Improving police use-of-force training is methodologically difficult. By providing a method for identifying the “expert” response to any given scenario, and by triangulating multiple methods, we aim to contribute towards police departments’ capacities to engage in more effective and targeted training. Forty-two police experts and 36 novices watched five scenarios taken from body-worn camera footage. The videos would pause at several points, and respondents gave both close-ended survey answers and open-ended written answers. Using a mixed-methods approach combining quantitative regression and natural-language-processing techniques, we triangulated our findings to reach conclusions regarding the differences between experts and novices. Relative to novices, expert police officers were more likely to report the importance of force mitigation opportunities to any given scenario in close-ended questions, and were more likely to use words associated with verbal de-escalation; novices were more likely to use words associated with physical control. The materials can be accessed at https://osf.io/wujkz/ .

中文翻译：

---

警察专业知识和武力使用：使用混合方法建模专家和新手使用武力的决策模型

从方法上讲，改进警察的使用武力训练是困难的。通过提供一种方法来识别对任何给定情况的“专家”响应，并通过三角测量多种方法，我们旨在提高警察部门的能力，使其能够进行更有效和更有针对性的培训。四十二名警察专家和三十六名新手观看了从身上的摄像头录像中截取的五个场景。视频将在几个点处暂停，受访者同时给出了封闭式调查答案和开放式书面答案。使用结合了定量回归和自然语言处理技术的混合方法，我们对研究结果进行了三角剖分，以得出有关专家和新手之间差异的结论。相对于新手，在封闭式问题中，专家警官更有可能在任何给定情况下报告减兵机会的重要性，并且更有可能使用与言语降级有关的词语；新手更有可能使用与物理控制相关的词语。可以在https://osf.io/wujkz/上访问这些材料。