The Colombian National Police and the Politics of Crime Control Evaluations

The Colombian National Police inaugurated a comprehensive operational model in 2010. Informed by evidence-based law enforcement models from the Global North, the MNVCC, or the National Quadrant Policing Model, integrates core features of procedural justice, hotspots, problem-oriented and community policing strategies. Just under a decade old, empirical assessments of the model’s impact vary in quality and availability. While the Colombian National Police presents the model as a successful intervention, there is little consensus on the degree to which the MNVCC has affected crime rates or perceptions of insecurity. The core purpose of this paper is to offer insight into the political dimensions that enable this contradictory narrative. Relying on privileged access to high-level administrators inside the Colombian National Police and other institutions, this study explains how structural features of official crime data—with political incentives specific to the Colombian context—provide the basis for contradicting claims surrounding the MNVCC’s impact.


Introduction
The Colombian National Police inaugurated a comprehensive operational model in 2010. Informed by evidence-based law enforcement models from the Global North, the MNVCC, or the National Quadrant Policing Model, 1 integrates core features of procedural justice, hotspots, problem-oriented and community policing strategies. The model has existed for just under a decade and empirical assessments of its impact vary in both quality and availability. While the Colombian National Police presents the model as a successful intervention, there is little consensus on the degree to which the MNVCC has affected observable crime rates or perceptions of insecurity. The core purpose of this paper is to offer insight into the political dimensions that enable this contradictory narrative. Relying on privileged access to high-level administrators inside the Colombian National Police and other Colombian institutions, the present study explains how structural features of official crime data-paired with political incentives specific to the Colombian context-provide the basis for contradicting claims surrounding the MNVCC's impact. Findings speak to structural contradictions and political incentives that undermine the integrity of criminal justice research that purports to be both evidence-based and objective.
Like many police forces in the Global South, the Colombian National Police is centralised and administered as a branch within the Ministry of Defence. The MNVCC offers a prescriptive framework that provides flexibility for a police force with significant variation in service needs across rural and metropolitan areas, and across distinctive regions, topographies, institutional arrangements and governance challenges (see Berents and Have 2017;Gómez et al. 2017;Higginbottom 2008;Moncada 2010;Román-Muñoz 2012;Ruiz-Vasquez and Páez 2012;Palacios 2006). The core focus of the present study is to explain a core conflict surrounding the effectiveness of the MNVCC. The Colombian National Police relied on studies by a third-party contractor to claim that the model 'has been successful in reducing crime across all major cities where it has been introduced and implemented' (Fundación Ideas Para La Paz 2012). However, independent research and media organisations have reported increases in the same crime categories, time periods and jurisdictions covered by the official MNVCC evaluations. 2 This contradiction remains consistent across the entire period of the MNVCC's existence. From 2010 until the time of writing, academic literature on Colombian policing are inconsistent with the claims associated with official impact evaluations (see Blattman et al. 2017;Gómez et al. 2017;Ruiz Vazquez 2012;Ruiz Vazquez et al. 2016). The most recent publication directly questioning the validity of the Colombian National Police evaluation methods was released in 2016 (see Ruiz-Vasquez and Páez 2016), undermining the investments made in the Colombian National Police's model. With ample descriptive coverage of these conflicting narratives, the contribution of this study is to provide an explanatory framework that can be useful to scholars interested in structural contradictions and the politicisation of crime and justice data from a Global South context. This paper is structured in a manner that integrates existing literature with fieldwork data. Part 1 presents the methodology and research design. Part 2 provides a descriptive overview of the Quadrant Model (MNVCC), demonstrating how the Colombian National Police have made substantive gains in operationalising policing strategies developed in the Global North. Part 3 highlights structural limitations inherent in official crime data, while Part 4 focuses on the political incentive structures observed in the Colombian context that problematise some of the ways criminal justice evaluations are communicated. The paper concludes with forward-looking approaches that account for the findings of this paper, and briefly outlines how researchers and practitioners alike might support the ongoing project of improving citizen security and public safety in Colombian society-a space in which the Colombian National Police will continue to play an important role. IJCJ&SD 20 Online version via www.crimejusticejournal.com

Research design
Grant-funded fieldwork was conducted during the summer months of 2015. 3 The study relied primarily on semi-structured interviews, supplemented by secondary data analyses on an exhaustive search for all records, reports and information regarding the MNVCC. Interviewees and police stations visited during this study are drawn from three jurisdictions: Bogotá, Cartagena and rural township Tercera, 4 located in the boundary area where the Departments of Cundinamarca and Boyacá adjoin. Bogotá -the capital -was chosen because it is the largest and most populous city. Cartagena was selected on the basis of the author's comparative advantage in accessing extended networks consisting of former elected officials and family members who remain active in public service. The third jurisdiction, Tercera, is the location of the author's incountry residence and was selected on the basis of their familiarity with and access to the local police.
