Evaluation of Baltimore's Safe Streets Program: Effects on Attitudes, Participants' Experiences, and Gun Violence

Full text

0
Evaluation of Baltimore’s Safe Streets Program:
Effects on Attitudes, Participants’ Experiences, and Gun Violence
Daniel W. Webster, ScD, MPH
Jennifer Mendel Whitehill, PhD
Jon S. Vernick, JD, MPH
Elizabeth M. Parker, MHS
Johns Hopkins Center for the Prevention of Youth Violence
Johns Hopkins Bloomberg School of Public Health
Baltimore, MD
January 11, 2012

1
EXECUTIVE SUMMARY
Background
In 1995, Dr. Gary Slutkin of the University of Illinois at Chicago developed the
CeaseFire program to reduce youth violence associated with firearms. CeaseFire is a
multifaceted intervention involving several different components. Most notably, street outreach
workers – often former gang members – develop relationships with high risk youth in high crime
urban areas. Outreach workers serve as positive role models for the young people, steering them
to resources such as job or educational training. Special outreach staff called violence
interrupters work to identify and resolve potentially dangerous conflicts before they escalate into
shootings. In addition, the program organizes community responses to shootings and attempts to
change social norms surrounding shootings, sending the message that using a gun to resolve
conflict is unacceptable. An independent evaluation by researchers at Northwestern University
found strong evidence that the program led to significant reductions in gun violence. A grant
from the U.S. Department of Justice enabled the Baltimore City Health Department (BCHD) to
attempt to replicate Chicago’s CeaseFire in Baltimore under the name Safe Streets.
Evaluation
The evaluation has four major components: 1) a review of implementation data for the
program; 2) an analysis of the effects of the program on homicides and nonfatal shootings; 3) a
community survey of attitudes toward gun violence; and 4) interviews with Safe Streets program
participants to ascertain their perceptions of the program’s effects on their lives.
Program Implementation
BCHD solicited proposals from community based organizations interested in
implementing the program in some of Baltimore’s most violent neighborhoods. Safe Streets was
initially launched in the McElderry Park neighborhood of East Baltimore in June 2007 and in the
Union Square neighborhood of Southwest Baltimore in August 2007. However, the Union
Square community group experienced substantial problems implementing the program, failing to
establish a stable group of outreach workers until March 2008. But program implementation
problems continued and Union Square’s contract was discontinued in July 2008. Additional
program sites were added latter. Elwood Park’s program was fully implemented as of March
2008, Madison-Eastend as of January 2009, and Cherry Hill as of January 2009.
Program staff were required to keep standard records of their activities including detailed
information about each incident mediated by outreach staff. Monthly totals and conflict
mediation forms were reported to BCHD and shared with the research team. After the initial
months of enrolling participants, program sites had 35 to 60 participants connected with outreach
workers at any given time and recorded 127 to 271 participant contacts per month.

2
A key approach to reducing violence was for program outreach workers to mediate
conflicts between individuals or groups in attempt at reaching a nonviolent resolution. From July
2007 through December 2010, Safe Streets outreach workers mediated 276 incidents. Nearly 9
out of 10 (88%) of these incidents involved individuals with a history of violence and 75%
involved gang members. Weapons were at the scene in nearly two thirds of the incidents. Based
on these conditions and other factors, outreach workers thought that 84% of the situations in
which they intervened would have either “very likely” (59.5%) or “likely” (24.6%) led to a
shooting. Outreach workers considered the situation to have been successfully resolved
(avoiding serious violence) in 69% of the incidents and at least temporarily resolved in an
additional 23% of the cases. The average number of incidents mediated per month ranged from
1.2 in Madison-Eastend to 4.0 in McElderry Park. Cherry Hill mediated an average of 3.2
incidents per month and Elwood Park mediated 1.4 incidents monthly.
Program Effects on Homicide and Nonfatal Shootings
We obtained data from the Baltimore Police Department for homicides and nonfatal
shootings from January 1, 2003 to December 31, 2010. We compared changes in the number of
homicide and nonfatal shooting incidents per month in the intervention neighborhoods with high-
crime comparison areas (police posts) without the intervention. To be a comparison area, the
police post must have been in the top 25% among all posts for the number of homicides and
nonfatal shootings from 2003 to 2006. Regression models were used to control for several
possible confounders including measures of police initiatives directed at reducing neighborhood
gun violence, arrests for weapon and drug violations, and baseline levels of homicide and
nonfatal shootings.
In Cherry Hill, Safe Streets was associated with statistically significant reductions of 56%
in homicide incidents and 34% in nonfatal shootings. Program effects in the three East
Baltimore sites varied. McElderry Park did not experience a homicide during the first 22 months
of program implementation (prior homicide levels in the area and citywide trends projected five
homicides in McElderry Park for that period without the intervention). However, homicides
increased during the period when program supervisors and staff also concerned themselves with
a new Safe Streets site in bordering Madison-Eastend where gang violence surged. During the
months McElderry Park’s program was running without the near-by Madison-Eastend program,
homicides were 53% lower than would have been expected without the intervention. However,
there were no program effects on homicides or nonfatal shootings in McElderry Park during the
months when Madison-Eastend’s program was operating. Both Elwood Park and Madison-
Eastend’s Safe Streets interventions were associated with statistically significant reductions in
nonfatal shootings (-34% and -44%, respectively). However, homicides were nearly three times
higher than would have been expected during the 18-month period the program was in operation
in Madison-Eastend. There was also evidence that positive programs extended into areas
bordering the neighborhoods that implemented Safe Streets.
Totaling statistically significant program effects across all the program sites and border
posts we estimate that the program was associated with 5.4 fewer homicide incidents and 34.6
fewer nonfatal shooting incidents during 112 cumulative months of intervention post
observations. There would have been more than 10 additional homicide incidents prevented had

