Andrea D. Lyon*, "Be Careful What You Wish For: An Examination of Arrest and Prosecution Patterns of Domestic Violence Cases in Two Cities in Michigan," 5 Mich. J. Gender & Law 253 (1999).

* Assistant Clinical Professor of Law at the University of Michigan Law School. I want to thank the University of Michigan Law School for its generous support of this research, as well as the following students who assisted me in the research and writing of this article: Lizabeth Boyer, Mya Bronson, Deborah M. Golden, Kaethe Hoffer, and David Kahng.

Introduction

I distinctly recall waiting one day in court for a law school clinic case to be called and noticing that of the six or seven domestic violence cases, three or four of the defendants were women. I thought that odd, but it had, after all, been a long time since I had spent any time in misdemeanor courts. Maybe women were doing more abusing than I remembered, or maybe it was just an aberrational day. But then I saw similar patterns on several other days.

I became curious: was this a pattern? If it was, why was it? I decided to look at six months of data on arrests for domestic violence. I chose the cities of Ypsilanti and Ann Arbor, because although they are the two largest cities in Washtenaw County n1 and both house major universities, n2 their populations are quite different. Ypsilanti has a much more significant minority population n3 than Ann Arbor. n4 Also Ann Arbor has a "must arrest" ordinance, n5 while Ypsilanti has the milder "should arrest" statute. n6 I looked at all of the domestic violence police reports from June through December of 1996.

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n1. According to the 1990 Census, the population of the City of Ann Arbor was 109,592. The population of Ypsilanti was 24,846. Unless stated otherwise, all population figures are from the 1990 Census. See U.S. Census Bureau: The Official Statistics [hereinafter Census Bureau], U.S. Gazeteer (visited October 3, 1998) <http://www.census.gov/cgi-bin/gazetteer>.

n2. Eastern Michigan University is in Ypsilanti and the much larger University of Michigan is in Ann Arbor.

n3. Ypsilanti's minority population was 31%, 7618 people. See Census Bureau, supra note 1, 1990 U.S. Census Data (visited October 3, 1998) <http://vemis.census.gov/cdrom/lookup>.

n4. Ann Arbor's minority population was 18%, 19,396 people. See 1990 U.S. Census Data, supra note 3.

n5. A "must arrest" statute requires police officers to make an arrest if there is probable cause that domestic assault occurred. See infra Part II, notes 55&ndash;63 and accompanying text; see also Ann Arbor, Mich., Code ch. 108, 9:68 (1995).

n6. A "should arrest" statute states that police "should" arrest a domestic violence perpetrator if there is probable cause to believe domestic assault has occurred. See infra Part II, notes 55&ndash;63 and accompanying text; see also Ypsilanti Police Department, Policy and Procedure #17 Domestic Violence. (March 13, 1991, revised October 14, 1991).

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

This Article will examine what this data revealed, but in order to be able to interpret what the data means, I did some other research. I was surprised by what I found - for example, although women tend to be injured most severely by domestic violence, they use violence in intimate relationships a little more often than men. n7 Part I of this Article traces a brief history of domestic violence and discusses the issue of who commits domestic violence, Part II discusses the "must arrest" and "should arrest" policies and their history, Part III lays out what the data received from Ann Arbor and Ypsilanti revealed, and Part IV addresses the policy implications of this information and proposes some changes.

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n7. See Murray A. Straus & Richard J. Gelles, Physical Violence In American Families 1-15 (1990) [hereinafter Physical Violence].

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

I. What is "Domestic Violence'? n8

Over one hundred years ago John Stuart Mill identified the abuse of women by their male partners as a significant social problem. n9 Still, male violence against women remains a threat in our society, costing too many women their dignity, their health, and sometimes their lives. Although women have made gains in their struggle for equality, and despite the fact that women are almost as likely as men to resort to violence in their relationships, women continue to be disproportionately vulnerable to, and harmed by, violence in their intimate relationships. n10

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n8. This Article will be dealing with violence in the context of heterosexual relationships. The particular issues raised by battering in the context of gay and lesbian relationships is beyond the scope of this Article, for which I apologize.

n9. The vilest malefactor has some wretched woman tied to him, against whom he can commit any atrocity except killing her, and, if tolerably cautious, can do that without much danger of legal penalty....When we consider how vast is the number of men, in any great country, who are little more than brutes and that this never prevents them being able, through the law of marriage, to obtain a victim, the breadth and depth of human misery...swells to something appalling.

John Stuart Mill, The Subjection of Women 62 (1869).

n10. See discussion infra Part I.B.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

Domestic violence is often defined in gender neutral terms. For example, domestic violence has been defined as: "The emotional, physical, psychological, or sexual abuse perpetrated against a person by that person's spouse, former spouse, partner, former partner or by the other parent of a minor child. Abuse may include threats, harm, injury, harassment, control, terrorism or damage to living beings or property.' n11 This definition obscures the fact that women are more frequently and more severely harmed as a result of violence in their homes and relationships. For that reason, many people advocate definitions of domestic violence that do not mask the gendered nature of the problem. n12

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n11. Margi Laird McCue, Domestic Violence: A Reference Handbook 2 (1995) (citing Laurie Hubbard, Portland, OR: Bureau of Community Development, From Harassment to Homicide: A Report on the Response to Domestic Violence in Multnomah County (1991)).

n12. See, e.g., Ann Jones, Next Time She'll Be Dead: Battery and How To Stop It 81-89 (1994).

