Spouse Assault Replication Program: Studies of Effects of Arrest on Domestic Violence
The Spouse Assault Replication Program, sponsored by the National Institute of Justice, attempts to replicate Sherman and Berk's 1984 Minneapolis Domestic Violence Experiment. Their study was the first to employ randomization to compare the effectiveness of different police responses to domestic violence. They found, according to victims' and official reports, that arrest for misdemeanor domestic violence was almost twice as effective as other police actions in reducing repeated violence. Their study received a great deal of attention and influenced public policies.
After detailed analysis of the Minneapolis study, criticisms began to emerge. Some of the criticisms related to selection of cases, low level of participation by police officers, and possible lack of follow‑through in truly randomizing the police responses.
The replication studies were conducted in Omaha, Charlotte, Milwaukee, Metro‑Dade (Miami, Florida area), Colorado Springs, and Atlanta from 1985 to 1990 to determine whether the Minneapolis study's results could be reproduced in other settings. To date, results from all the studies except Atlanta have been made public.
Methodologies of the Studies
All the studies dealt with cases in which there was probable cause for a misdemeanor arrest, and the studies usually had some limitations in selection of cases for random assignment. Most of the studies excluded cases where the suspect was gone or the victim wanted an arrest, where the police thought arrest was necessary to ensure safety of victim, or where a restraining order or warrant had been issued. They varied somewhat in whether the parties involved must have been married or have lived together for some part of the year preceding the incident and in whether both the victim and the abuser had to be present when police came. In Colorado Springs, unlike other settings, only 38% of cases involved an assault.
In each study, cases were randomly assigned to a variety of responses (see Table 1). The Omaha study included a sub‑study on effects of issuing a warrant for offenders who were absent when the police arrived. In Milwaukee, police were required to warn the couple that if they were called again, they would have to make an arrest. In Charlotte, police frequently ""advised"" the couple along with other randomized responses, and one police response consisted of issuing a citation for the offender to appear in court. In Colorado Springs, three out of four officer options included issuing emergency protective orders.
The consequences of arrest varied. In Charlotte, abusers averaged 9.4 hours in jail after an arrest. In the Metro‑Dade study, the average time an abuser was held in jail after arrest was 14.6 hours. Rates of prosecution of abusers varied from 1% to 35%. In Milwaukee, 5% of the offenders were charged, and 1% were convicted. In Charlotte, 35.5% of the citation and arrest cases were prosecuted, with 80% pleading guilty or being found guilty.
Findings of the Studies
The studies vary greatly in the types of outcome measures used (see Table 1). A simplified summary of the results suggests that, while arrest was sometimes related to an initial deterrent effect, it often faded by the end of one year. In Omaha, Charlotte, and Milwaukee, arrest was not more effective than other options in reducing recidivism by abusers. In Metro‑Dade and Colorado Springs, arrest had deterrent effects according to victim data but not according to official arrest reports. In Omaha issuing a warrant when the offender was absent was a deterrent. There are indications in some of the studies that arrest was a deterrent only for offenders who were employed.
Critiques of Studies
The major criticism of the studies is that they fail to take into account the context in which arrest takes place. The studies did not include the cultural background of the couples, what might have occurred between the couple before the call, the behavior of police during the call, or the consequences of arrest. Previous histories of offenders were not consistently analyzed. For example, in Charlotte, 69.4% of the offenders had previous criminal histories, so being arrested again might not have had much effect on them. The studies also could not measure the effects of police coming to victimsí houses during a violent incident. Police arrival interrupts ongoing violence and may give victims time to escape or get in touch with help.
In several of the studies, arrested abusers spent little or no time in jail. Few arrested men were prosecuted or found guilty, and the reports did not include information about whether any offenders were involved in court‑ordered batterers treatment. Although some studies mentioned that police gave out victim information cards, there was no analysis of the types of formal or informal support that the victims received. These studies were clearly not focused on coordinated criminal justice responses to domestic violence.
Other limitations of the studies include the failure to take into account the effects of the different police options used. For example, in Colorado Springs, the issuance of emergency protective orders might have confused the results. In Milwaukee, warnings from police might have frightened victims away from calling for further help.
These replication studies did not show that arrest definitely deters future violence by all types of domestic abusers. However, looking only at arrest and only at levels of repeated violence is an oversimplification. Recidivism is an inadequate measure of the effects of arrest. It is important to look at victims' perspectives on police interventions and at the message that arrest (or failure to arrest) gives to the victim, the abuser, their children, and to the community. The results of giving this message may not be immediately measurable. Since no alternative intervention by police has been consistently shown to be more effective than arrest, arrest for domestic violence seems morally correct. It treats domestic violence as a serious crime that is comparable to other crimes.
TABLE 1: Published Spouse Assault Replication Program Studies: Interventions and Findings
Follow-Through & Offender Characteristics
2. Separation by forcing offender to leave or convincing victim to leave;
3. Mediation; random assignment to compare effects of warrant issuance if offender absent at police arrival.
Covered whole city; only certain police shifts.
Arrest related to fewer assaults w/in 6 mos.; within 1 year arrest related to same level of recidivism as non-arrest; when offender absent at police arrival & received warrant, recidivism was lower according to official & victim interview data*; unemployed suspects more likely to reoffend than employed.
Mean jail time following arrest was about 16 hours.
65% offenders previously arrested, 64% offenders sentenced to jail, probation or fines.
Excluded cases where offender was absent or victim insisted on immediate arrest;
1. Advising & sometimes separating;
2. Issuing citation for offender to appear in court;
3. Arresting offender at incident.
All cases in city.
