10 February 2014 Lawrence Sherman & Heather Strang University of Cambridge
Feb 23, 2016
10 February 2014Lawrence Sherman & Heather Strang
University of Cambridge
Attempted Murder in Australia
• Canberra, 1997• Teeth Knocked Out• Loss of Blood• Arrest, admissions• Police divert from
prosecution• Police convene RJ
conference• Offender agreed • Victim had raped
assailant’s girl friend
RJ Conference,No Prosecution
Who Was There?
• Victim of Assault• Offender• Friends of Each• Priest of both• Police Officer
What Happened?
• Victim raped friend• Offender revenge• Teeth replacement?• Compensation asked• Refused • Priest suggestion • Stay away 1,000 metres• No arrests for 5 years
Criminal Law Status
• No charges filed (prosecution)• No criminal record (conference noted)• No contact with prosecutor• No probation, surveillance• Minimal followup by police
No further crime
What is Face-to-Face Restorative Justice Conference (RJC)?
– A process whereby all the parties with a stake in a particular offence come together to resolve collectively how to deal with the aftermath of the offence and its implications for the future
– A response to crime that emphasises repair of harm (restoration) over infliction of more harm (retribution)
– A major event done with advance planning
How can restorative justice be used?
• As a diversion from prosecution, normal court justice• In addition to normal court justice
– Pre-sentence– As a sentence (if victims consent, offenders safe)– Post-sentence
For juveniles and adultsFor trivial crime and very serious crimeFor direct and indirect victimsAt different points in the justice system
New Law: England-Wales
• November 2013• Royal Assent, Statutory Authority• Adjournment of sentencing for RJC • Crime and Courts Act 2013 chap. 22, schedule
16, Part II: “Deferring the passing of sentence to allow for restorative justice.”
• Home Secretary gave £3 million to PCCs to fund RJ—same month
Grievous Bodily Harm 2002: London Cab Driver Beaten, Robbed
HMP Holloway: Prison for Women A Tale of Two Victims
Holloway Women’s Prison, London
Two Histories
Natalie• Then Age 21• Raped age 19• Raped age 21• Sexual assault age 8• 25 prior arrests• 4 robberies• 1 prison term
Carol• Then age 56• No prior
victimizations• No prior crimes• Held on to purse• Hit on head• 70 stitches
One Victim Helped
• Back to work as a nurse (after 5 months lost)
• Resumed her normal life
One Victim Not Helped
• 5-year prison sentence
• Mother died• Released at 2.5
years• 47 days later re-
arrested• Charged, again,
with robbery
But these are stories..
• ANECDOTE, not evidence
But these are stories..
• ANECDOTE, not evidence
• For evidence we need experiments
But these are stories..
• ANECDOTE, not evidence
• For evidence we need experiments
• For experiments we need fair comparisons
But these are stories..
• ANECDOTE, not evidence
• For evidence we need experiments
• For experiments we need fair comparisons
• Comparisons tell us cost-effectiveness
Cost-Effective Justice7 UK (English) Experiments
SITE RJ COST Total Benefit Total Ratio London 598,848 8,261,028 1:14 Northumbria 275,411 320,125 1:1.2 Thames Valley 222,463 461,455 1:2 Total 1,096,722 9,042,608 1:8 * Computed from Shapland et al, 2008. All amounts expressed in Pounds Sterling**CJ benefit estimated at an average 22% of total costs of crime
How Do You Calculate Benefit?
• Compare cost of crime after RJ
• To cost of crime after not having RJ
• Identical kinds of cases
• Not “similar”—IDENTICAL
Canberra:Did RJ Program Cause Crime Drop?
71% 71%
49%
100%
20%
40%
60%
80%
100%
120%
140%
Yr (-2) Yr (-1) Yr (+1) Yr (+2)
Line 1
DCOffenders (n =62)
“Post hoc, ergo propter hoc”
• Logical Fallacy
• Known since Classical era
You Are Just Saying:
After this, therefore because of this
Canberra:Did RJ Program Cause Crime Drop?
71% 71%
49%
100%
20%
40%
60%
80%
100%
120%
140%
Yr (-2) Yr (-1) Yr (+1) Yr (+2)
Line 1
DCOffenders (n =62)
Cause lies not in a Trend,but in aCOMPARISON or NET difference
101%
121%
71% 71%
28%
49%
100%
20%
40%
60%
80%
100%
120%
140%
Yr (-2) Yr (-1) Yr (+1) Yr (+2)
CourtOffenders (n=59)
DCOffenders (n =62)
Randomized Controlled Trials RCTs: COMPARISON to WHAT ?
• Identical Conditions without program
• But no other conditions are truly identical
• Comparison units may differ in major ways
• How to make them as similar as possible?eliminate ”plausible rival hypotheses”
• Large samples, randomized controlled trials (RCTs)
RCT Experiments versus Observations
• Manipulate one thing at a time
• Control all other things
• Compare two different manipulations
• Repeat comparisons again and again
“Evidence-Based” Practice
• Originally described new rigour in medicine
• Limited to RCT evidence
• Not just “observed” correlations--coincidences(1950s divorce rates rose as imports of apples rose—strong correlation)
• But “manipulated” between two identical groups
What sort of RJ has been tested with randomised controlled trials?