Interviews (N = 24) included a major general of the Colombian National Police who oversaw the creation and implementation of the MNVCC in its earlier years. Other participants included entrylevel patrol officers and mid-tier ranking officers from all three jurisdictions. A federal judge, business executives, community leaders, candidates for local office, a former city mayor, an administrator from the Universidad de los Andes and analysts from research institutions were also interviewed. Consistent with the study's Institutional Review Board (IRB) designation, names and identities are omitted in statements that may present liabilities to the study's participants. 5 While the semi-structured interview instrument was tailored to the type of respondent, questions common to all respondents included: • Have you observed any noticeable changes or impacts stemming from the MNVCC's implementation?
• What is the scope and purpose of the MNVCC as understood by practitioners and those whose careers have direct intersections with the model?
• Why is there a contradiction between official discourse pertaining to trends in crime and public safety and other independent localised research organisations?
• How should one understand or neutralise this contradiction in messaging?
• What political considerations affect the activity of the Colombian National Police?
• What activities of the Colombian National Police appear 'exportable' as best practices? In other words, what appears to be most effective?
Interview respondents were recruited based on referrals and availability sampling. This decision is consistent with research designs in other high-risk fieldwork involving hard-to-access or elite populations (Babbie 2007;Brinkmann 2018;Heckathorn 1997;Fashola 2015;Ocejo 2013). With the exception of three interviews that took place via Skype, all interviews were conducted in person and lasted between 45 minutes and two hours. Interview data were recorded through contemporaneous notetaking, paired with computer-typed summaries in the immediate aftermath of each meeting. Respondent incentives were not uniformly provided, although participants were provided with token gifts (e.g., university-branded pens and informational pamphlets), consistent with local and customary business practices. Views expressed by the study's contributors do not reflect the positions or views of their respective organisations or public office(s). Interview data are presented in translated form and excerpts chosen on the basis of their representativeness (see Anderson 2010;Nikander 2008).

The Colombian National Police and the MNVCC
Comparative and cross-national research on criminal justice (i.e., police, courts and correctional systems) is continuously growing (Bennett 2004;Pridemore 2014;Sung 2006;Tonry 1998Tonry , 2015. For several decades, policing scholars in the Global North have developed a robust IJCJ&SD 21 Online version via www.crimejusticejournal.com literature base for the paradigms of hotspots policing (Braga et al. 2014; National Institute of Justice 2009), community-oriented policing (Gill et al. 2014;Maguire et al. 2019;Skogang 2006), problem-oriented policing (Rogers 2010;Weisburd et al. 2010) and procedural justice (Johnson et al. 2017;Maguire et al. 2016;Mazerrole 2013). Many of these same law enforcement strategies-often developed and studied in the United States (US)-have been applied and evaluated in cross-national contexts, including Australia (Dunn et al. 2015), Colombia Mejia et al. 2013;Fundación Ideas Para La Paz 2012), Russia (Gilinskiy 2005), Trinidad and Tobago (Maguire et al. 2017) and the United Kingdom (Weinborn et al. 2017). Consistent with the importation of these strategies by several Latin American police forces (Frühling 2012;Ungar and Salomon 2011), the MNVCC reflects a hybrid combination of these practices.
