10.1 Basic Analysis and Reporting Process

The focus of the RSVP is on conducting basic, straightforward analysis — procedures that prosecutors’ offices (and their partners) can implement without advanced statistical skills. The primary purpose is to help prosecutors’ offices and their partners to continually improve the effectiveness, efficiency, and equity of their response to sexual violence.

The data will not by itself explain why outcome values have changed or what needs to be done to improve them.100 However, it provides the foundation upon which prosecution performance can be accurately and thoughtfully discussed, and reveals useful clues if basic analytic steps are used. Each step helps identify patterns, and suggests issues that prosecutors’ offices and/or their partners need to address, even if only to explain outcome data that is concerning, e.g., high declination rates. Outcome measurement and analysis can also help identify potential training and technical assistance needs.

Performance reports should be prepared and reported at regular intervals — monthly, quarterly, or annually.101 The process also should permit access to the latest performance data at any time, as specific issues arise.

 

10.1-A. An Important Early Step is to Identify Who Will Design and Then Undertake the Ongoing Analytic Process.

This person could be someone from the prosecutor’s office who is tasked with managing data or working with office data systems, e.g., someone from the IT department or a partner agency with expertise in data management and analysis. Criminal justice professionals may also want to consider partnering with a local university’s criminal justice, sociology, psychology or even computer science department to automate data collection and analysis.102 The analytic tasks would include:

a. Developing the process to minimize calculation efforts required of prosecutors and partnered staff;
b. Summarizing the findings from each reporting period for the prosecutor’s office and its partners; and
c. Highlighting the findings that warrant additional attention.

This resource is targeted to prosecutors’ offices’ collection and review of their own data. Given trends in reporting data to outside agencies and the public,103 each partner should undertake its own analysis, focusing on those outcome measurements particularly relevant to their work. This will allow the prosecutor’s office and its partners to identify the extent to which progress has or has not been made — which practices are working well and which are not, and for whom. It should also help explain and compare progress for various victim groups and for various types of cases.


 

100 To obtain such information would require in-depth program evaluations, involving more advanced data collection and mathematical techniques.

101 The frequency of reporting may depend on the size of the office and the sexual violence caseload.

102 See, e.g., Bruke Mammo, Praveer Narwelkar, and Roshan Gianani, Towards Evaluating the Complexity of Sexual Assault Cases with Machine Learning, LU-CSE-19-002 (May 31, 2019), available at  https://engineering.lehigh.edu/sites/engineering.lehigh.edu/files/_DEPARTMENTS/cse/research/tech-reports/2019/LU-CSE-19-002.pdf. In the spring of 2019, students in the Lehigh University graduate course, “Artificial Intelligence for Social Good” commenced development of  a computer programming system to automate data collection and analysis related to sexual violence cases. The students’ work demonstrates that automated performance management systems are attainable in the near future. For more information, contact AEquitas.

103 See, e.g., Measures for Justice, available at https://measuresforjustice.org; see also San Francisco District Attorney’s Office DA Stat Dashboards, available at https://sfdistrictattorney.org/da-stat-dashboards.