1. Youth Violence Prevention and
Intervention: An overview
Chris Melde, Ph.D.
School of Criminal Justice
Michigan State University

2. Violence in Perspective
 The Long View of Crime
 National versus State and Local Trends
 Distribution of Risk
 Not random
 Perceptual Indicators and Resulting Behaviors
 Fear and Avoidance
 What makes youth violence unique?

3. Overall Violent Crime Rate (per
100,000): 1960 to 2010 UCR
Violent Crime Rate 1960 to 2010
800
700
600
500
Rate per 100,000
400
300
200
100
0

4. Murder Rate (per 100,000):
1960 to 2010 UCR
Murder Rate: 1960 to 2010
12
10
8
Rate per 100,000
6
4
2
0

5. Overall Violent Crime in Illinois:
2001 to 2010 UCR Data
Overall Violent Crime in Illinois
1600
1400
1200
1000
Rate per 100,000
State of Illinois
800 City A
City B
600 City C
City D
400
200
0
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010

6. Overall Violent Crime in Select Illinois
Cities (not named Chicago)
Overall Violent Crime in Illinois by City: 2006 to 2010 UCR
1600
1400
1200
1000
Rate per 100,000
800
600
400
200
0
2006 2007 2008 2009 2010

7. Crime in the Media
Milwaukee posts 4th largest drop in crime
12% decline in violence near top for big cities
From the June 8, 2005 editions of the Milwaukee Journal Sentinel
By CHASE DAVIS
cdavis@journalsentinel.com
Violent crime in Milwaukee plunged 12.3% last year, the fourth largest drop among the
country's 33 largest cities, according to a new FBI report.
More than 4,600 violent crimes - homicides, aggravated assaults, rapes and robberies -
were reported in Milwaukee in 2004, down from nearly 5,300 in 2003, the FBI said.

8. Crime in the Media: One year later
Sunday, Dec. 03, 2006
Middle America�s Crime Wave
By Kathleen Kingsbury
It's as if Milwaukee, Wis., had reverted to a state of lethal chaos. A Special Olympian is
killed for his wallet as he waits for a bus. An 11-year-old girl is gang-raped by as many as 19
men. A woman is strangled, her body found burning in a city-owned garbage cart. Twenty-
eight people are shot, four fatally, over a holiday weekend.
These are the kinds of crimes American cities expected never to see in high numbers again.
In the 1990s police departments nationwide began applying the so-called broken-windows
theory: arrest the bad guys for minor offenses, and they wouldn't be around to commit more
serious ones. This zero-tolerance approach--combined with more cops on the street to
enforce it, a strong economy and a fortuitous demographic change that reduced the
population of young men who typically cause the most trouble--lowered the rates of murder,
robbery and rape for 10 consecutive years. Until last year. Not only did crime suddenly
begin to rise in 2005, but the most violent crimes led the trend. Homicides shot up 3.4%.
Robberies, 3.9%. Aggravated assaults, 1.8%. Hardest hit were not metropolises like New
York City and Los Angeles but cities with populations between 400,000 and 1 million--such
as Baltimore, Md.; Charlotte, N.C.; St. Louis, Mo.; and Oakland, Calif.--and this year looks
to see similar rates of increase, if not worse.
Few places have suffered more than Milwaukee. The homicide count for the city of 590,000
fell from 130 in 1996 to just 88 in 2004. But last year, according to FBI figures, Milwaukee
saw the country's largest jump in homicides--up 40%, to 121.

9. The Reality:
Back to the 10-year average (20.1)

10. Distribution of Risk
The non-random nature of the distribution of violence risk.

11. Pin map of all homicides in Newark, New
Jersey, 1982-2008

12. The spread of homicide risk in
Newark, New Jersey, 1982-2008

13. Percent Black by Census Tract in
Newark, New Jersey

14. Violent Victimization by Sex
Overall Violent Victimization Rate (per 1,000) by Sex: NCVS 2009
20
19
18
17
16
15 Male
Female
14
13
12
11
10
Male

15. Violent Victimization by Race
Overall Violent Victimization Rate (per 1,000) by Race: NCVS 2009
45
40
35
30
25
Total Violent Victimization
20
15
10
5
0
White Black Other Multi-Racial