While there is controversy regarding the actual impact of the MNVCC and the legitimacy of some of the claims surrounding its impact, the MNVCC-as a prescriptive framework-effectively incorporates elements of four operational frameworks found across police institutions in the Global North: hotspots policing, community policing, problem-oriented policing and procedural justice policing. While there is considerable variation in how each of these practices are operationalised, there are nuances that differentiate their core purpose. Hotspots policing emphasises data-driven resource and personnel allocation according to service needs. This is conducted with the understanding that the distribution of police resources should mirror the distribution of crime and public safety-related demands. This policing model gained traction as researchers found evidence that micro-geographic units of space (e.g., specific city blocks) accounted for a disproportionate amount of police service calls . Instead of conceptualising the spatial distribution of crime with a neighbourhood or barrio-level unit of analysis, measurement innovations (e.g., geographic information systems) in hotspots policing helped researchers uncover how crime can be a function of human activity in hyper-local spaces. This emphasis on empirically auditing the micro-geographic (e.g., street segment-specific) variation in recorded crime is central to the contemporary criminology of place literature (see Weisburd et al. 2012) and is a primary theme for how the Colombian National Police structure their model. The MNVCC uses several layers of data inputs to generate the boundaries of a geographic unit (i.e., a quadrant) and this unit of analysis is central to the model's branding (i.e., Modelo Cuadrantes translates to 'quadrant model').
Problem-oriented policing is-as the name suggests-a mode of policing that allows for the customisability and tailoring of police practices to differentiated priorities and place-specific demands in any given jurisdiction. It is highly compatible with hotspots policing. One way of understanding the relationship between problem-oriented and hot spots policing is that the former might represent a broad strategy and the latter a series of tactics consistent with that strategy. Whereas hotspots policing need not involve inter-institutional partnerships, problemoriented policing is most effective when the police are part of broader coalition of organisations targeting a clearly defined problem. For example, if the specific issue is youth gun violence, hotspots policing can be a tactic used a part of a multi-prong strategy, in which officers focus on targeted outcomes -which are specific to youth gun violence -while community-based organisations simultaneously target broader factors and processes that precipitate youth violence and delinquency. In Colombia, policing challenges vary across a national police force. In Cartagena, barrios such as Olaya Herrera required a specific focus on youth gangs and homicides, whereas certain areas of Bogotá required an emphasis on reducing traffic-related accidents and petty crimes against property. Tercera, the rural jurisdiction, had a disproportionate share of inter-familial conflicts and petty thefts. The MNVCC is thus designed to allow for problemoriented allocation of personnel and resources across various urban and rural settings.
Community-oriented policing is a general framework that prioritizes community involvement in -and community support for -crime prevention efforts. In the Latin American context, relationships between national police forces and citizens can be strained or even hostile, reflecting local histories of political violence, state repression, and civil conflict (see Berents and IJCJ&SD 22 Online version via www.crimejusticejournal.com Have 2017; Gordon 2017;Higginbottom 2008;Schultze-Kraft 2017). In settings where the public is fearful or otherwise contemptuous of the state, community-oriented policing can help foster prosocial relationships-and importantly, opportunities for communication exchange and intelligence gathering-between police and community members. In Cartagena, the Colombian National Police maintained a variety of programs that are directly illustrative of a commitment to improved community ties, including the Puerta-a-Puerta or door-to-door program, in which the police distribute flyers containing quadrant information and officer contact details. The Colombian National Police of Cartagena also maintained over 36 WhatsApp chat groups that connect community leaders with quadrant officers. These WhatsApp groups complement the Community Front Program, 6 or the functional equivalent of a neighbourhood watch or constable, providing local resident leaders with a direct communication line to quadrant-level patrol officers. There is a separate community feedback hotline through which callers can register complaints and suggestions, as well as a radio station maintained by the Colombian National Police that allows callers to voice both compliments and discontent much in the way that one would observe a public radio talk show. Town hall meetings held seasonally by the Cartagena mayor's office were publicly announced so that citizens could raise issues or suggestions. Smartphone applications have also been developed to provide relevant quadrant and police information that is specific to the user's GPS location.
