Smart Stops - 2013-MU-CX-0012 summary report draft.pdf

2013-MU-CX-0012 summary report draft.pdf

1. 

   Smart Stops:An Inquiry into Proactive PolicingRobert E. WordenSarah J. McLeanAndrew P. WheelerDanielle L. ReynoldsCaitlin DoleHannah CochranSummary Report to the National Institute of JusticeDecember, 2018This research was supported by Award No. 2013-MU-CX-0012, awarded by the National Institute of Justice, Office of Justice Programs, U.S. Department of Justice, to the John F. Finn Institute for Public Safety, Inc. The opinions, findings, and conclusions are those of the authors and do not necessarily reflect those of the Department of Justice.FI N NThe John F. Finn Institutefor Public Safety, Inc.Robert E. Worden, Ph.D., Director • Sarah J. McLean, Ph.D., Associate Director421 New Karner RoadSuite 12Albany, NY 12205518.456.6323
2. 

   The John F. Finn Institute for Public Safety, Inc., is an independent, not-for-profit and non-partisan corporation, whose work is dedicated to the development of criminal justice strategies, programs, and practices that are effective, lawful, and procedurally fair, through the application of social science findings and methods. The Institute conducts social research on matters of public safety and security –crime, public disorder, and the management of criminal justice agencies and partnerships –in collaboration with municipal, county, state, and federal criminal justice agencies, and for their direct benefit. The findings of the Institute’s research are also disseminated through other media to criminal justice professionals, academicians, elected public officials, and other interested parties, so that those findings may contribute to a broader body of knowledge about criminal justice and to the practical application of those findings in other settings. The Finn Institute was established in 2007, building on a set of collaborative projects and relationships with criminal justice agencies dating to 1998. The first of those projects, for which we partnered with the Albany Police Department (APD), was initiated by John Finn, who was at that time the sergeant who commanded the APD’s Juvenile Unit. Later promoted to lieutenant and assigned to the department’s Administrative Services Bureau, he spearheaded efforts to implement problem-oriented policing, and to develop an institutional capability for analysis that would support problem-solving. The APD’s capacity for applying social science methods and results thereupon expanded exponentially, based on Lt. Finn’s appreciation for the value of research, his keen aptitude for analysis, and his vision of policing, which entailed the formulation of proactive, data-driven, and –as needed –unconventional strategies to address problems of public safety. Lt. Finn was fatally shot in the line of duty in 2003. The Institute that bears his name honors his life and career by fostering the more effective use of research and analysis within criminal justice agencies, just as Lt. Finn did in the APD.
3. 

   Smart StopsiiiContentsINTRODUCTION..........................................................................................................................1THE PARTNERSHIP.....................................................................................................................2Institutionalization......................................................................................................................3TOWARD SMART STOPS...........................................................................................................4Proactive Policing: Benefits.......................................................................................................5Proactive Policing: Costs...........................................................................................................7Smarter Stops..............................................................................................................................9High-Value Stopsin Albany......................................................................................................10Stops in High-Crime Places..................................................................................................11Stops of High-Risk People....................................................................................................17Positive Enforcement Outcomes...........................................................................................20High-Value Stops..................................................................................................................20Feedback...................................................................................................................................23PUBLIC ATTITUDES.................................................................................................................25Experience with Police -Stops.................................................................................................27Vicarious Experience with Police.............................................................................................27TRENDS IN PROACTIVE POLICING.......................................................................................29Cross-Sectional Patterns..........................................................................................................30Longitudinal Patterns...............................................................................................................31Officers’ Perceptions................................................................................................................32Conclusions...............................................................................................................................32SKILLS OF PROACTIVE POLICING........................................................................................33Sample and Method...................................................................................................................33Interview Sample..................................................................................................................33Results.......................................................................................................................................35Forming Suspicion................................................................................................................35Learning the Skills of Proactive Policing.............................................................................38Styles of Proactive Policing..................................................................................................40Summary of Findings................................................................................................................42CONCLUSIONS...........................................................................................................................43BIBLIOGRAPHY.........................................................................................................................44
4. 

