Illinois, Vermont, and West Virginia SACs Recognized for Innovative Projects
Three Statistical Analysis Centers-Illinois, Vermont, and West Virginia-received the 2013 SAC Innovation Award for projects that created new approaches to analyzing and using their state's justice data. The award was renamed this year (formerly the SAC Technical Innovation Recognition award) after a decision by the SAC Website and Innovation Awards Committee to broaden the award's scope. The award now recognizes statistical, analytical, and methodological innovations in addition to technical innovations. The awards were announced during the JRSA Annual Business Meeting webinar on October 9.
(l. to r.) Illinois SAC Innovation Award winners
Christine Devitt Westley, Mark Powers, Mark Myrent,
and Christopher Schweda.
Illinois - Illinois Criminal History and Recidivism Analysis Tool
Illinois recidivism data have historically been limited to three-year reincarceration rates published by the Illinois Department of Corrections (IDOC). That one statistic provides no insight into the nature of the recidivism of previously incarcerated individuals, and does not include such recidivism outcomes as rearrest or reconviction. Little is known about the prior criminal careers of those entering prison, as IDOC only offers data on the current offense and some demographic indicators of prisoners being admitted.
The Illinois SAC was awarded a series of State Justice Statistics (SJS) grants from the Bureau of Justice Statistics to build a data tool that allows users (the general public) to obtain a more detailed picture of the criminal histories of those being incarcerated and the recidivism outcomes for those exiting. SAC staff often receive information requests on the demographics of prison admissions and exits, which can also be handled by the tool more rapidly and efficiently. Additionally, the user can choose to restrict the recidivism period to a three-year window, or view results out as far as June 30, 2012. This range of data allows users to obtain long-term recidivism data (five years or more from exit) for the earliest cohorts. This tool gives users perhaps the first ever picture of long-term recidivism in Illinois.
The data tool is publicly available on the Illinois SAC website as part of a suite of data analysis tools. The users select from options in a form to construct a particular prison admission or release cohort. Once those options are selected, a button is clicked and additional tabs become visible and active to allow users to select the type of criminal justice events to be included in the criminal history or recidivism analysis. Users can then obtain statistical and graphical information for the chosen cohort analysis. The statistics are returned in tables that can be copied and pasted into spreadsheets. Criminal history and recidivism events (in the form of arrests, convictions, probation sentences, and prison admissions) are shown as an overall total, as well as by offense type. The statistical estimates that are produced include the total number and percentage in the cohort with the specified recidivism event, and the average number of events experienced. Chart options include the selection of cohort subgroups and a choice of several statistical estimates.
With the exception of the initial collection of IDOC and criminal history (CHRI) data and a web-based plug-in, staff at the SAC created everything for the tool from scratch. The tool is still relatively new, but the underlying database that was constructed in the project has been used to inform legislative analyses on offenders with offenses from a set of statutes. Additionally, the prison and probation cohorts were used to fill in parameters in the WSIPP/Results First Cost-Benefit Analysis model that the SAC is populating with Illinois-specific data. Long-term recidivism data (both convictions and probability of resource use) were needed for the model to provide estimates of recidivism reduction and costs avoided due to evidence-based programming. In the past, SAC staff often had to use case-level data files to obtain aggregate statistical information on prison admissions and exits for counties or the state as whole, requiring the development or modification of syntax to be used across multiple files. The tool can now be used to obtain the same information in seconds.
In making the award to the Illinois SAC the Award Committee called the tool "innovative and easy to use. Displays are easy to understand, and complex analyses are made simple with this tool."
(l. to r.) Vermont SAC Innovation Award winners
Max Schlueter and Peter Wicklund.
Vermont - The VCJR-XCT Tool: A Software Tool to Analyze XML-Encoded Criminal History Records
Few statistical analysis tools allow analysis to be done on XML-based data - and almost none allow analysis on XML-based data that are hierarchical in nature. The Vermont Center for Justice Research (VCJR) developed an internal tool that allows XML-encoded criminal history databases to be analyzed more easily using existing statistical analysis tools such as SPSS, which by themselves do not have native XML database capability.
These XML criminal history databases are characterized by data records organized in a hierarchical fashion - for example, a "rap sheet" for a person, and within the rap sheet subrecords for each offense, and within each offense, subrecords for each court appearance and the case dispositions. Well-known statistical analysis tools handle tables of like-kind records extremely well, but have little or no support for XML, especially when it consists of records within records within records. VCJR is currently unaware of any existing commercially available software that makes analysis of this type of data possible.
As a result, VCJR developed a software tool called VCJR-XCT to perform a translation of XML-encoded criminal history records into a set of tables of records, where each record can be related to others for a given hierarchical grouping. These tables can then easily be analyzed using a variety of existing tools for many criminal research projects and reports, such as recidivism studies, outcome evaluations, cost-benefit analysis, and longitudinal studies of offender crime patterns. (To view reports generated using the VCJJR-XCT tool, visit the VCJR report page.)
The VCJR-XCT tool was written in a combination of XSLT (an XML reformatting language), which performs the task of reading and processing the XML data into a set of tables for each type of sub-record; and SQL to combine the various sub-records into a single table. In addition to combining the sub-records into a single table, the tool also performs normalization of XML textual fields - for example:
- Translation of textual date fields (which are hand-entered in a variety of formats) into normalized well-formatted dates.
