Abstract
This study used observations of crime strategy meetings and interviews with police commanders to “get inside the black box of hot spots policing.” The findings focus on what the studied police commanders believed they were doing and why they believed those tactics would be effective during hot spots policing implemented under non-experimental conditions. An example causal model for the effectiveness of hot spots policing that emerged from the data is presented. While the commanders’ views aligned with commonly used policing tactics and crime control theories, their underlying theoretical rationale is complex. The presented model provides one causal model that could be tested in future hot spots policing evaluations, and a discussion is presented of how the study’s methodology can be applied in other jurisdictions to define localized causal models and improve hot spot policing evaluations.
Introduction
The disproportionate concentration of crime in a small number of microplaces (Braga, Hureau, & Papachristos, 2011; Braga, Papachristos, & Hureau, 2010; Sherman, Gartin, & Buerger, 1989; Weisburd, 2015; Weisburd, Groff, & Yang, 2012) has facilitated the development and adoption of hot spots policing—or focusing police resources in those few high crime microplaces (Braga & Weisburd, 2010; Skogan & Frydl, 2004; Weisburd & Braga, 2006). Reviews of hot spots policing evaluations have repeatedly found it reduces crime, but the implementation of hot spots policing has varied across jurisdictions and what exactly police should be doing in crime hot spots remains an open question (Braga, 2001; Braga, Papachristos, & Hureau, 2012; Telep & Weisburd, 2012; Weisburd & Braga, 2006). 1 To improve future hot spots policing policy and evaluations, there remains a need to open up hot spots policing’s “black box” to determine which actions work through which mechanisms and drive hot spots policing’s effectiveness (Goldkamp, 2010; Rossi, Lipsey, & Freeman, 2004). Therefore, this study used observations of police crime strategy meetings and interviews with police commanders to gain insight into what police commanders from one, large municipal police department believed they were doing when allocating resources to crime hot spots under non-experimental conditions and why they believe those tactics would be effective. The results are then discussed in terms of their implications for hot spots policing theory, practice, and evaluation.
Hot Spots Policing
Sherman and Weisburd (1995) conducted the first hot spots policing evaluation—the Minneapolis Hot Spots Policing Experiment. According to Sherman and Weisburd (1995), the experiment was motivated by two factors. First, the null findings of the Kansas City Preventive Patrol Experiment (Kelling, Pate, Dieckman, & Brown, 1974) were hypothesized to be due to an inadequate dosage of police presence rather than the police’s inability to deter crime. Second, drawing on the then new findings from Sherman et al. (1989), it was reasoned that one way to adequately increase the dosage of police presence would be to focus police resources in the roughly 3% of places that experience the majority of crime. Sherman and Weisburd (1995) ultimately found that increasing police presence in micro, high crime places reduced total calls for service, “soft” crime calls for service, and observed disorder. Since then, three additional studies have also found that increasing police presence in crime hot spots by having officers stationed on drug corners 24 hours a day (Lawton, Taylor, & Luongo, 2005), patrolling by foot Tuesday through Saturday for 16 hours a day (Ratcliffe, Taniguchi, Groff, & Wood, 2011), or randomly visiting hot spots for 15 minutes every two hours can control crime (Telep, Mitchell, & Weisburd, 2012).
Additional studies have found increased police enforcement can effectively impact hot spots. Weisburd and Green (1995) found increased patrols and arrests by squads of narcotics and patrol officers during crackdowns with an additional maintenance period to ensure drug dealing did not resurface, or what Braga and Weisburd (2006) referred to as “enforcement problem-oriented policing”, reduced both total disorder and some specific categories of disorder in Jersey City (NJ) drug markets (Weisburd & Green, 1995). 2 Likewise, directed motor patrol used to generate high numbers of vehicle and pedestrian stops, traffic citations, gun seizures, and arrests in one high crime police beat reduced gun crime in Kansas City (MO; Sherman & Rogan, 1995). Groff et al. (2015) found that aggressive enforcement and other actions (e.g., frequent conversations) focused on high-risk offenders reduced violent crime in Philadelphia (PA) hot spots. Finally, arrests and traffic enforcement generated during directed patrols reduced non-domestic firearm assaults in St. Louis (MO) hot spots (Rosenfeld, Deckard, & Blackburn, 2014).
Problem-oriented policing (POP) has also been used to effectively address hot spots (see Eck & Spelman, 1987; Goldstein, 1979). For example, the implementation of POP—mostly aggressive order-maintenance policing (i.e., foot and motor patrols, dispersing loiterers, enforcing public drinking and narcotics laws, and conducting field investigations) and efforts to improve physical conditions—in Jersey City (NJ) hot spots reduced the total incidents and calls for service as well as some specific crime categories (Braga et al., 1999). Braga and Bond (2008) also found a POP treatment in Lowell (MA) disorder crime hot spots reduced total calls for service and robbery, non-domestic assault, and burglary respectively. The POP treatment consisted of aggressive order-maintenance policing (i.e., foot and motor patrols, field investigations, and disorder enforcement) and a range of situational and social-service interventions, albeit the crime reductions were attributed to the situational interventions. Finally, POP using a range of interventions alongside high levels of patrol and enforcement reduced hot spot violent crime levels during the post-treatment period in Jacksonville (FL; Taylor, Koper, & Woods, 2011).
The “Black Box” of Hot Spots Policing
Despite the consistent finding that hot spots policing can effectively control crime, the range of tactics used within and across hot spots policing evaluations creates the classic “black box” problem. In sum, hot spots policing is commonly captured with a dummy variable (i.e., hot spots policing vs. no hot spots policing), but in practice it is comprised many tenets that vary across evaluations and may or may not reduce crime through different theoretical mechanisms (for an exception, see Braga & Bond, 2008).