   Smart Stops1INTRODUCTIONOver the last four decades many police agencies, and the executives who direct them, have exhibited a remarkable willingness not only to innovate but to subject themselves to the potentially harsh light of social scientific inquiry. For their part, many researchers have seized (or created) opportunities to partner with police agencies to conduct scientific investigations that have broadened and deepened our knowledge of what the police do, to and for whom, and with what consequences, from the micro-level of police-citizen encounters on the street to the macro-level of police strategy. Yet the utilization of research methods and research findings by police agencies is far from what it could be (see Weisburd and Neyroud, 2011), and the supply of research (and researchers) to guide police in making strategic, operational, and tactical decisions is far less than optimal (Rojek, Smith, and Alpert, 2012). One need not espouse the most demanding forms of evidence-based policing (Sherman, 1998) to believe that research could be more useful to, and better utilized by, police practitioners (see Moore, 2006; Sparrow, 2011). Differences in occupational orientations and incentive systems continue to hamper the formation and maintenance of constructive working relationships between police practitioners and researchers, despite efforts at federal and state levels to encourage and sometimes fund practitioner-researcher partnerships. Some existing or recent partnerships are notable for their apparent success, but they seemto be exceptions to an unfortunate rule. While we have learned some lessons about such partnerships over the last ten to fifteen years, we have much to learn about how these partnerships can be sustained and put to their greatest use.Among the many questions that practitioner-researcher collaborations might address are those that concern proactive policing. Social science has established that proactive policing can reduce the incidence of violent crime, particularly when police attention is strategically focused on the places and/or the people associated with high levels of violence. Several studies show that higher levels of police-initiated contacts with citizens, achieved through routine patrol or dedicated, directed patrol,are associated with lowerlevels of crime. Conventional wisdom holds, however,that proactive policing antagonizes the community, and while some research shows that such negative outcomes are not inevitable, other research shows that stops tendto yield less favorable attitudes toward police, detractingfrom police legitimacy. Police stops have been controversial in American cities, prompting public concerns about racial profiling, legislation mandating traffic stop data collection, civil litigation alleging civil rights violations, and organizational reforms implemented under consent decrees. It appears, then, that proactive policing poses a trade-off between crime control and legitimacy.Aside from the issue of racial biases, research has neglected questions about patterns of stops: where and when stops are made (i.e., the proportion of stops that are made in high-crime places at high-crime times), how many of those who are stopped are at demonstrably high risk for crime, how officers make judgments about whom to stop and in what post-stop activities to engage, and what factors are associated with successful stop outcomes. In some or all of these respects, it may be possible to increase the efficiency of proactive policing, by increasing the ratio of successful or other “high-value” stops to all stops. In this way, proactive policing would be conducted more surgically, such that the people who are stopped by police are those most likely to offend, and the stops that are made would have the greatest potential crime-reduction benefits, minimizingthe numbers of ‘innocent’ people who are caught–even if only momentarily –in the police net. Insofar as proactive policing can be conducted more efficiently,
5. 

   Smart Stops2then it may be possible to side-step the trade-off between crime control and legitimacy.This project provided for steps that would strengthen and enhance an existing police-researcher partnership, focused around analyses of proactive policing. The Albany Police Department (APD) and the John Finn Institute (the Institute) had an existing partnership that the partners sought to better institutionalizethrough a project whose substantive focus wasproactive policing. We plannedto undertake collaborative analysis that promised to make proactive policing in Albany better targeted andmore efficient, with more stops of high-value for crimecontrol, and correspondingly fewer lower-value stops, and thus less likely to have collateral consequences for community relations. Thus our purpose was for police to make “smart stops.”The project consisted principally of two distinct but related analytical undertakings. First, in collaboration with APD partners, we analyzed stops to differentiate between high-and lower-value stops, assess the potential for increased efficiency, and formulate an analytic product (or “bulletin) that would provide feedback on the nature of officers’ stops to operational commanders and supervisors. We planned thereupon to examine the nature and extent of changes in patterns of stops. Second, we identified officers who were skilled in proactive policing and sought to learn from them how they performed this task. Following a description of the partnership between APD and the Institute, we summarize the first of these undertakings and two related analyses, one of public attitudes and the other of the trend in proactive policing over time in Albany. Then we summarize the findings from our study of skilled officers. Our report ends with our conclusions.THE PARTNERSHIPAlbany is a city of nearly 100,000 residents, though as the home of a large state government, its daytime population swells. The APD has an authorized strength of approximately 345sworn officers. Uniformed patrol is organized in two divisions, each overseen by a commander,and within each division, three squads each commanded by a lieutenant; the city’s 22square miles are divided into 19patrol zones. In addition, the Neighborhood Engagement Unit (NEU) deploys officers to one of 33 narrowly circumscribed beats (which overlap with the larger patrol zones), in which NEU officers practice problem-solving and community policing. The Special Operations division includes a specialized traffic unit. The Information Coordination Unit (ICU) performs crime analysis, and the Training Unit develops and delivers in-service training. APD has made a large investment in its information infrastructure, with NIBRS-compliant crime reporting, and mobile access to crime and other information in the field. The APD Records Management System (RMS) contains information on crimes reported to the police and recorded (electronically) by officers on incident report forms, as well as information on arrests and field interviews. The Computer-Aided Dispatch (CAD) system includes records of field-initiated incidents. Each of these sources of electronic information supports tabular extraction of records for analytical purposes. APD has a history of partnering with the Institute researchers, one that even predates the establishment of the Institute; and the Institute is named in honor of the fallen Albany police lieutenant who initiated the partnership. In 1998, with a COPS grant to apply problem-oriented policing to the problem of Albany High School students bullied on their way to and from the school, then-Sergeant Finncommanded APD’s Juvenile Unit and reached out to Worden at SUNY-Albany for assistance with the analysis. By 2000, the research partnership expanded its
6. 

