Description of original award (Fiscal Year 2021, $499,056)
Statement of the problem: Research and reporting on the local jail system is inhibited by substantial barriers related to dispersed record management spread across more than 3,200 facilities. The majority of current data collection efforts focus on support of aggregate statistical reporting which severely narrows the utility of the data for other statistical purposes. Collecting person-level administrative records from the local jail system will advance opportunities for both statistical reporting and research. Subjects: Information will be collected from jail personnel about their data managements systems as well as person-level administrative records on local jail populations. Data collected from jail personnel will include the 30 largest jails in the United States. Person-level administrative data will be collected from 10 of the 30 largest jails. Partnerships: The Criminal Justice Administrative Records System (CJARS) is a partnership between the University of Michigan and the U.S. Census Bureau. Data collection, processing, and harmonization for the data infrastructure takes place at the University of Michigan. The University of Michigan is responding to this solicitation independently of the U.S. Census Bureau. Research design and methods: Data collected for this project will involve both semi-structured interviews with jail personnel and extraction of person-level administrative jail records. The raw person-level administrative records will be processed and harmonized to generate analytical files that will include the following files: admissions/release information, charge categories, and the jail population at midyear and end of year. Data processing and harmonization will leverage the CJARS data infrastructure and the innovative data processing techniques that it has developed. Analysis: Extensive data quality assessments will be conducted on analytical files to ensure data quality. Data quality assessments will include the evaluation of missing data, invalid data entries, improbable/impossible dates, illogical sequences of events, and outliers. Our data quality assessments will also benchmark the analytic files against the data received from jails to other published jail statistics. We will conduct benchmarking exercises to compare several metrics including demographic characteristics of inmates, charges, dispositions, detention status, midyear and average daily population counts, annual admissions counts, incarceration rates, and average time spent in jail. Products and reports: There are a number of products and reports that will be produced as part of this project. These will include a data specifications report, a report summarizing findings from interviews with jail personnel, raw data files and documentation, analytical data files, data assessment reports, and final data documentation.