The “black box” problem impacts even the “simplest” hot spots policing evaluations. For example, when George Kelling visited the treatment areas of the Minneapolis experiment, he noted “some [officers] were reading newspapers or sunning themselves while sitting on the patrol car, while others were engaging citizens in friendly interaction in community-policing style” (Sherman & Weisburd, 1995, p. 634). Other evaluations of police presence in hot spots have also noted the complexity of “simply” increasing police presence and concerns about being unable to explain which actions actually produced the observed crime reductions in which ways (Lawton et al., 2005, p. 495; Ratcliffe et al., 2011, p. 801; Telep et al., 2012, p. 929). As the tactics used during hot spots policing become more complex, hot spots policing’s “black box” becomes more complex. For example, in problem-oriented policing evaluations as many as 283 different tactics have been implemented in the treatment sites (Taylor et al., 2011, p. 158).
Of course, the hot spots policing “black box” problem limits our ability to explain why different hot spots policing tactics produce crime reductions (Goldkamp, 2010; Rossi et al., 2004). For example, evaluations of police presence may be framed with deterrence theory (e.g., see the title of Sherman & Weisburd, 1995), but when the police are also drawing on increased enforcement (i.e., arrests) it begs the question of whether incapacitation mechanisms are also at work. Similarly, police presence may generate deterrence by actually changing potential offenders’ perceptions of punishment certainty, but, in the terminology of situational crime prevention, it may also merely deny offenders access to attractive offending locations in a given situation (Clarke, 1980, 1995, 2008). The former instance changes offenders’ perceptions, perhaps for the long term, whereas, the latter may have no impact on offenders’ actual mindset and merely dissuade them for a short time. Without actually changing offenders’ long-term perceptions, hot spots policing effects will likely decay over time (Sorg, Haberman, Ratcliffe, & Groff, 2013). As the number and range of hot spots policing tactics increase, particular under the guise of POP, the theoretical mechanisms underpinning crime reductions also increase and become more complex (e.g., see Braga et al., 1999, pp. 569–571).
Rational, evidence-based policy, among other things, requires an understanding of “if,” “how,” and “why” different actions produce crime control benefits (Mears, 2007). Stated differently, if we do not know how and why some tactics were originally successful, then we cannot replicate those tactics in other jurisdictions or translate those evaluations into broader policies. We simply do not know what to tell other agencies to do. Agencies can easily overlook a valuable component of the original treatment. Criminal justice agencies, especially police departments, rarely rigorously evaluate their own initiatives and policies (Haberman & King, 2011), so they likely will not know that their version of hot spots policing was ineffective due to implementation problems. Therefore, following Goldkamp’s (2010) advice, hot spots policing researchers and practitioners at minimum need to (a) clearly explicate the theoretical reasons for expecting specific hot spots policing tactics to reduce crime, (b) directly measure the dosage of each implemented tactic with valid and reliable measures, (c) directly measure any additional (intervening) theoretical mechanisms hypothesized to underpin the tactics’ effects, and (d) conduct appropriate statistical tests that empirically evaluate and compare the effectiveness of different tactics and their underlying theoretical mechanisms (see Rossi et al., 2004).
The present study focused on the first point by examining police commanders’ perspectives on hot spots policing in order to “open hot spots policing’s black box.” Specifically, the study addressed two inter-related questions. First, when police commanders allocated resources to crime hot spots, what did they think they were doing? Second, why did police commanders believe what they were doing in crime hot spots would be effective?
Using police commanders to answer these two research questions has at least four advantages (see Goldkamp, 2010, pp. 464–466). First, police departments often rely on different tactics in practice (e.g., enforcement) than commonly advocated for by reformers and academics (e.g., situational crime prevention). Second, involving practitioners in the development of theory and policy may ultimately produce more buy-in among skeptics (see Goldkamp & Vîlciciã, 2009). Third, if police reformers and academics truly want police departments to do something other than what they typically do, then understanding and evaluating their perspectives should provide valuable insight for future reforms. In other words, demonstrating that police commanders’ assumptions about crime control are not empirically supported may convince them to change their beliefs and practices. Fourth, it would be short-sighted at best and arrogant at worse to presume police commanders, with extensive experience, should not be consulted when considering how hot spots policing works in practice.
Data and Methods
Data
This study was conducted with the Philadelphia Police Department (PPD). The PPD’s roughly 6,600 sworn officers and 800 civilians (fourth largest municipal police department in the United States) provide policing services to the City of Philadelphia’s (PA) roughly 1.5 million residents across 142 square miles (fifth most populous city in the United States). Philadelphians are predominantly Black or White (roughly 43% and 41%, respectively) with approximately 12% identifying as Hispanic or Latino (US Census Bureau, 2010). In 2011, roughly 26% of residents reported living in poverty and Philadelphia’s median income was $34,207 whereas the national median income was $50,502 (US Census Bureau, 2011). Despite recent declines, Philadelphia has typically experienced higher crime rates than other large United States cities over the past couple decades (Zimring, 2011).
In 2007, Michael A. Nutter was elected as Philadelphia’s mayor. During his campaign, Mayor Nutter promised to address the city’s rising homicide rate by reforming the PPD (Nutter, 2007). Mayor Nutter hired Police Commissioner Charles Ramsey, cochair of President Obama’s Task Force on 21st Century Policing, to reform the PPD due to his stellar reputation. In January 2008, Commissioner Ramsey released the PPD’s five-year strategic plan, which vowed to “aggressively attack concentrated pockets of crime” (Ramsey, 2008, p. 8). Under Commissioner Ramsey’s leadership, the PPD has operated under a city-wide hot spots policing strategy, conducted two hot spots policing experiments in collaboration with academic researchers (Groff et al., 2015; Ratcliffe et al., 2011), and worked to improve the department’s crime and intelligence analysis capacity. Thus, the PPD provided an adequate site for studying hot spots policing.