   Smart Stops3purview, seeking to extend the availability of analytic results for operational decision-making, and with Finn’s transfer to the Administrative Services Bureau and later promotion to lieutenant, APD took long strides toward building a robust analytic capacity and an appreciation for research. In 2001, with City funding, Worden conducted a survey of Albanyresidents regarding their perceptions of police and quality of life; such a survey had not previously been done in Albany. Later that same year, with City funds and APD’s cooperation, Worden commenced two on-going surveys for the Citizens’ Police Review Board (CPRB), one of complainants and one of “clients,” i.e., people who had a recent contact with APD officers by calling for service, being stopped, or being arrested. Drawing bi-weekly samples and interviewing approximately 800 clients each year, the surveys continued for three years, until 2004.The introduction of New York State’s Operation IMPACT1provided funding and a renewed focus on crime in 2004, whereupon Worden, McLean, and several graduate research assistants: assisted with the development of a gang database and conducted social network analyses of gangs; linked data on field interviews with probation data to better monitor probationers’ contacts with police; conducted a process evaluation of “senior staff meetings,” which served as Compstat-like management accountability meetings; analyzed patterns of violent crime; summarized research on chronic offender strategies and repeat burglaries for strategic consideration; and conducted an outcome evaluation of Operation Safe Corridor. In 2009-2010 we updated our survey of complainants for the CPRB, and also conducted a mail survey of residents. Throughout this period (2004-2010), and despite regular attendance at senior staff meetings, the researchers’ main points of contact were the chief and the assistant chief for administrative services, who was the primary grant-writer. At the time that this project was proposed, in 2013, this partnership had been sustained across six chiefs of police, who brought to their role varying levels of appreciation for research, and despite the untimely death in 2004 of a widely respected internal advocate for research. In retrospect, however, it was clear that the partnership lacked an institutionalized infrastructure. While we made concerted efforts to make the acquaintance of and remain in contact with many members of the command staff, fromchief to deputy and assistant chiefs to commanders and lieutenants, we had not established an organizational vehicle for systematically and regularly bringing research to bear on operational and managerial issues, either by reviews of extant research or by formulating plans for original research in APD. With this recognition, we planned to take appropriate steps to establish an infrastructure, as described below. InstitutionalizationFor this project, the partnership was formalized in terms of a steering committee –the Research Advisory Council (RAC) –and one or moreworking groups.The RACwascomprised of APDexecutive staff, determined by the chief, and Institute researchers. The RACsetsdirection for and exercisesoversight over partnership work, monitoring the progress of projectworking groupsand distilling implications from research findingsas they are generated.It also serves as a forum for formal exchanges about research, with APD personnel bringing operational and administrative issues for discussion about how extant research can help to address them, and the researchers bringing new theory and empirical findings from the field for discussion about their implications for policy and practice. In this way the RAC canfoster greater ownership of 1Operation IMPACT (Integrated Municipal Police Anti-Crime Teams) was a New York State initiative that resembledProject Safe Neighborhoods, and which providedfunding for strategic crime-reduction interventions by multi-agency task forces in each of seventeen jurisdictions across New York State.
7. 