16. Violent Victimization by Ethnicity
Overall Violent Victimization Rate (per 1,000) by Ethnicity: NCVS
2009
20
19
18
17
16
15 Hispanic
Non-Hispanic
14
13
12
11
10
Hispanic

17. Violent Victimization by Age-Group
Total Violent Victimization rate (per 1,000) by Age: NCVS 2009
40
35
30
25
20
15 Total Violent Victimization
10
5
0
12-15 16-19 20-24 25-34 35-49 50-64 65 and
over

18. Focusing on Youth Violence
 Why?
 Risk - Power Few
 Because the public is especially concerned
• Cycle of Juvenile Justice
 Malleability/Potential for Change
 Focus on ages 12 to 24
 While we often think of “youth” from a legalistic
standpoint (i.e., under age 18), that is not a realistic
representation.
 Adolescence

19. The Cycle of Juvenile Justice
 The primary philosophy guiding the creation of the
juvenile justice system was rehabilitation and
reintegration.
 Juveniles are not wholly responsible for their actions due
to their maturity level.
 There is time to reform their behavior, and create
productive citizens.
 This philosophy is easy to convey when juvenile
crime is not in the spotlight.

20. The Cycle of Juvenile Justice
 What happens when juvenile crime is perceived
to be extraordinarily high?
◦ A large contingent of people lobby for harsher
treatment of juvenile delinquents.
◦ The common sentiment: leniency “encourages
juveniles to laugh at the system, to believe they will not
be punished no matter what they do, and to feel free to
commit more frequent and serious crimes”
(Bernard, 1992, p. 37).

21. The Cycle of Juvenile Justice
 The Result
 Lawmakers impose harsher penalties, including
mandatory transfer to adult court for serious
crimes, and extended sentences; a „get tough‟ on
crime agenda.
 What typically happens to the perceived
juvenile crime rate in the face of „get tough‟
policies?

22. The Cycle of Juvenile Justice
 Chasing our tails
 Because our juvenile crime rate is, and always has
been, relatively high compared with adults (with the
exception of young adults 18-24), we change the
system once again.
 The juvenile justice philosophy returns to its‟ roots, by
focusing on rehabilitation and reintegration.
 The question isn‟t whether we‟ll have another moral
panic, according to Bernard (1992), but rather when.

23. Significant Supreme Court Rulings
 Roper v. Simmons (2005)
 No Juvenile Death Penalty
 Juveniles fundamentally different than adults
 Graham v. Florida (2010)
 Holding: Sentencing a juvenile to life in prison
without parole for crimes other than murder violates
the Eighth Amendment‟s ban on cruel and unusual•
punishment.

24. Supreme Court: No more life sentences
for juvenile killers
 Monday, June 25, 2012
 The Supreme Court says it's unconstitutional to
sentence juveniles to life in prison without parole for
murder, siding with the petitioner in case of Miller v.
Alabama.
 The high court on Monday threw out Americans'
ability to send children to prison for the rest of their
lives with no chance of ever getting out. The 5-4
decision is in line with others the court has
made, including ruling out the death penalty for
juveniles and life without parole for young people
whose crimes did not involve killing.

25. The Cycle of Juvenile Justice
 Back to Prevention and Intervention
 Prevention programs have a long, but not so
glorious, history.
 Most prevention programs do not “work.”
• “But if it helps just one child”
• Difficult for any single program to produce change.
 On average, schools provide roughly 14 different
delinquency prevention programs in any given
year.
 Schools provide ready access to youth
 Progression: Didactic models to skill development

26. Why Get Involved in School-Based
Prevention?
The Risks and Rewards of School-Based Programs

27. NCVS Youth Violence Data
(ages 12-18): 1992 to 2010
Youth Violence at School vs. Away from School (per 1,000)
80
70
60
50
Rate per 1,000
Non-School Serious
40
Non-school total
School Serious
30 School total
20
10
0
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010

