Objectives

Police departments around the country are implementing Early Intervention Systems (EIS) to identify officers who may be exhibiting problematic or unprofessional behaviors. The goal of EIS is to minimize officer misconduct and increase officer accountability. To evaluate whether EIS can actually differentiate “problem” from “non-problem” officers, we analyzed the performance of officers from incident reports of police–citizen interactions.

Methods

Using a blind scoring method, we evaluated performance from 1000 police reports; 500 randomly selected reports from EIS-flagged officers (treatment group) and 500 randomly selected reports from non-flagged officers (control group). Six hundred and sixty-seven reports contained relevant performance data. The interval-level metrics used to score officer performance were developed by Vila and colleagues (2016, 2018) to assess performance—expressed as a percentage—across a range of police–citizen encounters.

Results

The overall performance score assigned to officers across all 667 incident reports was 80.46% (SD = 8.75%). When separated into EIS-flagged and non-EIS–flagged incidents, performance scores were 80.63% (SD = 8.58%) compared to 80.27% (SD = 8.95). There was not a statistically significant difference between EIS-flagged and non-EIS–flagged performance.

Conclusions

The EIS evaluated does not appear to be differentiating between problem behavior and non-problem behavior. This suggests that the “thresholds” used to identify problem officers are not working effectively.

中文翻译：

---

评估警察局早期干预系统的有效性

目标

全国各地的警察部门正在实施早期干预系统 (EIS)，以识别可能表现出问题或不专业行为的警官。EIS 的目标是尽量减少官员的不当行为并增加官员的责任感。为了评估 EIS 是否能够真正区分“问题”和“非问题”警官，我们从警民互动的事件报告中分析了警官的表现。

方法

使用盲打分法，我们评估了 1000 份警方报告的表现；500 份来自 EIS 标志人员（处理组）的随机选择报告和 500 份非标志人员（对照组）的随机选择报告。667 份报告包含相关的绩效数据。Vila 及其同事（2016 年、2018 年）开发了用于对警官表现进行评分的间隔级别指标，以评估一系列警察与公民遭遇的表现（以百分比表示）。

结果

在所有 667 份事件报告中，分配给警官的总体绩效得分为 80.46% (SD = 8.75%)。当分为 EIS 标记和非 EIS 标记事件时，性能得分为 80.63% (SD = 8.58%) 与 80.27% (SD = 8.95)。EIS 标记和非 EIS 标记的性能之间没有统计学上的显着差异。

结论

评估的 EIS 似乎没有区分问题行为和非问题行为。这表明用于识别问题官员的“阈值”没有有效工作。