   Smart Stops4and involvement in research by APD personnel (see Weisburd andNeyroud, 2011). In our view, the integration of research into police practice cannot be sustained without an organizational mechanism that is both catalyst and guide, and the RAC is designed toserve this purpose. Working groups, in general,arecomprised of APDpersonnelfrom the units whose functions pertain to the projects, respectively, and so this initial project’s working group included patrol lieutenants and sergeants in addition to the researchers. The responsibilities of working groups includereviewing and refiningdata analysis plansand introducingprojectsto APDpersonnel who will participate as research subjects. The groupsreview and interpret initial research resultsand guide further analysis, and from the findingsthe groupsdistill implications for policy and practice and prepare strategic recommendations for the RAC’sconsideration. In this context, the partnership has continued to prosper, despite continued turnover at executive levels. The chief who endorsed this project departed in 2015, replaced by the deputy chief, who retired in 2017. His successor held the title of acting chief throughout his tenure, from April of 2017 through August of this year. Nevertheless, in addition to the smart stops project, Institute researchers have worked with APD on several evaluations of Gun Involved Violence Initiative (GIVE) strategies, a process evaluation of LEAD, and a project currently inprogress on police supervision.2TOWARD SMART STOPSInvestigatory stops are simultaneously one of the most fundamental tools thatpolice officershaveto enforce the lawand one of the most controversial.The jurisprudence of proactive policing is complicated (see, e.g., Fagan, 2016; Friedman, 2017), but it is fair to assert that the authorityto detain a person, at their discretion without regard to other agents in the criminal justice system, is one of the defining characteristics of modern police (Bittner, 1979).Conventional police wisdom holds that proactive policing –that is, officer-initiated contacts with violators or suspicious persons –is an effective crime control tactic, and the findings of social research have for the most part supported this proposition. Studiesusing different methodologies with different strengths and weaknesseshave found thatthe incidence of some types of crime declines, or is lower than one would otherwise predict, whenand where the police frequently make field stops of vehicles and/or pedestrians.Conventional wisdom also holds that the crime control benefits of proactive policing comes potentially at the price of police legitimacy, as stops may detract from public trust in the police. While social science evidence does not support strong causal inferences on this question, we should be cognizant of the trade-off between crime control and public trust that proactive policing may entail.Strategically focusing police resources onhotspots of crime has become a well-accepted police tactic, one whose effectiveness may turn at least partly on proactive tactics (Rosenfeld, Deckard, and Blackburn, 2014;Groff, Ratcliffe, Haberman, Sorg, Joyce, and Taylor, 2015). The concentration of crime in particular places is well known (Eck, Clarke, andGuerette, 2007; Sherman Gartin, andBuerger, 1989; Spelman, 1995; Weisburd, 2015), and targetedinterventions at such hotspots have enjoyedmoderate success in reducing crime (Braga, Papachristos, andHureau, 2014; Sherman andWeisburd, 1995).Insofar as hotspots policing involves police stops of suspected offenders, it represents a spatial refinement of a time-honored police tactic.2GIVE is the successor to New York State’s Operation IMPACT.
8. 

   Smart Stops5Wesuggest that what is good for hotspots policing may be good for proactive policing more generally: police agencies may be able tobe more efficientin crime control and also minimize the detrimental effects of proactive policingby bettertargetingtheir activity, making “smart stops.” By differentiating among stops in terms of their potential for crime reduction and managing them accordingly, it may be possible to realize the crime control value of proactive policing and at the same time reduce the strain that it can place on police-community relations.Using data on over 200,000 discretionary police stops by the Albany Police Department from2004 through 2013,we analyze three characteristics of stops: their locations; the people who are stopped; and their outcomes. We posit that stops have greater presumptive value for crime reduction when they are conducted in high-crime places, involve high-risk people, or produce enforcement outcomes. Such stops are “high-value” stops. Proactive Policing: BenefitsEmpirical inquiry into the crime reduction effects of proactive policing has taken various forms, running the gamut from randomized controlled trials to quasi-experimental designs to non-experimental analyses of naturally occurring variation. In 1973-1974, San Diego police experimented with field interrogation (FI) practices, assessing the effects of “traditional” FI practices and ofspecial training designed to reduce “friction” between officers and citizens (Boydstun, 1975). In the no-FI area, FIs were suspended for nine months. In the special-FI area, only the specially trained officers conducted FIs. The cessation of FIs was associated with an increase of nearly 40 percent in suppressible crimes, such as robbery and burglary, while their subsequent resumption was associated with a decrease. The special training did not appear to affect citizen attitudes.Evaluations of proactive gun patrols in Kansas City, Indianapolis, and Pittsburgh found crime reduction effects.Kansas City experimented with gun patrols about 20 years ago with a couple of units dedicated to proactive patrol in the form of vehicle stops and field interrogations, checking for guns in known gun hotspots (Sherman andRogan, 1995).The intervention was applied to a single police beat with a high level of gun crime, lastinga bit longer than six months, and it appeared to be successful in reducing gun crime in that beat, compared to another similar beat that did not receive enhanced enforcement efforts.These effects, however, were limited to gun crime, and did not extend to other types of crime.Indianapolis experimented with such patrols in two of its police districts, each for a period of 90 days,though the tacticsdiffered some from one another(McGarrell,Chermak, Weiss, and Wilson, 2001).In the East District the objective was to maximize vehicle stops, and thus to increase police visibility and also seize illicit drugs and guns; this approach was not effective in reducing gun crime, however.The North District implemented what they called a targeted offender approach, with greater selectivity in making stops and with a more thorough investigation, thereby maximizing seizures of illegal guns anddrugs; this approach was associated with decreases in gun crime. Pittsburgh’s Firearm Suppression Program comprised a directed patrol strategy, with stepped up patrols in two areas on high-crime days at high-crime times (Cohen and Ludwig, 2003). Althoughthe dosage was fairly low (one additional three-vehicle patrol team was assigned to each area) an evaluation indicated that gun crime (shots fired and aggravated assault) was affected.Both areas had gun crime trends similar to control neighborhoods during the “off” days, suggesting that the deterrent effect of the patrols was limited to the “on” days when the program was in effect.
9. 