After submitting a research proposal to the PPD executive command staff, the researcher was granted permission to conduct the study. The researcher had previously worked on research projects with the PPD and had developed professional relationships and trust with the command staff who approved the study. Standard institutional review board protocols were used to gain consent from the specific commanders who participated in the study.
The researcher was granted permission to collect data using two complementary methods: (a) field observations of PPD crime strategy meetings and (b) semistructured interviews with PPD commanders. PPD crime strategy meetings involve police executives (above the rank of captain) and local-level commanders (rank of district captain and below) using crime and intelligence analysis to determine operational strategy and resource deployments. These meetings are analogous to CompStat meetings but have been tailored for implementing the PPD’s hot spots policing strategy (Bratton, 1998; Maple, 1999; Willis, Mastrofski, & Weisburd, 2004). Therefore, observing the crime strategy meetings provided insight into how hot spots policing is actually implemented in Philadelphia relative to how it might be intended to be implemented (Rosaline, 2008). Semistructured interviews were then conducted with PPD district captains. District captains are responsible for overseeing the implementation of hot spots policing at the local level. The interviews provided data on commanders’ understanding of, experiences with, and attitudes toward hot spots policing (Flick, 2007, p. 79). Together, both methods provided data on police commanders’ implementation of and perspectives on hot spots policing under non-experimental conditions. 3
This study focused specifically on hot spots policing in violent crime hot spots. Since violent crime typically increases during the summer in Philadelphia, summer crime strategy meetings were observed to increase the likelihood violent crime problems would be the meetings’ focus. Purposive sampling is commonly used in qualitative research to ensure that the data collected are relevant for the study’s research questions (Flick, 2007). The author attended every crime strategy meeting possible between May and August of 2014. In total, eight meetings were observed (three in May, two in June, two in July, and one in August). Crime strategy meetings generally lasted 1.5 hours for a total of approximately 12 hours of observation. Using Spradley’s (1980) nine dimensions of social situations as a guide, handwritten field notes were recorded during observations. All field notes were typed up and expanded upon shortly after each observation (usually within 24 hours) to ensure the validity of the data (Fife, 2005; Spradley, 1979).
Next, the six district captains (out of a possible 21) with the greatest number of hot spots in his or her district were purposively sampled to be interviewed. Hot spots were defined as street blocks or intersections with at least five homicides, aggravated assaults, or street robberies during 2008—the first full year of Commissioner Ramsey’s tenure. The five violent crime minimum was based on the author’s knowledge from prior projects that commanders did not believe they had the resources to focus on places with fewer crimes. In other words, the five crime minimum was essentially the working definition of a hot spot in the PPD. This sampling approach ensured the captains interviewed would have the relevant hot spots policing experience necessary to provide insight into the research questions compared with districts with less violence where commanders may focus on other types of crime problems (i.e., property crime; Flick, 2007). All captains sampled agreed to participate in the interviews. The interviews were conducted between November, 2014 and February, 2015. The interviews, which lasted between 20 and 45 minutes, were audio-recorded. The audio recordings were transcribed after the interviews. During transcription, all personally identifying information was redacted or altered (e.g., the name of a church would be replaced with “church”). Subjects were also assigned pseudonyms to ensure confidentiality. Despite resource and practical limitations on the sample sizes of the observations and interviews, on-going analysis of the field notes and interviews during the data collection returned redundant themes and suggested saturation was achieved. 4
Analysis
The field notes and interview transcripts were imported into Atlas.ti for analysis. A multicycle coding process was used (Saldaña, 2009). First, descriptive codes were applied to inventory the data by applying simple labels to data segments (Saldaña, 2009, p. 66). The descriptive codes mostly described the types of crime problems discussed, the hot spots policing tactics commanders reported using, or data segments that directly answered the study’s two research questions. The second coding cycle used versus and in vivo coding to further refine the emerging codes and themes. Versus coding is commonly used in studies examining policies when there is a clear conflict in ideas, beliefs, or perspectives among study subjects (Saldaña, 2009, p. 94). The versus codes were applied to passages discussing the importance of the quantity versus the quality of actions taken in hot spots. In vivo codes use the verbatim wording of qualitative data for coding (Saldaña, 2009, p. 74). 5 Finally, the emerging codes were further refined in the third cycle of coding, and the consistency in emerging themes across the field notes and interview transcripts suggested saturation was achieved and all emerging themes were identified. 6
In the discussion of the results later, the number of interviews and field observations in which each theme appeared is also noted. The quantification of themes in qualitative research is a controversial practice (see Maxwell, 2010), but given the relatively small number of interviews and observations used in this study it was important to demonstrate the themes were observed in more than half of the individual data pieces and therefore can be considered “true” themes for the police department understudy. As discussed further later, like all qualitative research, readers should not interpret these findings to necessarily be generalizable to all police departments simply because they were quantified, but rather the numbers demonstrate the themes were common in the data and not selectively chosen.
Results
Before discussing the results, it is important to provide some additional context to the PPD’s hot spots policing strategizing, particularly the crime strategy meetings. The PPD operates under a consent decree regarding its field investigation practices (“Bailey, et al. v. City of Philadelphia, et al.,” 2011). Among other tenets, the consent decree requires the PPD to electronically record data on all field investigations (commonly referred to as stop, question, and frisks 7 ) and permit periodic auditing and analysis of those data by an external reviewer. Additionally, the PPD agreed to review its policies on training, supervision, and discipline as well as retrain its personnel on the legal requirements of field investigations. Thus, the PPD has made extensive efforts to address and eliminate criticisms, warranted or unwarranted, of its practices of investigating pedestrians and automobiles and confirm its practices meet all legal standards. As one high ranking commander explained to me during the formulation of this study, the PPD operates under the assumption that when its personnel develop operational strategies their number one priority is to ensure all laws and departmental policies are followed. 8
The size of Philadelphia and the PPD also means that is imperative that meetings operate efficiently. Given the external pressures on the PPD to also address violent crime, the PPD must move quickly during the crime strategy meetings in order to ensure hot spots policing strategies are considered and developed for all areas of the city. This means that police executives may not always stop to tell the local commanders to ensure their tactics are legal, but rather it is an operating assumption based on the reestablished precedent that stemmed from the PPD’s agreement to enter into the consent decree and recent training efforts. Overall, the purpose of raising these points prior to discussing the results is to explicitly point out to the critical reader that the importance of legal policing tactics is not lost on PPD personnel and to argue any of the findings later suggest otherwise would be to take the findings out of the context in which the data were collected.