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

However, researchers in the past fifteen years have revealed that women are almost as likely as men to resort to violence in their conflicts with intimate partners. n13 Many people cite this research to support the position that domestic violence is gender neutral. Some argue that violence and assault are criminal acts regardless of the gender of the assailant, and that because of this, we should not support definitions of, or policies regarding, domestic violence that may avoid holding women accountable for the violence they perpetrate. n14

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n13. See Physical Violence, supra note 7.

n14. See Murray A. Straus, Physical Assaults by Wives: A Major Social Problem, in Current Controversies on Family Violence 67, 80 (Richard J. Gelles & Donileen R. Loseke eds., 1993).

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

A. Who is Resorting to Violence?

1. Both Men and Women

According to two massive studies conducted in 1975 and 1985, men and women resort to violence at almost the same rate during conflicts in their interpersonal relationships. n15 These National Family Violence Surveys were based on interviews with nationally representative samples of married and cohabiting couples. In 1975, 2,143 couples were surveyed, and in 1985 data was gathered from 6,002 couples. n16 Although research has indicated that men are less likely to admit having committed assaults, the data from the 1975 and 1985 studies revealed that women admit to engaging in a surprisingly high rate of violence. n17 In response to questions that specifically asked who initiated the physical violence in conflicts that became violent, women reported that they had struck the first blow in 53.1% of the fights, their partner hit first 42.3% of the time, and they were unsure about 3.1% of the fights. n18

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n15. See Jeffrey L. Edleson & Mary P. Brygger, Gender Differences in Reporting of Battering Incidences, 35 Family Rel. 377 (1986); Ernest N. Jouriles & K. Daniel O'Leary, Interspousal Reliability of Reports of Marital Violence, 53 J. Consulting & Clinical Psychol. 419 (1985); Maximiliane E. Szinovacz, Using Couple Data as a Methodological Tool: The Case of Marital Violence, 45 J. Marriage & Fam. 633 (1983).

n16. See Straus, supra note 14, at 68.

n17. See Jouriles & O'Leary, supra note 15, at 420.

n18. See Straus, supra note 14, at 74 (citing the 1985 National Family Violence survey).

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

A large number of studies and reports have confirmed the fact that women are as likely, and in some cases more likely, to assault their male partners. For example, a 1989 report by Sugarman and Hotaling, which dealt with the results of twenty-one studies of assault among dating couples, 393 per 1,000 women in dating relationships resort to violence, as compared to 329 per 1,000 men. n19

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n19. See Straus, supra note 14, at 71 (citing Sugarman & Hotaling study).

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

This data may seem to contradict the conclusion that the problem of domestic violence is a problem of men abusing women. The "who" of domestic violence only tells one part of the story, however. In order to see the whole truth, it is critical to study the "why" and the "how" of the violence that is carried out in the context of intimate relationships.

2. Men Use Violence to Exert Control

Historically, men were legally entitled to physically assault their wives, n20 and up until the 1980s, men could not be legally charged with raping their wives. n21 American society was not willing to make such intense violation of women by their husbands illegal until quite recently. n22 In addition to the legal and social traditions that support male aggression against their intimates, contemporary studies reveal that men continue to feel justified in exerting control over the women with whom they are involved. n23 According to a report based on interviews with 109 women survivors of battery at the hands of male partners, violence against them was just one way in which batterers exercised control. n24 "Interviews with batterers show that men believe they are justified in their use of violence, particularly when their wives do not conform to the ideal of the "good wife.'' n25 There are many reasons that men cite for the violence that they inflicted on "their women': the house was a mess, the food tasted bad, her clothing was suggestive, or she looked at another man. n26 One survivor of male assault identified the reasons that she was battered: "Because I wouldn't act like he wanted me to act, talk like he wanted me to talk, be who he wanted me to be.' n27

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n20. English common law, as well as the case law of many of the States, recognized the "right' and "privilege' of men to control their wives through physical abuse (subject to certain moderation, of course) until the late 1800's. See, e.g., Demie Kurz, Physical Assaults By Husbands: A Major Social Problem, in Current Controversies on Family Violence, supra note 14, at 88, 90-91 (quoting Angela Browne, When Battered Women Kill 167 (1987)).

n21. See, e.g., People v. Scott, 227 N.E.2d 72, 74 (1967) (stating indictment was proper since victim was not the defendant's wife).

n22. See, e.g., Me. Rev. Stat. Ann. tit. 17-A, 253 (West 1983 & Supp. 1997); State v. Smith, 426 A.2d 38 (1981) (examining history of marital exception in New Jersey).

n23. See Kurz, supra note 20, at 92 (citing R. Emerson Dobash & Russell P. Dobash, Violence Against Wives: A Case Against Patriarchy (1979) (men use anger and psychological abuse in addition to physical abuse to control women); David Adams, Treatment Models of Men Who Batter: A Profeminist Analysis, in Feminist Perspectives on Wife Abuse 176, 191-192 (Kersti Yll<um o> & Michele Bograd eds., 1988); James Ptacek, Why Do Men Batter Their Wives?, in Feminist Perspectives on Wife Abuse, supra, at 133).

n24. See Kurz, supra note 20, at 91&ndash;92 (citing Dobash & Dobash, supra note 23); see also Mildred Daley Pagelow, Woman-battering: Victims and Their Experiences (1981); Lenore E. A. Walker, The Battered Woman Syndrome (1984).

n25. Kurz, supra note 20, at 92 (citing Adams, supra note 23 and citing Ptacek, supra note 23); see also Dobash & Dobash, supra note 23.

n26. See, e.g., Jones, supra note 12, at 92.

n27. Jones, supra note 12, at 94 (citation omitted).