   Smart Stops6Non-experimental analyses of policing and crime in cities and neighborhoods have also foundthat proactive policing reduces seriouscrime, especially robbery.For many such studies the measures of proactivity leave something to be desired, and of course the analyses are subject to caveats about model specification, estimation, and causal inference. Since crime and enforcement activity have, a priori, reciprocal effects, isolating and estimating the effect of the latter on the former confronts serious analytical challenges. Wilson and Boland (1978) contrast “aggressive” police strategy, which provides for frequent contacts with the community, with a passive strategy; an aggressive or proactive strategy, resembling the legalistic style that Wilson (1968) previously described, involves frequent vehicle stops for moving violations and to check for stolen cars or wanted persons, and frequent stops to question suspicious persons. Treating traffic citations as a proxy measure of a proactive strategy, Wilson and Boland analyzed three types of crime –robbery, burglary, and auto theft –in 35 large cities in 1975, finding that police proactivity has a beneficial effect on robbery, but on neither burglary nor auto theft. Whitaker, Phillips, Haas, and Worden, (1985)analyzed proactivity and crime at the neighborhood level, measuring four forms of proactivity through systematic social observations of police patrol and crime through a victimization survey. Their analysis found evidence that suspicion stops have effects on robbery, burglary, auto theft, and vandalism (but not on thefts from cars).Sampson and Cohen (1988) replicated Wilson and Boland’s analysis and extended it to include enforcement activity directed toward disorder. They also disaggregated enforcement and offending by race and age. Their indicator of proactive policing differed –arrests per officer for disorderly conduct or driving under the influence –and their sample of cities was larger (171); they analyzed 1980 crime levels. They found that police proactivity affects robbery and (more weakly) burglary, in the aggregate and also, for the most part, when the analysis is disaggregated by race and age. Kubrin, Messner, Deane, McGeever, and Stucky (2010) further improved on this line of analysis, forming a better specified model of robbery and employing panel data for 2000-2003. They too found that proactive policing reduces robbery. More recently, in their examination of stop, question, and frisk (SQF) in New York City, Weisburd, Telep, and Lawton (2014) found that stops were spatially concentrated in a small fraction of micro-places, i.e., street segments and intersections; they also found that stops were strongly correlated with crime, as they would be if they were part of a hotspots policing strategy. Weisburd, Wooditch, Weisburd, andYang(2016) applied two analytic approaches to disentangle the reciprocal effects of stops and crime in New York City between 2006 and 2011; both approaches yielded evidence that indicated a “significant deterrent impact” (p. 33). Other researchers have also found that intensive stop activity by the NYPD reduced crime (Rosenfeld andFornango, 2012; Rosenfeld andFornango, 2017; MacDonald, Fagan andGeller, 2016), but typically describe the effects as very small. For example, MacDonald, et al. (2016) find that it would take 100 probable cause stops to prevent 3 Part I crimes. Recent hotspots policing experiments in St. Louis and Philadelphia corroborate the quasi-and non-experimental findings.In St. Louis, Rosenfeld, et al.(2014) tested the effects of two forms of policing on firearm crime: one was directed patrol, meaning a police presencein hotspots but with a minimum of officer-initiated activity, while the other was directed patrol plus officer-initiated activity.Proactivity included occupied vehicle checks (vehicle stops), pedestrian checks, building checks, unoccupied vehicle checks, foot patrol, and problem-solving.Each of these treatments was delivered to 8 randomly assigned hotspots, while both were withheld from 16 control hotspots, which received policing as usual.Directed patrol by itself had no detectable
10. 

