Scholarly studies from 1985-2011 were located using a number of databases, including PSYCH-INFO, National Criminal Justice Reference Service, Criminal Justice Abstracts, Criminal Justice Periodicals, and Sociological Abstracts. This led to a total of about 2035 sources. After screening out irrelevant articles, a total of 16 were included for protective orders, 39 for arrest, 24 for prosecution, and 27 for jury decision making. An evaluation of the methodologies employed found that some studies have used one or multiple types of official data sources, such as police reports (incident forms, narratives or affidavits), court records, and criminal histories, while other studies have used interviews or surveys of victims or suspects; but few have used triangulation of both official sources and surveys or interviews. Almost all studies of differential decision making in jury verdicts have been experimental studies of simulated situations.
Most studies on differential treatment in arrest and prosecution have focused on gender and then race, while studies on differential decision making in the issuance of protective orders and jury decision making have focused primarily on gender. In the few studies that examined protective orders, judges were overwhelmingly more likely to issue them to women than to men seeking them (particularly in cases of less severe violence histories), to impose greater restrictions on male defendants, and to defer cases of male plaintiffs, and deny requests at 10-day hearings. The overwhelming number of studies that examined differential arrest by gender found that male suspects are more likely to be arrested than female suspects; however, the difference in arrest rates was mitigated by dual arrests, which contribute to a significant increase in the number and greater likelihood of arrests of women. Greater arrest rates for women also seem to be affected by higher SES, and the presence of weapons and witnesses (legal factors).
Many students of IPV have argued that when examining the context and history associated with the arrest of women, particularly in dual arrest incidents, that women were engaged in self-defense. However, when official action was taken against women, there was greater leniency by citing instead of taking them into custody or by charging them with less serious offenses. While a small number of studies have not found evidence of differential treatment by prosecutors regarding the gender of the offender or victim, most studies with smaller community samples, and some with larger samples, found that males were consistently treated more severely at every stage of the prosecution process, particularly regarding the decision to prosecute, even when controlling for other variables (e.g., the presence of physical injuries) and when examined under different conditions.
The gender discrepancy decreases somewhat with the decision to file felony or misdemeanor charges. In the few studies of gender differences in conviction and sentencing, most have found that male defendants are more likely to be convicted and to receive more severe sentences than female defendants. Subjects in experimental studies of jury decision making in IPV cases have stronger reactions to abuse committed against female victims and abuse committed by male perpetrators, with blame and responsibility often attributed to male perpetrators of assault at higher rates than female perpetrators.
In studies of IPV that simply look at arrest rates in both single and dual arrest cases, as well as the general police handling of IPV incidents, there appears to be little differential treatment against racial minorities. Mandatory arrest policies appear to reduce the importance of victim race to insignificance. Most studies of decisions to prosecute, as well as the few studies on conviction and sentencing, have found that race and ethnicity were not statistically significant when specifying relevant conditions or controlling for other variables. Thus, in terms of the differential criminal justice response regarding demographic categories, it appears that the less favorable treatment of males regarding the issuance of protection orders, arrest, and prosecution is most salient. However, before recommending new policies, interview and survey studies of legal actors should be undertaken in order to better understand the differential treatment found in studies using official sources, as well as how the concepts of “probable cause” and “primary aggressor” are interpreted when applied.
Moreover, future research studies on differential criminal justice response based on demographic and SES factors should include greater specifications and distinctions in the variables examined and broaden the scope of the samples and methods employed. First, studies need to better define the specific types of criminal justice response and distinguish them from related types of responses. For instance, since arrest in IPV is supposed to stop the violence, it is important to distinguish between arrest, which involves taking a suspect into custody and lodging him/her, and citation, which requires appearance in court at a later date. In order to understand whether a differential response is isolated or systemic, it is also important to examine other related police and prosecutor decisions, such as the decision regarding what to charge the suspect.
Secondly, there is a need to specify the characteristics of those groups and both the situational and historic context where a differential criminal justice response has been found. Therefore, it is important to examine the interaction effects of other socio-demographic factors (particularly SES and age), relationship status, and situational conditions (e.g., presence of weapons or alcohol) with primary socio-demographic factors. Besides triangulating official sources with interviews and surveys in studies on protective orders, arrest, and prosecution, experimental studies of jury decision should be complemented by using focus groups to better understand the group dynamics of juries; and with court records to compare findings from simulations with actual cases. Finally, since community studies with small samples are not generalizable, there is a need for more multi-site or national studies.