28. Adolescent Fear of Crime by Location:
1995-2009
Percentage of students age 12-18 who reported being afraid of
attack or harm, by location, 1995-2009
14
12
10
Percent of students
8
Total At School
6 Total Away from School
4
2
0
1995 1999 2001 2003 2005 2007 2009

29. Student Avoidance Behaviors
Avoidance Behaviors by Students at School (percent of students):
NCVS
10
9
8
7
Percent of students
6 Total
Avoided School activities
5 Any activities
Any class
4
Stayed home from school
3 Avoided one or more places in school
2
1
0
1995 1999 2001 2003 2005 2007 2009

30. Popular School-Based Programs
 Olweus Bully Prevention
 Blueprints Program
 http://www.stopbullying.gov/laws/index.html
• Page with information on Illinois State Laws on Bullying
 Gang Resistance Education and Training (G.R.E.A.T.)
 National Evaluation (2006-2012)
 Slight reduction in gang membership
 Improved attitudes about the police
 More negative views about gangs
 Drug Abuse Resistance Education (DARE)
 Has not worked
• possible negative effects for low risk, positive effects for high risk
youth.
 New iteration of the program

31. Risks of School-Based Programs
 Iatrogenic Effects
 Mixing high and low risk youth in programs has been
associated with more anti-social behavior on the part
of youth considered to be at low-risk.
• Evidence of this in community and correctional setting as well
(Hennigan and Maxson, 2012; Dodge et al., 2006).
 Priorities of the School
 Education versus Prevention
• Especially in high risk schools
 Long-term commitments are difficult

32. Module 2: Community Interventions
Background on Research Partnerships
Evidence of Impact
32

33. Continuum of Promising Practices for
Comprehensive Gang Intervention
SUPPRESSION SOCIAL SERVICE
DETERRENCE CAGI INFORMAL
Boston Ceasefire “Spergel Model” Chicago Ceasefire
“Pulling Levers” “Violence Interrupters”

34. Traditional Research Model
 Researchers were outsiders in problem-solving
process
 Not involved in problem identification
 Observers, not participants, in program development
and implementation
 Involved only as independent evaluators of impact
• What went wrong? (i.e., the complainers)
– You should have done x, y, and z.
34

35. Action Research Model
 Active, ongoing partnership between researchers
and practitioner agencies
 Use research process to help solve local
problems
 Data collection to identify and understand problems
 Strategic analysis to develop targeted interventions
 Program monitoring and feedback for refinement
 Assessment of impact
35

36. Data-Driven Problem Analysis
 Gather data on the selected crime
problem, including its
sources, victims, offenders, and settings
 Analyze the data to identify specific aspects and
components of the problem
36

37. Focused Interventions and
Linking to Evidence-Based Practice
Research facilitates:
 Developing focused interventions aimed at
reducing the specific sources and components of
the crime problem
 Implementing these focused intervention
strategies utilizing the resources and expertise of
the working group partners
 Basing interventions on “best” practices and
“promising” strategies
37

38. Monitoring, Feedback, and Evaluation
 Monitor the implementation of the
interventions
 Provide constant assessment and feedback
on the conduct and effects of the interventions
 Modify and refine the interventions based on
feedback assessments
 Evaluate the impacts of the interventions on
the service delivery system and on the
targeted crime problem
38

39. Background on Research Partnerships
Evolution from
 Boston Ceasefire (The Boston Miracle)
 Strategic Approaches to Community Safety
Initiative (SACSI)
 Project Safe Neighborhoods (PSN)
 Drug Market Initiative (DMI)
 Smart Policing and Intelligence Led Policing
39

40. Boston Ceasefire Foundation
Two Aspects to Boston Ceasefire
 Focused deterrence, “pulling levers” strategy
 Systematic problem solving process
 Multi-agency working group
 Problem solving model
 Police-researcher partnership
40

41. Boston Ceasefire: Focus on Youth Violence
Small Proportion Strategies
Youths focused on
Problem Analysis
Involved street gangs those at highest
and crews risk for violence
41

42. Intelligence Led Problem Solving
Problem
Analysis
Assessment Violence Strategy
and Feedback
Problem
Implemen-
tation
42

