Clinical Versus Actuarial Judgments in Criminal Justice Decisions: Should One replace the Other?
How Much Risk Can We Take? The Misuse of Risk Assessment in Corrections
Clinical Versus Actuarial Judgments in Criminal Justice Decisions: Should One replace the Other?
1 Portions of the preceding discussion adapted from Gottfredson and Gottfredson (1986).
2 Historical-Clinical-Risk-Management-20 (HRC-20).
Validation of the Risk and Resiliency Assessment Tool for Juveniles in the Los Angeles County probation System
1 These estimates are based on unpublished analyses from Turner & Fain (2003).
2 Police refer cases to the District Attorney in Los Angeles for processing. Youths charged with offenses for which the District Attorney must file and those youths who are detained in juvenile hall are directed to the Court for arraignment. The SDRRC was administered at this pre-plea stage for these “court” cases. Youths not initially referred to court—those generally with more minor offenses—are referred to Probation to make a determination of how to handle the case. These “non-court” cases can received a number of possible outcomes, including having the case closed, the youth being placed on informal probation, or the case being referred to court. The SDRRC was administered to “noncourt” cases at this point.
3 Due to logistical restrictions, we were not able to pilot test the SDRRC in all area Probation Department offices.
4 For weighting purposes, ethnicity was divided into five categories: black, Hispanic, white, other race, and unknown. Age was categorized as less than 13, greater than 18, and single years of age for ages 13-18.
5 We were not able to use age in calculating weights because there were too few representatives in the sample for some combinations of gender, race/ethnicity, court type, and age.
6 This includes those with negative scores.
7 We include the square of age as a factor in the logistic regression because age has a curvilinear relationship with rearrest. Little (n.d.) used a similar analytic approach in her evaluation of the SDRRC.
How Much Risk Can We Take? The Misuse of Risk Assessment in Corrections
1 See http://bmj.bmjjournals.com/cgi/content/full/312/7023/71. Also see http://www.ahrq.gov/clinic/epc/ for a listing of the growing number of evidence-based practice medical and mental health centers in the U.S. and Canada.
2 Lowenkamp, Christopher T. and Edward J. Latessa. 2005. “Evaluation of Ohio’s CCA Funded Programs. Final Report.” Cincinnati, OH: University of Cincinnati, Center for Criminal Justice.
3 Austin, James. 2006. “What Should We Expect From Parole?” American Probation and Parole.
4 For a recent summary of the validity of the LSIR and its history see Girard, Lina and J. Stephen Wormith, 2004. “The Predictive Validity of the Level of Service Inventory- Ontario Revision on General and Violent Recidivism among Various Offender Groups” Criminal Justice and Behavior, Vol. 31, No.2:150-181 and Violent Recidivism among Various Offender Groups” Criminal Justice and Behavior, Vol. 31, No.2:150-181.
5 For more information on MHS, Inc see their website at http://www.mhs.com/index.htm.
6 For more information about Northpointe see their website at http://www.northpointeinc.com/contact.htm.
7 Washington State Institute for Public Policy. Washington’s Offender Accountability Act: An Analysis of the Department of Corrections’ Risk Assessment. December 2003. Olympia, Washington; James Austin, Dana Coleman, Kelly Dedel-Johnson, and Johnette Payton. 2003. Reliability and Validity of the LSI-R Risk Assessment Instrument . Washington, DC: The Institute on Crime, Justice and Corrections at the George Washington University; and James Austin, 2006. Vermont Parole Board Risk Based Guidelines, Technical Assistance Report #2. Washington, DC: National Institute of Corrections.
8 Washington State Public Policy Institute. 2003. p.4
9 Austin, James, Dana Coleman, and Kelly Johnson. 2002. “Reliability and Validity of the LSI-R for the Pennsylvania Board of Parole and Probation.” Washington, DC: The Institute on Crime, Justice and Corrections, The George Washington University.
From Theory to Practice: The Lifecycle Document for the Results-Based Management Framework for the Federal Probation and Pretrial Services System
1 The terms “outcome-based” and “results-based” are used interchangeably.
2 The term “evidence-based practice” implies that 1) there is a definable outcome(s); 2) it is measurable; and 3) it is defined according to practical realities (e.g. recidivism).
3 As shown by the framework development flow chart, the process is iterative. All references to “completion” refer to the initial development process.
4 While the focus of the measurement matrix is on outcomes, examples of measures for the inputs depicted in the logic model have been provided. It is expected that these will be revised and refined based on review and comment from system staff and stakeholders.
5 While most of the imputation-based procedures described are available through SPSS, other more advanced procedures may require additional statistical software, such as SAS or Stata.
6 For principal component analysis, variance maximizing (varimax) rotation should be selected as the extraction method. Additionally, different criterion (Kaiser criterion, scree test) should be examined to determine which solution makes the best sense, often one retaining more factors (Kaiser) than the other (scree).
7 There are over 25 goodness-of-fit calculations available through the SPSS add-on AMOS. The most common used are model chi-square (not significant indicates model fit), GFI (goodness-of-fit index) (greater than or equal to .90 to accept the model), and CFI (comparison-fit-index) (greater than or equal to .90 to accept the model). Any or all of these should be compared.
8 Path analysis and structural equation modeling can be conducted using SPSS and the SPSS add-on software AMOS as well as SAS. necessary to modify the model (add and/or remove arrows depicting relationships, add/or remove variables, etc.).
9 Most univariate analysis can be conducted using the Descriptive Statistics option within the Analyze function of SPSS.
10 Bivariate analysis can be conducted by using the Analyze function of SPSS and selecting Descriptive Statistics, Compare Means, and Correlate options.
11 When comparing means using ANOVA, multiple range tests are used, the most popular of which is Tukey’s HSD procedure.
12 ANOVA can be conducted by using the Analyze function in SPSS and selecting the Compare Means/One-Way ANOVA option.
13 This particular statistical technique is not available through SPSS. It would require the use of Stata statistical software.
14 Survival analysis can be performed using the Analyze function and selecting the Survival/Cox Regression option in SPSS.
15 While you can use SPSS to generate trend graphs, a software add-on, SPSS Trends, is required to conduct more sophisticated time series analysis.
16 When comparing across different populations, it is important to use b’s (regression coefficients) rather than $’s (standardized regression coefficients) because they are more sensitive to fluctuations in variances and covariances across populations.