Criminal Justice Outcomes in Intimate and Non-intimate Partner Homicide Cases

3. THE PRESENT STUDY (cont'd)

3.7 Analytic Procedures

Results from the bivariate analyses are presented first, allowing for an examination of patterns that exist between the type of victim-accused relationship and the criminal justice outcomes as well as the other independent and control variables. An examination of these associations is important because such patterns can begin to identify how intimate partner homicides might differ from non-intimate partner homicides in the characteristics of the incidents or the individuals involved. Second, multivariate analyses using logistic regression and ordinary least squares regression are used to assess the extent to which the victim-accused relationship affects the criminal justice outcomes, controlling for relevant legal and extra-legal factors. Put more simply, the multivariate analysis further isolates the independent effects of intimacy on criminal justice decision-making by holding constant the effects of other variables.[25] It is recognized that, as accused persons move through the criminal process, various individuals will drop out of the system at various stages and, thus, the effects of sample selection bias need to be considered.[26] Additional analyses examine in more detail the effects of intimacy over time on criminal justice outcomes as well as the association between gender and intimacy. While qualitative information was not available consistently in all cases, where possible, case narratives that provide more descriptive information on individual cases are provided to further illustrate the context surrounding intimate and non-intimate partner homicides.