Perceptions of officer-involved shootings by police officers versus civilians

ABSTRACT

In two preregistered experiments, we explore how perception and memory for a use-of-force incident differ between officers who participated in the incident live and civilians who later viewed a Body-Worn Camera (BWC) video of the incident. In Experiment 1, responses were compared between online civilians and officers who had participated live in a shooting simulator. Responses to event memory and state of mind questions revealed numerous differences between these two groups. Experiment 2 assessed specific mechanisms underlying these effects with an additional group of officers who participated online. Our results have important implications for the application of Graham v. Connor, 1989. This US Supreme Court decision provides the superordinate legal context for determining whether the force used by an officer was justified. It is important to acknowledge that the perspective of an officer is likely to differ from that of civilians who only view the officer’s BWC recording afterward.

KEYWORDS:

• Body-worn camera
• police use-of-force
• Graham v. Connor
• stress and memory

Author contributions

KP developed the study concept and obtained funding for this research. KP and TS drafted the research design. TS and JB collected the data from the live police officers. TS built the Qualtrics survey and collected the MTurk data from civilians and online officers. All authors participated in analyzing and interpreting the data. KP drafted the manuscript, and TS participated in the revision process. All authors read and approved the final manuscript.

Author note

This research was funded by a grant to Kathy Pezdek from the National Science Foundation, Program in Law and Social Sciences. We have complied with APA ethical standards in the treatment of subjects in this study, and these data have not previously been disseminated or published. The authors report there are no competing interests to declare.

This experiment was formally preregistered on Open Science Framework and can be accessed at https://osf.io/dk253/?view_only=7c4c751bff86487aa80e3050a36be747. All anonymized data are publicly available on OSF.

We are grateful to Erik Fritsvold, University of San Diego, Michael Scott, Arizona State University, and Seth Stoughton, University of South Carolina Law School, for helping us recruit the officers who participated in Experiment 2.

Disclosure statement

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

Notes

1 Shane et al. (2017) reported that fatal police shootings in the U.S. in 2015–2016, although always unfortunate, are relatively uncommon, averaging about 0.336 per 100,000 citizens and are relatively stable between these two years.

2 For the purposes of this study, we consider civilians to be anyone except the participating officers; this includes members of the public, jurors, legal professionals.

3 For a review of the psychological research on the use of BWCs, see Pezdek (2022).

4 While the Federal ‘reasonableness standard’ established by Graham is the minimum protection guaranteed, states may impose a higher standard.

5 Marr et al. (2021) suggested that the effects of stress may differ between tasks typical of fundamental memory studies and those typical of eyewitness memory studies. However, this view seems at odds with the empirical evidence.

6 There are separate issues that concern the credibility of video evidence in general and also whether triers of fact are likely to rely more on an officer’s account or civilians’ interpretation of video evidence. These issues are different than those considered in this study but have been discussed elsewhere. See for example Jones et al. (2017) and Pezdek and Reisberg (2022).

7 The COVID-19 pandemic necessitated that the civilian sample participate on MTurk rather than in-person.

8 This study was preregistered on OSF as a 2-groups design, referring only to the two groups of civilians. Comparisons of the civilian participants with police officers were then included as exploratory analyses. This is because OSF has strict guidelines for the use of previously collected data. Therefore, although conducting an analysis that compared the previously collected officer data with the new civilian data was the most conceptually appropriate approach, this needed to be pre-registered as an exploratory analysis by OSF standards.

9 Effect size estimated from Taylor (2019).

10 See: https://osf.io/dk253/?view_only=7c4c751bff86487aa80e3050a36be747

11 In the study by Pezdek et al. (2022), the officers responded to the same questions later at two points in time. Importantly, in that study, the officers had viewed their own BWC video prior to answering the subsequent questions for one scenario but not for the other scenario. In the current study, none of the officers had viewed their BWC video prior to answering the Time 1 test questions.

12 Based on a report from the FBI’s Law Enforcement Officers Killed and Assaulted database, 8.5% of all officers killed on duty between 2011 and 2020 were responding to domestic disturbance or domestic violence calls. See: https://www.fbi.gov/services/cjis/ucr/leoka

13 In this analysis, and all analyses in this study, across test questions different numbers of participants were excluded because of missing data, primarily related to some ambiguous responses that could not be coded.

14 For all statistical results see: https://osf.io/dk253/?view_only=7c4c751bff86487aa80e3050a36be747

15 Throughout, 95% confidence intervals are provided in brackets.

16 This includes question 8 for which responses of officers significantly differed from the civilians without the dispatch call but not for civilians with the dispatch call.

17 Effect size estimated from Taylor (2019).

18 All analyses involving police officer samples (in-person or online officers) were, strictly speaking, exploratory analyses for the reasons discussed in Footnote 6.

19 As a reminder, in this analysis, and all analyses in this study, across test questions different numbers of participants were excluded because of missing data, primarily related some ambiguous responses that could not be coded.

20 Pair-wise statistical comparisons between officers who participated live and civilians with dispatch call will not be reported here because this contrast was reported for these same conditions in Experiment 1.

21 For all statistical results see: https://osf.io/dk253/?view_only=7c4c751bff86487aa80e3050a36be747

22 95% confidence intervals are provided in brackets.

23 We did not, however, assess physiological levels of stress experienced by officers in this study.