Date created: | Last Updated:

Category: Project

Description: Previous research has shown that high-profile incidents of police misconduct can reduce trust and confidence in the police among the wider public (Kochel, 2019; Nägel & Lutter, 2021; Reny & Newman, 2021). However, the effects of these ‘vicarious’ experiences can be heterogeneous and even null depending on the institutional and societal context, sub-population, and specific circumstances of the event. There is relatively little research on to what extent incidents of police violence in one country resonate across the globe, influencing individual attitudes towards local police. The study ‘Phantom Pains’ in the British Journal of Political Science aimed to address this research gap by assessing the impact of the death of Eric Garner in the United States on Londoner’s attitudes towards the London Metropolitan Police in the United Kingdom (Laniyonu, 2021). His study takes advantage of the overlap between fieldwork conducted for the Metropolitan Police Attitudinal Survey and the event, utilizing the interview date as an instrumental variable to assign respondents to treatment (post-event) or control (pre-event). This type of quasi-experimental design, known as unexpected event during survey design [UESD], assumes that because the event is ‘unexpected’ assignment to treatment and control is as-good-as-random, conditional on a set of assumptions. Laniyonu (2021) analysed treatment effects using regression discontinuity design [RDD], and found that Black Londoners had significantly lower support for the Metropolitan Police following the death of Eric Garner. White and South Asian respondents showed little to no change following the event. However, recent research on UESD has highlighted a number of potential issues embedded in the design that can threaten internal validity and the robustness of results. These issues include the extent to which all key assumptions related to ignorability and excludability are addressed, evaluating robustness across methodological choices and model specifications, and checking for heteroskedasticity due to highly unbalanced groups. Analyzing robustness across models is particularly important as there can be many potential analytical choices or “forking paths” available when executing empirical research. For example, in relation to UESD, researchers may opt for different methodological designs (e.g. regression discontinuity, difference-in-differences or straightforward before/after comparisons), select or compute the optimal bandwidth(s), and calculate different standard errors. In addition, as Laniyonu (2021) rightly states, the generalizability of UESD studies is often limited due to the focus on a single case and context. The current study aims to reproduce and replicate the effect of Eric Garner’s death in New York City on attitudes towards the police in London. As a first step, using the data and statistical code provided by Laniyonu (2021), we aim to reproduce the results presented in the paper. Here we will also evaluate the robustness and reliability of results by assessing the ignorability and excludability assumptions, as well as alternative design strategies and specifications, as recommended in the UESD literature (Muñoz et al., 2020; Nägel & Nivette, 2022). In doing so, we plan to implement the full range of ‘good-practice-recommendations’ discussed in the literature, thus providing a comprehensive and rigorous understanding of the hypothesized effect. In the second step we aim to conduct a replication of another high-profile event (i.e. the death of George Floyd in 2020) on attitudes towards the police using the same data from the Metropolitan Police Attitudinal Survey in London. This pre-registration focuses specifically on outlining the hypotheses, research design, and analyses planned for the George Floyd replication.