The research base for community-oriented policing has generated important insights for procedural justice. Procedural justice refers to the perception of citizens that police requests and demands are anchored in a legitimate process, thus increasing the likelihood of compliance with police directives (Mazerolle 2013;Nagin and Telep 2017). Concepts like trust-in-the-police and likelihood of obeying the law are central to procedural justice. At a basic level, procedural justice represents the degree to which citizens view police actions as fair and legitimate, as citizens are likely to cooperate with police if they view police actions as legitimate in specific police-citizen encounters (Johnson et al. 2017;Maguire et al. 2016). The concept of legitimacy, or process fairness, is intuitively critical to any police system, particularly if one considers the correlated harms of a police force whose actions and members are viewed as fundamentally unfair and illegitimate by the policed population (see Arias and Ungar 2009;Johnson et al. 2017;Nivette and Eisner 2012). The Colombian National Police emphasised a robust curriculum used to train new officers, which includes training on how to respectfully address citizens and validate their concerns during interpersonal exchanges. 7 In light of these prioritized modes of police service, there remains a conflict in whether the model is responsible for influencing both crime rates and citizen perceptions of insecurity. The following section explains the structural basis for the conflicting interpretations of the MNVCC's impact.

Official crime data and structural contradictions
In the US, introductory criminal justice textbooks often provide fundamental critiques of crime data, often in the first two chapters. 8 Manning (2009: 452) captures part of the broader significance of such critiques: There is no body of data that is more consistently and brilliantly critiqued than officially gathered and processed crime data, and yet it is repeatedly used without apology in every major journal that publishes work on crime and crime control. (see also Baer and Chambliss 1997) The problematic nature of relying on official crime data is well known to criminologists, yet researchers continue 'skating on thin ice' as though biased criminal justice data are issues to be corrected 'as statistical techniques become more sophisticated and survey methods improve' (Xu 2014: 162;see Young 2011). The political and bureaucratic interests that have historically shaped the integrity of official crime statistics have been well documented (see Andreas and Greenhill 2010;Black 1970;Butterfield 2002;NL 2014;Seidman and Couzens 1974;Yeung et al. 2018) but IJCJ&SD 23 Online version via www.crimejusticejournal.com are seldom accounted for in mainstream criminological literature (Chambliss 2004(Chambliss , 2001(Chambliss , 1994see Copson 2014;Manning 2009).
In Colombia, crime data are privately maintained by the Colombian National Police, and special permissions must be granted to obtain limited amounts of information. A high-ranking police official articulated that this was a de facto loophole for restricting information that is technically required to be publicly available. The studies that inform the official narrative are based on selectively chosen data. The discourse surrounding these conflicting claims provide a basis for understanding abstract empiricism and its relevance to the present study. Originally coined by Mills (1959) in The Sociological Imagination, abstract empiricism captures a series of incisive 9 critiques of major trends in social science, which include a fetishisation of quantitative research techniques and a disciplinary fixation with branding sociology as empirical, and by extension, objective or scientific. Half a century later, late criminologist Jock Young published The Criminological Imagination (2011) to apply many of the same critiques to his primary field. As Young (2011) perceived it, conventional criminological scholarship 10 was overly fixated on further refining statistical methods for hypothesis testing at the expense of being able to question the degree to which the field has been laundering the practice of using restrictive methods on structurally flawed biased crime data under the guise and clout of scientific objectivity.
The evaluation of the MNVCC (see FiP 2012) contains methodological detail of the following kind. The article specifies that the evaluation team 'conducted more than 17,000 surveys, interviews, focus groups, workshops, field visits to stations and quadrants, and interviews with local authorities and community organizations' (FiP 2012). However, the treatment effect-and the results that are used to justify the claim that the MNVCC has reduced crime-rests exclusively on a difference-in-difference 11 analysis that compares four months of post-treatment with the same four months of the previous year (see FiP 2012: 16;Mejia et al. 2013: 12). Cartagena experienced the most significant increases in violent crime following the MNVCC's rollout, but has never been included in any evaluation since its 2010 inception due to 'improper implementation' FiP 2012). Omitting this city significantly affects statistical estimates, and there is no qualitative context (or data from a study that included focus groups and interviews) on why or how the model was not properly implemented in this jurisdiction. IJCJ&SD 24 Online version via www.crimejusticejournal.com A full-time analyst and technology consultant familiar with the institutional contracts between the Colombian National Police and FiP provided additional context in a private interview. When asked about how to interpret these conflicting reports, they responded: The reports are designed to show that if the MNVCC model is implemented properly, then an average decrease in crime will occur. The evaluations are not meant to imply that real decreases in crime have been actualised due to the model's existence.