43. Evidence of Impact: Boston Ceasefire
 Two one-half years without youth homicide
 Homicides decreased 63%
 Calls for Service decreased 32%
 Gun assaults decreased 25%
43

44. Evidence of Impact: SACSI
Ten city initiative
 SACSI cities experience a significant decline in
violent crime when compared to non-SACSI
cities
 Indianapolis showed a decrease in homicide and
gun assaults of 35-40%
44

45. Reducing Homicide Risk: Indianapolis
Homicide victimization risk by Group
per 10,000 residents
152.1
160
140
112.9
120
100
80 66.4
60 45.6
40 26.1
14.8 14.9 18.2
11.5
20 5.1 2.2 4.5 3.5 2.6
0
All 15-24 year Young white Young white Young black Young black Young black All other victims
old victims female victims male victims female victims male victims male victims in
five hotspots
Pre-IVRP Post-IVRP
45

46. Boston and Indianapolis as Examples
 Research identified the highest risk
individuals, groups, and contexts
 Hard work by task force partners (criminal justice
officials, social services, community groups) then
focused strategies on highest risk
46

47. Findings from Boston and SACSI
Successes resulted in development and incorporation of
strategic problem solving model in a series of major
DOJ initiatives since 2001
 Project Safe Neighborhoods (PSN)
 Comprehensive Anti-Gang Initiative (CAGI)
 Drug Market Initiative (DMI)
 Smart Policing Initiative (SPI)
47

48. Project Safe Neighborhoods
 Nation-wide DOJ program intended to reduce gun
crime in America
 94 separate programs, one for each US Attorney
Office in the 50 states and territories
 Based on the Action Research/Strategic Problem-
Solving Model
 Funding provided for a local research partner to
work with each PSN task force
48

49. PSN Impact – Stage One
Series of site specific case studies
 Ten tests of impact on gun crime
 Case studies
 Reductions in gun crime in all ten sites
 Impact in two of these studies was equivocal
49

50. Summary of Case Studies
Project Exile Strategic Problem Solving
• Montgomery • Lowell
• Mobile • Omaha
• Greensboro
• Winston-Salem
• Chicago (Papachristos et al.)
• Stockton (Braga)
• St. Louis
– (significant but also drop in
comparison sites)
• Raleigh
– (reduction but not significant)
50

51. Summary of Case Studies
 Sites chosen because of evidence that PSN was
implemented in rigorous fashion
 Thus, results not generalizable but suggest that PSN
may have an impact where effectively implemented
51

52. PSN Impact – Stage Two
 Assess impact of PSN in all U.S. cities with
populations of 100,000+
 Trend in violent crime 2000-01 compared to 2002-06
 Compare PSN target cities with non-target cities
 Compare cities by level of PSN implementation
dosage
52

53. Measuring Implementation
 Step One – composite measure
 Research integration
 Partnerships
 Federal prosecution
 Step Two – focus on level of federal prosecution
 per capita and level increase
53

54. Step One
 PSN target cities in high implementation districts
experienced significant declines in violent crime in
comparison to cities in low implementation districts
and non-target cities
54

55. Step Two -Violent Crime Trends in PSN Target
Cities by Federal Prosecution Level
1150
1100
Medium Prosecution
Violent Crime Rate per 100,000 Population
1050
1000
Low Prosecution
950
900 High Prosecution
850
2000 2001 2002 2003 2004 2005 2006
High prosecution sites (n = 26) Medium prosecution sites (n = 29) Low prosecution sites (n = 27)
55

56. Violent Crime Trends in Non-PSN Sites by
Federal Prosecution Level
800
Medium Prosecution
750
Violent Crime Rate per 100,000 Population
700
Low Prosecution
650
600
550
500
High Prosecution
450
400
2000 2001 2002 2003 2004 2005 2006
High prosecution sites (n = 90) Medium prosecution sites (n = 42) Low prosecution sites (n = 38)
56

57. PSN Impact
Level of PSN PSN Target Cities Non-target Cities
Dosage
Low -5.3% +7.8%
Medium -3.1% <-1.0%
High -13.1% -4.9%
-8.89% -0.25%
57

58. What‟s in a Name?
Being a target city and having a higher level of dosage
was significantly related to a reduction in violent
crime controlling for:
 Concentrated disadvantage
 Population density
 Police resources
 Correctional population
58
Source: Journal of Quantitative Criminology (2010) 26:165-190.

