Consistent with their emphasis on crime control, it is expected that the use of discretion will be significantly correlated with those variables that are related to the defendant�s likelihood to recidivate. This model, which evaluates the decision to strike a strike against variables which measure various aspects of recidivism, is in fact highly significant at the 99.9999% level (X² = 30.72). Individual variables that are significantly different from zero include the evaluation factors prior strikes remote in time and no recent criminal history These results indicate that prosecutors are over three times more likely to strike a strike when prior strikes remote in time is cited and two-and-a-half times more likely to strike a strike when no recent criminal history is cited.²⁵

Previous studies, including a major analysis of felony offenders by Gottfredson and Gottfredson, indicate that the best statistical prediction of a defendant�s likelihood to recidivate is the length and activity of his own criminal career (Gottfredson and Gottfredson 1986; Greenwood and Abrahamse 1982). It appears as if prosecutors in this study are corroborating this assessment as well. By justifying the decision to strike a strike because the defendant has managed to remain relatively crime free (misdemeanor offenses are generally not counted as noteworthy offenses in this evaluation process), prosecutors are essentially stating their confidence in the defendant�s returning to a law-abiding lifestyle once the prison term for the current offense has ended.

Interestingly, the variable calculating the number of years between the last strike and the current offense is not significantly different from zero. This may be due to the fact that this variable is also capturing the effect of the defendant�s age on the decision to strike a strike, since the well-seasoned career criminal would have theoretically have a longer time span between the last strike and the current offense, whereas a younger defendant would have less intervening time between the two offenses. Age has been a significant predictor of recidivism in other studies (Tonry 1996), as has the age at which criminal activity began (Gottfredson and Gottfredson 1986; Greenwood and Abrahamse 1982), but could not be directly. tested with the information available in this dataset. Although not significant, this variable is negatively correlated with the decision to strike a strike as expected. The longer the criminal career, as measured here by the intervening time between the last strike and the current offense, the more likely it is that the defendant will continue to re-offend, and the less likely the chances that the prosecutor will exercise discretion to strike a strike.

Another surprising result is that the variable drug convictions also failed to vary significantly from zero, nor was it negatively correlated with the decision to strike a strike as expected. In previous

²⁴Logistic coefficients can be changed into odds ratios by subtracting the coefficient from 1 and then converting the number into a percentage. For example, a one unit increase in the independent variable years between last strike and current offense yields a 4% decrease in the odds of the dependent variable (I)A Discretion) taking the value of 1 (1.00 - .9603 = .0397 or 4%). Likewise, the variable drug conviction produces a 47% increase in the odds that prosecutors will strike a prior strike.

²⁵Odds ratios were convicted using the previous formula: When the independent variable prior strikes remote in time is increased by one unit (from 0 to 1 there is a 326% increase in the odds that prosecutorial discretion will be exercised (4.26 � 1 .00 3.26 or 326%).

or go to...
1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11
12 | 13 | 14 | 15 | 16 | 17 | 18 | 19