Skip to main content
. Author manuscript; available in PMC: 2023 Mar 8.
Published in final edited form as: J Crim Justice. 2022 Mar 8;79:101900. doi: 10.1016/j.jcrimjus.2022.101900

Table 5.

Regression Estimates for the Effect of SB416 on Conviction Recidivism

Any Convictiona Count of Convictionsb
Est. SE p OR 95% CIOR Est. SE p ER 95% CIER
Adjusted for Length of Follow-Upc
 Overall Effect
   Intercept −0.638 0.179 <0.001 0.53 [0.37, 0.75] −0.644 0.113 <0.001 0.53 [0.42, 0.66]
   SB416 −0.191 0.246 0.439 0.83 [0.51, 1.34] 0.025 0.157 0.871 1.03 [0.75, 1.39]
 Effect by County
   Intercept −0.298 0.247 0.227 0.74 [0.46, 1.20] −0.578 0.145 <0.001 0.56 [0.42, 0.75]
   SB416 −0.148 0.327 0.651 0.86 [0.45, 1.64] 0.193 0.192 0.315 1.21 [0.83, 1.77]
   County B −0.779 0.368 0.035 0.46 [0.22, 0.94] −0.172 0.233 0.460 0.84 [0.53, 1.33]
   SB416 × County B −0.421 0.532 0.428 0.66 [0.23, 1.86] −0.957 0.362 0.008 0.38 [0.19, 0.78]
  EMM: County Bd
   SB416 vs. PAU −0.092 0.067 0.170 0.57 e −0.250 0.102 0.013 0.47 e
Adjusted for Length of Follow-Up and Time Incarceratedf
 Overall Effect
   Intercept 0.003 0.190 0.985 1.00 [0.69, 1.45] 0.176 0.122 0.151 1.19 [0.94, 1.51]
   SB416 −0.404 0.258 0.117 0.67 [0.40, 1.11] −0.216 0.166 0.193 0.81 [0.58, 1.12]
 Effect by County
   Intercept 0.243 0.261 0.353 1.27 [0.76, 2.13] −0.002 0.152 0.988 1.00 [0.74, 1.34]
   SB416 −0.306 0.341 0.371 0.74 [0.38, 1.44] 0.159 0.199 0.424 1.17 [0.79, 1.73]
   County B −0.539 0.388 0.165 0.58 [0.27, 1.25] 0.470 0.250 0.060 1.60 [0.98, 2.61]
   SB416 × County B −0.533 0.553 0.335 0.59 [0.20, 1.74] −1.461 0.382 <0.001 0.23 [0.11, 0.49]
  EMM: County Bd
   SB416 vs. PAU −0.180 0.093 0.048 0.43 e −1.160 0.336 <0.001 0.27 e

Note. SB416= Oregon Senate Bill (SB) 416 Program. PAU = Probation as Usual. For Overall Effect, SB416 is the estimated intervention effect across both counties. For Effect by County, SB416 is the estimated effect within County A. Within County B, the effect of SB416 is provided by the Estimated Marginal Mean (EMM).

a

Binary logistic regression model (logit link).

b

Negative binomial regression model (log link).

c

Natural log transformed exposure term for length of follow-up.

d

EMM = Estimated Marginal Mean; the difference in probability/count between SB416 and PAU in County B.

e

95% CIs for OR/ER could not be computed; the formulation did not estimate the SE for the between-group difference in log odds/log event rate.

f

Natural log transformed exposure term that removes time incarcerated from length of follow-up.