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

3. Women Use Violence to Express Frustration and to Fight Back

Men "typically hit or threaten to hit to force some specific behavior on pain of injury," and "a woman may typically slap a partner or pound on his chest as an expression of outrage or in frustration because of his having turned a deaf ear to repeated attempts to discuss some critical issue.' n28 Research has also shown that a significant difference between male and female aggression is that women are much less likely to engage in aggression if it would cause harm to their target. n29

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n28. Straus, supra note 14, at 78.

n29. See Straus, supra note 14, at 78 (citing Alice H. Eagly & Valerie J. Steffan, Gender and Aggressive Behavior: A Meta-Analytic Review of the Social Psychological Literature, 100 Psychol. Bull. 309 (1986)).

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

A great deal of research has also suggested that women engage in violence only in response to perceived threats of physical or sexual violence against them. n30 Many women use violent tactics as a first-strike defense against getting hit themselves, although this is an ill-conceived tactic, as minor violence by women significantly increases the likelihood that they will be severely assaulted by the men they attack. n31 Women also often identify their violence as responsive to efforts by their partners to dominate, or in retaliation for their partners' violent behavior. n32 Specifically, one 1984 study found that many women hit first due to "well-founded fears" of being raped or beaten, n33 and a 1989 study found that women read certain types of psychological abuse by men as clear early-warning signs of impending violence and lash out in response. n34

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n30. See Kurz, supra note 20, at 96 (citing Browne, supra note 20; Jalna Hanmer & Sheila Saunders, Well-founded Fear: A Community Study of Violence to Women (1984); Daniel G. Saunders, Who Hits First and Who Hurts Most? Evidence for the Greater Victimization of Women in Intimate Relationships. Paper presented at the 41st Annual meeting of the American Society of Criminology, Reno, Nev.).

n31. See Straus, supra note 14, at 79 (citing Lee H. Bowker, Beating Wife-Beating (1983); Scott L. Feld & Murray A. Straus, Escalation and Desistance of Wife Assault in Marriage, 17 Criminology 141-161 (1989); Richard J. Gelles & Murray A. Straus, Physical Violence, Child Abuse, and Child Homicide: A Continuum of Violence or Distinct Behaviors?, 2 Hum. Nature 59-72 (1988)).

n32. See Straus, supra note 14, at 79; Kurz, supra note 20 (citing B. Emery et al., Why Women Hit: A Feminist Perspective, Paper presented at the Annual Meeting of the National Council on Family Relations, New Orleans, La. (1989)).

n33. See Kurz, supra note 20, at 96 (quoting Hanmer & Saunders, supra note 30).

n34. See Kurz, supra note 20, at 96 (citing Saunders, supra note 30).

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

B. Who Is Being Hurt?

1. Usually Women

Domestic violence causes more injuries to women than auto accidents, muggings, and rapes combined. n35 Overwhelmingly, women end up bruised, bleeding, or dead because violence explodes in their homes. Even the Family Violence Surveys, which are most often cited for the finding that women are almost as likely as men to engage in violence, found that women are six times more likely than men to end up injured because of domestic violence at the hands of their partners. n36 Studies show that acts of violence committed by men cause more injury than the same acts committed by women. n37 Perhaps more significantly, studies also show that men are more likely to engage in the severest types of abuse - the kinds that break bones, cause concussions, and land women in hospitals.

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n35. See Not Just A "Family Matter": Hearings on Domestic Violence Before the Subcomm. on Criminal Justice of the House Judiciary Comm., 103d Cong., 2d Sess. 9 (1994) (testimony of Vicki Coffey, Executive Director, Chicago Abused Women Coalition) [hereinafter Hearings on Domestic Violence].

n36. See Straus, supra note 14, at 69.

n37. See Straus, supra note 14 (citing Janice E. Stets & Murray A. Straus, Gender Differences in Reporting Marital Violence and Its Medical and Psychological Consequences, in Physical Violence, supra note 7, at 151&ndash;66; Murray A. Straus, The Conflict Tactics Scales and Its Critics: An Evaluation and New Data on Validity and Reliability, in Physical Violence, supra note 7, at 49-73; Murray A. Straus, Injury and Frequency of Assault and the "Representative Sample Fallacy" in Measuring Wife Beating and Child Abuse, in Physical Violence, supra note 7, at 75&ndash;91).

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

Abuse by intimates may be the single largest cause of severe injury to women. A review of 3,676 hospital records randomly selected from among female patients who came to the hospital with an injury revealed that forty percent of the injuries came from deliberate assaults by partners, and nineteen percent of those women had a previous history of abuse-caused injury. n38 Based on such information, the American Medical Association estimates that one out of every four women is likely to be abused by a boyfriend or husband at some point in her life. n39 Each year, almost four million women will be the victims of severe assaults by boyfriends and husbands. n40 Based only on assaults that are so severe that they come to the attention of police or medical personnel, the American College of Obstetricians and Gynecologists believes that the number of women severely assaulted in any year ranges between three and four million. n41 It may be especially appropriate for an organization of obstetricians and gynecologists to speak out against domestic violence, given that many battered women's organizations believe that women face an increased risk of assault when they are pregnant. Studies of battered women place the number of battered women who are assaulted during pregnancy between twenty-five percent and sixty-three percent, and many medical professionals have noted the propensity of batterers to direct their assaults specifically at the pregnant woman's stomach and reproductive organs. n42

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n38. See McCue, supra note 11, at 81.

n39. See McCue, supra note 11, at 79.

n40. See McCue, supra note 11, at 79.

n41. See McCue, supra note 11, at 79.

n42. See McCue, supra note 11, at 83.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