Figure 1 shows the study’s results graphically. The first column in Figure 1 shows the specific actions the commanders reported taking to address violent crime hot spots. The second column in Figure 1 shows what commanders believed they were doing when taking those in actions in crime hot spots. Even though police commanders exercised a degree of discretion and autonomy when conceptualizing and implementing hot spots policing, commanders expressed the common beliefs that when allocating resources to violent crime hot spots, they were (a) making crime hot spots unattractive, (b) disrupting high-risk offenders, and (c) educating potential victims. The third column in Figure 1 shows the commanders’ expressed beliefs of how those actions work to control crime by (a) making potential offenders “think twice,” (b) removing potential offenders’ and victims’ opportunities to offend or be victimized, (c) getting potential offenders “off the street,” or (d) target hardening.
Police commanders’ expressed causal model for hot spots policing.
Making hot spots unattractive
All studied police commanders emphasized the importance of making hot spots unattractive places to offend. While the importance of geography is implied, all studied police commanders were interested in both short-term- and long-term hot spots. Long-term hot spots were places with historically high crime levels. For example, Captain V stressed the importance of addressing a long-term violent crime hot spot in his district. There are certain areas in this district we’re never going to move out of. We’re just going to be there because that’s where the history has been. And if it stops, if it drops down there, we're always going to be there because we know that it’s always been there …
There was also a continuous emphasis on addressing short-term hot spots amongst all commanders. Short-term hot spots were defined as places where two or more crime events (e.g., robberies) occurred close in space and time (i.e., within a few blocks and a few days to a few weeks). For example, during an observation, a captain was questioned about what she was doing to address street robberies and she replied, “there have been two robberies on [street block] about a week apart” and indicated that area was her priority (Observation 3). Short-term hot spots could have been but were not necessarily always long-term hot spots.
Regardless of the temporal extent of the hot spots, for all commanders studied, making hot spots unattractive places involved using two tactics simultaneously: (a) increasing police presence and (b) generating activity. When Captain V described his most recent success addressing a violent crime hot spot—an area with historically high levels of gang violence, he explained: Deployment wise, we increased our deployment, we increased our visibility in that area, we got some feet on the ground … We got foot beats down there and constantly hit that area, [intersection]… and also officers assigned that PSA [Police Service Area] know if they are not answering priority jobs [Part I crimes in progress or officer assist calls], routine patrol, routine patrol that area. We want a more visible presence … [I]f you’re heading back to headquarters and you got a couple minutes, make a run through it. The more visibility [the better]. There’s a [fast food restaurant] there, right at [Intersection], you want to get a cup of coffee, go in there. I'm being serious. It gives a more visible presence, so people get to see an increase … So the visible presence but with a calculated move [in a historically violent area] …
When officers are deployed to make hot spots unattractive by being more visible, they are expected to simultaneously “generate activity.” In police parlance, generating activity means stopping reasonably suspicious pedestrians and cars, enforcing traffic laws, and arresting law breakers (i.e., enforcement). The importance of generating activity was apparent in every field observation. The following field notes excerpt represents common exchanges regarding the generation of activity that took place during all observed crime strategy meetings. One executive commander then stated, ‘I don’t see the increase in activity we talked about.’ The captain noted, ‘we had an increase in Part I arrests, but I think I’m going to be moving some guys around [re-deploy officers from other areas to his grid areas] to get the results we want.’ An executive commander responded, ‘Here’s the thing. This place is firing up. I don’t have time to wait. These areas need to be locked down. They needed to be locked down last week. You need to get your guys banging these areas [generating activity].’ (Observation 8)
It is important to note, however, that the commanders did not utilize “quotas” for emphasizing the generation of activity. In fact, there was a clear emphasis on balancing the quantity of activity with what they called “quality activity.” In other words, all of the studied commanders, including the executive commanders, emphasized the importance of at least some activity being generated. But how much activity should be generated was subjective and situational, and the executive commanders seemed to just want assurance that the local commanders were attempting to address the crime problems in their districts by investigating reasonably suspicious pedestrians and automobiles and arresting law breakers. At the same time, however, all of the commanders studied stressed the importance of the quality of police activity—police enforcement focused on the “right places” or crime hot spots and the “right people” or known offenders.
For example, during the crime strategy meetings, executive commanders would frequently point out when police commanders’ activity levels were “insufficient,” and local commanders would deflect this criticism by emphasizing the quality of their officers’ activity. For example, after a captain was questioned by an executive commanders for having low activity levels, she replied, “I’d rather my guys only make 100 stops of the offenders on my list than 500 any ol’ body’s. Like the good people. The church going ladies. My people are trying to stop the right people (Observation 5).”
