A. Piquero – Handbook of Quantitative Criminology (2010)

Автор: A. Piquero
Название книги: Handbook of Quantitative Criminology
Формат: PDF
Жанр: Юриспруденция и право
Страницы: 787
Качество: Изначально компьютерное, E-book

This is the authoritative volume on methodological and statistical issues in the criminology/criminal justice field. It provides overviews of the issues with examples and figures, as warranted, and it covers new and advanced methods in the field.

Quantitative criminology has certainly come a long way since I was first introduced to a
largely qualitative criminology some 40 years ago, when I was recruited to lead a task
force on science and technology for the President’s Commission on Law Enforcement and
Administration of Justice. At that time, criminology was a very limited activity, depending
almost exclusively on the Uniform Crime Reports (UCR) initiated by the FBI in 1929 for
measurement of crime based on victim reports to the police and on police arrests. A typical
mode of analysis was simple bivariate correlation. MarvinWolfgang and colleagues were
making an important advance by tracking longitudinal data on arrests in Philadelphia, an innovation
that was widely appreciated. And the field was very small: I remember attending my
first meeting of the American Society of Criminology in about 1968 in an anteroom at New
York University; there were about 25–30 people in attendance, mostly sociologists with a
few lawyers thrown in. That Society today has over 3,000 members, mostly now drawn from
criminology which has established its own clear identity, but augmented by a wide variety of
disciplines that include statisticians, economists, demographers, and even a few engineers.
This Handbook provides a remarkable testimony to the growth of that field. Following
the maxim that “if you can’t measure it, you can’t understand it,” we have seen the early
dissatisfaction with the UCR replaced by a wide variety of new approaches to measuring
crime victimization and offending. There have been a large number of longitudinal self-report
studies that provided information on offending and on offenders and their characteristics to
augment the limited information previously available from only arrest data. The National
Crime Victimization Survey (NCVS, formerly the NCS) was initiated in 1973 as an outgrowth
of the Commission’s recommendation to provide a measure of the “dark figure of crime”
that did not get reported to the police. These initiatives had to be augmented by analytic
innovations that strengthen the quality of their data. Inevitably, some data would be missing
and imputation methods had to be developed to fill the gaps. Self-reports were hindered by
recall limitations, and life calendars were introduced to facilitate memory recall.
Economists became interested in crime shortly after Garry Becker, building on the notion
that the “demand” for crime would be reduced by increasing the punishment, or “price.” He
proposed an initial model of deterrence and his successors brought multivariate regression as
a standard tool in criminology. That opened the door to variations such as logistic or probit
models, for analysis of natural experiments when randomized design was not feasible, and
for the use of propensity scores to better match treated and control populations. That brought
time series models and hierarchical models into criminology also. Experimentationwas used to a limited degree early in criminology, but those experiments
were largely limited to the kinds of psychological treatments that could be tested on a randomly
separated treatment and control groups of offenders. Largely under the initiative of
Lawrence Sherman, who led with the Kansas City Preventive Patrol experiment, we have
seen a striking variety of randomized social experiments testing various means of operating
elements of the criminal justice system, including police or courts as well as corrections,
and new methods had to be developed to enhance the validity of those experiments and to
compensate for the difficulty of incorporating a placebo into a social experiment.
Since there were limits to the degree to which one could experimentally manipulate the
criminal justice system, a wide variety of modeling approaches developed. These include
simulation models to analyze the flow of offenders through the system, models of criminal
careers, and their dynamics from initiation to termination. Daniel Nagin introduced
trajectory models as an important means of aggregating the dynamics of hundreds of individual
longitudinal trajectories into a small number of distinct patterns that could capture the
essential characteristics of longitudinal phenomena. Other models included spatial models of
the diffusion of criminal activity within a community or across communities, network models
characterizing the linkages among groups of offenders, and many more.
These are just a sampling of the many analytic innovations that Alex Piquero and David
Weisburd have admirably assembled in this Handbook. This allows someone seeking an
appropriate and innovative method for collecting some new data or for analyzing a particular
set of data to explore a wide variety of approaches that have already been used, and hopefully
to build on them in new ways that will provide an additional chapter for a future edition of the
Handbook.