Like the percentage of assaults that force women to seek medical treatment, it is only a fraction of the abuse that women suffer at the hands of men in their lives that comes to the attention of the criminal justice system. Ninety-one percent of all violent crimes between spouses consisted of aggression by husbands or ex-husbands. n43 In contrast, only five percent of these crimes were perpetrated by women against their male partners. n44 As recently as 1995, nearly thirty percent of women who were murdered were known to have been killed by husbands, former husbands, or boyfriends. n45 Additionally, a publication by the Committee of the Judiciary of the U.S. Senate reported that in 1991, not less than 21,000 crimes against women by intimates were reported to the police every week in 1991. n46 This figure is twice the number of reported robberies. And even though over 1.13 million women each year are reported as victims of serious domestic violence, including assaults, aggravated assaults, murders, and rape against women in the home, the Senate Committee asserted that the real figure was probably three times as high. n47

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n43. See Kurz, supra note 20, at 89-90 (citing National Crime Victimization Survey of 1982 (an annual study conducted by the Department of Justice that surveys approximately 60,000 households)).

n44. See Kurz, supra note 20, at 89-90 (citing National Crime Victimization Survey of 1982, supra note 43).

n45. See Bureau of Justice Statistics, National Crime Victimization Survey, (last modified July 22, 1996) <http://www.usdoj.gov/vawo/manual/who.htm>.

n46. See McCue, supra note 11, at 79.

n47. See McCue, supra note 11, at 79-80.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

2. When Men are Harmed

Men suffer only a fraction of the violence women suffer at the hands of their intimate partners. When men are injured by violence at the hands of their female partners, their injuries tend to be three times less severe than their partners' injuries. n48 Of men who were murdered in 1995, only three percent were killed by their intimate partners. n49 To date, most studies of men who were murdered by women with whom they were involved indicate that a substantial proportion of their deaths were the result of acts of self-defense or retaliation by the women, generally after extended histories of brutal victimization by the men. n50

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n48. See Kurz, supra note 20, at 90 (citing Richard A. Berk et al., Mutual Combat and Other Family Violence Myths, in The Dark Side of Families: Current Family Violence Research 197-212 (David Finkelhor et al. eds., 1983)).

n49. See Bureau of Justice Statistics, supra note 45.

n50. See Straus, supra note 14, at 73 (citing Browne, supra note 20; Angela Browne & Kirk R. Williams, Exploring the Effect of Resource Availability and the Likelihood of Female-Perpetrated Homicides, 23 Law & Soc'y Rev. 75 (1989); Nancy C. Jurik, Women Who Kill and the Reasonable Man: The Legal Issues Surrounding Female Perpetrated Homicide, Paper presented at the 41st Annual Meeting of the American Society of Criminology, Reno, Nev. (November 1989)).

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

C. Why Don't Women Just Leave?

Given the disproportionate harm that women suffer from violence in their intimate relationships, many people wonder why they do not simply leave their abusive partners.

First, women do leave. They sneak out, walk out, get a divorce, move to another town, leave the state, desert the communities in which they are entitled to live, say good-bye to their friends and families. Despite these efforts, they are often still not safe. They are tracked down, dragged back, made to understand that they cannot leave, or are simply killed. The danger that women face jumps astronomically when they attempt to, or do, leave their abusive partners. Women who are separated from their partners are more vulnerable to domestic violence than any other group of women; they suffer from intimate-offender attacks at a rate about three times higher than divorced women and about twenty-five times higher than that of married women. n51 Although separated and divorced women only constitute ten percent of all women, they suffer seventy-five percent of reported spousal violence. n52

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n51. See U.S. Department of Justice, Women Usually Victimized By Offenders They Know (advance for release, Aug. 16, 1995) <http://aspensys.aspensys.com:209/RO-4656-range/ncjrs/data/viowomen.txt> (However, because the survey records a respondent's marital status only at the time of the interview, it is possible in some instances that separation or divorce followed the violence.).

n52. See McCue, supra note 11, at 114 (citation omitted).

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

Second, it is hard for women to get up and walk away. Women overwhelmingly are responsible for taking care of their children. Women have less access to well-paid jobs and are frequently economically dependent on the men in their lives. When a battered woman does manage to walk away, she faces a fifty percent chance that her standard of living will drop below the poverty line. n53

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n53. See McCue, supra note 11, at 113 (citation omitted).

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

Abusive men often succeed in destroying women's self-esteem and their relationships with potential support networks. Frequently, abused women lose their jobs because they are being stalked and harassed by their abuser. Therefore, creating an effective escape plan may be well beyond a battered woman's capacities. n54

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n54. See generally Jones, supra note 12. See also, Beth Sipe & Evelyn J. Hall, I Am Not Your Victim (1996) (first-hand account by Sipe, a survivor of domestic violence).

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

The problem of domestic violence is not a gender neutral problem. It is a problem that disproportionately harms women, interfering with their health, well-being, life choices, and lives. To the extent that men are endangered by domestic violence, they are in danger because women do resort to violence when they are left (or believe they are left) with no alternative ways of escaping from the damage that male violence does to them. At this point, women often become lethal.

II. The "Must Arrest" and "Should Arrest" Policies and Their History

A mandatory arrest ("must arrest") law requires police to arrest a suspect in a domestic violence situation if there is probable cause to believe that the suspect has assaulted his or her partner. n55 A pro-arrest ("should arrest") law states that the police "should" arrest a domestic violence perpetrator as long as there is probable cause to believe a domestic assault has occurred. n56 The mandatory and pro-arrest laws limit police discretion, remove the decision to press charges from the victim of the domestic assault, and give this decision to the prosecutor's office.

