Do Shootings Lead to More Shootings and Can a Parked Police Car Prevent Them?

Evidence from Philadelphia's Near-Repeat Violence Research
AI Summary
Research question: Does a shooting increase the risk of another shooting occurring nearby in the following days or weeks?
Bottom line: Yes. In a study of 3,785 shootings in Philadelphia, Ratcliffe and Rengert (2008) found that shootings were significantly more likely to cluster within 400 feet (approximately one city block) and 14 days of a previous shooting. During the period examined after a police intervention began, there were 33% more near-repeat shooting pairs than expected from a simulated random distribution. The findings suggest that retaliation, coercion, and escalation may create short-term concentrations of gun violence.
However, a Philadelphia police initiative designed to prevent retaliatory shootings did not eliminate the pattern, raising questions about the resources and precision needed for effective prevention.
George Rengert and I published the following study in Security Journal in 2008. We wanted to establish whether shootings demonstrated what criminologists call a near-repeat pattern, where one crime is followed by an increased likelihood of another crime occurring nearby within a relatively short period.
The results provided quantitative evidence for something many experienced police officers already suspected. Shootings were not simply clustered in the usual high-crime neighborhoods. They also demonstrated a significant relationship across space and time.
What Is a Near-Repeat Shooting?
The concept of near-repeat victimization originally emerged from research into residential burglary. Researchers discovered that when a house was burgled, nearby houses experienced an elevated risk of burglary for a limited period afterward.
The idea is that crime risk can spread temporarily from one location to nearby locations.
We wondered whether a similar pattern might exist for shootings. The underlying explanation, however, would probably be quite different. Burglars may return to familiar areas because they know the opportunities available. Shootings, by contrast, may be connected through retaliation, intimidation, coercion, or the escalation of disputes.
For our research, we defined a near-repeat shooting as a subsequent shooting occurring close to an earlier incident and within a short period. We focused particularly on events occurring within approximately one city block (400 feet) and within 14 days.
This allowed us to examine whether shootings were more closely connected in space and time than we would expect by chance.
Examining Nearly 3,800 Shootings in Philadelphia
We obtained detailed shooting records from the Philadelphia Police Department covering two 13-month periods between August 2003 and September 2005. The database included 3,785 confirmed shooting incidents in which someone had been struck by gunfire. These included both fatal and nonfatal shootings. We deliberately excluded reports of shots fired where no victim was identified. We then used a statistical technique based on the Knox test, originally developed to examine the clustering of disease cases in space and time.
The approach compared the actual distances and time intervals between shootings with patterns generated through computer simulations. By randomly rearranging the dates of shootings while keeping their locations fixed, we could estimate how frequently nearby shootings would occur if there were no meaningful relationship between their timing and location. We repeated this process 999 times for each study period.
The question was straightforward: Were there more shootings occurring close together in both space and time than would be expected from the underlying geographic distribution of violence?
The Results: A Two-Week Window of Elevated Risk
In both study periods, we found statistically significant evidence that shootings clustered within 14 days and 400 feet of a previous incident. During the second study period, our simulations suggested that approximately 81.5 shooting pairs would be expected within this space-time window if shooting dates were randomly distributed. The actual number was 109.
That represents approximately 33% more near-repeat shooting pairs than expected, a statistically significant difference.
The finding matters because it suggests that shootings are not entirely independent events. Although some violence may reflect the persistent characteristics of high-crime locations, there appears to be an additional short-term risk associated with a recent shooting. A shooting may therefore be more than an incident requiring investigation. It may also be an indicator of an increased risk of further violence nearby.
Did Philadelphia's Priority Corners Program Prevent Retaliatory Shootings?
An important aspect of the research was that it took place around the introduction of a Philadelphia Police Department initiative called Priority Corners. The program followed Operation Safe Streets, which had concentrated officers at street-corner drug markets. Priority Corners adopted a more mobile approach, directing police attention toward locations where investigators believed retaliatory shootings were likely.
The idea was sensible. If a shooting created an elevated risk of further violence nearby, police could concentrate resources in that area during the period of greatest danger. However, our analysis found that the near-repeat shooting pattern remained statistically significant after the program was introduced.
This does not necessarily mean the underlying strategy was wrong.
Philadelphia was experiencing substantial gun violence, and the police department had limited resources. The number of locations requiring attention frequently exceeded the department's capacity to provide sustained coverage.
The program may have prevented some retaliatory shootings at particular corners without producing a sufficiently large effect to alter the citywide pattern. We could not establish that from the available data.
The more important lesson is that a theoretically sound policing strategy may still struggle if the resources available are insufficient to deliver the necessary intervention.
Not All Shooting Hotspots Are the Same
One of the more interesting findings came from examining what we called initiator events—shootings that were followed by at least one near-repeat shooting. Across the 26-month study period, we identified 642 such shootings. We then examined where these initiating incidents occurred relative to the general distribution of shootings across Philadelphia.
The geography was not identical.
Some police sectors had disproportionately high numbers of shootings that initiated near-repeat sequences, even though they were not necessarily the areas with the highest overall shooting counts.
The correlation between the concentration of initiator events and shooting frequency was relatively weak, at 0.296.
This suggests an important distinction for crime analysts. The locations with the most shootings are not necessarily the locations where shootings are most likely to be followed by additional nearby violence.
For police commanders, this creates an opportunity to think beyond conventional hotspot maps. Identifying places where shootings frequently initiate further violence may provide a more precise basis for prevention.
Why This Research Matters
The broader contribution of the study was to demonstrate that near-repeat patterns are not confined to property crime. The same statistical techniques previously used to examine burglary could identify meaningful short-term patterns in serious violence. That finding has implications for crime analysis, intelligence-led policing, and the design of violence prevention strategies.
It also highlights the difference between identifying a crime pattern and successfully intervening to disrupt it. Our research demonstrated a measurable near-repeat shooting problem in Philadelphia. It did not demonstrate that the Priority Corners intervention successfully reduced that problem.
Understanding when and where violence is most likely to recur provides an opportunity for more targeted prevention. But exploiting that opportunity requires reliable intelligence, sufficient resources, and interventions capable of addressing the mechanisms driving the violence.
A shooting is not necessarily the end of an incident. Sometimes it is the beginning of the next one.
Citation
Ratcliffe, J. H., & Rengert, G. F. (2008). Near-repeat patterns in Philadelphia shootings. Security Journal, 21, 58–76. https://doi.org/10.1057/palgrave.sj.8350068
Full paper link: https://www.jerryratcliffe.net/_files/ugd/f5df24_b69e363ef9a242e6ae538decf2f6ed62.pdf
Keywords
Primary: Near-repeat shootings, retaliatory gun violence, Philadelphia shootings, near-repeat crime analysis, gun violence prevention.
Secondary: Evidence-based policing, intelligence-led policing, shooting hotspots, spatial crime analysis, temporal crime patterns, Priority Corners Philadelphia, crime mapping, violence prevention, Jerry Ratcliffe, George Rengert.
Long-tail search terms: Do shootings lead to more shootings? How long is the risk of retaliatory violence elevated after a shooting? What is the near-repeat shooting phenomenon? Can police predict retaliatory shootings? What does research show about shooting retaliation in Philadelphia? Can near-repeat analysis improve gun violence prevention?