    Smart Stops7effects on firearm crime relative to the control condition, but directed patrol that included officer-initiated activity reduced nondomestic firearm assaults (but not firearm robberies).Further analysis suggests that it was to some extent proactive arrests and also occupied vehicle checks that accounted for the impact of this intervention.In Philadelphia, two experiments illuminate the effects of proactive policing.First, Philadelphia’s foot patrol experiment provided for the deployment of new police academy graduates to 60 foot beats in violent crime hotspots (Ratcliffe, Taniguchi, Groff, and Wood, 2011).The foot patrol treatment proved to be effective in reducing violent crime, particularlyin higher-crime hotspots; moreover, it appeared that proactive policing –pedestrian stops, vehicle stops, and police involvement in incidents that tends to be officer initiated (disturbances, and narcotics and disorder incidents) –was instrumental in achieving these crime reduction effects.Second, the Philadelphia policing tactics experiment tested the crime reduction effects of three forms of hotspots policing: foot patrol; problem-oriented policing; and offender-focused policing(Groff, et al., 2015).Offender-focused policing, which provided for extra attention to identified repeat violent offenders, proved to be effective in reducing violent crime.The extra attention consisted of surveillance and aggressive patrol: “officers made frequent contact with these prolific offenders rangingfrom making small talk with a known offender to serving arrest warrants for a recentlycommitted offense (Groff, et al., 2015: 34).Insofar as proactive policing has crime control benefits,the crime reduction is apparently achieved not only or primarily through incapacitationbut through deterrence, directly increasingthe risk of apprehension perceived by would-be offenders. Onlya small fraction of stops resultsin arrests or in the seizure of guns or illicit drugs (see, e.g., Boydstun, 1975; Fyfe, 1998; McGarrell, et al., 2001; Rosenfeld, et al., 2014; Sampson and Cohen, 1988; Whitaker, et al., 1985; Wilson and Boland, 1978).Thus even a stop that does not culminate in legal action might serve to underscore the riskof apprehension for the person(s) stopped (and perhaps others through word of mouth).In Kansas City, dedicated gun patrols seized 29 firearms in as many weeks of operations totaling 4,512 manhours, during which police made 616 arrests and issued 1,090 tickets (Sherman and Rogan, 1995).This is not unambiguously inefficient.It might be inefficient if enforcement activity were disproportionate to crime, but as Weisburd, et al. (2014) show, the numbers of SQFs in New York are strongly correlated with crime at the level of street blocks.The crime reduction effects of proactive policing appear to hold more consistently for some types of offenses –violent crime, and especially gun crime –than for other types of crime, such as burglary.3But even for gun crime the supportive evidence is not uniform; Rosenfeld, et al. (2014) detected impacts on firearm assaults but not on firearm robberies, which they acknowledged as puzzling. Be that as it may, we take the accumulated evidence to indicate that proactive policing is an evidence-based strategy, with demonstrable social benefits. An account of these benefits must be juxtaposed with one of the various costs and potential costs.Proactive Policing: CostsPolice stops of citizens involve the exercise of police authority, and as Mark Moore observes, “All other things being equal, we would like the police to use the authority we grant them sparingly. .... In an important accounting sense, we have to recognize the grant of authority to the police as an asset, and count its use in police operations as a cost to be weighed against the 3In addition to the studies summarized in the text, see MacDonald (2002). However, Weiss and Freels (1996) report contradictory results from Dayton.
11. 