- Translation of textual description strings (e.g. case disposition text, event text, etc.) converted into well-formatted codes.
- Adding additional summary fields which summarize the list of charges into an easier-to-analyze text string.
- Translation of free-form textual sentencing fields into a normalized sentencing time for each type of sentencing option, which makes analysis queries much easier to perform.
The VCJR uses the tool to perform studies on criminal data for justice agencies in the state. VCJR obtains criminal history data in an XML-encoded format from the Vermont Criminal Information Center at the Vermont Department of Public Safety and converts the data using the VCJR-XCT tool into a form that makes it easy to perform the required analysis. After the analysis is completed, criminal justice agency clients receive a written report with data tables presented in Excel or other graphic applications.
Without VCJR-XCT, analysis on existing XML-based criminal history records data would either be impossible or, at the very least, extremely costly to perform because of the extensive manual effort that would be required to analyze the data using manual methods. In the 10-year period prior to the development of the VCJR-XCT tool, VCJR only conducted two research projects that involved measuring the recidivism of various offender types. The lack of research was in large part due to the difficulty of identifying funding sources necessary to conduct criminal history research that involved considerable manual intervention. Since the development of the tool VCJR has conducted 11 policy-related and program evaluation projects that focused on recidivism of offenders or program participants. The ability to eliminate manual coding of criminal history data has increased the accuracy of the data while reducing the cost and shortening the time necessary to complete projects. The development of the tool has been instrumental in the VCJR's ability to emphasize evidence-based practices and data driven policy in Vermont by offering low-cost program evaluations that can be completed quickly. The tool also facilitated Vermont's ability to participate in the Pew-MacArthur Results First initiative.
In making the award to the Vermont SAC the Award Committee cited the tool's innovative capability that allows criminal justice agencies to access data themselves and enables the Vermont Center for Justice Research to do new types of research.
Stephen Haas and Erica Turley with West Virginias
SAC Innovation Award.
West Virginia - Project IDA: Making Statistics Accessible Through Interactive Data Analysis
Project IDA (Interactive Data Analysis) is the first state-level, online, query-based application for obtaining statistics from multiple data sources. Initially the application contained law enforcement and adult corrections population data; community corrections program data have since been added. In addition to a variety of data sources, the system can also include data over multiple years to illustrate trends. The application utilizes a user-friendly interface that allows users to obtain information through a series of data filter options -anything from aggregate crime totals at the state or county level to very detailed information on the number of offenders with specific characteristics. The application can be accessed through the WVSAC website at https://apps.wv.gov/dcjs/sac/.
In the past, West Virginia had no centralized location for obtaining such a wide variety of statistics. Persons in need of information would first have to determine which agency or organization might have the appropriate data, then someone within that agency would have to analyze the data to respond to the request. With the launch of Project IDA, visitors to the WVSAC website can interact with the data sets on their own and instantaneously generate statistics tailored to their individual needs. From the West Virginia Incident-Based Reporting System (WVIBRS) data, users can obtain statistics on all reported offenses, victimizations, and arrests reported to state, county, and local law enforcement agencies statewide by simply using the drop-down menus associated with the WVIBRS victim segment data. If the user needs information pertaining to inmates committed to, released from, or currently in the custody of the Division of Corrections, they can query the Adult Correctional Population Data (ACPD). These can be used to provide answers to some very policy-relevant, pragmatic questions of interest. For instance, users could find an answer to the question, "How many inmates are currently in DOC custody for life with mercy and what crimes were they sentenced for?" Project IDA is intended to encourage the use of data in decision-making situations and can be a valuable resource for its users.
Use of a query-based approach, as utilized in Project IDA, has a number of strengths. First and foremost, it enhances the accuracy and validity of the data from the different sources, while eliminating the need for users to possess specialized analytical skills or software programs such as SPSS. The data were cleaned, analyzed, and prepared for the case-level database, creating a subset of records which ultimately supply the information generated from the queries. Second, the use of backend data to inform the queries protects the confidentiality and anonymity of individuals. Lastly, Project IDA provides detailed descriptions of the data sources and definitions of all variables. This lessens the likelihood that data will be misinterpreted. While a query-based approach provides a great deal of structure for the end user, it continues to be very flexible and allows users to perform a multitude of queries to answer important questions related to policy and practice.
The WVSAC was responsible for the conceptualization and development of Project IDA and is solely responsible for its maintenance. The WVSAC partnered with WV Interactive, as part of the WV.gov Portal Modernization Project, and was involved in every facet of this effort from conceptualization through the launch of the application.
Project IDA has many specific criminal justice-related impacts and enables the general public, legislators, academicians, agency heads, and other stakeholders to utilize data in making decisions. It is accessible to anyone with an Internet connection, and eliminates the need for a WVSAC analyst to spend time answering data requests. Instead, people can access the website (http://apps.wv.gov/dcjs/sac) and obtain the information faster and more easily than before.
The WVSAC will continue expanding the capabilities of Project IDA, and anticipates adding juvenile court statistics in the next phase of the project. The WVSAC also plans to add updated data files annually for WVIBRS and the ACPD to allow users to examine trends over time. Through Project IDA the WVSAC seeks to act as a clearinghouse for crime- and justice-related data for the state.
In making the award to the West Virginia SAC the Award Committee noted that this innovative application is very easy to use and they liked that it incorporates a lot of NIBRS categories into the format.