The Effects of Sanction Intensity on Criminal Conduct - JDAI Helpdesk
The Effects of Sanction Intensity on Criminal Conduct - JDAI Helpdesk
The Effects of Sanction Intensity on Criminal Conduct - JDAI Helpdesk
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
‘typical’ survivor functi<strong>on</strong> for the model.<br />
Figure 2.4 shows the covariate-adjusted<br />
comparative survivor functi<strong>on</strong>s for the treatment and c<strong>on</strong>trol groups. In this graph, the<br />
c<strong>on</strong>tinuous covariates are held at their means and categorical covariates are set to the<br />
modal category. Thus, the ‘typical’ <str<strong>on</strong>g>of</str<strong>on</strong>g>fender is male, n<strong>on</strong>white, about 41 years old,<br />
supervised in the West regi<strong>on</strong>, lives in a zip code area with an average household income<br />
between $20,000 and $29,999, and was not incarcerated post-random assignment. <str<strong>on</strong>g>The</str<strong>on</strong>g><br />
graph shows no difference in the comparative survivor functi<strong>on</strong>s for the average<br />
treatment group <str<strong>on</strong>g>of</str<strong>on</strong>g>fender compared to the average c<strong>on</strong>trol group <str<strong>on</strong>g>of</str<strong>on</strong>g>fender, c<strong>on</strong>firming<br />
earlier findings. <str<strong>on</strong>g>The</str<strong>on</strong>g> ‘typical’ <str<strong>on</strong>g>of</str<strong>on</strong>g>fender has a slightly lower risk <str<strong>on</strong>g>of</str<strong>on</strong>g> failure over time than<br />
the unc<strong>on</strong>trolled sample average. Only around 15 per cent had failed after two years.<br />
<str<strong>on</strong>g>The</str<strong>on</strong>g> results <str<strong>on</strong>g>of</str<strong>on</strong>g> the instrumental variables model used to explore whether<br />
differential treatment take-up impacts the effect <str<strong>on</strong>g>of</str<strong>on</strong>g> LIS are presented in Table 2.5. <str<strong>on</strong>g>The</str<strong>on</strong>g><br />
first part <str<strong>on</strong>g>of</str<strong>on</strong>g> the table shows the results <str<strong>on</strong>g>of</str<strong>on</strong>g> the reduced form OLS regressi<strong>on</strong>, from which<br />
we obtain the probability <str<strong>on</strong>g>of</str<strong>on</strong>g> <str<strong>on</strong>g>of</str<strong>on</strong>g>fending by assigned treatment, and the first stage <str<strong>on</strong>g>of</str<strong>on</strong>g> the<br />
2SLS regressi<strong>on</strong>, where assigned treatment and its interacti<strong>on</strong> with <str<strong>on</strong>g>of</str<strong>on</strong>g>fender<br />
characteristics are used to predict treatment take-up. <str<strong>on</strong>g>The</str<strong>on</strong>g> sec<strong>on</strong>d part <str<strong>on</strong>g>of</str<strong>on</strong>g> the table shows<br />
the outcomes from the sec<strong>on</strong>d stage regressi<strong>on</strong>, which shows the actual effect <str<strong>on</strong>g>of</str<strong>on</strong>g><br />
treatment <strong>on</strong> those who comply with random assignment.<br />
<str<strong>on</strong>g>The</str<strong>on</strong>g> first result to note is that the reduced form (ITT) model is largely c<strong>on</strong>sistent<br />
with our earlier findings about the impact <str<strong>on</strong>g>of</str<strong>on</strong>g> assigned treatment and other covariates.<br />
<str<strong>on</strong>g>The</str<strong>on</strong>g>re are slight differences because we include interacti<strong>on</strong> terms in this model to adjust<br />
for heterogeneity in treatment assignment, but we still see little impact <str<strong>on</strong>g>of</str<strong>on</strong>g> LIS <strong>on</strong> the<br />
probability <str<strong>on</strong>g>of</str<strong>on</strong>g> recidivism. SES, regi<strong>on</strong>, and m<strong>on</strong>ths in jail were again related to the<br />
91