    Smart Stops8benefits of lowering crime” (2002: 23).Even when stops are effected within Constitutional requirements, they represent intrusions into citizens’ lives –social costs that should be borne at no more than socially optimal levels.In addition, field stops and the actions that officers take in these encounters might occasionally infringe on citizens’ civil liberties.The use of discretionary police stops is not infrequently the subject of civil litigation alleging violations of civil liberties and/or racial profiling (Tillyer, Engel, andWooldredge, 2008).Like many police-citizen encounters, field stops are of “low visibility” (Goldstein, 1960); they are typically subject to limited review by administrative or judicial officials, and police misconduct can seldom be established because even when it eventuates in a complaint, the typical complainant is of dubious credibility and the claims cannot be corroborated by neutral third parties.When officers are encouraged to be proactive, one might expect that the number of stops for which officers lack reasonable suspicion would rise disproportionately. The law governing “Terry stops” is complex (Fagan, 2016), so much so that it is likely that sometimes officers will unwittingly make legally insufficient stops. One might also expect that the number of occasions on which officers further intrude unjustifiably into citizens’ private affairs–e.g., by conducting an improper search–would increase at least proportionately, even if unintentionally. Gould and Mastrofski (2004) found that searches were fairly infrequent–about one every ten hours in the field –but nearlyone-thirdof the searches were unconstitutional; the officers most prone to conduct illegal searches were in general good cops whose searches were seen as “normal and necessary” in the war on drugs.Furthermore, stops may detract from public trust in the police, especially among the segments of society whose regard for police is already the most tenuous.Stops would seem to inevitably cast police and citizens into adversarial roles, and one might expect that many citizens would come away from such an experience with less favorable views of the police.Moreover, insofar as minority citizens areand/or perceive that they are more likely to be the subjects of field stops, they are likely to regard proactive policing as a form of harassment.Although findings based on cross-sectional surveys surely overstate the effects of a personal experience of a stop, even panel survey findings show that stops detractfrom police legitimacy, as theytend to yield less favorable attitudes toward policeindependent of prior attitudes(Brandl, Frank, Worden, andBynum, 1994; Skogan, 2005, 2006a; Tyler andFagan, 2008;Tyler, Fagan andGeller, 2014).Negative experiences tend to influence more general attitudes toward the police more than positive experiences do (Skogan, 2006a). Furthermore, it appears that such contacts might have indirect as well as direct effects,as vicarious experiences also affect citizens’ attitudes toward the police (Rosenbaum, Schuck, Costello, Hawkins, andRing, 2005).Thus the potentially adverse impact of proactive policing extends beyond those who are stopped to other members of the community. Proactive policing might, then,entail costs in the form of degraded police-community relations.However, some previous research has contradicted the expectation that higher levels of police proactivity lead to adverse community reactions (Boydstun,1975; Chermak, McGarrell, andWeiss, 2001; Ratcliffe, Groff, Sorg, and Haberman, 2015; Shaw, 1995; Whitaker, Phillips, andWorden, 1984). Indeed,residents of the most crime-ridden neighborhoods often call for increased police presence (Kennedy, 1998), and even minority communities that are disproportionately subject topolice enforcement still tend to have an overall positive perception of the police (Skogan, 2006b).Even so, we would be properly concerned about the accumulation of such experiences among substantial segments of minority communities, as many young men
12. 