Simultaneous to holding local commanders accountable for minimal effort during the crime strategy meetings, executive commanders also stressed the importance of generating quality activity. The executive commander responded that the district needed to get its weekend activity up, adding ‘Make sure your people are doing what they should be doing. A car stop at 9 A.M. on a Sunday ain’t going to cut it’. (Observation 7)
The importance of quality activity was also emphasized during all of the interviews. Captain L noted: I'd rather have lower activity and have it done the right way than somebody just producing numbers, which in effect, you know, are really useless … I could stop every grandmother walking up and down [Avenue] and have a whole bunch of ped [pedestrian] stops, but in reality, what did I do, other than waste time and paper? [W]e just don’t want to be stopping the old lady going to church on Sunday morning. It counts as a car stop, and you look at the numbers it says a car stop, they [the executive commanders] don’t know if there’s a car stop, the old lady going to church, or is it a car stop that got four knuckle heads out there probably committing a crime. So it’s just not making the stop … Who are you stopping? And what cars are you stopping? [Are] these stops then going to make the quality of life better for the people of the [District]? … We’re not just stopping people to generate numbers; we’re stopping people to prevent crime … [T]he cops are there for a reason typically. Might be some manager like me says, ‘I got guys hanging out on [street block] and robberies over there, you know, I want you to stop guys who are hanging out [on street block].’ So I'm giving cops direction and they’re following up by saying, ‘Hey, here's a guy on the corner of [street], so I’m going to stop him.’… you need probable cause, or actually you don’t, you can just have mere encounters and talk with people … But it really should be supported by information [about the ‘right places’ and ‘right people’] … We’re not going to stop a regular person walking down the street. We’re going to make sure we have descriptions of those people in the same area [where robberies had hypothetically just occurred]
10
and I expect my cops to stop them … but like I said before I want quality stops. I don’t want to stop people just walking down the street … It could be two kids walking down the street going to play basketball … I don’t want my cops stopping people for that … I’d rather have quality car stops and ped stops that may result in an arrest, or results in a bench warrant, or something, rather than just stopping people for the sake of numbers. That doesn’t help me because people sometimes have bad encounters with police and I don’t want them to have that. Stop the right people. That’s all I ask … I think there's somewhere in the area between 40 and 50,000 people who are wanted on bench warrants for failing to show up for court and things like that, so you also have that residual, you know, that crossover, you’re able to get people who, you know, are wanted off the streets also. The commanding officer for the [special unit] then detailed how some officers had made a ‘quality ped stop’ in the area of some recent shootings. The officers stopped a guy who was believed to have been involved in the shooting based on intelligence. After a foot pursuit, the guy was caught with two guns. Ballistic evidence was then used to link the guns to the recent shootings … (Observation 1)
Nonetheless, disrupting crime hot spots with police presence and activity was rationalized to be effective for addressing violent crime hot spots in two ways. First, potential offenders may be denied access to crime places thereby reducing potential offending and victimization opportunities at those locations. Four of the six interviewed commanders provided specific examples of how increasing police presence and activity would encourage suspicious persons to leave an area and reduce future offending and victimization opportunities (similar examples were also discussed during five-field observations). Captain O noted, “When they [potential offenders] see the cops all around here, cops on a bicycle, cops on foot, they may go elsewhere.” Captain Y gave a more general example of how police presence may remove suspicious people from an area who may later offend or be victimized: “My foot beats were off yesterday, I drove by [intersection], all the junkies were on the corner, all the drunks were on the corner. You know, you go by when the foot beats are in there, they’re gone.” Alternatively, Captain O explained how pedestrian stops can be used to get drug dealers suspected to be driving violence to leave hot spots. We tell them, basically, ‘you can’t stay here, go hang in front of your own house … if you don’t live around here, hang in front of your house. You don’t hang in front of somebody else’s house …’
Second, four commanders directly expressed the belief that police presence and activity can disrupt hot spots by making offenders “think twice” about offending there during their interviews (direct statements for the same theme were also made during two-field observations). Captain Y explained in a general sense how police presence and activity work together to achieve this effect. You know, maybe they’ve seen a cop three more times than they usually see it, but the perception is, ‘Hey man, the cops are around [the hot spot].’ [A]nd I just think the more active officers are, and if it’s channeled the right way in the right place … it has to be good and focused, channeled activity. It just can’t be empty activity for the numbers … [But] ‘that 30 minutes you’re in an area on a car stop, you know, there’s no bad guys around. You know nothing bad is going to happen.’ It’s that presence, and the more active you are, you know? … [I]f you’re a guy who likes traffic enforcement, ‘God bless you, I want you to do that. But can we target that in our high crime, high problem areas where I have more of a presence?’ They are still going to blow the stop signs, they’re still going to make illegal turns, but now I’m getting more bang for my buck because you’re producing activity for me and then you’re actually maybe deterring crime. That’s the way I kind of look at it …
Furthermore, if quality activity is generated, then those potential offenders know they specifically should “think twice” about offending. As Captain O put it, “When they [suspected offenders] see a lot of cops around there, they get a little scared of something like that … And the other thing is stopping them … so that they know by stopping them we’re looking [at them].” Because, as Captain Y stressed, “if we can get a sustained presence in an area to where the guys know, ‘Hey the cops are around. Hey, they’re going to stop us!’ Maybe they won’t carry that gun.” Or, perhaps, offend more broadly. Therefore, the generation of quality activity means that specific offenders are being informed they should “think twice” about offending in crime hot spots.
Disrupting high-risk offenders
In addition to disrupting the place through the mechanisms described earlier, quality activity was also believed to be important for generating arrests. Generating arrests were then central to police commanders’ belief that they were also disrupting high-risk offenders by getting them “off the street” through incarceration. Given the study’s focus on violent crime, high-risk offenders included repeat robbers and people (potentially) engaged in retaliatory violence. 11 In short, police commanders used crime and intelligence analysis to identify perceived high-risk places and people in order to deploy officers to generate quality activity, make arrests, and incarcerate them. Disrupting high-risk offenders through arrest and incarceration was believed to be imperative because it was also believed high-risk offenders would continue to offend until they were “off the street.”
