Book contents
- Frontmatter
- Contents
- Contributors
- Foreword
- Preface
- 1 Experimenting with Crime and Criminal Justice
- Part I Experimenting with Crime
- Part II Experimenting with Crime Prevention and Criminal Justice
- 6 Randomized Experiments in Criminology
- 7 Increasing Equivalence in Small-Sample Place-Based Experiments
- 8 Multisite Randomized Trials in Criminology
- 9 Using Regression Discontinuity Designs in Crime Research
- 10 Quasi-Experimentation When Random Assignment Is Not Possible
- Part III Assessing Research Evidence and Future Directions
- Index
- References
6 - Randomized Experiments in Criminology
What Has Been Learned from Long-Term Follow-Ups?
Published online by Cambridge University Press: 05 June 2014
- Frontmatter
- Contents
- Contributors
- Foreword
- Preface
- 1 Experimenting with Crime and Criminal Justice
- Part I Experimenting with Crime
- Part II Experimenting with Crime Prevention and Criminal Justice
- 6 Randomized Experiments in Criminology
- 7 Increasing Equivalence in Small-Sample Place-Based Experiments
- 8 Multisite Randomized Trials in Criminology
- 9 Using Regression Discontinuity Designs in Crime Research
- 10 Quasi-Experimentation When Random Assignment Is Not Possible
- Part III Assessing Research Evidence and Future Directions
- Index
- References
Summary
LONGITUDINAL-EXPERIMENTAL STUDIES
Farrington (1983) and Farrington and Welsh (2005, 2006) reviewed randomized experiments in criminology with the following features: (1) at least fifty units (e.g., persons or areas) were initially assigned to each condition, or at least one hundred units were initially assigned to two experimental conditions; (2) there was an outcome measure of offending; and (3) the experiment was published in English. They found that 122 different experiments of this kind had been published up to 2004.
Farrington (1979) reviewed longitudinal studies in criminology with the following features: (1) at least several hundred persons were initially studied; (2) there were at least two personal contacts with the participants and/or their families, separated by at least five years; (3) there was a measure of offending; and (4) the study was published in English. At that time, only eleven studies of this kind had been published.
Farrington, Ohlin, and Wilson (1986) reviewed the advantages and problems of both experimental and longitudinal studies in criminology and recommended that combined longitudinal- experimental studies should be carried out. For a number of reasons (specified in more detail by Tonry, Ohlin, and Farrington 1991), they recommended that the ideal study should have several assessments (personal contacts) over several years, followed by an experimental intervention, followed by several more assessments over several years. No study of that kind had ever been conducted in criminology, although there had been studies that compared officially recorded offending in a few years before the intervention with officially recorded offending in a few years after (e.g., Empey and Erickson 1972; Empey and Lubeck 1971).
- Type
- Chapter
- Information
- Experimental CriminologyProspects for Advancing Science and Public Policy, pp. 111 - 140Publisher: Cambridge University PressPrint publication year: 2013
References
- 22
- Cited by