59. Drug Market Intervention (DMI)
The problem of open air drug markets

60. Drug Market Intervention (DMI)
 Same principles and strategies applied to open, overt
drug markets
 Four Goals
 Eliminate the open-air drug market
 Return the neighborhood to the residents
 Reduce crime and disorder
 Improve the public’s safety as well as their quality of life
60

61. Drug Market Intervention Sites
Early “Adopter” Cities BJA DMI Training Cities
High Point, NC Baltimore, MD Memphis, TN
Winston-Salem, NC Chicago, IL Middletown, OH
Raleigh, NC Cook County, IL Milwaukee, WI
Providence, RI Durham, NC Montgomery County, MD
Rockford, IL Flint, MI New Haven, CT
Hempstead, NY Gary, IN New Orleans, LA
Nashville, TN Guntersville, AL Peoria, IL
Berlin, MD Indianapolis, IN Roanoke, VA
Jacksonville, FL Ocala, FL
Jefferson Parish, LA Seattle, WA
Lansing, MI
61

62. High Point
Impact on crime in the target area
Offense Type Offenses Per Offenses Per Percent ARIMA Models
Month - Pre Month – Post Change (significance)
Violent 1.83 1.27 -30.6% <.10
Property 9.24 8.54 -7.5% NS
Nuisance/Drug 2.67 1.81 -32.2% <.10
62

63. Rockford, IL
Target Area 12 month comparison pre- HGLM Comparison to Trend in
and post-DMI Remainder of City
Non-violent crime -24% <.10
Violent crime -14% NS
Non-Violent Crime Rate
Per 1,000 Residents
20
15
10
5
0
Target Area Remainder of City
63
Source: Corsaro, Brunson, and McGarrell. Forthcoming. Crime and Delinquency.

64. Nashville, TN
Target Surrounding City Wide ARIMA
Area Area Models
Drug Equipment -39.5% -52.1% -9.3% <.05
Narcotics Violations -49.7% -51.0% 5.5% <.05
Violent Crimes -23.6% -24.0% -7.4% <.15
Property Crimes -28.4% -25.6% -7.0% <.05
Call for Police Service -26.1% -6.2% -5.9% <.15
The remainder of city either experienced no decline or an increase in these offenses.
64
Source: Corsaro, Brunson, McGarrell, (2010) Evaluation Review, 34, 6:513-548.

65. Summary
 Strongest evidence in terms of reducing drug related
crime
 Interviews with residents in three sites indicate:
 Neighborhood perceived as much safer and better place to
live
 Appreciation for police
65

66. Implications
 Appears to be promising approach to addressing
open-air drug markets
 Change in trajectory of neighborhood
 Neighborhood begins to look like “normal neighborhood”
66

67. Promising Practices
 Some combination of…
 focused deterrence
 communication
 data-driven problem solving
 and linkage to opportunities,
…appears promising in reducing gun crime
67

68. “There is strong research evidence that the more
focused and specific the strategies of the police, the
more they are tailored to the problems they seek to
address, the more effective the police will be in
controlling crime and disorder.”
68
Source: National Academy of Sciences, Fairness & Effectiveness in Policing (2004: 5)

69. 15 Years of Suggestive Evidence
on Reducing Gun Crime
Directed Police Project Exile Strategic Problem Equivocal Evidence
Patrol Solving
Kansas City Richmond Boston St. Louis
Indianapolis Montgomery Indianapolis Durham
Pittsburgh Mobile Los Angeles
Stockton
Lowell
Pre-PSN
Omaha
Greensboro
Winston-Salem
Chicago
Mixed Model (Combination of Above Strategies)
PSN National Assessment (all cities over 100,000 population)
69

70. Common Ingredients
 Focused enforcement, focused deterrence
 Systematic problem solving process
 Multi-agency working group
 Problem solving model
 Police-researcher partnership
70

71. An Evaluation of the
Comprehensive Anti-Gang
Initiative (CAGI)
Edmund McGarrell, Ph.D.
Michigan State University
Chris Melde, Ph.D. Timothy
Bynum, Ph.D.
Michigan State University Michigan State
University
Nicholas Corsaro, Ph.D.
University of Cincinnati

72. Disclaimer
 This project was supported by Award
#2007-IJ-CX-0035 awarded by the
National Institute of Justice,
Office of Justice Programs, U.S.
Department of Justice. The opinions,
findings and conclusions are those
of the authors and do not
necessarily reflect those of the
Department of Justice.