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n55. See, e.g., Colo. Rev. Stat. 18-6-803.6 (Supp. 1996) (making arrest mandatory if probable cause to believe any crime against a person or a felony crime against property occurred); Nev. Rev. Stat. Ann. 171.137 (Michie 1992) (making arrest mandatory if probable cause to believe battery has been committed within preceding four hours unless mitigating circumstances exist); see also Marion Wanless, Note, Mandatory Arrest: A Step Toward Eradicating Domestic Violence, But Is It Enough?, 1996 U. Ill. L. Rev. 533, app., tbl.A.2 (1996) (describing arrest thresholds of each state).

n56. See, e.g., Mich. Comp. Laws Ann. 776.22 (Supp. 1998); Ann Arbor, Mich., Code ch. 108, 9:68 (1995); Washtenaw County Sheriff Dep't, Policy and Procedure (1995) ("Arrest and prosecution of the suspect(s) involved shall be regarded as the most appropriate law enforcement response when deputies determine that probable cause exists in domestic violence situations.").

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

An example of a mandatory arrest law is a municipal ordinance passed by the city of Ann Arbor: "(1)A Police Officer shall arrest a person when the Officer has reasonable cause to believe that person has, within the previous 24 hours, assaulted a spouse ... if the victim has visible signs of injury from the assault or if the assailant used or threaten to use a dangerous weapon....' n57 The Ann Arbor mandatory arrest ordinance follows a trend in several states. Thus far, fifteen states and the District of Columbia have enacted mandatory arrest laws for domestic violence situations. n58 An example of a pro-arrest law is the Michigan statute concerning domestic violence calls:

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n57. Ann Arbor, Mich., Code ch. 108, 9:68 (1995).

n58. See Wanless, supra note 55, at 534. The fifteen states are Arizona, Colorado, Connecticut, Hawaii, Iowa, Maine, Missouri, Nevada, New Jersey, Oregon, Rhode Island, South Dakota, Utah, Washington, and Wisconsin.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

Sec. 22. (1) Each police agency in this state shall, by January 1, 1995, develop, adopt, and implement written policies for police officers responding to domestic violence calls. The policies shall reflect that domestic violence is criminal conduct.

....

(3) The policies shall address, but not be limited to addressing, all of the following:

(a) Procedures for conducting a criminal investigation with specific standards for misdemeanor and felony arrests.

(b) Procedures for making a criminal arrest. The procedures shall emphasize all of the following:

(i) In most circumstances, an officer should arrest and take an individual into custody if the officer has probable cause to believe the individual is committing or has committed domestic violence and his or her actions constitute a crime. n59

This section examines the motivations that influenced the passage of mandatory and pro-arrest laws. First, responding to inadequate police response to domestic violence, the women's movement in a number of different states successfully lobbied for mandatory and pro-arrest laws. Originally, empirical studies showing a deterrent value of arrest in domestic violence situations influenced lawmakers to pass mandatory and pro-arrest laws. n60 Since then, however, these studies have been heavily criticized, and indeed, there is little credible evidence that the mandatory or pro-arrest statutes are in fact a deterrent at all. n61 Second, some courts have held that police departments and municipalities can be civilly liable for not responding appropriately to domestic violence situations. n62 In order to avoid civil liability, many states, municipalities, and police departments have enacted mandatory or pro-arrest laws. n63

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n59. Mich. Comp. Laws Ann. 776.22 (West 1998); see also Illinois Domestic Violence Act of 1986, 750 Ill. Comp. Stat. 60/301(a) (West 1994) (allowing, but not mandating, law enforcement personnel to make a warrantless arrest if domestic violence has occurred).

n60. See Lawrence W. Sherman & Richard A. Berk, The Specific Deterrent Effects of Arrest for Domestic Assault, 49 Am. Soc. Rev. 261 (1984); see also Att'y Gen.'s Task Force on Family Violence, Final Report 17, 24 (1984).

n61. See Developments in the Law - Legal Responses to Domestic Violence, 106 Harv. L. Rev. 1528, 1539-1540 (1993).

n62. See, e.g., Thurman v. City of Torrington, 595 F. Supp. 1521, 1531 (D. Conn. 1984); see also Bruno v. Codd, 396 N.Y.S.2d 974, 979 (N.Y. Sup. Ct. 1977) (settling lawsuit with results favorable to battered women).

n63. See Eve S. Buzawa & Carl G. Buzawa, Domestic Violence: The Criminal Justice Response 101&ndash;04 (1996). Examples of cities that have enacted mandatory or pro-arrest laws to avoid liability include Atlanta, Georgia; New Haven, Connecticut; Chicago, Illinois; and Dallas, Texas. See Joan Zorza, The Criminal Law of Misdemeanor Domestic Violence, 1970-1990, 83 J. Crim. L. & Criminology 46, 59&ndash;60 (1992) [hereinafter Zorza, Criminal Law].

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

A. The Women's Movement and Domestic Violence Arrests

The women's movement lobbied for mandatory and pro-arrest laws due to the generally inadequate response of the police to domestic violence situations. The police often ignored domestic violence calls or purposely delayed responding. n64 They considered domestic violence a private, family matter and so minimized public intervention, particularly arrest. n65 In the 1970s, police departments that had a policy on dealing with domestic violence usually had a non-arrest policy. n66 Finally, in most jurisdictions, police officers could not make a warrantless arrest for a misdemeanor assault unless the assault occurred in the officer's presence. n67 Because most police charge domestic violence assaults as misdemeanor assaults, the police had to obtain an arrest warrant in order to arrest the abuser. n68 Since most domestic assaults occur before the police arrive, the added burden of obtaining an arrest warrant probably contributed to the reluctance of the police in arresting the abuser.

