urban institute nonprofit social and economic policy research

Development of an Empirically-Based Risk Assessment Instrument

Read complete document: PDF


PrintPrint this page
Share:
Share on Facebook Share on Twitter Share on LinkedIn Share on Digg Share on Reddit
| Email this pageE-mail
Document date: April 01, 2003
Released online: April 01, 2003

The nonpartisan Urban Institute publishes studies, reports, and books on timely topics worthy of public consideration. The views expressed are those of the authors and should not be attributed to the Urban Institute, its trustees, or its funders.

Note: This report is available in its entirety in the Portable Document Format (PDF).


Overview

PURPOSE OF STUDY

In 2001, the Urban Institute was commissioned by the District of Columbia Pretrial Services Agency (PSA) to develop a risk assessment instrument to assist its diagnosticians in recommending conditions of pretrial release for the thousands of defendants they process each year. This is the final report on the second and final phase of the instrument development. This phase of the research built on the earlier work by extending the period for observing outcomes by 14 months (from nearly two years to three years or longer), expanding the set of predictor items considered for inclusion on the instrument, and searching for combinations of items for inclusion.

The resulting instrument is intended to serve two primary goals. The first goal is to make the development of release recommendations more objective and consistent across defendants. The attainment of this goal should improve the transparency of PSA assessment and recommendation processes to observers both inside and outside the agency. The second goal is to improve the accuracy of decision-making based on risk assessment. Improved accuracy should increase public safety, reduce court costs associated with non-appearance, and reduce the number of low-risk defendants whose liberty is restricted.

The instrument is designed to predict two outcomes, risk of failure-to-appear, or FTA (indicated by issuance of a bench warrant for failure-to-appear), and risk of rearrest (which included either a new arrest record or a citation). Measures that might predict either or both of these two outcomes (i.e., FTA or arrest under supervision) were created from the Automated Bail Agency Data Base (ABADABA) and Drug Testing Management System (DTMS) data. The data included information about the criminal histories, demographics, health, employment, and drug use of all defendants processed by PSA during the study period.

Candidate predictors were constructed from these data. They included items on a list provided by PSA, based on institutional understanding of the characteristics of defendants who fail, and items suggested by UI, based on knowledge of the research literature on the prediction of criminal outcomes. The list was constrained, however, to data in the existing ABADABA and DTMS; some candidate predictors could not be measured with available information. The significant predictors of either outcome (FTA or rearrest) are included in the final instrument.

The assessment instrument was developed using data on a cohort of defendants processed by PSA between January 1, 1999 and June 30, 1999. This time period was selected jointly with PSA because it was: 1) sufficiently recent that changes in the defendant population between 1999 and the present were expected to be of minor importance to the form and function of the instrument, but also 2) sufficiently long ago that nearly all of the defendants would have completed their PSA supervision before the data were provided to UI for analysis.

The scores on the instrument range from 0 to 100 for each risk outcome. The instrument scoring is designed to assist PSA diagnosticians in prospectively assessing the appearance and safety risk posed by individual defendants. Once a diagnostician has answered all of the questions on the instrument, the instrument weights the answers to compute two risk scores (one each for appearance risk and for safety risk) that range from 0 to 100.1 These scores are then used to classify the level of risk into five categories that could be used when making a release recommendation for court in the bail report. The risk score and risk category could be included in the bail report as well to provide the judge with the benefit of this information.

SUMMARY FINDINGS ON INSTRUMENT

UI is submitting the instrument in the form of a Microsoft Excel spreadsheet that can be used to compute risk scores based on the answers to the questions input into the appropriate cells. Delivering the instrument as a functioning spreadsheet is the most concise, comprehensive explanation of how the questions, answers, and corresponding weights relate to each other to produce the risk scores. The spreadsheet allows PSA administrators to explore the consequences of adjusting the cut-point values used to assign one of five risk categories (i.e., Low, Condition Monitoring, Moderate, High, or Severe) to defendants based on the risk scores, which range from 0-100, computed by the instrument. The spreadsheet instrument may also be printed to hard copy, complete with instructions for answering each question.

Note: This report is available in its entirety in the Portable Document Format (PDF).


1 Although the computation of the risk scores from the answers and weights is purely arithmetic, it is not simple. Even a user who felt competent to perform the computations would shortly find it tedious to perform them with a hand calculator. Ideally, the instrument should be integrated into the management information system used by PSA so that the computer can calculate the scores with little or no input from the user.


Topics/Tags: | Crime/Justice


Usage and reprints: Most publications may be downloaded free of charge from the web site and may be used and copies made for research, academic, policy or other non-commercial purposes. Proper attribution is required. Posting UI research papers on other websites is permitted subject to prior approval from the Urban Institute—contact publicaffairs@urban.org.

If you are unable to access or print the PDF document please contact us or call the Publications Office at (202) 261-5687.

Disclaimer: The nonpartisan Urban Institute publishes studies, reports, and books on timely topics worthy of public consideration. The views expressed are those of the authors and should not be attributed to the Urban Institute, its trustees, or its funders. Copyright of the written materials contained within the Urban Institute website is owned or controlled by the Urban Institute.

Email this Page