Probably not, says a recent report by the District of Columbia Crime Policy Institute. A Bayesian Meta-Analysis of Drug Court Cost-Effectiveness finds that:
On average, drug courts cost $5,000 more per participant than is yielded in benefits, and there is only a 14 percent chance that the benefits will exceed the costs. …There is a 1 percent chance that the benefits could be as high as $23,000 per participant, which across 150 participants is an aggregate social gain of $3.4 million.
The study simulates the costs and benefits of an average drug court using Bayesian analysis, a statistical technique that, the authors note, “provides more flexibility to estimate complicated models than any alternative” and “better describes true reality.” Rather than yielding a point estimate like a cost-benefit ratio, Bayesian analysis measures the likelihood that the benefits of a policy or program will outweigh the costs.
For more information about Bayesian analysis and its use in CBA, read the report, and watch the accompanying presentations, Meta-Analytic Cost-Benefit Estimates and A Bayesian, Meta Cost-Benefit Model.
For general information on Bayesian statistics, see:
- Andrew Gelman, et al. Bayesian Data Analysis. Boca Raton, FL: Chapman & Hall/CRC, 1995.
- Sean Eddy. “What is Bayesian Statistics?” Nature Biotechnology 22, no. 9 (September 2004): 1177-1178.
No related posts.