73. Comprehensive Anti-Gang Initiative
 What is the Comprehensive Anti-Gang Initiative?
 Project Safe Neighborhoods(94 US Attorney’s Offices)
 Spergel/OJJDP Comprehensive Model
 A lesson in paradigm shifts and paradigm clashes.
 Successes and failures of CAGI
 WWTJDILTP41? (Klein or Maxson, ? through Davidson (MSU), ?)
 Painful lessons learned from Ignaz Semmelweis and Dan
Ariely

74. Stopping the Cycle of Gang Violence
Youth Exposed to Incarceration/
Violence Re-entry
Involvement with
Getting into Guns, Violence,
Gangs Drugs

75. Continuum of Promising Practices for
Comprehensive Gang Intervention
SUPPRESSION SOCIAL SERVICE
DETERRENCE CAGI INFORMAL
Boston Ceasefire “Spergel Model” Chicago Ceasefire
“Pulling Levers” “Violence Interrupters”

76. Components of the Spergel Model
 The Comprehensive Community-Wide Gang
Program Model
 1) Community Mobilization
• Development of a community-wide working group
 2) Social Intervention
• Street workers that mentor youth. Bridge to institutions.
 3) Opportunity Provision
• Employment, education, service
 4) Organizational Change
• Policies and practices routinely fail “at-risk” youth
 5) Suppression
• Specialized gang units, intensive probation, prosecution

77. CAGI Details: Funding Areas
 1) Enforcement and Prosecution ($1 million)
 Collaboration of local and federal law enforcement
• Intelligence sharing and case coordination
 Collaboration of local and federal prosecution
 2) Prevention and Intervention ($1 million)
 The prevention and intervention strategy provides
gang focused programming to youth at high risk for
gang involvement
 3) Prisoner Reentry ($500,000)
 The prisoner reentry strategy provides mentoring,
social services, and treatment to gang-involved
offenders returning to “the” community from prison.

78. Participating CAGI Sites
 Original Six (2006)
 Cleveland (OH-ND)
 Dallas/Fort Worth (TX-ND)
 Los Angeles (CA-CD)
 Milwaukee (WI-ED)
 Pennsylvania Corridor (PA-ED)
 Tampa (FL-MD)
 Subsequent Four (2007)
 Indianapolis (IN-SD)
 Oklahoma City (OK-ED)
 Rochester (NY-WD)
 Raleigh/Durham (NC-ED and NC-MD)
 Final Two (2008)
 Chicago (IL-ND)
 Detroit (MI-ED)

79. Average Violent Crime Trends between 2002-2009
(Violent Crime Rate per 100,000)

80. Law Enforcement Results
 Process Evaluation Results suggest that law
enforcement was quick to implement, and
included 5 sites that engaged in policing
strategies highly consistent with the model.
 Data Driven
 Targeted
 Solid Cases
 No usable gang data in 10 of 12 sites.
 No overall program effect on violent crime.
 Coefficient in the right direction, not significant

81. Law Enforcement Implementation Ranking and Changes in
Violent Crime (High Imp. = 15% reduction in violent crime)

82. Reentry?
 The reentry component of the intervention was
not suitable for an outcome analysis.
 Poor planning led to the inability to identify and serve
the intended population.
• Six sites did not become “operational” until 2009 or after.
• Only 5 sites met targeted number of clients.
 Program assumed inmates would return to
original neighborhood, which was not realized.
 No data driven planning; based on perception
 A common complaint was that the program did
not always deliver on promises.
 Economic collapse and job market in 2008.

