance with evidence. (Bonta & Andrews, 2003, p. 215)

Here we see theoreticism operating at its best. Ignore the evidence that reductions in criminogenic needs are associated with reduced criminal behaviour, turn a blind eye to the fact that there is not a shred of evidence that psychodynamic interventions reduce recidivism and simply assert that your approach makes the most sense.

Ward and Stewart appear to be arguing for a return to the good old days when treatment providers relied on nondirective, relationship-oriented techniques to build feelings of well-being. (Bonta & Andrews, 2003, p. 217)

Impact of Appropriate Vs. Inappropriate Treatment (Andrews, 1998)

0.35 0.3 0.25 0.2 0.15 0.1 0.05 0 -0.05 -0.1

Combined Tx Appropriate Inappropriate Sanctions

Type of Treatment
0.3 0.25 0.2 0.15 0.1 0.05 0 Non Behavioral Cognitive Behavioral

Andrew, 1994

Type of Treatment & Young Offenders

0.25 0.2 0.15 0.1 0.05 0 Non Behavioral Cognitive Behavioral

Dowden & Andrews, 1999

Appropriate Treatment Targets

Sexual deviancy Sexual pre-occupation Low self-control Grievance thinking Lack of meaningful adult relationships (Hanson & Morton-Bourgon, 2004)

Proponents of Personal Distress Variables

Some combinations of these experiences lead the emerging sexual offender to acquire: a low sense of self-worth; a failure to internalize the confidence, attitudes, and skills necessary to meet his needs prosocially; and a selfinterested disposition or a sense of entitlement. (Marshall, Marshall, Serran, & Fernandez, 2006, p. 15)

Theory of Low Self-Esteem

Low Self-esteem = Seeking sex from non-threatening partners
Seeking coerced sex

Research on Low Self-Esteem

Study 1 Study 2 Child molesters lower Child molesters normal range (Marshall et al., 2003)

Contrary Research
People with low self-esteem Do not typically undertake novel activities That require persistence (Baumeister et al., 1989)

Research on Self-Esteem
Rapists and nonsexual offenders No differences (Fernandez & Marshall, 2003)

With self-esteem we encourage and facilitate the expression of behaviors such as engaging in social and pleasurable activities, as well as verbal (or subvocal) behaviors such as complementing themselves when they do things that deserve rewards and repeating positive self-statements throughout each day. (Marshall et al., 2006, pp. 28-29).

Outcome of Self-Esteem Treatment

Increased self-esteem on inventories Reductions in loneliness Increases in intimacy skills No increase in victim empathy (Marshall et al., 1997)

Thornton , Beech & Marshall 2004

Pre-treatment self esteem correlated with recidivism All offenders in community sample in treatment 85% graduation rate Self-esteem improved during treatment

Why Treat Self Esteem?

Self-esteem after treatment Self-esteem before treatment ? Recidivism

Conclusions: Treating self-esteem did not change correlation between pre-treatment selfesteem and recidivism (Info not available on incarcerated sample.)

Does Treatment of Self-Esteem Reduce Deviant Arousal

Offenders with low self-esteem & deviant arousal This required the processing of a substantial number of offenders before a sufficient number of participants was detected who met criteria. (Marshal, 1997, p. 88)

Our theory, then, suggests that the sexual aspects of child molestation may not be central to the motivational forces that drive these offenders. (Marshall, p. 87)

Subjects
Female

victims only

Pre-pubescent

Out

of home 2 had more than 3 victims

Only

Self-esteem Reduce Deviant Arousal?

Attending treatment program With Cognitive distortions, relapse prevention, self-esteem & other interventions Minus deviant arousal (Marshall, 1997)

What Causes Recidivism? Low Self-esteem or High

Today, it is common to propose that low selfesteem causes violence, but the evidence shows plainly that this idea is false. Violent acts follow from high self-esteem, not from low self-esteem. This is true across a broad spectrum of violence, from playground bullying to national tyranny, from domestic abuse to genocide, from warfare to murder and rape. Perpetrators of violence are typically people who think very highly of themselves. (Baumeister, 1997, pp. 25-26

Self-Esteem & Violence Baumeister Argument

Contradictory Claims Toch Self-doubts and a sense of inadequacy Exaggerated self-esteem

Self-Esteem & Violence Baumeister Argument

Higher in men than women but women < violence Depressions is correlated with self-esteem Depressed individuals do not commit more violence Psychopathy correlated with violence Grossly inflated self-esteem

Self-Esteem & Violence Baumeister Argument

High self-esteem not correlated with violence Grossly inflated and unstable self-esteem

Stability of Self-esteem

High but unstable self-esteem = highest level of self-reported angry/hostile responses High but stable self-esteem = lowest rates (Kernis et al., 1989)

Grandiosity & Aggressiveness

Grandiosity correlated with aggressiveness (Wink, 1991)

Dominance and hostility correlated with grandiosity and narcissism (Novacek & Hogan, 1991)

High self-esteem based on grandiosity (not achievement) Disconfirming evidence = Negative reaction

(Baumeister,1997)

High but unstable self-esteem most likely to respond defensively to negative feedback. (Kernis, Cornell, Sun, Berry and Harlow, 1993)

Why Do We?

Despite empirical evidence still treat Self-esteem Social skills Offense responsibility

Tough or Tender Minded William James

Tender minded

Tough Minded

Principled Idealistic Optimistic Dogmatical

Empirical Pessimistic Skeptical

Lost Souls or Predators

What to Treat?
Self-esteem?

Instability

of self-esteem? Mood deterioration? Sexualizing mood deterioration?

What to Treat
Sexual Deviancy Behavioral reconditioning Minimal arousal conditioning Covert sensitization Olfactory satiation Medication

What to Treat
Antisocial Attitudes & Beliefs