Recidivism

2001

Preface to the Internet Edition
I started working in the general area of recidivism and correctional measures over 20 years ago, under the assumption that I would be able to make some marginal changes in the measurement of recidivism. What has happened has been quite gratifying. The primary yardstick of the time, the “one-year recidivism rate,” has generally (but not completely) been replaced with measures based on survival analysis, as suggested by this book. It was also gratifying to have the book receive awards in two fields, criminal justice (the Leslie T. Wilkins Book Award) and operations research (the Lanchester Prize). Despite this, the book has been out of print for a number of years, prompting me to retrieve the copyright from Academic Press and publish it on the Internet.

During this period the software environment has changed to a great extent: the FORTRAN programs listed on pp. 206-222 are included more for the sake of historical curiosity than out of any expectation that they will be used. In addition, the use of Cox regression procedures in survival analysis has made the search for covariates simpler than as described in Chapter 10. [However, those who use Cox regression techniques should be aware that, even though it is a nonparametric method, it makes the implicit assumption that the hazard rate of each individual is the same as for that of all other individuals, except for a multiplicative constant.]

One concern that I had at the time I initially wrote this book, and which I still have, is the penchant of some social scientists to use only the standard methods they learned in graduate school, without considering the conditions under which they can and cannot be used. They appear to be wary of going beyond the “tried and true” (or in some instances, “tired and false”) methods.

This is in some sense equivalent to a shoe salesman attempting to sell a size 9 shoe to a person with a size 8 foot, because it’s the only size available. My growing concern about the misuse of methods was detailed in a paper published in November 1994, in the Journal of Research in Crime and Delinquency, “Deviating from the Mean: The Declining Significance of Significance,” and elaborated on in subsequent articles in the Journal of Quantitative Criminology. This book can be used as a primer on how to deal with a measurement problem, by first considering how it has been dealt with in the past, and then attempting to fashion a method based on the characteristics of the problem.

A note on this version: I have tried to keep the same text on this version as on the printed page, so that the index would not have to be redone. This has meant that some of the pages are longer than others. I’ve also taken the opportunity to correct some of the typos; however, as anyone who has scanned material into a computer, new errors are often introduced – one cannot expect the character recognition to be correct everywhere. I would appreciate it if those who find mistakes would communicate them to me Other than that, this version is the same as the original. ..more.. by Michael D. Maltz, Department of Criminal Justice and Department of Information and Decision Sciences, University of Illinois at Chicago