Perspectives on policing

Citation

Dunham, J.R. (2009), "Perspectives on policing", Policing: An International Journal, Vol. 32 No. 2. https://doi.org/10.1108/pijpsm.2009.18132bae.001

Perspectives on policing

Article Type: Perspectives on policing From: Policing: An International Journal of Police Strategies & Management, Volume 32, Issue 2

Owing to the discretionary nature of the work of actors in the criminal justice system (Gottfredson and Gottfredson, 1988), disparity, or at least the perception of, in and across parts of the criminal justice system is likely to exist. One type, disparity in police action, especially police use of lethal and nonlethal force, has been a continuous topic of concern for researchers (Engel, 2008; Hickman et al., 2008). Upon her examination of the state of the police use-of-force literature, Engel’s (2008) assessment resulted in four overall conclusions. First, research within this realm lacks consistent and well-operationalized measures, which results in difficulty when trying to compare and assess findings across studies. Second, similar to the overall rare statistical nature of violent crime, police use-of-force is a statistically rare event. Third, researchers studying police use-of-force have failed to adequately assess the more important issue of excessive use-of-force. Finally, use-of-force scholars have, more recently, assessed past research in an effort to develop a national system for reporting police use-of-force.

A review of the most recent literature resulted in multiple works revisiting the topic of police use of force and the evaluation of officer use-of-force. The following will provide an overview of a few of these academic works.

Hickman, M., Piquero, A. and Garner, J. (2008), “Toward a national estimate of police use of nonlethal force”, Criminology and Public Policy, Vol. 7 No. 4, pp. 563-604

A vast number of police agencies and officers exist within the USA. Owing to this fact and the discretionary nature of the application of police power, the opportunity for the use and misuse of both lethal and nonlethal force is abundant. Even though this potential is present, Hickman et al. (2008) argue that minimal data is available to assess the nature of police use-of-force at the national level. In their analysis, they sought to assess police use-of-force at the national level through the combination of two existing national surveys. In doing so, they attempted to provide an improvement to the current state of measurement of police use-of-force research.

The authors’ assessment of the literature revealed 36 studies addressing police use-of-force from an incident-level. However, the nature of the measurement and operationalization of each study varied considerably by data source, unit of analysis, geographical focus, and timeframe, upholding Engel’s (2008) assessment that measurement within this realm of research lacks consistency. Data sources included arrest reports, household surveys, participant observations, police surveys, suspect surveys, and use-of-force forms. The studies utilized a range of units of analysis, as well, including arrests, police stops, calls for service, and household, suspect, and citizen contacts. Additionally, sampled data for the analyses ranged from a single city to multiple jurisdictions and from a single year to seven years. Only three of the studies utilized a nationally-representative sample.

Hickman et al. (2008), in an effort to refine the variability and provide additional comparisons, utilized two Bureau of Justice Statistics surveys in their national-level assessment: the Police-Public Contact Survey (PPCS) and the Survey of Inmates in Local Jails (SILJ). The PPCS is the only national survey providing police use-of-force indicators; however, critics argue that it fails to provide an adequate assessment of the amount of use-of-force because it does not include data from individuals who were recently incarcerated. In this assessment, the authors used data from both the 2002 PPCS and SILJ. The PPCS sample included 76,910 persons aged 16 years or older, while data from the SILJ included 6,982 interviews. Following weighting of both samples, data from the PPCS represented 215.5 million persons aged 16 and older, and data from the SILJ represented approximately 438,000 inmates.

Demographic analyses found that both sources of data provided fairly consistent results. Of all individuals reporting an occurrence to the PPCS, 78.3 percent were male, 56.3 percent were Caucasian, and 34.6 percent were individuals aged 20-29. Of those arrested, males (88.9 percent of the PPCS, 93.1 percent of the SILJ) and individuals aged 20-29 (42.8 percent of the PPCS, 38.8 percent of the SILJ) reported police threat or use-of-force the most. The two sources diverged when race was taken into account. The majority of those arrested and reporting police use-of-force according to the PPCS were Caucasian (50.4 percent); however, the majority of inmates experiencing police use-of-force according to the SILJ were African American (46.8 percent). Although the reported magnitudes varied, the overall likelihood of experiencing police threat or use-of-force was greatest for males (21.8 percent PPCS and 23.3 percent SILJ), African Americans (24.0 percent and 26.7 percent), and adolescents (31.2 percent and 33.1 percent; Hickman et al., 2008).

Multivariate analyses conducted by the authors of each source and in combination mirrored the demographic results above. For all individuals, the odds of a male reporting an instance of police use-of-force were greater than three times that of the odds of a female. African Americans were more than three times as likely and Hispanics were approximately two times as likely as Caucasians to report experiencing police use-of-force, and younger individuals were more likely to experience force than older individuals. Additionally, the authors found that the most significant predictor was provocation on the part of the individual. Across all predictors, the existence of a form of provocation significantly increased the likelihood of an individual experiencing force compared to an incident with no form of provocation. Forms of provocation included verbal insults, cursing, arguing, threatening, fleeing, physical assault on the officer, and resisting arrest. Furthermore, the severity of the force used by the officer followed the same predictors as use-of-force, in general.

