What Works in Corrections: Reducing the Criminal Activities of Offenders and Delinquents.

What Works in Corrections: Reducing the Criminal Activities of Offenders and Delinquents by Doris Layton MacKenzie, New York, Cambridge University Press, 2005, 390 pp.

Doris Layton MacKenzie demonstrates with scientific measures that Robert Martinson's "nothing works" mantra was not only wrong but wrong on many fronts. But it is not this revelation that is important for those reading this book; those of us working in the field of corrections in 1974 knew Martinson was wrong, we just could not prove it. Studies reviewed by Martinson were of such poor scientific quality, he could come to no other conclusion. MacKenzie's book is not only important for what it proves--there are other works, which do the same--but also for showing correctional administrators the absolute necessity of empirically proving the effectiveness of various programs. If for no other reason, every correctional administrator should read the first part of this book.

Today, as it requires more and more fiscal resources, corrections has fallen under much greater scrutiny, and this has caused stress upon organizations responsible for providing care, custody and correction. There has been nothing under more scrutiny than programs designed to reduce recidivism, which fair or not, has become an outcome measure by which success is gauged.

MacKenzie's book, an outgrowth of her 2000 article, "Evidence-Based Corrections: Identifying What Works," in Crime and Delinquency, is an important read for anyone who desires to provide programming to inmates. It demonstrates not only that many things do in fact work, but that administrators and program designers must take those steps to create evaluation criteria, which will stand methodological examination and scientific rigor. Most of us who work in the field are only just now coming to terms with performance management systems and empirical evidence. MacKenzie demonstrates the absolute necessity of learning to ask relevant evaluation questions about program design and program hypothesis. It has been a long time since we entered our elementary school science classes and created our first project for the local science fair. It would serve us all well to remember these basic lessons of hypothesis, methodology, control groups, results and conclusions when establishing and designing a program meant to reduce recidivism or increase institutional efficiency.

In What Works in Corrections, the first step for MacKenzie was to identify programs that could be evaluated. That is to say, programs designed well enough to allow for an evaluation to take place. She categorizes five different levels of studies of programs, which have been described either qualitatively or quantitatively, and a Level 3 study is required as the minimum baseline for a review to be meaningful. Level 3 studies compare two or more groups--at least one with the program being reviewed and one without the program. "The design of the study and the statistical analysis assured reasonable similarity between the treated group and the comparison(s)," MacKenzie wrote.

The book also provides an exhaustive analysis of those programs impacting recidivism as categorized by such definitions as academic education, sex offender treatment, drug treatment and so forth. The real benefit of these chapters is to describe for program designers the things that have been tried and worked, the things that are promising, and the things that the evidence shows do not work. Data is presented and analysis provided to demonstrate conclusions. The bibliography provided and the research presented give examples in a systemic manner. The bibliography and research part of this book is very valuable as a meta-analysis for recidivism-based research. There are other avenues to get this information such as Gendreu, Little and Goggin's work in their 1996 journal article, "A Meta-Analysis of the Predictors of Adult Recidivism: What Works!" found in Criminology. But what MacKenzie does, which is different from this and other works examining evidence-based corrections, is explain the methodology and differences in a manner understandable by all. It takes each category, such as academic education, and lists all of those studies reviewed against the minimum benchmark to describe their effectiveness. What a wonderful resource--a resource that does not require an extensive academic literature search. The author has done it for you.

MacKenzie admits that while she has reviewed many outcome measures and not just recidivism, she has concentrated on it for two reasons. First and probably most important, for legislatures, recidivism has become the primary outcome measure for correctional programming. I would argue, as MacKenzie does, this is short-sighted on the part of those providing funding, as there are many other valid reasons to develop and evaluate programs, but I am also bound by the practical realism that many do not look beyond this primary measure. The second reason MacKenzie has chosen recidivism as the measure of what works is that to measure other program indices at the same time would muddy the water. Here I disagree. If placed in the proper context, I believe there would be great benefit to having the other considerations such as cost, safety and offender needs measured and evaluated in a similar work. Joan Petersilia in a 2004 article in Federal Probation offers that such a narrow evaluation of "what works" does not account for psychological indices nor for what is attractive to those working in corrections, and as such, Petersilia is critical of MacKenzie's narrow view of focusing on sociological variables.

This makes interesting academic discourse, but it does not in any way detract from this work. As stated at the beginning of this review, this work is important not just because of what it says, but because, in a very readable form, it provides administrators and program designers rationales and methods for measuring the effectiveness of correctional programs looking to reduce recidivism. It articulates that we must become comfortable asking such questions if we are to be considered professional in the programs we implement.

Reviewed by Art Beeler, warden at the Federal Correctional Complex in Butner, N.C.