FORUM on Corrections Research

Measuring what prisons do: A synopsis

Gerald G. Gaes¹

Office of Research and Evaluation, National Institute of Justice

The opinions expressed in this article are those of the author and are not intended to represent the policy of the National Institute of Justice or the United States Department of Justice.

Introduction

Prisons are complex organizations. They serve many purposes, and have a multitude of symbolic meanings to diverse audiences. How do you create a prison performance yardstick that encompasses both this inherent complexity and the multiplicity of purpose and interest? These are some of the questions raised in our forthcoming book Prison Performance: Laying the Groundwork to Compare Public and Private Prisons,² co-authored by Gerald Gaes, Scott Camp, Julianne Nelson, and William Saylor. This book is a culmination of research conducted by the co-authors on the problem of prison performance measurement. Some of this work was a byproduct of our interest in comparing publicly and privately operated prisons while serving as either employees or contractors studying this topic for the Federal Bureau of Prisons. In writing this book, however, we have tried to place our analysis in a broader context. The book is also intended to be a paradigm for the performance measurement of any government agency, and a resource for anyone interested in methods to evaluate whether a government service should be privatized. What follows is a chronological presentation of the book’s contents.

Performance measurement perspectives

In chapter 1, we concern ourselves with the purpose of prison. This, we argue, is the place to begin if a system is to have a coherent framework for performance measurement. In this chapter, we represent the deliberation of other scholars who have offered different perspectives on either the purpose of the criminal justice system in general, or prison in particular. We consider arguments by John DiIulio, James Q. Wilson, and Charles Logan. However, we describe Logan’s proposal in great detail, since we believe his work was seminal in outlining the logical steps leading from prison purpose to prison performance measurement. While we reject Logan’s thesis for what prisons ought to do, we give him credit for his thoughtful analysis of the problem. Instead, we argue that jurisdictions must define their own purpose, and this is usually expressed in the agency’s mission. However, once defined, it is incumbent upon jurisdictions to translate missions into objectives, and objectives into measurable dimensions. If one of the missions of a prison system is to promote successful offender reintegration, then that should be transformed into objectives such as increasing the offender’s skills. This, in turn, must be expressed as measurable performance indicators, such as the actual level of skill achievement. In this chapter, we also take up recidivism as a key performance measurement indicator. We explore the complexity and difficulty in measuring recidivism at both the individual and institutional level. Measuring individual recidivism has many inherent difficulties, but we also consider the additional intricacies in measuring prison recidivism rates. We consider recidivism to be such an important dimension of prison performance that we take up the issue again in a later chapter.

Audit

In chapters 2, 3, 4, and 5 we consider different ways prison jurisdictions can monitor prison performance by focusing on numerous approaches that have been used by a variety of jurisdictions. In chapter 2, we discuss the prison audit. In many jurisdictions, the governing agency will send in teams of inspectors or auditors to examine the way a particular prison function is being conducted, and to determine whether or not the prison is in compliance with policy and law. We show how these kinds of procedures can easily be translated into hard scientific data by focusing on a model prison security audit developed with the support of the National Institute of Corrections.

Assessment

In chapter 3, we consider “qualitative assessment.” This is intended to have a broad meaning encompassing different kinds of reports or presentations that do not necessarily include large amounts of data or detailed analyses. Agencies often conduct “after action” reports when a significant event has occurred, such as a prison riot, or a particularly egregious prison homicide. There are also many scholarly examples in the literature, books devoted to understanding the broad context and social history of a prison. Jim Jacob’s book on the Stateville prison in Illinois is a classic example. These kinds of broad, often historically based, treatments give texture and contour to the understanding and analysis of prison performance.

Performance measurement

Chapter 4 is devoted to behavioural performance measurement. Broadly conceived, this is analysis based on the recording of an inmate’s behaviour. Sometimes these data are collected in the normal, day-to-day context of prison operations. Other times these data are collected using special assessments for a particular study. One of the most common forms of behavioural data collected in prison is misconduct incident reports. While such data may have the appearance of objectivity, we discuss why the context of behavioural measurement must be understood before we can make conclusions about its veracity.

Chapter 5 discusses survey measurement as a performance tool. Both the Correctional Service of Canada and the Federal Bureau of Prisons have conducted large scale surveys of both inmates and staff. These data can envelop any topic from the most sensitive to the mundane. In this chapter, we show how such data can provide a great deal of insight into prison performance.

