Craig V. D. Thornton

Crime, Deterrence, and Rational Choice

This study examines the deterrent effect of formal sanctions on criminal behavior. While most research on deterrence assumes a rational-choice model of criminal decision-making, few studies consider all of the major elements of the model. In particular, three critical limitations characterize the empirical literature on deterrence: the failure to establish a causal ordering of sanctions and crime consistent with their temporal ordering; the focus on conventional populations and nonserious criminal acts, which are of less interest to the question of how society controls its members; and the inattention to the return or reward component of the decision-making process. To address these issues, we specify, estimate, and test a rational-choice model of crime on data that were collected on individuals, gathered within a longitudinal design, and derived from three distinct populations of persons at high risk of formal sanction. The results support the reward component of the rational-choice model, but fail to support the cost or deterrent component, as measured by perceived risks of formal sanctions. (abstract Adapted from Source: American Sociological Review, 1986. Copyright

Evaluating the benefits and costs of the job corps

Social programs have a wide variety of effects and often have the explicit objective of improving the economic status of the people they serve. In order to be useful to policymakers, benefit-cost analysis of social programs should explicitly take account of these two important program features. The approach used in this analysis of the benefits and costs of the Job Corps does this and provides a useful methodology for evaluating other social programs. According to the analysis, the program has substantial net value for society as a whole as well as for the average Corpsmember.

Transitional Aid for Released Prisoners: Evidence from the Life Experiment

Findings are presented from a controlled experiment designed to test the effectiveness of transitional aid programs for ex-prisoners in reducing theft crimes. Upon release from prison, a sample of men with high (ex ante) probabilities of committing theft crimes were enrolled in treatment and control groups. One year after release, a group receiving financial aid had significantly fewer arrests for theft crimes than did the controls. Calculations of the social benefit/cost ratio show that the benefits of transitional income maintenance substantially outweighed the costs. In contrast, the provision of job-placement assistance turned out to have no significant effect on post-release behavior.

Measuring Program Outcomes

Administrators are faced daily with questions from various constituents regarding how well their program is doing and whether it has really had an effect on the participants. Answering these questions requires clearly defined and measured outcomes. In many programs, the time hasn’t been taken to implement the necessary steps to have relevant outcome measures on participants, despite their importance. Part of the reason is the focus on process analysis; part is due to ignorance of the literature regarding currently accepted outcome measures; and part of the reason is simply lack of time to get the data on outcomes or to develop the techniques to capture outcome measures. Our general feeling is that administrators can operationalize outcome measures, but they simply haven’t, due to financial, time, or “perpetual crisis” restraints. Thus, our attempt in this chapter is to provide you with a “broad brush” approach to selecting and measuring appropriate outcome measures that can be used for a number of purposes, including reporting, program evaluation, and as the basis for impact and benefit-cost analyses.

Risk Selection among SSI Enrollees in TennCare

The issue of risk selection is especially important for states that enroll blind and disabled beneficiaries of Supplemental Security Income (SSI) in Medicaid managed care. SSI beneficiaries have persistent needs for care, have a wide variety of chronic conditions, and often need atypical and complex services. Risk selection occurs when the health care needs of beneficiaries enrolled in a specific plan differ systematically from the needs of the overall beneficiary population and payments do not reflect those needs. We assess the extent of risk selection among managed care plans for SSI beneficiaries over the first three years of Tennessees Medicaid managed care program, TennCare. Using claims data containing fee-for-service expenditures prior to enrollment in managed care, we find substantial evidence of persistent risk selection among plans. Results are robust to most alternative measures of risk selection for most plans.

Participants in the Medicaid Buy-In Program, 2000–2004 Characteristics, Earnings, and Medical Expenditures

The Medicaid Buy-In program, a state Medicaid option since the late 1990s, is designed to encourage adults with disabilities to work by allowing them to buy into Medicaid when their earnings are above the standard Medicaid eligibility limits. This study provides a descriptive profile of enrollment, expenditures, and earnings for Buy-In participants in 27 states between 2000 and 2004. Analyses indicate that younger participants receiving no federal income support when they enroll in the Buy-In program have higher earnings compared with older participants receiving Social Security Disability Insurance payments. The innovative process used in this study to link data from multiple agencies could be adapted for exploring other policy issues related to the employment of individuals with disabilities.

Effective employment services for persons with mental retardation

Evidence from the Social Security Administrations Transitional Employment Training Demonstration indicates that the overall performance of supported employment programs can be improved through program models that emphasize flexible and persistent services.

Describing Your Program

The process of evaluation begins with specifying exactly what is being evaluated. This careful specification is necessary in order to make the structured comparisons that lie at the heart of any evaluation. It begins with a statement of the problem or problems being addressed by the program under evaluation; that is, it reviews the program goals. It then outlines the population being served and the services provided. Finally, it considers the process by which the services provided will produce the desired effects on the persons being served.

Estimating Program Impacts

We talked about comparisons in the previous chapter and indicated that evaluation involves structured comparisons. Now the issue is really before us. This is the point where we get to the tough issues we mentioned regarding the need to measure outcomes on both sides of the comparison (participants and nonparticipants) and then contrast these outcomes. In the simplest and most ideal sense, impacts of a program are estimated by contrasting the outcome measures for a group of participants with those for a comparison group. Administrators need to make many decisions that involve comparisons and choices among alternatives. Throughout the book we have stressed that administrators need to make these decisions based on valid evaluation data, and that evaluation involves structured comparisons. In this chapter, we discuss various ways to structure comparisons and thereby be in a better position to estimate program impacts.

Back-of-the-Envelope Benefit-Cost Analysis

We stressed throughout Chapter 9 that benefit-cost analysis should be perceived as a process for organizing information rather than as an inflexible rule for making decisions. In particular, benefit-cost analysis provides a convenient summary measure for those impacts that can be measured and valued in dollars and a framework for assessing the potential importance of impacts that cannot be valued in dollars.