The above excerpt reveals the core issue with the actual text of the official evaluations, and how such analyses are communicated to the public. The nuance reflected in this interview excerpt is absent from the evaluation studies, prompting the reader to interpret the publications as indicative of the MNVCC's actual-not hypothetical-success. Importantly, at the outset of the conversation, the analyst made a clear appeal to the methodology employed by the original research team. The analyst began their statements with a reminder that 'the findings were based on robust and specific methodologies', indirectly appealing to the rigor of econometric method, and leaving out any reference to triangulation, qualitative follow-up or how such results are received by police leadership. The subtext of the conversation was thus a manifestation of abstract empiricism, in which 'robust methodologies' are conflated with accurate or valid results, and the veracity of a claim is implicitly viewed as a function of how in vogue the methods were for its generation. This reflects the consequences of abstract empiricism that goes unexamined. As Young (2011: 13) wrote, abstract empiricism comes with substantive implications, given that 'the more quasi-scientific the rhetoric, the more sophisticated the statistic, the more that they are distanced from what they are studying, the more secure they feel'.

The politics of crime control in Colombia
In a candid exchange concerning the political dimensions of Colombian National Police operations, a high-ranking public official from the Ministry of Defence stated during an interview that 'the Mayor of Bogotá is the de facto press secretary of the Colombian National Police, but that such a relationship works in a symbiotic fashion'. This was especially instructive, since mayoral elections for Bogotá were to be held three months after the fieldwork period (2015). This underscores the political importance of how crime and public safety indicators are reported in the capital city. Further, this statement was made in the context of discussing the Colombian National Police's influence on the political election cycle and the campaign messaging regarding issues of crime and safety-topics of extreme salience and import in Colombian public discourse and consistent with literature in the US concerning 'governing through crime' -or how crimerelated claims are central to the ways in which political leaders frame their electoral objectives (Simon 2007;see Kaste 2016).
Various sociolegal factors underlie the political climate in which crime statistics are both generated and communicated. A federal judge who was interviewed for the study emphasised the role that laws have in shaping the very nature of criminal justice data. In 2007, Colombia passed Law 1153, or The Law of Small Causes, 12 restricting the prosecution of crimes that do not correspond to a minimum of four years' incarceration (see Cuartas 2008). Through this legislation, arrests for lower-order crimes would generally not appear before a judge or receive sentencing under traditional guidelines. The interviewed judge emphasised that this law significantly 'improved crime statistics for the Mayor and the police, but in an artificial manner'. This respondent was particularly critical of whether any law enforcement model can be said to have a true effect on reducing crime if the broader legal framework is ignored: Even if examining the numbers of tried cases and convictions, the change is artificial due to changes in how cases are filtered. So, to answer your question specifically and simplistically, it is my opinion and professional observation that IJCJ&SD 25 Online version via www.crimejusticejournal.com the MNVCC is completely irrelevant to the intent and actual practice of reducing crime.
This Law of Small Causes can be viewed as an attempt to alleviate a backlog in criminal courts but can also represent the ways crime statistics can be selectively modified. According to this same respondent, the 2007 law was a rational strategy for reducing the judicial caseload, while also advancing the narrative that crime levels -as measured through judicial caseloads -are declining. A media executive from RCN (a major television network) provided a relevant analogy concerning the MNVCC: 'it is more aptly understood as a collection of normative guidelines for how to structure policy deployment and efficiently allocate resources'. This kind of sociolegal context is what might be expected from a research team that conducts 17,000 interviews, focus groups and workshops. It is also the kind of data that are more appropriate for a major policing model rollout, as opposed to an estimation technique that relied on the narrow window of a four-month period to measure and assess the MNVCC's impact via a difference-in-difference estimation (see FiP 2012; Mejia et al. 2013).
Data from the 17,000 interviews, focus groups and workshops cited in the official evaluationsyet omitted entirely from any publicly available literature-may be a better source of generating evidence-based avenues for improving Colombian National Police operations. While a critical position might view police-generated crime data as public relations instruments and are selfserving, self-auditing tools (Manning 2009), the official narrative reflects an inadvertent form of abstract empiricism that is perpetuated not only in this specific Colombian context, but reflective of broader themes involving contracted crime and justice research partnerships more broadly (see Young 2011).
