In domestic-violence literature, “dual arrest rates” are defined in different ways and often not described fully. Dual arrest means that two persons are arrested for mutual domestic violence. Because roughly half of domestic-violence incidents result in an arrest, dual arrests as a share of all incidents is about half of dual arrests as a share of arrest incidents. Dual-arrest incidents as a share of total arrest incidents is slightly more than half of arrests in dual-arrest incidents as a share of total arrests. The former statistic is an incident-based measure. The latter statistic is an arrests-based measure that recognizes that dual arrests produce two arrests. As a matter of general concern about persons caught up in the criminal justice system, dual arrests as a share of arrests is more conceptually relevant.

The best U.S. national estimate is that domestic-violence dual arrests amount to about 8% of arrests for domestic violence. Arrests made in dual-arrest domestic-violence incidents accounted for an estimated 7.4% and 7.5% of all arrests for domestic violence according to the U.S. National Incident-Based Reporting System (NIBRS) in 2000 and 2010, respectively. The corresponding dual-arrest arrest shares for intimate-partner-violence arrests in NIBRS 2000 and 2010 are 7.3% and 7.2%, respectively. In addition to coverage skewed away from the largest cities, NIBRS under-reports arrests for relatively minor crimes. That’s likely to bias downward the share of arrests made in dual-arrests circumstances relative to total arrests for intimate-partner violence. A reasonable overall national estimate for dual arrests based on NIBRS data is 8%.

The share of domestic-violence arrests resulting from dual-arrests varies widely across jurisdictions. Domestic-violence arrests resulting from multi-person arrest incidents are predominately dual arrests. The best state-reported domestic-violence dual-arrest statistics are for Connecticut, Rhode Island, and Arizona. In Connecticut from 1992 to 2011, 35% of domestic violence arrests were arrests made in multi-arrest domestic-violence incidents. Dual-arrest statistics for Arizona, 2003 to 2004, and for Rhode Island, 1999 to 2002, are 11.6% and 8.7%, respectively. The dual-arrest share for Alaskan State Police in 2004 was 8%. In one jurisdiction in Colorado in 1994, the aggregate share of domestic violence arrests that occurred in dual-arrest incidents was 8.0%.^ In New York across eight research sites surveyed in 1996-1997, the aggregate share of domestic violence arrests that occurred in dual-arrest incidents was 8.9%. In a survey of New York State prosecutors in 2000, about half indicated dual arrests amount to 1-5% of cases, while 30% indicated that dual arrest cases amounted to over 11% of domestic-violence cases.^

Some claims about dual-arrest rates in the domestic-violence literature aren’t credible. In 1997, a scholarly article reported in a peer-reviewed journal:

The state of Washington initially reported a dual arrest rate of 50% and Oregon, 11% (Epstein, 1987).^

The source “(Epstein, 1987)” apparently in turn cited as its source an article in 1985 in the National Lawyers Guild Anti-Sexism Newsletter.^ ^ That’s not a commonly known statistical publication, nor a publication with much statistical authority. The source of its statistics isn’t clear. In any case, the cited statistics are not adequately defined. While the credibility of these statistics are merely superficial, they have been frequently repeated in domestic-violence literature considered to be authoritative.

The domestic-violence literature on dual arrests nearly uniformly emphasizes the need to reduce the frequency of dual arrests. One such article reported that Dallas, TX, had a dual arrest rate of 6%, and then the dual arrest rate fell to 1% after police were trained to arrest the person “whose culpability is greatest.” ^ Another study reported:

Boulder County, Colorado has had a “pro-arrest policy” since 1986; in 1994, Dual Arrests accounted for 4.3% of the total (Jones and Belknap 1999), but by 1999 this portion had risen to 10% (Boulder County Domestic Abuse Prevention Project: Summary 2000). In Lincoln and Lancaster Counties, Nebraska, through a coordinated community effort, the Dual Arrest rate was reduced by 55%, from 12.1%, in the years following the implementation of mandatory arrest, to 6.2% more recently (personal communication, Moyer, B., January 2, 2001).^

This study reports on specific efforts in New York State to reduce dual arrests:

many efforts have been made by the NYPD, the DA’s offices and domestic violence advocates to decrease the numbers of Dual Arrests in the five counties. Though there are no publicly released statistics kept by the NYPD on Dual Arrests, a joint subcommittee of two major domestic violence tasks forces has met several times since 1998 to address this issue. This subcommittee is comprised of representatives of all of the DA’s offices in the City, the Director of NYPD’s Domestic Violence Unit, and staff from the Helpline at the UJC. The committee has coordinated several “snap shots” evaluations whereby DA staff track the numbers and kinds of Dual Arrests that enter the five DA’s offices during a particular week. From this, we have calculated estimates of the numbers of Dual Arrests in NYC. In 1998, we estimated that approximately 2,500 Dual Arrest cases occurred annually. By December 1999, we estimated these numbers had dropped to approximately 1,500 Dual Arrest cases annually.^

Credible social-scientific data indicates that about 40% of violence between intimate partners is mutual violence.^ Domestic-violence literature and policy has largely been built upon gender stereotypes. Efforts to reduce dual arrests have meant efforts to decrease arrests of women for domestic violence by gender-profiling men for arrest. Domestic-violence dual arrest rates are typically much lower than credible shares of mutual violence. Comparison of dual arrest rates for same-sex and opposite-sex violence indicates that gender profiling affects dual arrest rates. Gender profiling men for arrest for domestic violence helps to account for domestic-violence dual arrest rates being low relative to a relevant, factual benchmark.