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n64. See Joan Zorza, Mandatory Arrest for Domestic Violence: Why It May Prove the Best First Step in Curbing Repeat Abuse, 10 Crim. Just., Fall 1995, at 2, 2-3 [hereinafter Zorza, Mandatory Arrest]; see also Zorza, Criminal Law, supra note 63, at 48 (citations omitted); see also infra note 115 and accompanying text.

n65. See Zorza, Mandatory Arrest, supra note 64, at 2-3.

n66. See, e.g., Zorza, Criminal Law, supra note 63, at 49 (citing Michigan police policy as taught in its police training academy):

a. Avoid arrest if possible. Appeal to [complaintant's][sic] vanity. b. Explain the procedure of obtaining a warrant. (1) Complainant must sign complaint. (2) Must appear in court. (3) Consider loss of time. (4) Cost of court. c. State that your only interest is to prevent a breach of the peace. d. Explain that attitudes usually change by court time. e. Recommend a postponement. (1) Court not in session. (2) No judge available. f. Don't be too harsh or critical.

n67. See Zorza, Criminal Law, supra note 63, at 61 (citation omitted).

n68. See Zorza, Criminal Law, supra note 63, at 61.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

Other myths which caused police nonresponse to domestic violence were that domestic violence calls were particularly dangerous, battered women can safely leave their abusers, and battered women do not follow through on criminal charges against their abusers. n69 However, although domestic violence calls account for thirty percent of all police calls in the United States, these calls account only for 5.7% of police deaths. n70 Also, battered women cannot easily leave an abusive relationship because the violence escalates and becomes most severe when they try to leave their abusers. n71 "Battered women, who are frequently threatened and often beaten by their abusers while the criminal cases are pending, are no more likely to try to drop charges than any other victims of violent crimes who are threatened or harmed by the perpetrator of the crime.' n72

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n69. See Zorza, Mandatory Arrest, supra note 64, at 3.

n70. See Zorza, Mandatory Arrest, supra note 64, at 3.

n71. See Zorza, Mandatory Arrest, supra note 64, at 3; see also Andrea D. Lyon & Helen Morrison, Self-Defense and Battered Spouse Syndrome: A Legal and Psychological Perspective, Criminal Defense, May&ndash;June 1989, at 8.

n72. Zorza, Mandatory Arrest, supra note 64, at 3.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

Women's and battered women's advocates, tired of the lack of effective police response, became the force in some states behind the passage of mandatory or pro-arrest laws. For example, in 1977, Oregon passed the first mandatory arrest law, after it was proposed by the Oregon Coalition Against Domestic and Sexual Violence. n73 In Iowa, the Iowa Coalition Against Domestic Violence drafted Iowa's mandatory arrest law and lobbied for its passage. n74 In New Jersey, the New Jersey Coalition Against Domestic Violence played an important part in the enactment of that state's mandatory arrest law. n75 Similarly, advocates against domestic violence in Missouri were influential in the passage of Missouri's mandatory arrest law. n76

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n73. See Zorza, Criminal Law, supra note 63, at 63.

n74. See Wanless, supra note 55, at 539 (citing Telephone Interview with Laurie Schipper, Executive Director, Iowa Coalition Against Domestic Violence (Feb. 14, 1995)).

n75. See Wanless, supra note 55, at 539 (citing Telephone Interview with Marcia Seal, Program Coordinator, New Jersey Coalition Against Domestic Violence (Feb. 21, 1995)).

n76. See Wanless, supra note 55, at 539.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

B. Empirical Studies and Domestic Violence Arrests

An empirical study on deterring domestic violence was conducted in the early 1980s in Minneapolis, Minnesota which supported the proposition that arrest was an effective deterrent to domestic violence. n77 The Minneapolis Police Department and the Police Foundation, funded by the National Institute for Justice, conducted a scientific study on 314 cases involving domestic violence. n78 In the study, the police responded to actual domestic violence situations in one of three ways: (1) arrested the abuser, (2) removed the abuser from the household, or (3) mediated or advised on the situation. n79 Based on official police reports, the study showed that six months after police responded to the domestic violence, "10% of those arrested, 19% of those advised, and 24% of those removed from the scene had subsequently repeated their violence.' n80 Victim reports also confirmed that arresting the abuser had a better deterrent effect than the other two methods: nineteen percent of those arrested, thirty-seven percent of those advised, and thirty-three percent of those removed repeated their violence. n81 Further studies have been done which support the findings of the Minneapolis study. n82

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n77. See Sherman & Berk, supra note 60, at 263.

n78. See Joan Zorza, Must We Stop Arresting Batterers?: Analysis and Policy Implications of New Police Domestic Violence Studies, 28 New Eng. L. Rev. 929, 934 (1994) [hereinafter Zorza, Must We Stop].

n79. See Zorza, Must We Stop, supra note 78, at 934 (citing Lawrence W. Sherman & Richard A. Berk, The Minneapolis Domestic Violence Experiments, Police Found. Rep. Apr. 1984, at 1).

n80. Zorza, Must We Stop, supra note 78, at 934 (citation omitted).

n81. See Zorza, Must We Stop, supra note 78, at 934 (citation omitted).

n82. See, e.g., Richard M. Tolman & Arlene Weisz, Coordinated Community Intervention for Domestic Violence: The Effects of Arrest and Prosecution on Recidivism of Woman Abuse Perpetrators, 41 Crime & Delinq. 481 (1995) (finding arrest to be a successful long-term deterrant in one Illinois county).