Evaluations of Colombian public policy and policing stand to benefit from research designs that are informed by -or at least account for -relevant political, historical and cultural context. Colombian policing has undergone four major branding changes in the past two decades, broadly reflecting similar trends in US law enforcement evolution (see Lane 1992;Sherman 1998). Akin to the ways in which acronyms are part of a branding process by successive administrations and transitions in elected office, the MNVCC is the latest iteration 13 of community policing, but significantly enhanced by the latest evidence-based methods from the Global North (e.g., hotspots and problem-oriented strategies). Similar to the claims made by a former New York City Mayor, the evaluation of crime control policies (e.g., 'broken-windows policing') can significantly influence how the public perceives safety and security, and also how they might cast their vote in a future election (Vedantam et al. 2016). In the U.S., for example, initiatives like Operation Ceasefire and the Boston Gun Project take on a life of their own once they are associated with favourable results, even if such results do not technically make any causal claims (see Kennedy et al. 2001).
The processes reflected by the term abstract empiricism 14 provide cover for a state's security industry to justify its activities and investments, no matter what the evidence may be (Manning 2009). In an interview with El Tiempo, a flagship national Colombian newspaper, a former general was asked to comment on higher crime rates in Bogotá, to which they responded, 'there have been increases in reporting to the police and that shows increased trust in the police ... we are working towards a better understanding of the crime rates' (El Tiempo 2015, web). To the extent that official crime narratives are interpreted as valid, this renders the MNVCC tautologically successful. If crime rates decrease, it is a function of the problem-oriented components of the model; if crime rates increase, it is a function of gains in community policing. 15 The end result, no matter what researchers find, is that successful public safety interventions depend on policing models, decreasing the possibility that social scientists ask whether crime and social harm could be addressed without implicitly legitimising an increasingly robust and ever-growing security apparatus. In short, abstract empiricism provides ammunition for critiques like those of Stan IJCJ&SD 26 Online version via www.crimejusticejournal.com Cohen of the London School of Economics, who quipped that the terms 'criminology ' and 'imagination' are unconventional pairings. 16 Studies that emphasise their 'objective' nature tend to inadvertently sanitise the implicit structural biases embedded in such claims, while also being largely correlational and ahistorical (Hall and Winlow 2015). In a broad sense, researchers are not truly independent if their own professional incentive structures (e.g., promotion and tenure, securing major contracts with the Colombian National Police) can be significantly conditioned by how grant-funded opportunities are intertwined with institutional interests (Loader and Sparks 2010;Schept et al. 2015; see also Ornstein and Thomas 2018;Shelden and Brown 2000;Ziai et al. 2018). These incentives can be particularly influential in studies that claim policy relevance. For instance, in a study that seeks to understand how higher basal areas of trees in Bogotá are associated with fewer homicides, the 'objective' evidence-based findings concerning the role that taller trees might play in mitigating violence fail to mention the incentives 17 that the authors have in the very nature of the research question (see Escobedo et al. 2018). Consistent with the Millsian tradition of sociology, it may be reasonable to view any claims of objectivity with a strong degree of scepticism, particularly when one can empirically examine how studies of a major police model are selectively communicated by a contracting think tank 18 during an election year. 19

Future directions
The current paper provides a critical explanation for the mixed messages surrounding the impact of the Colombian National Police's operational strategy by focusing on the politics of crime reporting and the structural interests embedded in discussions of public safety interventions. The quality of the MVCC's structural features and design are, however, separate from the claims regarding its impact on real crime measures. While the current study raises questions about the impact of the MNVCC, the time spent in the field yielded data that support a positive outlook for the Colombian National Police. Future studies regarding the Colombian National Police might incorporate mixed method approaches or find avenues for using the data ostensibly collected from 17,000 surveys (see FiP 2012; Mejia et al. 2013). Research designs might include interviews with personnel who design and conduct law enforcement functions and directly engage with those responsible for data entry and data management to identify potential structural trends in crime data generation (see Butterfield 2002;Copson 2014;Manning 2005). This, in turn, would provide a more comprehensive understanding of Colombian policing and address the potential critiques concerning abstract empiricism, where econometric models are applied to questions that might be better suited for mixed-methods or qualitative research designs.