The captain noted, ‘there have been two robberies on that block about a week apart.’ The robberies happened on a Friday and a Saturday, so his plan is to have his officers in the area this Friday/Saturday to make an arrest … (Observation 3) I actually will put a crime bulletin [analytical report] out … I’ll say, ‘Listen, we got this guy, these two guys doing robberies in [police beat].’ And I’ll say, ‘This is the description, the times are from five o’clock in the afternoon till one o’clock in the morning. And then this is the descriptions, so I want a car in that area.’ Tell the line squad that they have to have a car in that area along with my two tactical teams I’ll put in that area … Sometimes when they see all the cops around they stop robbing. Unfortunately they’ll probably go elsewhere, but I’m hoping that my cops see two guys that match that description walking down the street, we’ll stop and get ID from them … we’ll bring them in, maybe get them fingerprinted and photographed, find out who they are, and see if we can match them up to other robberies in that area. When discussing a cluster of robberies, the executive commander then noted that the [district] was able to arrest two guys in the area for multiple robberies and that those offenders may even ‘go for jobs all over the city’, so the district should be credited with two good arrests. (Observation 2.) The executive commander then instructed the captain to get his [the repeat robber’s] picture out to his patrol officers to make sure they know to get him ‘off the street’. The captain then noted, ‘We will. That’s how it works. Get one of these guys off the street and the pockets disappear.’ (Observation 3). They’ll feed in that area and they’ll keep hitting it, hitting it, and hitting it until they're caught. Just had one, two weeks ago, you know, where the guy was arrested for two robberies, and, so, just reading over the crime reports that I do every day, as soon as I saw he was arrested, his description, the location, you know, I got right on the phone with the captain from [divisional detectives] and I said, ‘I know there's at least four other ones’. And then here as they looked … now they're up to about 16 or 17, not just here but, you know, up in [neighborhood] where the guy’s from.
So what we did with that is we identified the gang members … We do kind of a biographical profile of these gang members … Then we made the cops [in patrol] … and the tactical teams I use a lot, and we say this is where we need you to be because there’s gang friction here. I need you guys to be in the area of [street block] and I want you stopping people there, and not just people, but the right people. These are the gang members [gesturing towards a pile of papers on his desk like they were a list of gang members]. These are the guys on the corner, these are the guys hanging out in the alley, these are the guy on the steps, you know. And I need you as officers to go out and stop those guys based upon probable cause and interfere with their operation [narcotics sales] so to speak … then in doing so … we get gun recoveries, we get narcotics recoveries, and you know even minor violations like traffic violations, and all that, all builds our information base … And in doing so, one we have officer presence there. So we reduce the chance for a crime to happen. Two, we know these guys. If you look at the history of violence particularly in gangs, gangs are kind of easy in that because if there’s a history of violence it’s, we’re pretty certain that it’s going to continue because they don’t stop and go home. So … the presence there and in doing those enforcement efforts denies them the opportunity to go out and shoot someone else or get shot themselves while they're on the corner … because another gang is not going to come over with the cops hanging around all the time, you know, they say it’s ‘hot out’, or whatever, you know, ‘the heat is on’, that kind of stuff …
Additional remarks by Captain M further illustrate the shared belief that retaliatory violence will continue until the police disrupt the offenders involved. If you look at gang violence, you can even do it for regular shootings, sometimes you can trace shootings back 10 years … We've had shootings so long that people can’t remember what the original argument was over, but they do remember the last shooting was a guy from [street block] shot by a guy from [street block]. So, ‘we don’t like them guys in general, we're going to go to [street block] and shoot them …’ We’re going to get in that area right after the shooting … [W]e're conducting ped stops, we’re conducting vehicle investigation, we gain intel, we're working with [the detectives], we’re trying to get that information out there to lock it down … We don’t want a retaliation shooting. We don’t want this guy who is shot, now he’s going to come back. [W]e’re able to get access to the victim’s Twitters page, or whatever page, we can find out who he runs with. We’ll get information about, ‘yeah, such and such was shot and we want to retaliate, or whatever’… Now we know that ‘such and such’ was shot by ‘such and such’ and ‘such and such’ is planning on retaliating at this particular place, so it’s huge. Now we are using that knowledge and we’re using it wisely. So you're preventing, now you’re being proactive with it, and not just waiting for the shoe to drop …
Educating potential victims
The final emergent component of police commanders’ efforts to address violent crime was their shared belief that educating the public on how to protect themselves from victimization was important. While educating potential shooting victims was not discussed, education emerged as a common approach for addressing robbery during three interviews and all but one-field observation. 12 The commanders’ rationale for educating victims centered on “target hardening.” It was believed that potential victims would take precautionary actions to protect themselves from victimization if they were educated on possible actions they could take.
Educating potential victims involved various activities that took place across numerous mediums, such as talking with citizens at community meetings, talking with students at schools and local universities, distributing flyers around neighborhoods, calling residences via the reverse 911 system, and even writing articles for local newspapers. Captain L described using a variety of means, such as handing out fliers to residents and businesses, to educate citizens on how to better prevent their own victimization. [E]ach lieutenant has their own PSA [Police Service Areas] meetings with the community. And I have a town hall meeting, and a workshop that we have here [in the District] … We like to give people information, we try to educate them on what's going on in their area … And I've also, several times, put out, robo calls, reverse 911 calls out to the people in the district, you know? Education is a lot worth it. That's why I try to do an article in the [neighborhood paper], that's why I do the reverse 911 calls, community meetings … Anything that we can do to get the message out … A lot of times we try to tell the kids [students at a local university] [via email], ‘know your environment, walk in pairs, don’t be out by yourself. Certain areas where it's dicey, maybe have three or four. Don’t, maybe not walk. Maybe jump in a cab. If you’ve been drinking don’t stumble down the street.’ [W]hat we’re trying to do is go to our schools, there's [number] schools here in the [district], go to our schools, we created posters about describing how not to be a victim … we hung them up in every school here in the [district], we made flyers and pamphlets to hand out to each student, saying the same things as on the posters, and then what we’re doing is going to the general assemblies that each school has normally a couple times a year, we're going to them and repeating the same thing. We think, you know, here that if we constantly repeat that message eventually somebody will hear …
Discussion
The present study used observations of and interviews with police commanders from one large municipal police department that had publicly adopted an organization-wide hot spots policing strategy to get inside hot spots policing’s “black box” and found which tactics police commanders used, what police commanders believed they were doing, and why they believed those actions would be effective during a non-experimental implementation of hot spots policing. The results revealed police commanders generally focused on increasing police presence and enforcement in high crime areas. The importance of focusing these actions on high-risk offenders was also stressed. The belief was that these actions would deny potential offenders and victims access to those locations in order to reduce opportunities for violent crime outright or make offenders “think twice” about offending in those areas. These effects were expected for both the general offender population as well as for specific high-risk offenders who came to the attention of the police. When high-risk offenders were the focus of police, the goal was that these offenders would eventually be arrested and incarcerated. Getting these high-risk offenders “off the street” was necessary because police commanders believed they would continue to offend otherwise. Finally, police commanders made efforts to educate potential (robbery) victims on how to protect themselves from victimization in hopes they would later engage in target hardening. These findings provide important insights for thinking about hot spots policing theory, policy, and evaluation.