• Face to face RJ conferences (RJC) between crime victim and offender
• In the presence of a trained facilitator• And of their supporters (family and friends) and
others affected• Offender must have accepted responsibility for
offence• Direct discussion between victim and offender
focused on feelings rather than facts• May be either instead of formal justice processes or
in addition to them
Testing RJC with Experiments
• Over ten years, 10 RCTs conducted on RJ in Aust + UK + US– Different offences– Different offenders– Different locations– Different points in the justice system
• Objective to field-test as broadly as possible• Equal probability of assignment: court as usual compared with
diversion to RJ (Australia) or court as usual compared with court plus RJ (United Kingdom)
• Outcomes measures: reoffending and victim satisfaction
10 RCTs Comparing Victim-Present RJCs with Conventional Justice
Offender s
1. Australia <30 years violence (diversion)121
2. Australia juvenile personal property (diversion) 249
3. US Indianapolis juvenile property/violence (diversion) 7824. UK juvenile property/violence (police Final Warning)
1655. UK adult property (Magistrates Courts)
636. UK adult assault (Magistrates Courts) 447. UK robbery (Crown Courts)
1068. UK burglary (Crown Courts)
1869.. UK violence – Probation
6310. UK violence – Prison
103
Total offender N = 1882
Free Download: Just Google “Strang Restorative Justice Campbell”
What Does Our Campbell Collaboration Review of These Studies Conclude?
Offenders – • Slows some down, others stop reoffending completely while others are unaffected• May be better for the most prolific offenders
On average 27% reduction in repeat convictions across British trials
Offence Types• Works better for violence than property offences• Wasted on minor offences?
Victims • Unequivocal evidence on greater benefit for most of those willing to meet their offenders
PART I: Effects for Victims
BENEFITS:
• participation • information• fairness and respect• apologies (and sometimes forgiveness)
EFFECTS:
• Fear, Anger, Sympathy for Offender• Post-Traumatic Stress measures• Desire for Retaliation• Satisfaction With Process
Interviews With Victims in 4 Sites
• After Restorative Conferences
• 1995-2003
• Robbery, • Assault, • Burglary, • some other property crimes
Percentage of victims angry before/after meeting. Preliminary UK & Australia
Percentage of victims sympathetic before/after meeting. Preliminary UK & Australia
Percentage of victims afraid before/after RJ meeting. Preliminary UK & Australia
Findings on Victim Post-traumatic Stress
• London Crown Courts
• Burglary & Robbery
• Most meetings in prisons
• Telephone interviews over several months
• Standard scale to measure psychological trauma
• Dr. Caroline Angel, forthcoming article, J Exp. Criminology
Average level of Victim Post Traumatic StressBoth Robbery & Burglary
p ≤ 0.010
RJ Helps Women Victims PTSS More
Victim Benefits
Compared with conventional justice, RJ provides
• significantly higher victim satisfaction than court justice
• significantly higher levels of apology• significant greater reduction in desire for
revenge • significantly greater reduction (approx 40%) in
post-traumatic stress symptoms (PTSS) for robbery and burglary victims.
PART II: What effect do face-to-face RJ conferences have on re-offending
• Frequency of reconvictions– across different points of criminal justice process– with personal victims intended to be there?
• ANSWER: 9 out of 10 tests with personal victims show less crime for RJ than CJ (Australian juvenile property experiment failed for RJ)
• NB especially results for prison and probation experiments
Personal Victims Present: % Change* in Reconviction Frequency
*per person, RJ Group compared with Control Group
46%
-52%
-8%-16%
-48%
-24%-41%
-33%
-55%
-23%
-80%
-60%
-40%
-20%
0%
20%
40%
60%
80%
CBR1 CBR3 LR5 LR6 NU7 NU8 NU9 TV10 TV11 IND12
What about RJC and violent crime?
• Frequency of Reconviction• In Violent Crime Experiments• Youth and Adult Combined• All levels of seriousness from simple assault to grievous
bodily harm
• ANSWER: 5 out of 5 violence tests show less crime for RJ
Percent Change* in the Frequency of Reconviction – Violence Experiments
*per person, RJ Group compared with Control Group
What about RJC and property crime?
What Effect Does RJ have on
• Frequency of Reconviction on• Property Crime Offenders• Youth and Adult Combined
• ANSWER: 3 out 4 tests show less crime for RJ• Effects not as big, or as prevalent, as for violence
• RJ WORKS BETTER FOR MORE SERIOUS CRIME
Percent Change* in the Frequency of Reconviction – Property Experiments
*per person, RJ Group compared with Control Group
46%
-16%-24%
-40%
-20%
0%
20%
40%
60%
80%
CBR1 LR6 NU8
What about RJC and Youth Crime?
What Effect Does Face-to-Face RJ Have on
• Youth Crime • Property and Violent• US, UK, Australia?
• ANSWER: 3 out of 4 tests show less crime for RJ
Percent Change* in the Frequency of Reconviction – Youth Experiments
*per person, RJ Group compared with Control Group
What about RJC and adult crime?
What Effect Does Face-to-Face RJ Have on
• Adult Crime
• Property and Violent
• US, UK, Australia?
• Answer:
• 6 out of 6 tests on adults = less crime after RJCs
Percent Change* in the Frequency of Reconviction – Adult Experiments
*per person, RJ Group compared with Control Group
Summary of RJ Effects, compared with conventional justice
• Violent Crime: Biggest, clearest effects of RJ •Property Crime: mixed, not so big
• RJ appears more effective than court alone post-sentence than pre-sentence
• RJ better for women victims than men (but good for both)
• UK: RJ Cost effective for government costs
• Across all studies: – significantly fewer crimes– Significantly better for victims
Policy Implications for RJCs with Serious Crimes
• Investment of RJC in more serious crimes (when victims want to do it)• Particularly Cost-effective in reducing reoffending after conviction in court and prior to sentencing (English law) • Need more tests of RJC at point of prison release (weak statistical power in our research) - but promising here
• Need a new RCT with serious crime for a new jurisdiction
• 10 February 2014
THANK YOU
Lawrence Sherman & Heather StrangUniversity of Cambridge