In Latin America, major political leaders and policymakers are often formally educated in the US, and practices from the Global North are readily imported to contexts in which they may or may not suitable or historically appropriate (Arias and Ungar 2009;Carrington et al. 2019;Tonry 2015). One of the highly qualified interviewees familiar with the project admitted that there are problematic elements to the practice of having crime-related phenomena evaluated solely by economists. This paper does not seek to promote a sense of hostility or fragmentation as to what kind of research base should be prioritised, or which disciplinary subfield should be most visible. However, with causal inference, randomised controlled trials, and quasi-experimental research designs constituting the current fixation for some of the most visible hubs of criminology and criminal justice research in the Global North, it would be worthwhile for the research community to consider the potential consequences of an uncritical acquiescence to this trend (Carrington et al. 2016;Hall and Winlow 2015; see also Knepper and Potter 2011). One such consequence is that researchers across both the Global North and South undermine their own independence, contributing to an applied managerial approach to studying issues of crime and safety that legitimises the use of problematic data sources to help advance instrumental goals (e.g., efficiency, risk management) of state institutions. Unchecked, this trend may disincentivise studying crime and social harm through alternative perspectives (see Brown and Schept 2017 In the present case, continued inquiry into the political incentives affecting Colombian crime and public safety policies are not only warranted, but a necessary first step in understanding justice policy not only as a function of interpreting coefficients from imperfect data, but as a study of how things actually are. Original notes, appointment records and documents revealing specific names of the study's contributors are omitted from publication as part of the 'high-risk' classification of the IRB protocol. The reviewers of this journal have been provided with corroborating data for politically sensitive interviews as a supplemental quality assurance measure. 6 Frente de la Comunidad. 7 A video example of the curriculum can be found at https://www.youtube.com/watch?v=8sv4xyhMUKI, where officers are instructed to avoid colloquialisms and slang terms that might be interpreted as patronising, sexist or unprofessional. 8 Including, for instance, not only the major instruments that generate crime data (e.g., Uniform Crime Report, National Incident-Based Reporting System) but their problematic features, like the UCR's hierarchy rule, or issues of selfreporting biases with the National Crime Victimization Survey. 9 Whereas the term incisive can mean insightful or analytically focused, it is fair to claim that many in the social science community, both at the time of Mills's writing and now, might use the term polemical to characterise the critiques. 10 This critique, and The Criminological Imagination, are primarily referencing criminology in the global north generally and the United States and United Kingdom specifically. 11 One of the fundamental assumptions of the difference-in-difference method is that there are no variations in the 'treatment effect' or nature of the intervention. A model that has been in a perpetual state of improvement and updating is generally incompatible with this method, and yet this is the driving mechanism for the official success narrative. Further, it might be unreasonable to expect any large-scale policing strategy to have robust effects in a four-month period. 12 La Ley de Pequeñas Causas. A legislative update under the same name was passed in 2017 by the Santos Administration (see El Tiempo 2018). 13 Community Participation (PARCO), Community Policing (POLCO), Community Surveillance (VICOM) and the Cuadrant Model (MNVCC) PNVCC are all police models of the Colombian National Police that emphasise community policing, and which have coincided with organisation-wide branding and training efforts (see Ovallos and Eugenia 2013). 14 That is, the disproportionate clout of hypothesis testing and the conflation of evidence-based (i.e., quantitative methodologies) with valid and objective results.
IJCJ&SD 28 Online version via www.crimejusticejournal.com 15 It is important to note that community policing may often have little, if anything, to do with affecting crime rates (Gill et al. 2014), and successful gains in community policing strategies may actually increase official crime rates if they are found to be a measure of increased reporting to the police (see Hickman et al. 2000). 16 Referenced in Xu (2014) and on back cover of The Criminological Imagination (Young 2011). 17 The institutional affiliations of the referenced study include entities with structural incentives, or stakes, in a) promoting trees in the city of Bogotá and b) securing funds for additional research connecting ecology and urban design to crime indicators. This is a structural trend concerning the potential for bias when researchers are incentivised to study questions that are of direct interest to their financiers. 18 For a general argument concerning the embedded political incentive structures and biases of ostensibly non-partisan think tanks, see Lipton and Williams (2016); Shepard (2011)