Official reports and victim reports at 6 mos show arrest not a deterrent.
Mean jail time following arrest was 9.4 hours.
69% of offenders had previous criminal histories; abuser's prior criminal history related to rate of recidivism; 35.5% offenders with citation or arrest were prosecuted; less than 1% were jailed after a guilty plea or finding.
Excluded cases where suspect was not locatable or had outstanding arrests or restraining orders against him;
1. Full arrest;
2. Short arrest;
3. Warning message that if police return someone will go to jail.
4 districts in city with most domestic violence.
According to police reports to hotline of all domestic violence incidents, arrest was not a deterrent. Researchers conclude arrest had a differential effect depending on whether abusersí employment or marital status.*
5% of offenders charged with a crime and 1% convicted.
Excluded cases where suspect or victim was not on the scene or suspect had outstanding arrests or restraining orders against him;
Stage 1: Random assignment to arrest.
Stage 2: Random assignment to Safe Streets Unit follow-up intervention (detectives visiting homes).
Initially included only married couples, then expanded to cohabitants.
Victim interviews and official data indicate less recidivism for arrested offenders at 6 mos.; victim reports at 6 mos. indicate less recidivism for employed suspects.*
However, employment status was significantly related to whether arrest was a deterrent.
Average suspect spent 14.6 hours under arrest.
Colorado Springs, CO
1. Emergency protective order & arrest of suspect.
2. Emergency protective order & crisis counseling for suspect.
3. Emergency protective order.
4. Restoring order at the scene.
Only 38% of cases involved an assault; 54% were harassment.
Police data shows slight difference for employed abusers.
Victim interviews show arrest as deterrent.
7% of victims & 24% of suspects were in the military.
*Bold print indicates findings have been reported as statistically significant.
Author of this document:
Arlene Weisz, P.h.D.
Wayne State University
School of Social Work
Distribution Rights: This Applied Research paper and In Brief may be reprinted in its entirety or excerpted with proper acknowledgement to the author(s) and VAWnet (www.vawnet.org), but may not be altered or sold for profit.
Suggested Citation: Weisz, A. (2001, November). Spouse Assault Replication Program: Studies of Effects of Arrest on Domestic Violence. Harrisburg, PA: VAWnet, a project of the National Resource Center on Domestic Violence/Pennsylvania Coalition Against Domestic Violence. Retrieved month/day/year, from: http://www.vawnet.org
Berk, R. A. (1993). What the scientific evidence shows: On the average, we can do no better than arrest. In R. A. Gelles & D. R. Loseke (Eds.). Current controversies in family violence, pp. 323‑336. Newbury Park, CA: Sage.
Berk, R. A., Campbell, A., Klap. R. & Western, B. (1992). The deterrent effect of arrest in incidents of domestic violence: A Bayesian analysis of four field experiments. American Sociological Review, 57, 698‑708.
Bowman, C. G. (1993). The arrest experiments: A feminist critique. The Journal of Criminal Law and Criminology, 83 (1),201‑208.
Buzawa, E. S. & Buzawa, C. G. (Eds.). (1996). Do arrests and restraining orders work? Thousand Oaks, CA: Sage.
Dunford, F. W., Huizinga, D., Elliott, D. S. (1990). The role of arrest in domestic assault: The Omaha police experiment. Criminology, 28 (2), 183‑206.
Frisch, L. A. (1992). Research that succeeds, policies that fail. The Journal of Criminal Law and Criminology, 83 (1),209‑216.
Garner, J., Fagan, J., & Maxwell, C. (1995). Published findings from the spouse assault replication program: A critical review. Journal of Quantitative Criminology, 11, 3‑28.
Hirschel, J. D., Hutchinson, I. W., & Dean, C. W. (1992). The failure of arrest to deter spouse abuse. Journal of Research in Crime and Delinquency, 29 (1), 7‑33.
Hirschel, J.D. & Hutchinson, I.W. (1996). Realities and implications of the Charlotte spousal abuse experiment. In E.S. Buzawa & C.G. Buzawa (Eds.). Do arrests and restraining orders work? Thousand Oaks, CA: Sage.
Lerman, L. G. (1992). The decontextualization of domestic violence. The Journal of Criminal Law and Criminology, 83 (1), 217‑240.
Pate, A. M. & Hamilton, E E (1992). Formal and informal deterrents: Dade County experiment. American Sociological Review, 57 (5), 691‑697.
Sherman, L. W. & Berk, R. A. (1984). The specific deterrent effects of arrest for domestic assault. American Sociological Review, 49, 261‑272.
Sherman, L.W., Smith, D. A, Schmidt, J. D., and Rogan, D. P. (1992). Crime, punishment, and stake in conformity: Legal and informal control of domestic violence. American Sociological Review, 57, 680-690.
Sherman, L. W. (1992). Policing domestic violence: Experiments and dilemmas. NY: Free Press.
Zorza, J. (1994). Must we stop arresting batterers?: Analysis and policy implications of new police domestic violence studies. New England Law Review, 28 (4), 929‑990.
* The production and dissemination of this publication was supported by Cooperative Agreement Number U1V/CCU324010-02 from the Centers for Disease Control and Prevention. Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the CDC, VAWnet, or the Pennsylvania Coalition Against Domestic Violence."
This Applied Research paper and In Brief may be reprinted in its entirety or excerpted with proper acknowledgement to the author(s) and VAWnet, a project of the National Resource Center on Domestic Violence, but may not be altered or sold for profit.