    Smart Stops9of color have personal and/or vicarious experiences with being stopped by police (Brunson, 2007; Carr, NapolitanoandKeating, 2007; Gau andBrunson, 2010).While actuarial approaches to identifying and policing crime hotspots has been readily adopted by many police departments (Braga et al., 2014),somehave questioned whether this has come at the priceof disenfranchising minority communities (Tonry, 2011). Such concerns turn not only on the costsand benefits of stops but also on those ofalternative strategies,such as problem oriented policing or crime prevention through environmental design, whichare thought not to beas invasive into individual liberties (and do not generate charges of racial discrimination). But police-initiated contacts with suspected offenders or wrong-doers are among the tactics that comprise those strategies. Increased foot patrol often improves public perception of the police (Wilson andKelling, 1982), but one of its purported virtues is that officers on foot are more likely to intervene to address “broken windows,” including people who are committing low-level crimes or otherwise engaging in disorderly behavior (Ratcliffe, et al., 2011). While problem oriented policing strategies have enjoyed slightly more success in reducing crime than simply increased levels of traditional enforcement (Braga et al., 2014), increased levels of more traditional forms of policing are nevertheless common as one of the responsesto problems (Braga, Weisburd, Waring, Mazerolle, Spelman, andGajewski, 1999). Smarter StopsThat proactive policing has crime control benefits does not imply that more proactive policing is better, nor does it imply that each stop has the same marginalbenefits regardless of where, when, and of whom they are made.Stops in high-crime places –and especially at high-crime times –would probably be more valuable than stops in low-crime places. Certainly that would appear to be a lesson to be drawn from the literature on hotspots policing.In addition, stops of high-risk people would probably be more valuable than stops of low-risk people.We note that the more effective directed patrol operation in Indianapolis was the more “selective” and targeted toward known offenders, with longer and more thorough investigations (McGarrell, et al., 2001); fewer citizen contacts were made but greater effects were achieved.Enforcement attention focused on high-risk violent offenders in Philadelphia proved effective in reducing violent crime (Groff et al., 2015).MacDonald, et al. (2016), examining stop activity in New York City, finds that probable cause stops are associated with subsequent crime reductions, but stops based on merely suspicious behavior had no associationwith decreases in crime.4However, extant research has mainly neglected questions about patterns of police proactivity, other than looking for evidence of racial bias in stops and post-stop enforcement(Weisburd, et al., 2016 being a recent exception).We believe that police could do better in guiding and encouraging officers to make “higher-value” stops –those that are more likely to serve crime control purposes. So we tentatively define higher-value stops to include:stops with successful enforcement outcomes: arrests, drug or gun seizures;stops of people with significant criminal histories (previous and recent arrests for predatory offenses) or other identified risk factors (e.g., gang involvement); orstops in high-crime places.4This conclusion is contingent on their classification of the reasons for stops, which should not be accepted uncritically.
13. 

    Smart Stops10We treat stops with enforcement outcomes as high-value only advisedly, and because some studies show that arrests mediate or amplify the effects of proactivity(Jang, Lee, andHoover, 2012; Rosenfeld, et al., 2014).But extant research indicates that the effects of proactive policing on crime do not turn on making arrests.Moreover, since legal stops are based on the low standard of reasonable suspicion, false positives are to be expected.False positives are difficult to identify even in retrospect, and still more challengingto identify from police records.A person who is contemplating a crime –say a strong-arm robbery –when s/he is stopped and thereby dissuaded from its commission may be released without any further legal action by police.Furthermore, police often do notinvoke the law even when they have the legal justification to do so (Terrill andPaoline, 2007).Hence we cannot estimate the false positive rate from stop outcomes, and valuable public safety purposes can be served without an arrest or a citation (and might be better served that wayin many instances). We do not mean to imply that stops of high-risk people, or stops in high-crime places, should be made without regard to their legal justification.We do mean that in allocating their scarce resources of police time and effort, preference should be given to acting on suspicion when the person in question is a known offender, or when the location is a high-crime place.In effect, under circumstances that are not well beyond the (ill-defined) threshold of reasonable suspicion, doubts might be resolved in favor of not making a stop if it is not a high-risk person or a high-crime location.High-Value Stops in AlbanyWe applied the three criteria for high-value stops to Albany’s stops: stops in high-crime places; stops of high-risk people; and stops with successful enforcement outcomes, to stops occurring from 2005 through the end of 2014, a ten year period. Going back to 2005 was a matter of convenience, as this was the earliest that all necessary information was available electronically, including the previous 5-year arrest history for individuals stopped (as electronic arrest records date to 2000).Data on stops come from APD’scomputer-aided dispatch (CAD) and record management systems. The CAD system includesafield thatindicatesfor each recorded incident whether itwas initiatedby anofficer, along with codes that capture the nature of the incident. We endeavored to differentiate investigatory stops from officer-initiated events that either involve no contact with a person (such as building checks) or are not of an enforcement nature. We also endeavored to distinguish investigatory stops from stops made only for the purpose of traffic enforcement, though the distinction is not clear in practice and difficult to effect based on the data that are available. Vehicle stops can serve both purposes simultaneously, and a vehicle stop prompted only by a traffic infraction can become investigatory in nature. We consulted with APD personnel in the selection of incident types among officer-initiated incidents that should be treated as investigatory stops. In addition, we excluded stops conducted by traffic units on the premise that they are predominantly for the more limited purpose of traffic enforcement.Additional information about stops was drawn from arrest reports, field interview cards, and traffic tickets. We analyze stops at the incident level;a single vehicle stop of a driver and two passengers,for example,is treated as one vehiclestop rather thanas three separate stops.More than238,000 stops were recorded in the ten year period, 45percentof which were vehiclestops and 41percentpedestrianstops. (The remaining12percentcould not be classified as either a vehicle or pedestrian stop based on the CAD information.)