In both chapters 4 and 5, we also delve into the statistical and methodological complexities of measuring performance at the level of the organization — in this case, the prison. New statistical tools have been developed and the software is now widely available to conduct multilevel analysis of data. These tools have been used to investigate school achievement, hospital quality, and prison performance. The breakthrough in these techniques is that they allow the investigator to take into account features of the organization and the individual simultaneously. To measure the performance of a hospital, the analyst must account for the individual characteristics of the patients. When measuring schools, you need to account for different compositions of students. Prison performance can understandably vary with the attributes of individual inmates. We explain this methodology from the perspective of an unsophisticated reader. Our intention is to make this understandable to the policy maker, the administrator, and the student. There are many technical explanations of multilevel models, and in the book we refer to some of these in our discussion. We also demonstrate how the results of these complex multilevel models can easily be depicted in an assortment of graphic forms. These depictions show how institutions rank from best to worst on prison performance dimensions.

The cost of prisons

Chapters 6 and 7 are devoted to the costing of government services and some of the economic theory behind the assumed efficiencies in prison privatization. These chapters rely heavily on the work of one of our co-authors, Julianne Nelson, who has spent a great deal of her career studying cost benefit and cost analysis. She has also conducted studies of the relative costs of publicly and privately provided prison operations and prison medical services. We incorporate Julie’s prescription for how to do a cost analysis of private and public provision. We also consider the economic theory lurking behind the private-public debate about any government service, but particularly prison services. We discuss the work of economic theorists who have taken different sides of the argument — those that believe private provision is inherently more efficient, and those who argue that there are some services that may not produce greater efficiencies when privatized. In chapter 7, we also show how cost can and should be incorporated into the analysis of prison performance. This is particularly decisive when comparing publicly and privately operated prisons. While one prison may be less expensive than another, what happens to quality when cost goes down? In this chapter, we also focus on labor. By far, the greatest proportion of prison expenses comes from the cost of correctional workers. In this context, we raise the possibility that the private provision of prison services depends on the “McDonaldization” of prison work. This is the idea, borrowed from the organizational sociology literature, that prison work can be turned into highly regimented and specific routines analogous to the way the McDonalds restaurant chain has developed its service. This makes labor more easily replaceable, or fungible as the economist would argue. But, what does this do to prison quality? We discuss the evidence.

In chapter 8, we raise the possibility of making performance measurement comparisons across jurisdictions. Can we compare the performance of the Correctional Service of Canada with the Federal Bureau of Prisons? Can we compare the quality of work conducted by the departments of corrections in Texas, California, and Michigan? While these comparisons are fraught with problems, we also recognize that there are some events and processes that may only be understood, if, and when, we perform these kinds of analyses.

Recidivism vs. desistance

Chapter 9 returns to the measurement of recidivism by embedding the topic in an exciting area of criminology called the criminal life course literature. This kind of research is being conducted in Canada, the United Kingdom, Germany, the United States, and other countries. The perspective is developmental and tries to account for the occurrence of criminal propensities at some time in the life course of the individual. Some researchers are even studying prenatal and perinatal causes of criminal propensity. Life course work also examines the factors that lead to the process of criminal desistance — the termination of a criminal career. We give a brief, broad overview of this exciting new work, and show how it may be fundamental to an understanding of recidivism and how we can incorporate these ideas into prison performance.

Summary

In chapter 10, we take all of the ideas from the previous chapters and show precisely how one can measure all of the prison functions. We also show how these can be incorporated into user-friendly tools for administrators. Then in chapter 11, we broaden the horizon of this work by showing how prison performance fits into the larger framework of government performance and accountability. We do this by reviewing the most recently published literature in the field of public administration. We demonstrate how the prior 10 chapters can be considered a paradigm for investigating the performance and cost efficiency of any government service, from collecting trash to providing social welfare. In this chapter, we also explore unintended consequences of performance measurement, including those that may undermine the accuracy of the system.

The last chapter summarizes the major themes of the book and lays out an agenda for future work. Our intent in writing this book was to demonstrate how it is possible to measure prison performance, to show how this paradigm generalizes to any government service, and to point out that most of the prior analyses of prison privatization have been either poorly conceived or poorly executed.

1   13505 Ivy Way, Bowie, MD 20715

2   Gaes, Gerald G; Camp, Scott D; Nelson, Julianne, & Saylor, William G. Prison Performance: Laying the Groundwork to Compare Public and Private Prisons, Alta Mira, CA: AltaMira Press, forthcoming.