83. Prevention and Intervention
 Weed and Seed Philosophy
 Identifying those most at-risk for gang
membership and violence.
 Gang members, while a small fraction of the
population, account for roughly 50 to 75 percent of
violent crime in some U.S. locations.
 Just target the bad apples (simple right?)
 How do we identify them?
 Who do we ask?
 Where do we get our information?
 Extant research suggested problems in targeting
the correct population.

84. A Unique Opportunity: Cleveland, OH
• African American Male School Dropout Rate = 75%
• Overall graduation rate = 43%
• School District wanted to know the needs of students.
• Cleveland awarded grant to implement CAGI
• Wanted a standardized instrument to help identify youth
most at-risk for gang membership.
• The Global Risk Assessment Device (GRAD) (Gavazzi
et al., 2003)
– 1) Those targeted as part of the CAGI (n = 146)
– 2) A general sample of African American males
attending public high school (n = 1,438)

85. Risk Domains
• Four Risk Domains were evaluated across the two
samples:
– Disrupted Family Processes (17 items)
– Sub-Domains: Conflict; Parental Tiptoeing; Hardship
– Mental Health (26 items)
– Sub-Domains: Internalizing; Externalizing; ADHD
– Educational Risks (12 items)
– Sub-Domains: Disruptive Classroom Behavior; Threats to
Educational Progress; Learning Difficulties
– Delinquent Peers (sub-scale) (3 items)
• Associate with Gangs, Involved with Gangs, Delinquent Friends

86. Hypothesis
• Those targeted as part of the CAGI will report
more risk than the non-targeted general school-
based sample in each domain.
– The targeted sample was suppose to represent the
youth most at-risk of gang membership from high gang
areas.
– The non-targeted sample represents African American
males attending public high schools across the city.

87. Sample Description
 CAGI Sample
 146 African American males
 Ages 14 to 17 (mean = 15.9, s.d. = .9)
 General School-Based Sample
 1,438 African American males
 Ages 14 to 17 (mean = 15.5, s.d. = .8)

88. Results Red = CAGI Blue = School
Level of Risk by Domain and Target Group
13.56*
10.83
6.66*
5.61 5.25
5.12
2.03 2.33*
Education Family Mental Health Peer Group

89. WWTJDILTP41? (Klein or Maxson, ?)
Common Hurdles to Program Fidelity
 1) The idea that “something is better than
nothing.”
 Reality: Half-hearted implementation of programs
designed for high-risk youth fair no better than no
implementation at all (Kovaleski et. al., 1999).
 2) Insufficient “buy in” on the part of
program providers.
 Effect: Program providers deviate from the plan;
implementation slippage leads to little dosage (Ruiz-
Primo, 2005).

90. Painful Lesson from Ignaz and Ariely
 3) “Buy in” is simply  4) Must be easily
not enough. integrated into routine!
 Many people bought in to  “What about us?”
CAGI
 Working with high risk
 Ignaz Semmelweis, M.D. youth is difficult.
(1818 – 1865)
 Dan Ariely, Ph.D.
 A simple solution
 (Duke University)

91. Paradigm Shifts and Paradigm Clashes
RISK NEED
 Law enforcement  Social service
have an easier time providers view the
implementing risk world differently.
focused interventions.  Identify subjects in
 It is how they view the need, and do their
world; it is what they best to remedy those
do. needs.
 Many youth are in
 Implementation of need, but few youth
data driven programs are at risk.
is now commonplace.

93. Not Implemented As Intended
Law
Enforcement
Prevention/Intervention
Re-entry

94. The State of the Spergel Model
 When done with sufficient fidelity, the program
appears to reduce violence.
 The ability of jurisdictions to implement, let alone
sustain, such a model is severely limited.

95. Chicago Ceasefire:
Too Big to Fail?
Understanding the spread of a “popular”
violence intervention

96. The Origins of Chicago Ceasefire
How it works
Understanding Violence as a “Behavior”
 How do we learn behavior?
 Unconscious imitation/modeling
 Social Pressure
 Symbolic Interactionism (“Looking Glass Self”)
 “Fear of being called chicken is almost certainly the leading
cause of death and injury from youth violence in the United
States” (Zimring, 1998: 80).
 Shame vs. Guilt (Code of the Street, Made in America)
Understanding Violence as an “Epidemic”
 If violence acts like an epidemic, why not treat it as an
epidemic?