The hot spots policing tactics employed by the studied commanders are not necessarily innovative (see Braga & Weisburd, 2006, pp. 145–149). The police have long focused on increasing presence and enforcement, albeit in geographically large areas (Weisburd & Eck, 2004). Further, critics of hot spots policing have criticized the police for relying on these tactics (Rosenbaum, 2006), albeit Braga and Weisburd (2006) argued these tactics may be “good enough” to generate crime control gains. The commanders’ expressed rationales for the effectiveness of their tactics also closely align with commonly cited crime control theories. For example, making the broader offender population or specific offenders “think twice” might be characterized as general and specific deterrence. Getting specific offenders “off the street” is consistent with selective incapacitation. Denying access to attractive crime locations and target hardening are components of situational crime prevention (Clarke, 1980, 1995, 2008).
On the other hand, the results suggest police are more calculated and sophisticated than their critics sometimes give them credit for. First, the police commanders studied have perspectives that are aligned with highly regarded criminological research. It was already noted how the commanders’ rationales for their hot spots policing tactics aligned with popular crime control theories. Likewise, police commanders also recognized and strategized based on the notion that only a small number of offenders are responsible for a disproportionate amount of crime (Piquero, Farrington, & Blumstein, 2003; Wolfgang, Figlio, & Sellin, 1972) as well as the fact that desistance is a complicated process (Maruna, 2001), perhaps even involving significant life events (Laub & Sampson, 2009), that will not occur overnight. In other words, their focus on high risk, repeat offenders are consistent with how many leading criminologists view criminal careers. Of course, the police must be responsive to the public who expect an instantaneous response to crime problems rather than waiting for offenders to complete the desistance process.
Second, the police commanders studied are quite sophisticated theoreticians. Rather than emphasizing a single rationale, the police commanders simultaneously emphasized numerous mechanisms by which the tactics they employed may work to control crime. This perspective is different than academics’ emphasis on finding the single “best” theoretical explanation of phenomena or police scholars’ emphasis on and search for the “best” way to conduct hot spots policing (Braga et al., 2012). In other words, the findings suggest that police commanders’ professional beliefs are that multiple theoretical mechanisms can work simultaneously to reduce crime. In practice, it may be necessary for future hot spots policing strategies to be designed to use tactics that work through multiple theoretical mechanisms simultaneously in order to achieve optimal crime control.
In the context of the previous two points, in effect, the police commanders’ perspective on hot spots policing is aligned with what is essentially the most comprehensive academic theory of hot spots policing to date (Nagin, Solow, & Lum, 2015). The police commanders embraced generating deterrence through both the sentinel and apprehender police roles in their views of police presence and enforcement (Nagin, 2013). Likewise, the studied commanders’ emphasis on encouraging target hardening via citizen education is consistent with Nagin et al.’s (2015, p. 80) recognition that increasing non-sanction costs can be an effective policing tactic. On the other hand, the police commanders’ belief that “getting high risk offenders off the street” can disrupt crime problems equates to the hypothesis that incapacitation via the police apprehender role may play a more salient role in determining crime rates than was emphasized by Nagin et al. (2015, p. 79). Nonetheless, the present findings suggest crime policies derived from the authors’ theory, with possible further refinement, may interest police commanders thereby, perhaps, increasing the likelihood of successful adoption and implementation (for a critique of the theory, see Pickett & Roche, in press).
The police commanders’ perspective on hot spots policing, however, could also be critiqued by drawing on the criminological literature. A reliance on police presence and enforcement will likely always raise concerns about racial disproportionality or procedural justice (e.g., see Brunson, 2007; Gelman, Fagan, & Kiss, 2007; Greene, 1999; Rosenbaum, 2006). The difficulty of accurately predicting high-risk offenders is also well known and may be even more difficult to do when a city’s entire population is considered rather than just adjudicated offenders (Auerhahn, 1999). The criminal justice system’s lack of resources may make it difficult to both obtain the proper dosages of police presence and enforcement as well as incarcerate high-risk offenders prior to conviction. Of course, the previous two examples ignore the ethical arguments against these practices (Auerhahn, 1999; Rosenbaum, 2006). Finally, high-risk offenders may merely be replaced by other offenders (Blumstein, 1993; Piquero & Blumstein, 2007; Zimring & Hawkins, 1995). Quantitative evaluations of the present model and its respective critiques will be needed to assess its practical value.
