studies drug offenders have been statistically linked with high rates of recidivism (Greenwood and Abrahamse 1982; Wish and Johnson 1986), and it was expected to be a significant predictor in this study as well. One possible explanation for its non-significance here is the high number of three-strike candidates in this study which have had at least one prior drug conviction. Out of 254 cases (2 cases had missing data), 149 or 58.7% indicated the presence of a drug conviction. Because so many three-strike offenders have had prior drug problems, it is very likely that prosecutors are not considering drug abuse to be an important criterion in determining which defendants should qualify for discretionary treatment.
Model 2
The second sub grouping of variables concentrated in Model 2 tests the use of discretion against variables related to the mitigating characteristics of the defendant and the case as presented through the presence of specific moderating factors Once again, the dependent variable is the prosecutor�s decision to strike a strike. This time, however, the independent variables include: no history of violence (coded by the prosecutor based upon an assessment of the defendant�s criminal history), no history of weapons possession or use (also coded by the prosecutor based upon criminal history information), documented history of mental illness (coded by the prosecutor if a legal determination of mental instability has been made), proof problems (noted by the prosecutor if the defendant�s prior strikes are not properly documented, and therefore are unable to be proven in court), serious medical condition (noted by the prosecutor if the defendant appears to be suffering from a chronic or terminal debilitating condition), and mitigating personal characteristics (coded by the prosecutor if the defendant appears to have a supportive family, steady employment record, or a pattern of successful completion of parole). The hypothesis, variable list, and expected directions of the logit coefficients for Model 2 are summarized in Figure 3.
| Hypothesis | Variables |
|---|---|
| In using their discretionary authority to strike a prior conviction, prosecutors will be more likely to strike a prior strike if there are mitigating factors present in either the case or in relation to the defendant. | Dependent variable: DA Discretion Independent variables: |
 |
� Years between last strike and current offense (+) � No history of violence (+) � No history of weapons use (+) � Mental illness (+) � Proof problems (+) � Serious medical condition (+) � Mitigating personal characteristics (+) |
| Variable | Logit Coefficienta |
Standard Error |
Logistic Coefficient24 |
Expected Direction |
|---|---|---|---|---|
| Defendant has no history of violence | 1.0160 | .6587 | 2.2438 | Yes |
| Defendant has no history of weapons possession or use | 2.6050* | .7427 | 12.9976 | Yes |
| Defendant has documented history of mental illness | 2.5350** | .9366 | 13.2849 | Yes |
| Prior strikes have proof problems | 3.0521* | .8189 | 21.1592 | Yes |
| Defendant currently has serious medical problems | 1.0481 | 1.2696 | 4.8473 | Yes |
| Defendant has mitigating personal characteristics | -0.5253 | .5435 | .5307 | No |
| constant | -1.6676 | .2053 | - | - |
|
Number of observations | 249 | X2 56.92 | |
|
Log-likelihood | -117.53 | ||
| Figures are unstandardized logit coefficients. The decision to use discretion is coded 1 if prosecutorial discretion is used to strike a strike, 0 if the prosecutor chose not to strike a strike. | ||||
| Numbers in bold are significant as follows: * Significant at p<.001. ** Significant at p<.005. 1Significant at p < .01. 1Significant at p < .05. Observed probability levels based on a one-tailed test. | ||||
In working with mandatory sentences like the California three-strikes law, prosecutors likely recognize that �rules without discretion cannot fully take into account the need for tailoring results to unique facts and circumstances of particular cases� (Davis 1969, 17). Adjusting the requirements of the law to meet this need is not a new practice for prosecutors; discretion within the criminal justice system is pervasive (Misner 1996; Rhynhart 1985; Walker 1993). However, given the purpose of three-strikes, prosecutors are expected to be mindful of the crime control impact that their decisions will have upon the community at large by only striking prior strikes for those defendants and cases that present mitigating circumstances. This expectation is confirmed through the significance of this model at the 99 9999% level (X2 = 56.92).
or go to...
1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11
12 | 13 | 14 | 15 | 16 | 17 | 18 | 19
Return to Home Page