Hickman et al. (2008, p. 588) hold that the “PPCS accounts for about 87 per cent of the total force events derived from both the PPCS and the SILJ data”. The SILJ data only account for 13 percent of the force events. However, the SILJ events, according to the analyses, have a greater likelihood of use-of-force than those in the PPCS and a greater level of severity of force. Owing to the nature of these events, it would follow that any studies utilizing only the PPCS would fail to account for a key population of individuals who report experiencing the statistically rare force events.

Thus, Hickman et al. (2008) provide an extension to the existing literature through the combination of the PPCS and the SILJ for a national-level analysis of police use-of-force. This is the first study that has assessed the nature of force events for individuals who are incarcerated. Additionally, the authors call for the establishment of a national police use-of-force reporting system. Furthermore, since the greatest predictor of a force event is provocation by the suspect, their results suggest that the key mechanism for limiting police threat or use-of-force will require the manipulation of suspect behavior, as opposed to manipulating officer behavior, such as through training or the establishment of departmental policy.

Klinger, D. (2008), “On the importance of sound measures of forceful police actions”, Criminology and Public Policy, Vol. 7 No 4, pp. 605-17

Following a review of Hickman et al. (2008), Klinger (2008) reiterated the necessity of utilizing adequate and complete measures of police use-of-force and provided a critique of Hickman et al.’s analysis. Additionally, he argued that social scientists must conduct due diligence in the use of data sources because the use of “unsound and incomplete data [can] easily […] lead to [an] unsound and incomplete understanding” of the concept in question, in this case police use-of-force (p. 605).

According to Klinger (2008), one of the main measurement issues social scientists must deal with is providing measures that are theoretically sound. He argues that researchers often utilize measures from prior research without adequately assessing their conceptual validity. Research within the police use-of-force arena has not surpassed this issue. For example, counts of lethal use-of-force are often acquired through the Supplemental Homicide Report (SHR) produced by the Federal Bureau of Investigation, even though studies have shown that these counts are inaccurate.

In his analysis, Klinger (2008) outlined his concerns of the use of arrest-based measures of nonlethal force. His initial concern rests with the results from Hickman et al. (2008) because they are estimates and not exact values of police force incidents; thus, consumers of the article need to remain aware of this fact. Also, researchers using the PPCS and SILJ in the future should be cautious of the validity of the data from these sources. This is due to the potential for over-counting police use-of-force. Over-counting could occur because reports of arrestees may be inaccurate due to the adversarial nature of the contact. Additionally, the author failed to find validity in results from the SILJ, as described in Hickman et al. (2008); he found that they are grossly overestimated, according to empirical, theoretical, and anecdotal reasons.

Thus, even though sound and complete measures are necessary for sound and complete analyses, Klinger (2008) cautions against the blind use of official data sources, especially the SHR. He argues that researchers need to begin to assess the validity of measures utilized in other studies prior to applying them to future research. Therefore, the author provides much discussion against the use of existing data sources and measures for the study of police use-of-force and calls for additional research to determine the validity of national data sources for and the operationalized measures of the police use-of-force concept.

Smith, M. (2008), “Toward a national use-of-force data collection system: one small (and focused) step is better than a giant leap”, Criminology and Public Policy, Vol. 7 No. 4, pp. 619-27

Responding to the call by Hickman et al. (2008) for the development of a national reporting system to track police use-of-force incidents, Smith (2008) discusses the potential avenues for the development of such a system, along with extant factors that could potentially have an impact on its establishment. Additionally, he provides an assessment of potential limitations and the factors a national system should include to be most effective.

Three potential avenues for the development of a national system are suggested by Smith (2008). First, even though congress is limited by the constitution in its ability to mandate a state or its municipalities to submit data to the federal government, it can attach the mandate to the distribution of federal aid. Thus, a national system could be developed requiring the submission of police use-of-force data due to the receipt of certain federal monies. Second, a voluntary reporting system could be established, similar to the Uniform Crime Reports (UCR). A third type of system would provide incentives, via direct grants, for agencies that provided police use-of-force data to the program.

Smith (2008) suggests that certain limitations of a national reporting system need to be taken into account. First, the author questioned the use of such a system that would provide the same cross-sectional data currently available through the PPCS and the SILJ. Second, the establishment of a national use-of-force reporting system would fail to provide an avenue for the study of the main missing element of use-of-force studies – excessive force. This is due to the fact that key components necessary for the determination of excessive force are not readily available in police department data collection systems. Finally, the establishment of a national system to assess excessive force would require additional steps by local police departments and a more qualitative data collection effort that would not readily fit at the national level.

Thus, according to Smith (2008), the most effective method for the establishment of a national police use-of-force reporting system is through the use of direct incentives for participation. Furthermore, he suggests that an effective system of data collection should be framed as a harm reduction or police-citizen violence reduction data collection initiative to provide data for use in the development of policy, training procedures, and police practices.

