Offender Profile and Recidivism among Domestic Violence Offenders in Ontario

8. Logistic Regression Analysis of Variables Influencing Recidivism

Table 27 presents logistic regression coefficients for the strongest variables influencing recidivism as well as odds-ratio results. This analysis was undertaken to identify those variables that have the strongest relationship to the likelihood of recidivism after controlling for all other statistically significant variables at the same time. The variables highlighted in this table show the four variables that have the most influence in recidivism. In sum, odds-ratio results show that offenders who had at least one prior conviction were twice as likely as offenders who did not have a prior conviction to be reconvicted after the index domestic violence conviction (2.3). The results are similar for offenders who had previously been sentenced to prison (2.2) and offenders who received a prison sentence for the index domestic violence conviction (1.9). And finally, offenders aged 18–34 were 1.6 times more likely than offenders aged 35 and over to be reconvicted following the index domestic violence conviction. It is noteworthy that having appeared in a DVC for the index domestic violence conviction did not play a role in the likelihood of recidivism for the offenders in this sample.

Table 27: Logistic Regression for Variables Influencing Recidivism, 2001
  Pr > ChiSq Exp (Est)
Number of Concurrent Convictions (Index Domestic Violence Conviction) 0.2997 1.070
Existence of Prior Convictions 0.0012 2.282
Existence of Prior Prison Sentences <.0001 2.225
Existence of Prison Sentence for Index Domestic Violence Conviction 0.0003 1.901
Offender Age 0.0022 1.580
Existence of Prior Convictions for Spousal Offences 0.1211 1.588
Existence of Prior Convictions for Other Violent Offences 0.8434 0.963
Seriousness of Index Domestic Violence Conviction 0.8742 0.976
Gender 0.4665 1.233
Domestic Violence Court 0.6840 0.940

1. R-Square=0.1555; Max-rescaled R-Square=0.2180

Table 28 presents logistic regression coefficients for the strongest variables influencing the severity[14] of the post-index domestic violence conviction as well as odds-ratio results. This final analysis was undertaken to identify those variables that have the strongest relationship to the severity of the reconviction after controlling for all other statistically significant variables at the same time. The variables highlighted in this table show the four variables that have the most influence on the severity of the reconviction. In sum, odds-ratio results show that offenders who had at least one prior conviction were 2.3 times more likely than offenders who did not have a prior conviction to be reconvicted of a violent offence after the index domestic violence conviction. As well, offenders who received a prison sentence for the index domestic violence conviction and offenders who received a prior prison sentence were almost twice as likely as offenders who did not receive such sentences to be reconvicted of a violent offence following the index domestic violence conviction (1.9 and 1.8 respectively). And finally, offenders aged 18–34 were 1.7 times more likely than offenders aged 35 and over to be reconvicted of a violent offence following the index domestic violence conviction. Again, it is noteworthy that having appeared in a DVC for the index domestic violence conviction did not play a role in the likelihood of violent recidivism for the offenders in this sample.

Table 28: Logistic Regression for Variables Influencing Severity of Reconvictions, 2001
Pr > ChiSq Exp (Est)
Number of Concurrent Convictions to Index Domestic Violence Conviction 0.1071 1.112
Existence of Prior Convictions 0.0018 2.322
Existence of Prior Prison Sentences 0.0026 1.837
Existence of Prison Sentence for Index Domestic Violence Conviction 0.0004 1.925
Offender Age 0.0009 1.674
Existence of Prior Convictions for Spousal Offences 0.7319 1.108
Existence of Prior Convictions for Other Violent Offences 0.8460 0.963
Seriousness of Index Domestic Violence Conviction 0.8502 1.031
Gender 0.5524 1.197
Domestic Violence Court 0.7539 1.051

1. R-Square=0.1250; Max-rescaled R-Square=0.1824

[14] Included in this analysis are Category 1 (Serious Violence) and Category 2 (Violence) offences in the seriousness index presented in the Methodology Section.