3
there not been significant increases in Madison-Eastend and in the area bordering Elwood Park
that coincided with program implementation.
Youths’ Attitudes About Gun Violence
To assess the effects of Safe Streets on attitudes about the appropriateness of using a gun
to resolve conflicts, we conducted surveys in three Baltimore neighborhoods – McElderry Park,
Union Square, and Oliver. A first wave of surveys was conducted in November/December 2007
– after implementation had begun in McElderry Park but prior to a largely failed program
implementation in Union Square. Oliver, which had unsuccessfully applied for Safe Streets
funding, served as another nonintervention comparison neighborhood with baseline levels of gun
violence similar to that of McElderry Park. For the second wave of surveys, conducted in Spring
2009, we excluded Union Square due to implementation problems which led to discontinuance
of the program.
For each survey wave, young men ages 18 to 24 were recruited on the street and in public
places to complete a brief, anonymous, self-administered survey. The survey contained
hypothetical scenarios based on common sparks for shootings. One set of survey questions
asked whether the respondent thought it was okay to either “threaten” or “shoot” the antagonist.
Another set of questions asked respondents whether they thought their friends would think it was
okay to threaten or shoot the antagonist in the same situations.
For survey Waves 1 and 2, youth in McElderry Park were much less likely than youth in
the other neighborhoods to believe that it was okay to use a gun to resolve disputes in our
scenarios. In fact, youth in McElderry Park were 4 times more likely to have the lowest level of
support (“little or no”) for using violence than were youth in Union Square. Regression models
showed that Wave 1 respondents in McElderrry Park were less likely to support using guns to
settle disputes (p<.001) after controlling for confounders. In the models for Wave 2, McElderry
Park respondents were less likely to be in the “strong” support for gun violence category
(p<.001), but there was no longer a significant neighborhood difference for being in the
“moderate” support category.
Program Participants’ Experiences and Views of Program Impact
In May 2011, we conducted anonymous interviews with program participants in Cherry
Hill and McElderry Park to learn about their experiences with Safe Streets. Outreach staff
provided information about the survey to each adult (age 18+) program participant and directed
those who were interested to come to the program office at designated times when research
interview staff would be available to conduct interviews. A total of 32 program participants in
Cherry Hill and 33 in McElderry Park were interviewed.
As the Safe Streets program envisions, program participants are at high risk. Nearly half
of program participants (48%) had ever been shot at.
Program outreach workers appear to be important parts of the lives of these young
people. Two-thirds of participants saw their outreach worker 3 or more times per week; for

4
three-quarters of participants, these meetings lasted an average of more than 1 hour. Outreach
workers provided program participants with various types of assistance. Participants who sought
assistance reported that outreach workers helped with activities including: finding a job (88%);
job interviewing skills (75%); job training (63%); getting into a school or GED program (95%);
and resolving family conflicts (100%).
Outreach workers also helped the majority (52%) of program participants settle an
average of two disputes. Twenty-eight percent of these disputes involved guns and 91% avoided
violence. Overall, 80% of program participants reported that their lives were “better” since
becoming program participant of Safe Streets.
Conclusions
Safe Streets was implemented in four of Baltimore’s most violent neighborhoods,
engaging hundreds of high-risk youth, promoting nonviolence through community events, and
mediating over 200 disputes with the potential to lead to a shooting. The program was associated
with less acceptance for using guns to settle grievances in the one intervention neighborhood
where attitudes were studied. Program participants reported benefiting from their connections to
outreach workers in numerous ways that could be protective against future involvement in
violence.
Three of the four program sites experienced large, statistically significant, program-
related reductions in homicides or nonfatal shootings without having a counter-balancing
significant increase in one of these outcome measures. Both program sites where Safe Streets
was linked to large reductions in homicides mediated about three times as many disputes per
month than did the other two program sites. Future efforts should focus on understanding and
improving program implementation and discovering the conditions under which the program can
be most effective in reducing violence.