97. Stopping Epidemics
 Public Health Model of Intervention
 1. Interrupt Transmission
• Limit Transmissibility
 2. Who is most “at-risk” for infection?
• Target the risks and needs associated with high risk people.
– Clean needle program in emergency rooms.
– Free condoms at health clinics.
 3. Change group norms?
• Much the same as smoking, condom use, seat belts etc.
– Every needle pack comes with literature on the effects of
drugs on the body.

98. Violence Interrupters
“Carefully” chosen interventionists
 Why do they have to be “carefully” chosen?
 Potential Political Fallout
How did they find their clients?
 Past programs that have tried to target “at-risk” youth
have failed to garner the correct clientele.
 Why?
 How did Chicago Ceasefire try to overcome this
deficiency?

99. Outreach Workers
 Behavior change in the long-term
 Violence interrupters and the outreach staff were not
one in the same.
 Modeling
 Service referral
 Ongoing guidance

100. Change Group Norms
 Multiple Messengers; Same Message
 How did Ceasefire Chicago deliver a unified
message?
 What were their tactics?

101. Science Necessitates Replication
 Wilson and Chermak (2011)
 Both professors at Michigan State University
 Evaluation: One Vision One Life Program
 Pittsburgh, PA
 Community-based intervention in the mold of
Ceasefire Chicago
• Utilized Violence Interrupters
• Utilized Outreach Workers
• DID NOT utilize police to the same extent as Chicago
– No legal deterrent message from police or prosecutors.

102. Results of the Evaluation
No impact on homicide in target
neighborhoods, relative to comparison
communities.
Aggravated assault and gun assault rates
increased in the target neighborhoods relative
to comparison neighborhoods.
There was evidence of a “spillover” effect of the
program on adjacent neighborhoods.
 Neighborhoods around Southside experienced
increases in aggravated and gun assaults.

103. What does this mean?
Regression Effect in Chicago?
How do we promote informal social control?
 How can we coerce residents to take control of their
neighborhood?
How can we document the “work” of the violence
interrupters?
 Needed for replication
Gang Cohesion and Violence
 External (real or perceived) pressure creates
cohesion.

104. Violence Interrupters
are NOT NEW
 Tita and Papachristos (2010) reviewed these
efforts in Chicago, Boston, and LA in the
1960s, wherein they were found to have no, or
even iatrogenic, effects.
 Sounds familiar
 Other replications of Chicago Ceasefire have
had mixed findings as well.
 Baltimore, MD- Mixed
 Newark, NJ- No effects

105. Is Chicago Ceasefire
“Too Big to Fail?”
Science and Policy move at different speeds.
 Chicago Ceasefire started over a decade ago.
 Policymakers wanted results quickly
 Without scientific outcome analyses, they based decisions on
word of mouth “success” stories.
Breaking down the evaluation of Skogan.
 Network data suggested two (3?) of seven target
neighborhoods had evidence of success.
 Papachristos (2011)
 Hot Spot analyses found three successful
neighborhoods.
 Block and Block mapping analysis (Papachristos, 2011)

106. Is Chicago Ceasefire
“Too Big to Fail?”
If we are treating homicide like an epidemic
(i.e., a public health model), why not treat the
results the same way?
 If HIV treatment was found to increase HIV cases in
some places, would we continue to fund such
programs?
 If cancer treatment was found to increase cancer
cases in some places, would we continue to fund such
programs?
Ceasefire is being replicated in numerous
places, with more trainings being requested.

107. Resources for Identifying What Works, and
What Does Not
 Office of Justice Programs: Crime Solutions
 http://www.crimesolutions.gov/default.aspx
 Blueprint Series: University of Colorado
 http://www.colorado.edu/cspv/blueprints/
 Office of Juvenile Justice and Delinquency
Prevention (OJJDP)
 http://www.ojjdp.gov/mpg/Default.aspx