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

However, two years after the release of the original Minneapolis police experiment, the National Institute of Justice funded six new studies to discover whether the results of the Minneapolis study could be replicated in six other cities. n83 The results from these new studies were much more ambivalent about the deterrent effect of arrest on domestic violence. n84 The most dramatic study was the Milwaukee experiment. In Milwaukee, Lawrence Sherman, the author of the original Minneapolis experiment, concluded that: "The evidence shows that, while arrest deters repeat domestic violence in the short run, arrests with brief custody increase the frequency of domestic violence in the long run among offenders in general.' n85

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n83. See Zorza, Mandatory Arrest, supra note 64, at 4. The six cities are Omaha, Nebraska; Milwaukee, Wisconsin; Charlotte, North Carolina; Colorado Springs, Colorado; Miami, Florida; and Atlanta, Georgia. The results have been released and made public for all of the studies except Atlanta.

n84. See Symposium on Domestic Violence, 83 U. Crim. L. & Criminology, Spring 1992.

n85. Lawrence W. Sherman et al., The Variable Effects of Arrest on Criminal Careers: The Milwaukee Domestic Violence Experiment, 83 J. Crim. L. & Criminology 137, 139 (1992).

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

The replication experiments have also been criticized for several reasons, including the failure to properly replicate the Minneapolis experiment. According to Joan Zorza:

The studies [replication experiments] were all conducted independently of each other, and with a lack of consistency that makes overall comparison and analysis difficult. Issues explored in one experiment were completely ignored by others. Methods utilized in one study differed from those used in others, and some studies never revealed their methodologies.... As Sherman admits, none of the police replication studies actually reproduced the research design of the Minneapolis experiment. A variety of police responses were tested on different distinct populations based on distinct domestic violence statutes and police arrest policies. n86

Thus, although far from conclusive, there is some empirical support that arrest has an effect in delaying, if not deterring, further incidences of domestic violence by these offenders.

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n86. See Zorza, Must We Stop, supra note 78, at 963&ndash;965.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

C. Police Liability and Domestic Violence Arrests

Some states may also have enacted mandatory or pro-arrest laws in order to avoid liability. n87 Courts have held police departments liable for failing to protect battered women. n88

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n87. See Wanless, supra note 55, at 538&ndash;39; Zorza, Criminal Law, supra note 64, at 53-60.

n88. See Catherine F. Klein & Leslie E. Orloff, Providing Legal Protection for Battered Women: An Analysis of State Statutes and Case Law, 21 Hofstra L. Rev. 801, 1016-18 (1993); Zorza, Criminal Law, supra note 64, at 53-60; Wanless, supra note 55, at 538&ndash;39.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

The most important case in this area is Thurman v. City of Torrington. n89 In Thurman, a federal jury awarded Tracey Thurman $ 2.3 million after the Torrington, Connecticut Police Department failed to protect Thurman from her abusive husband. n90 Not only did the Torrington Police Department fail to respond to more than twelve attempts by Tracey Thurman and a friend, Judy Bentley, to involve the police, whether by calling 911 or attempting to file police reports after incidents, but one police officer actually stood by while Charles Thurman kicked Tracey Thurman in the head. n91 This officer was responding to a 911 call placed approximately twenty-five minutes before his arrival. n92 Before the officer's arrival, Charles had stabbed Tracey repeatedly in the chest, neck, and throat. n93 This same officer then allowed Charles to enter into Ms. Bentley's residence and return from within with their minor child. n94 Charles then proceeded to drop the child on Tracey and kick Tracey in the head for a second time. n95 Soon after, three more officers arrived on the scene. n96 The police still permitted Charles to wander about the gathered crowd and threaten Tracey. n97 Charles was finally arrested and taken into custody after he approached Tracey one more time, this time while she was lying on a stretcher. n98 Charles, during interactions with members of the police department at his place of employment, had actually boasted to these police officers that he intended to "get" his wife and that he intended to kill her. n99 In trial, Tracey proved that the police department's policy of treating women and children abused by a male friend or relative differently from persons assaulted by strangers amounted to sex discrimination in violation of the Equal Protection Clause of the Fourteenth Amendment. n100 The Thurman case made the police behavior infamous through the wide coverage it received from both the mainstream press and academic journals. n101

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n89. Thurman v. City of Torrington, 595 F. Supp. 1521 (D. Conn. 1984).

n90. In Thurman, the federal district court declined to dismiss Tracey Thurman's complaint. When the case went to trial, the jury ".... didn't buy the argument that Tracey, who was partially paralyzed when her neck was broken in the ordeal, should never have opened the door. She was awarded $ 2.3 million in damages, a sum that was later reduced to $ 1.8 million." Real Protection for Abused Women, Patricia G. Barnes, editorial, Pittsburgh Post-Gazette, Sept. 21, 1997, at B1; see also, Eileen McNamara, Batterers Win Another Round, Boston Globe, May 13, 1998 at B1 (reporting the settlement to Tracey Thurman paid by the Torrington, Connecticut police department).

n91. See Thurman, 595 F. Supp. at 1524-26.

n92. See Thurman, 595 F. Supp. at 1525-26.

n93. See Thurman, 595 F. Supp. at 1525.

n94. See Thurman, 595 F. Supp. at 1526.

n95. See Thurman, 595 F. Supp. at 1526.

n96. See Thurman, 595 F. Supp. at 1526.

n97. See Thurman, 595 F. Supp. at 1526.

n98. See Thurman, 595 F. Supp. at 1526.

n99. See Thurman, 595 F. Supp. at 1526.

n100. See Thurman, 595 F. Supp. at 1526.

n101. See Zorza, Criminal Law, supra note 63, at 60.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

For all these reasons - police liability, the results of domestic violence studies, and the women's movement - both "should arrest" and "must arrest" statutes exist. Examples of both of these policies are found in the cities of Ypsilanti and Ann Arbor.