Nonetheless, the results contribute some additional insights for future hot spots policing policy and evaluation. The model that emerged from the data provides a starting point for further understanding the effectiveness of hot spots policing in the future. Specifically, researchers can make efforts to directly measure the dosage of each action in the model. Next, the links among the actions taken and changes in the intervening mechanisms can be assessed using appropriate data sources and measures. For example, community surveys might be used to measure the extent to which changes in police presence link to changes in residents’ perceptions of arrest risk or punishment certainty (see Corsaro & Brunson, 2013). Data on criminal justice case processing might be used to develop measures of the extent of incapacitation effects in crime hot spots. The different links in the model can then be manipulated (through randomization) to evaluate each component. Additional links can be added to the model to determine if additional mechanisms, such as situational crime prevention, can improve the model’s overall effectiveness (see Goldkamp, 2010, pp. 476–478).
Since this model may not generalize to how police commanders from other departments view hot spots policing, researchers can also use this study as a guide to undertake similar studies in other jurisdictions before impact evaluations in order to develop more robust hot spots policing evaluations. This approach may have a number of benefits. First, by directly testing a collaborating police department’s underlying causal model, researchers can likely increase buy-in during hot spots policing evaluations by designing evaluations that interest local decision makers. In other words, police departments may be more likely to enter into a long-term academic partnership or implement the treatment with integrity when the evaluation is directly testing their own views on police effectiveness. Second, explicitly outlining and testing police commanders’ beliefs and assumptions may increase the likelihood that police commanders will accept results that potentially contradict their long-held beliefs because they will be able to see how the research design and findings followed directly from their own beliefs. Further, outlining all assumptions a priori reduces police commanders’ ability to reject unfavorable findings because the researchers “did not evaluate it right.” Third, if evaluations repeatedly do not support police commanders’ assumptions of hot spot policing effectiveness, then they may be more open to additional (innovative) hot spots policing tactics or, in converse, police researchers may have to adjust their theories and perspectives. In short, hot spots policing evaluations may stimulate more extensive reform by (at least initially) aligning more closely with the police’s perspectives and beliefs.
Finally, the results demonstrate that hot spots policing is far more complex than the designs of previous evaluations have considered. This is not a criticism of the researchers, but rather typical of how evaluation research evolves. Early evaluations determine if an effect exists at all and later work seeks to understand any observed effects in more depth (Shadish, Cook, & Campbell, 2002). The present model illustrates the theoretical mechanisms underpinning tactics that are far more simple than many police reformers advocate for are still quite complex, and more rigorous evaluations are needed to develop a deeper understanding of how hot spots policing tactics link to specific theoretical mechanisms and produce crime control benefits. Hopefully, this study will demonstrate to hot spots policing evaluators that the careful development and testing of causal models for different hot spots policing tactics is necessary in order to fully understand what hot spots policing “is” and how it works.
The findings and implications of this study should be considered in light of its limitations. First, the study took place in a single city or police department at a single point in time, so the present findings may not generalize to other settings. Second, one might question the validity of the data for multiple reasons: (a) the possibility that reactivity occurred during the field observations (Padgett, 2008), (b) commanders could have been untruthful during the interviews (Mastrofski & Parks, 1990), (c) the interview sampling approach only included a relatively small number of commanders with a potentially unique subset of characteristics, or (d) hot spots policing strategizing may have taken place outside of the crime strategy meetings.
While readers should be aware of these limitations, the present study was carefully designed based on the rationales discussed in the methods section, with somewhat limited research resources, and with an effort to minimize disruptions to PPD’s operations. Overall, the goal of the study was not to generalize the findings to all police departments but to gain an in-depth understanding of hot spots policing in everyday practice perceived to be missing from the literature (i.e., “open the black box”).
Efforts were also made to anticipate and minimize the limitations. For example, crime strategy meetings participants frequently took notes, so the author dressed similar to meeting participants and attempted to take notes without drawing attention or disrupting the meetings (Padgett, 2008). It is probably unrealistic to believe the PPD would change their style of policing and jeopardize public safety or their jobs due to the author’s presence at the meetings. Efforts to minimize untruthfulness during the interviews included (a) informing subjects of their rights and institutional review board protections (i.e., promising confidentiality) as research subjects, (b) promising personal identifying information would be omitted from the data or subsequent research findings, and (c) using the author’s professional relationships with PPD personnel developed during previous projects to develop rapport with interview subjects. Finally, while the interview sample size was relatively small (albeit roughly 29% of a population is respectful), data were collected for all PPD commanders at least once during the field observations. The interview and field observation data also corroborated each other, and there was nothing to indicate the qualitative data were unrepresentative of how police commanders understood and implemented hot spots policing (including reviews of this article by PPD personnel).
In sum, there is no way to know for sure how much the limitations of the present study impacted the results and the findings should not necessarily be generalized to other contexts, but the findings are still useful for thinking about future hot spots policing strategies and evaluations. Similar studies in other jurisdictions, using additional methods (e.g., document reviews), will begin to help us refine the model or develop other models while assessing the extent to which different models generalize to other departments. Scholars could also expand on these findings by asking police personnel about their knowledge, receptivity, and perceptions of (evidence-based) actions they are not currently using as well (see Lum, Telep, Koper, Grieco, 2012; Telep & Lum, 2014; Telep & Winegar, in press). Nonetheless, for now, the present study clearly demonstrates that police commanders’ perspectives on hot spots policing are complex, but using systematic research to understand those perspectives can be useful for thinking about and improving future hot spots policing policy and evaluation.
Footnotes
Acknowledgments
The author is grateful for that funding, but the expressed findings, conclusions, and recommendations are the author’s and do not necessarily represent those of the National Institute of Justice, Office of Justice Programs, or U.S. Department of Justice. Jerry Ratcliffe, Ralph Taylor, Jen Wood, and the anonymous reviewers are thanked for their helpful comments on this work. The author would also like to thank Police Commissioner Charles Ramsey, Deputy Commissioner Kevin Bethel, and Deputy Commissioner Nola Joyce for supporting this study as well as the commanders who granted him access to study their efforts to make the City of Philadelphia safer.
Declaration of Conflicting Interests
The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the National Institute of Justice under award number 2014-IJ-CX-0007.