The role of social distance in the relationship between police-community engagement and police coercion

ABSTRACT

While there is a considerable body of evidence on the influence of individual and situational, and to a lesser extent community factors, on police use of force, little is known about the influence of police agency factors on use of force by police. Greater use of community policing approaches has recently been recommended for U.S. policing agencies, in the wake of a series of high-profile police-involved deaths of predominantly African American citizens. However, there is limited empirical evidence that community policing can influence or impact the use of force by police. The current study uses an online survey of Australian police officers to examine whether frequency of community engagement affects officers’ attitudes towards coercive and non-coercive policing responses to a vignette, and whether social distance might explain this relationship. Regression analyses demonstrated that community engagement was positively related to endorsement of non-coercive policing responses and negatively related to endorsement of coercive policing responses. Social distance mediated the former relationship, but not the latter. These results suggest that community policing approaches may increase the propensity for non-coercive policing responses, and that a reduction in social distance to the community is one pathway through which this occurs.

KEYWORDS:

• Police coercion
• community engagement
• community policing
• social distance
• police use of force

Acknowledgements

We would like to thank the Queensland Police Service (QPS) for their assistance with this research. The views expressed in this material are those of the authors and are not those of the QPS. Responsibility for any errors of omission or commission remains with the authors. The QPS expressly disclaims any liability for any damage resulting from the use of the material contained in this publication and will not be responsible for any loss, howsoever arising, from use of or reliance on this material.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 As noted later, ultimately officers from 51 Divisions participated in the survey.

2 It is not clear in total how many officers would have received the invitation to participate, as the number of frontline officers within all Divisions is unknown, and complete distribution of the email invitation to officers was not able to be tracked as the survey link was sent via local command. However, prior to survey dissemination, Officers-in-Charge in each of the targeted Divisions were asked to estimate the number of frontline officers in their Division. Investigative officers are attached to Districts rather than Divisions, so the number of investigative officers owned by each Division had to be estimated. This response rate has been estimated based on the number of responses from the targeted Divisions only (n=274), as total officer headcounts in non-targeted Divisions were unknown.

3 POLSIS Profiles are published by the Queensland Government Statistician’s Office and incorporate the use of concordance data to create estimates of population characteristics from Census data that map to each Division. POLSIS Profile reports were accessed at https://www.police.qld.gov.au/rti/published/about/orgStrct/Brisbane-Region.htm

4 Reported Crime Trend data were accessed at https://www.police.qld.gov.au/online/data/

5 The U.S. National Law Enforcement Applied Research and Data Platform survey, formerly the National Police Research Platform, is accessible at https://www.nationallawenforcementplatform.org/

6 These questions were adapted from questions used in a U.S. National Law Enforcement Applied Research and Data Platform survey, formerly the National Police Research Platform (https://www.nationallawenforcementplatform.org/).

7 Removing the 16 general duties officers who selected n/a on the three community engagement items did not change the main findings of the analyses.

8 The IRSD is developed from a synthesis of area-level Census indicators related to disadvantage, including income, education, unemployment, occupation, housing overcrowding, and household composition. An IRSD score is given to each census area and areas are ranked from most disadvantaged (low IRSD rank) to least disadvantaged (high IRSD rank).

9 Using the complex samples function in SPSS 25.

10 Note: When mediation analyses were run without the covariates results did not change substantially.