6th Milestones Meeting
Global Campaign for Violence Prevention
Mexico City, 13 November 2013
Towards Measurable Violence Prevention Targets:
Tracking Tests of Targeted
Interventions: Triple-T
Lawrence W. Sherman
Cambridge University &
University of Maryland
My Proposal
The United Nations should fund the WHO to
create a proactive global registry of every
rigorous test
of any intervention designed to prevent
violence, with equal emphasis on
1. What works
2. What does NOT work
3. Where, in context
Please Raise Your Hands
If you have ever heard
of the
Campbell
Collaboration
TESTS OF INTERVENTIONS
•
•
•
•
•
•
•
International Voluntary Effort
Hosted by Norwegian Government
Tiny Budget
High standards, Systematic Reviews
What Helps, What Harms?
Education, Social Services, Crime
www.campbellcollaboration.org
Many Interventions Prevent
Violence
• Many not even designed for violence
• Some that are opposed for violence
• Yet may work well on violence
• Such as restorative justice conferences
Restorative Justice Conference
2-3 Hour Meeting
•
•
•
•
•
Victims
Offenders
Family
Friends
Discuss three things:
1. What Happened?
2. Who was harmed and how?
3. What should offender do to repair?
Works Better for Violence
But Where Is The Evidence?
Where Do We Know, not just Assume,
that it Works?
• Mexico?
No
• Burma?
No
• Spain?
No
ENGLISH -SPEAKING COUNTRIES
ONLY:
Australia, UK, US
Are There Non-English Tests?
• Why don’t we know?
• 519 studies reviewed
• All in English
• Only10 met standards
• But others may exist
• Not in English
• Main obstacle:
$$$$$$$$$$$$$
My Proposal
The United Nations should fund the WHO to
create a proactive global registry of every
rigorous test
of any intervention designed to prevent
violence, with equal emphasis on
1. What works
2. What does NOT work
3. Where, in context
Definition of Evidence-Based
Internal Validity
External Validity
• Unbiased tests
• Replications
• Effective where tested • Different settings
• Well-specified model
• Same results?
What Global Means
• Not that interventions work everywhere
• But that they works somewhere
• Evidence on where they do, or NOT
• Develop understanding of WHY, e. size
• Frances Gardner, Oxford—cross national
First Domestic Violence Arrest
Experiment: 1980-84
• Minneapolis Police
• Controlled Test
• Arrest,
mediation,
separation
• Arrest reduced
repeat offending
by 50%
Domestic Violence:
1. Arrests deter employed men
2. Arrests make unemployed men more
violent
3. Arrests deter men in areas of low
unemployment
4. Arrests increase violence by men in
areas of high unemployment
Milwaukee: Repeat domestic
violence per 1,000 suspects per year
800
Employed
Subgroups Vary
Unemployed
Omaha: Repeat domestic violence
per 1,000 suspects per year
800
600
600
400
400
200
200
0
Arrested Warned
Arrested Warned
0
Employed
Arrested Warned
Miami: Percent of Offenders with
repeat domestic violence
20%
Employed
Unemployed
Arrested Warned
Arrested Warned
15%
10%
5%
0%
Unemployed
Arrested Warned
How Long Does it “Work”
• How long does it take to detect harm
• When we intervene with individuals
•
•
•
•
What should the standard of care be?
6 months?
2 Years?
23 Years?
What We Don’t Know...
•
•
•
•
•
•
•
Can kill us
And other people
Even when we are sincerely trying to help
But sincerity is not evidence
Fighting harmful practices is as important
As promoting helpful practices
Only TESTING can tell us the difference
Best—and Worst--Practices
Best Practices
• Nurse-family partners
• Pre-school partners
• Hot spots patrols
• Restorative justice
conferences
• Drug courts
• Juvenile diversion
• Problem-oriented
policing
Worst Practices
• Arrest for minor
domestic disputes
• Police second
responders to
domestic calls
• Scared Straight
• DARE
• CCTV to prevent
violence
Cambridge Somerville 1930s
•
•
•
•
•
•
•
10-12 yr-old boys
Random assignment
Mentors—monthly
Summer camp
Health care
30-year followup
Treatment kids died
sooner
• More crime
• More mental illness
Professor Joan McCord
Domestic abuse 2d Responders
•
•
•
•
•
•
•
•
•
After crisis call
3-10 days later
Police visit victim
With social worker
No offender home
Provide info
Follow-up
No effect on injury
More frequent DV
Chicago: Causing Crime
Effects of “Scared Straight” on
Crime: 7 RCTs
CCTV Effect on Violence:
Farrington Campbell Review
• Violence was reported in 23 evaluations,
but CCTV had a desirable effect in
reducing violence in only 3 cases (Airdrie,
Malmö, and Shire Town). Overall, there
was no effect of CCTV on violence.
• How much money spent on CCTV could
be spent on interventions that work?
Good News on Hot Spots
Policing—in US!!
Three new tests UK
• One under way in Caracas
• First district-level in Trinidad
• Marriage of epidemiology and intervention
• But still needs RCTs all over the world
• Trinidad murder rate 40 per100,000
Where Should Police Patrol:
Crime Peaks or Crime Valleys?
Triple-T Against Crime
Testing
Targeting
Tracking
EvidenceBased
Prevention
30
Homicides & Shootings in 40 Trinidad
Police Districts,
2 months Before and after September 2013:
20 Controls vs. 20 “TTT” Experimentals
My Proposal
The United Nations should fund the WHO to
create a proactive global registry of every
rigorous test
of any intervention designed to prevent
violence, with equal emphasis on
1. What works
2. What does NOT work
3. Where, in context
A PROACTIVE Registry
•
•
•
•
•
•
•
•
Seeking every day for newly reported tests
In 50-100 languages—(or at least 5!!)
Contacting authors, obtaining copies
Reviewing, coding, abstracting
Entering in data base
Studying external validity--generalizability
Making results available in 100 languages
A truly global project on what works
The Measureable Targets
1. Doubling the number of rigorous tests of
interventions to prevent violence every
five years, world-wide.
2. Doubling the number of tests every five
years in each country.
3. Doubling the number of countries that
have ever completed a rigorous test of
such interventions within five years.
Funding
• Much more than doubling
• 6 million deaths a year
• What we invest in world peace
• We can invest in violence prevention
6th Milestones Meeting
Global Campaign for Violence Prevention
Mexico City, 13 November 2013
Towards Measurable Violence Prevention Targets:
THANK YOU
Lawrence W. Sherman
Cambridge University &
University of Maryland
Descargar

rigorous test - World Health Organization