Gaps and limitations previous research Flashcards
How has baltimore’s safe streets been evaluated? Describe and critique Shani’s work.
four sites implemented before 2019 and all of the discontinued sites except for Union Square were evaluated using synthetic control models and log binomial regression models.
Models assessed program impacts on homicides and non-fatal shootings separately.
SSB was associated with increased homicides and non-fatal shootings at the three discontinued sites, a reduction in homicides but an increase in non-fatal shootings in three additional sites (McElderry Park, Cherry Hill, and Sandtown)
decrease in both homicides and non-fatal shootings in Park Heights.
No estimates however were statistically significant
Do you think Safe Streets has changed norms in Baltimore/NY?
Delgado and Butts John Jay, Bronx, 17 item violence index: would you use violence in a range of hypothetical conflict scenarios- found decrease in CV areas and no change in non-CV areas
Surveyed target pop: young men 18-30 even though norm change is at broader community level (2014-2016)
Baltimore: compared park heights (intervention) to SW baltimore. Randomly selected block faces, approached any male youth age 18-24
Measured exposure to Safe streets (seen fliers, been helped) and changes in attitudes toward violence using exploratory structural equation modeling (attitudes measured pre intervention and 1 year after implementation). Same violence index
Your friend is at a club with his girl and this guy is dancing with her. Would most of your friends think it was right to threaten the guy with a gun? Shoot the gun?
Saw declines in both communities, those who reporting seeing stop shooting less likely endorse violent norms
Larger improvement in intervention community
What qualitative work has been done on safe streets model? (albany and Baltimore)
Qualitative interviews of community residents impacted by a CV program in Albany highlighted the importance of the program addressing ‘the whole picture’ through meaningful community collaboration
Valued community events like neighborhood barbeques, shooting response events, appreciated consistent presence, talking about the program improving quality of life- helping with school and employment or training
qualitative work with CV staff in Baltimore detailed successful mediation techniques- a bit more of program mechanism questions. How staff do their job, what are the most effective tactics
Successful mediations: built on trust, immediate priorities to separate potential shooter and victim followed by persuading parties to resolve conflict peacefully.
Brokering peace: return of stolen property, emphasizing negataive consequences of violence (jail, death, increased police attention)
Some evaluations have shown SSB associated with increases in H or NFS. Why is that?
Sites receive same training, city contracts, technical assistance
Shows importance of strong implementation, appropriate oversight, meaningful support
Geographic and contextual factors: cherry hill geographically isolated (to benefit) vs. Neighborhoods west side
Ex. Of madison east end- gang/drug related conflict when 12 people shot at a cookout in 2009, retaliation
No one has suggested SS has mechanistically increased H in the way that police involvement can increase violence
Easier to intervene on homicide because more planned than NFS
Talk about chicago findings- how chicago implementation and evaluation same and different
Similar decentralized cbo model
Different that not limited to city of chicago- some scattered in broader region and illinois, mixed of primarily AA vs. Primarily Latino
Started decentralized, larger main office presence implementation- similar to MONSE structure and current centralization of CBO oversight
Similar challenges in staffing: “wanna work but still wanna be in a gang” –> accidental hires of current gang members, unstable funding, low salaries no benefits, issues of politics (although from the state)
Different: started large and got smaller (25 to 2 due to funding instability)
Different, separate ow and vi, vi through centralized model with weekly meetings for cross site collab
Street outreach now part of a more holistic approach “chicago cred” cure violence focused as national organization
Eval: time series analysis, Matched comparison areas (Actual neighborhoods)
ARIMA: Intervention analysis in time series refers to the analysis of how the mean level of a series changes after an intervention, when it is assumed that the same ARIMA structure for the series holds both before and after the intervention. AutoRegressive Integrated Moving Average- allows you to uncover patterns in the data e.g. “instant and permanent”
Showed stat signific decrease shootings and homicides in 3 of 7 assessed areas
Control areas less good than synthetic, many contiguous w each other and w intervention areas
Talk about new york findings- how new york implementation and evaluation same and different
Intervention: well funded, well integrated with HVIP and other violence prevention
Compared two intervention with two comparison (actual neighborhood)
Different: used hospital data for outcomes
Interrupted time series analysis ARIMA- accounts for prior trends and seasonality
Saw stat signif decline in all shootings at one site compared to control
Talk about phoenix findings- how phoenix implementation and evaluation same and different. What were process and impact eval metrics?
In Phoenix, a combined process and impact evaluation found the program was conducting some (e.g. conflict mediations, identification at risk individuals) but not all key program components (e.g. no collaboration with the faith-based community or other community groups), and the program was associated with a significant increase in the number of fatal and non fatal shootings
why bad? Not enough mediations, small pub ed campaign, no good connections for referrals, translates less well to suburban area (easier when more concentrated)
Can you expand on what you mean by implementation factors that may impact program success?
Staffing: fully staffed, secure budget, livable wage
Implementation and oversight, training (what gets measured gets done)
Dynamics CBO and MONSE, shifting political climate
Can you expand on what you mean by contextual factors that may impact program success?
Types of conflicts: more effective gang related than retaliatory/weapons
Likelihood of violence: the risk-level of individuals involved; whether the incident was related to retaliation; the number of people involved; and shooting likelihood.
Police- clearance rates
Chicago eval with geospatial spillover- talk about those findings and how advancing that knowledge
only half of the clients surveyed indicated that they lived or hung out in the targeted zone. (showed target area maps)
violence interrupters ranged even more widely, as they followed gang activities. Much of this was required as part of field operations, and this was not a neat laboratory experiment. When the staff completed their conflict mediation paperwork, CeaseFire was able to monitor this, and maps showed that programmatic activities were spread over broad areas of the city.
However, effects of those dispersed interventions could not be captured using the approach utilized here, and if they affected the comparison areas they could even mask the apparent impact of the program in targeted areas.
Countered with “two block gangsters”
Report defining their areas based on street boundaries, not ceasefire or the police. Started loose with geo until started eval- mixed feelings. Only half said they only intervened in conflicts tha happened in their official zones