Missing time-series data and the impact of sentencing guidelines in Minnesota: can the debate be adjudicated?
Article Abstract:
The issues raised by Stewart J. D'Alessio and Lisa Stolzenberg and by Carlisle Moody and Thomas Marvell pertain to the effect of missing data in measuring the impact of an innovation and how these may be handled. Both the methods used and recommended are flawed in that neither Moody and Marvell nor D'Alessio and Stolzenberg considered the issue of statistical power of a time series evaluation design. The power is optimized when the number of pre- and post intervention observations are equal. A more definite conclusion regarding the impact of the Minnesota intervention needs an augmented data base aside from the time series data.
Publication Name: Criminology
Subject: Law
ISSN: 0011-1384
Year: 1996
User Contributions:
Comment about this article or add new information about this topic:
The uncertain timing of innovations in time series: Minnesota sentencing guidelines and jail sentences - a comment
Article Abstract:
Stewart J. D'Alessio and Lisa Stolzenberg measured the effects of the sentencing guidelines and jail sentences innovations using an autoregressive integrated moving average analysis. However, their exclusion of 22 months of data prior to and four months of data immediately following the application of the innovation makes the findings spurious. One way to solve the problem of pinpointing with as much accuracy as possible when an innovation begins to take effect is by treating observations surrounding the innovation as missing data. Selection of which data to exclude must be done cautiously and must be justified.
Publication Name: Criminology
Subject: Law
ISSN: 0011-1384
Year: 1996
User Contributions:
Comment about this article or add new information about this topic:
Challenges for multilevel models of school disorder: response to Hoffmann and Johnson
Article Abstract:
The authors discuss individual-based variables as greater predictors of school misconduct than local- or community-based variables. Earlier criticism of their multilevel modeling called into question their use of a small sample size of schools. The authors refute this critism pointing out the importance of sampling individuals within the schools, and lending less credence to between-school effects on the concluding data.
Publication Name: Criminology
Subject: Law
ISSN: 0011-1384
Year: 2000
User Contributions:
Comment about this article or add new information about this topic:
- Abstracts: Environmental crime. Sentencing guidelines in Minnesota, 1978-2003. Youth justice in New Zealand
- Abstracts: Strange tale of the judge and the Hoover files. Judges' sentencing decisions bolstered
- Abstracts: Disciplinary rules: the relevance of disciplinary rules to the fairness of dismissal. Contract, confidentiality and the public interest
- Abstracts: Concentrated government efforts to prosecute fraud, and corporate sentencing guidelines, make compliance programs necessary for health care organizations
- Abstracts: Emerging combatants, war crimes and the future of international humanitarian law. Examining international responses to illicit arms trafficking