The chapter focuses on assessing the impact of the London Gangs Matrix—a controversial, longstanding policing tactic—and a mayoral oversight priority. However, this initiative seemingly presented insurmountable challenges in terms of traditional impact methodologies (i.e. retrospective nature, lack of single selection score, section bias, and pan-London status). What follows is the development of a novel quasi-experimental permutation approach called the ‘shuffle’ which was successfully used to explore impact. Results are described (e.g. offending reduced on the Matrix compared to the shuffled group) alongside a range of technical aspects such as the procedure and power analysis of the technique. Finally, reflections are offered on both the Matrix and the shuffle approach itself. What has been developed is a valuable addition to the methodological landscape—a theoretically grounded quasi-experimental statistical approach to examining impact.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Innovation in Impact Evaluation: Devising a Novel Quasi-experimental Approach to Understanding the Impact of the London Gangs Matrix

  • Paul Dawson,
  • Tom Davies,
  • Toby Davies

摘要

The chapter focuses on assessing the impact of the London Gangs Matrix—a controversial, longstanding policing tactic—and a mayoral oversight priority. However, this initiative seemingly presented insurmountable challenges in terms of traditional impact methodologies (i.e. retrospective nature, lack of single selection score, section bias, and pan-London status). What follows is the development of a novel quasi-experimental permutation approach called the ‘shuffle’ which was successfully used to explore impact. Results are described (e.g. offending reduced on the Matrix compared to the shuffled group) alongside a range of technical aspects such as the procedure and power analysis of the technique. Finally, reflections are offered on both the Matrix and the shuffle approach itself. What has been developed is a valuable addition to the methodological landscape—a theoretically grounded quasi-experimental statistical approach to examining impact.