John K. Roman

Reducing Drug Use and Crime among Offenders: The Impact of Graduated Sanctions

The evaluation of the Washington, D.C., Superior Court Drug Intervention Program (SCDIP) compared drug felony defendants randomly assigned to either a docket offering structured graduated sanctions in combination with drug testing and judicial monitoring, or a docket using drug tests and judicial monitoring only. Assignment to the graduated sanctions docket was found to reduce drug use prior to sentencing. Program participants were significantly less likely to use drugs prior to sentencing, and, in the year after sentencing, were significantly less likely to be arrested and and had significantly fewer arrests. This paper describes characteristics of the sanctioning program that appear highly correlated with positive outcomes.

Can Cost-Benefit Analysis Answer Criminal Justice Policy Questions, And If So, How?

Cost-benefit analysis (CBA) is commonly used in criminal justice research, but the results are rarely embraced with great confidence. The CBA method that is routinely applied in crime research is not a well-developed empirical method but rather an intuitive extension of economic theory appended onto an impact evaluation. Small errors resulting from common problems in evaluation, such as small sample sizes and selection effects, are magnified in the CBA framework. In particular, rare events with large costs or benefits can obscure all other program effects. Even with careful attention to these issues, CBA conducted in conjunction with an impact evaluation will, as a rule, omit critical information. Because public safety is at the core of crime control, it is critical that CBA captures the effects of crime-control policy at the community level. Moving the unit of analysis from individual to community level effects may resolve these problems.

Assessing the Costs and Benefits Accruing to the Public from a Graduated Sanctions Program for Drug‐Using Defendants

This paper presents a cost-benefit analysis of the returns to the public from reductions in recidivism associated with a graduated sanctioning program for drug felony defendants. Estimates of program costs for operating a court-based drug testing and sanctioning program are presented with estimates of the value of potential benefits of averted criminal incidents and crime control. The results, based on the evaluation of the Superior Court Drug Intervention Program in Washington, D.C., found that the program saved two dollars in averted crime-related costs for every dollar spent on the program. This paper presents explicit description of the methods used to derive these results so that they may be applied to the evaluation of other experimental/quasi-experimental programs.

Evaluating the Impact of Cybersecurity Information Sharing on Cyber Incidents and Their Consequences

The Department of Homeland Security (DHS) facilitates cybersecurity information sharing among federal government departments and agencies and critical infrastructure owners and operators to promote their security. Information sharing is deemed of critical importance to accomplish the department’s cybersecurity mission; indeed, information sharing is one of the central planks of Executive Order 13636: Improving Critical Infrastructure Cybersecurity, which calls for greater cybersecurity information sharing between the government — not least DHS — and the private sector. But while the importance of information sharing in cybersecurity is intuitive — information that is relevant, timely, and accurate should help cyber defenders reduce vulnerabilities and mitigate threats — the impact of information sharing has not been empirically assessed. The lack of empirical support for information sharing raises two notable issues. First, information-sharing partners, particularly those in the private sector, are sometimes reluctant to participate in government-sponsored initiatives because of concerns about liability, resource costs, and return on investment. Absent empirical demonstration of the value of cybersecurity information-sharing efforts, DHS may be unable to better incentivize participation. Second, information-sharing efforts may, for a variety of reasons, be ineffective (not least due to a lack of participation or the dissemination of irrelevant information). Without assessing the relationship between information sharing and the number and severity (i.e., consequences) of cyber incidents, DHS may be unable to identify and improve poorly performing information sharing efforts. A previous Homeland Security Studies and Analysis Institute (HSSAI) study recommended a suite of metrics to measure various relevant inputs, processes, outputs, and outcomes for cyber information-sharing efforts (Fleming and Goldstein 2012). It did not, however, seek to suggest ways to empirically test the hypothesis that information sharing reduces the number or severity of cyber incidents (it was assumed to do so, per DHS guidance). Accordingly, building on the previous HSSAI research, the present paper sets forth views on use of the dependent variable (some measure of cyber incidents), primary independent variable (some measure of information sharing), control variables, and model specifications.