Owing to the recent call within the literature for the development of a national-level police use-of-force data collection system, Smith (2008) provides key insights into both procedures for the establishment of the system and limitations surrounding this type of effort. It seems from his discussion that a national system could potentially provide a more adequate data source for comparison across jurisdictions. However, its establishment based on existing police department data would likely not provide the means to assess the more important nature of excessive use-of-force, which is a key missing element in the research.

Bazley, T., Mieczkowski, T. and Lersch, K.M. (2009), “Early intervention program criteria: evaluating officer use of force”, Justice Quarterly, Vol. 26 No. 1, pp. 107-24

Following citizen complaints of officer misconduct and a directive by the United States Civil Rights Commission, police agencies across the country began to establish early warning systems (a.k.a. early intervention programs) in an effort to prevent incidents of misuse of power. While the process and indicators for the established early intervention programs vary considerably, officer use-of-force is one of the indicators most often included. Additionally, the evaluation and measurement of indicators, especially use-of-force indicators, lacks consistency. Due to the variability across programs and the limited research available on this topic, Bazley et al. (2009) sought to assess the use and measurement of use-of-force indicators for early intervention programs. In doing so, they argue that this indicator should provide a measure of both the frequency of use-of-force and the use-of-force in relation to citizen resistance.

Even though limited research exists in this area, the authors review concluded that four key elements exist for the development of an early intervention program (Rahtz, 2003). First, the performance indicators used to evaluate the officers must be identified. Second, the indicators should be weighted according to importance. Third, a “threshold value at which officers are identified as problematic” must be determined (Bazley et al., 2009, p. 109). Finally, the system should provide a detailed response for those police officers who have been flagged with the potential for misuse of power. The authors also found through their review of the literature that indicators included in the evaluation should be weighted using one of three approaches (Walker, 2003): “(1) agency-wide or department-wide simple quantitative criteria thresholds; (2) peer-officer thresholds; and (3) factor and multifactor performance ratios” (Bazley et al., 2009, p. 110).

For this study, the authors assessed an early intervention program in a southeastern US municipal police department. Those officers identified by the department as potentially problematic during the study period (the year 2000) included 33 officers (32 male, 1 female) and were flagged due to use of a high level of force. Of those, only one officer’s conduct was considered unsuitable. All data used in this study was collected on a self-report basis from the officers, as mandated by their departmental policy.

In order to determine the accuracy of the department’s use-of-force indicator and threshold level, the authors utilized a weighted force factor analysis. The author’s calculated weighted force factor values for each officer and then used those as a predictor of early intervention program qualification in a logistic regression model. Force factor values were determined by subtracting the ordinal force level scale value corresponding to the most severe use-of-force action on a report to the ordinal force level scale value of the most severe resistance by the suspect. A value of zero corresponded to the use of appropriate force in response to the resistance encountered by the officer. A negative force factor value corresponded to the use of force exceeding the resistance level, whereas a positive force factor value corresponded to the use of less than necessary force in comparison to the suspect resistance level (for a detailed description of the weighted force factor calculation see Bazley et al., 2009).

The authors assessed a total of 26,883 reports of use-of-force by 810 officers (729 male, 81 female) for the study year. Of those, approximately half reported suspect resistance levels. The 13,617 reports with resistance levels involved 743 officers; however, the incidents are not mutually exclusive to individual officers.

Analyses by the authors provided significant results; however, they were theoretically inconsistent. Both the weighted force factor and the number of use-of-force reports significantly predicted surpassing the early intervention threshold. But, the weighted force factor was positively related to qualification for the program, which is theoretically inconsistent. Thus, officers who used less force than needed in comparison to the resistance they encountered were more likely than those who used greater than necessary force to qualify for the intervention program. Additionally, the number of use-of-force reports was also positively related to qualification for the program. In other words, officers with greater use-of-force incidents were more likely to surpass the threshold for the intervention program.

Thus, Bazley et al. (2009) concluded that their analysis provided support for the importance of selecting the appropriate threshold criteria for flagging an officer for early intervention. In doing so, they argued that the number of use-of-force incidents an officer accrues should not be the only use-of-force indicator. This is due to the fact that the raw number of use-of-force incidents fails to provide a baseline for the amount of force that could have been necessary in each event. Furthermore, use of a force factor or weighted force factor as a use-of-force indicator would provide a baseline for comparison. Results from this analysis found that the majority of officers who used force actually used less force than necessary when comparing the officer’s use-of-force to the resistance of the suspect.

Results from this analysis suggest that additional studies should be conducted of early intervention systems, especially those focusing on use-of-force indicators. If, in fact, simple quantitative count thresholds flag officers who are using less force than necessary, then police departments utilizing those systems are potentially missing those officers who are actually their potential problems. Thus, agencies with count thresholds would need to evaluate and reassess their criteria in order to adequately predict potentially problematic officers. Additionally, this analysis highlights the (weighted) force factor ratio, which could provide a baseline for measuring excessive use-of-force, which Engel (2008) argues is the more important use-of-force issue that has not been adequately tested in the literature.

Jessica R. DunhamUniversity of Cincinnati, Cincinnati, Hamilton, USA