III. Results of the Investigation

Ypsilanti has a "should arrest" policy, n102 while Ann Arbor has a "must arrest" policy. n103 The following research regarding the effect of these laws is based upon redacted police reports from Ann Arbor and Ypsilanti police departments from June through December 1996. n104

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n102. See Ypsilanti Police Department, Policy and Procedure #17, supra note 6.

n103. See Ann Arbor, Mich., Code ch. 108, 9:68 (1995).

n104. The police departments struck the names of the offenders and victims, as well as identifying information such as addresses, prior to providing us the redacted reports. Obtaining all of the records from Ypsilanti proved to be a time-consuming and laborious process. First, the small number of reports produced represented less than half the number of cases that SAFEhouse estimated emanated from Ypsilanti. After much pressure and many bureaucratic excuses, what appeared to be a complete number of reports was eventually produced. Nevertheless, nearly a quarter of these reports were not signed by any officer. This seemed highly suspicious. In my experience, no commanding or supervising officer would allow the filing of any police report without reference to its author, usually both on a signature page and in the body of the report. The reasons for this are obvious; how is the prosecutor or the police department to know who to contact to retrieve evidence or testify to events observed (including statements made by witnesses) if the department does not know who made the report? Michigan law requires police departments to make incident reports and mandates that each police department have procedures for preparing a written report, whether or not an arrest is made. See Mich. Comp. Laws Ann. 776.22 3(h)(West 1997). If the Ypsilanti Police Department either never had any such reports, or if the reports had been lost or destroyed, they may have become concerned when I made the FOIA request for reports that should have existed, and thus either created or recreated them to respond to our request.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

These results should be read with caution. First, there is no way to know if records were purposely hidden or if records were not coded as domestic violence because another crime took precedence, so that flaw is an intrinsic part of the study. n105 Second, there are significant differences between Ann Arbor and Ypsilanti arrest policies, which do not appear to be based on the difference in arrest standards. As I will discuss later, I believe another factor is creating the differences between the cities. Third, small sample size affects the results. In the studied time period, there were only 157 domestic violence calls, meaning police officers arrived at the scene, and out of those, only nineteen women were arrested. The data reveal that in slightly over twelve percent of the calls a woman was arrested. The breakdown is 9.5% in Ann Arbor and 15.9% in Ypsilanti. n106 While it appears that race does not seem to be a significant predictor for whether a man, a woman, or either is arrested, there is one exception. Men are less likely to be arrested if they are in a mixed race relationship, but more likely to be convicted once they are arrested. n107 The major statistically significant result is that if the police find out about a prior history of abuse, n108 they are less likely to arrest a woman; however, if they have actually been called to that residence before, they are more likely to arrest a woman.

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n105. To analyze the data, I used a statistical program called SPSS. The variables were all assigned names, some more self-explanatory than others, which I have tried to explain throughout the paper.

n106. All statistics and other information in connection with this study are on file with the author.

n107. The identity of the arresting officer appears not to have been a factor because no pattern emerged based on that variable. (The maximum number of arrests for any primary or secondary officer is five. Although some officers served as both primary and secondary officers at some point, the maximum number of arrests by anyone was eight, which seems reasonable in a six-month period.)

n108. "Prior history" or "Prior history of abuse" means self-reported history as opposed to prior arrests or prior officer contact with the same couple. See also infra Part III.A.1.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

A. Statistical Summaries-Arrest Patterns

What follows are charts which summarize the data and then an analysis of that data. The first charts examine who is getting arrested. There are some preliminary descriptive tables summarizing the data. Following those tables, I begin to break down the data into more meaningful statistics.

[SEE CHART IN ORIGINAL]

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n109. Later abbreviated as Ypsi for purposes of space.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

[SEE CHARTS IN ORIGINAL]

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n110. Describes the result in a dual arrest situation.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

[SEE CHART IN ORIGINAL]

[SEE CHARTS IN ORIGINAL]

n111

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n111. "Offenders" designates those people so classified by the responding officers.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

[SEE CHARTS IN ORIGINAL]

[SEE CHARTS IN ORIGINAL]
[SEE CHARTS IN ORIGINAL]

[SEE CHART IN ORIGINAL]

The next analyses done were two-by-two using a test called a chi-square. This test yields a number called Fisher's Exact Test, which tells if the distribution in the two-by-two matrix is due to something other than chance. If the number is less than 0.05, then the distribution is not dependent on chance. However, exactly which variable is affecting the result and how cannot be determined without a subsequent test, the logistic regression.

1. Prior Self-Reported History of Abuse

An analysis of arrests and self-reported history n112 of abuse shows that prior history affects the chances of male arrest, but not female arrest. The fact that there is no significantly cognizable difference in the effect of self-reported history for women might be explained by the number of women in the sample. Combining the women arrested in Ann Arbor and Ypsilanti gives a number closer to 0.05. n113 In other words, self-reported history of domestic violence may have the same effect on arrest rates for women, but it is hard to tell here without a larger sample size.

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n112. Self-reported history refers to notations in the police reports recording comments by the victim or offender.

n113. But see chart on page 292, showing that history is significant.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

On the other hand, one could argue that men are arrested if there is a previous self-reported history of domestic violence because that shows the police are doing their job. Whereas, when women are arrested, the police are not really arresting abusers, but rather arresting the women on the basis of some other motivation.

[SEE CHARTS IN ORIGINAL]

2. Prior Phone Calls

Overall, across all the data, when a woman is arrested, it does not matter if there have been calls to the police from that residence before. n114

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n114. But see chart on page 292, showing that calls are significant.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

[SEE CHART IN ORIGINAL]

However, when you separate the data for Ann Arbor and Ypsilanti, the results change. In Ypsilanti, the effect of prior calls is not significant, similar to the overall result, while in Ann Arbor, prior phone calls to the police becomes a significant factor. Possibly, there is either conscious or unconscious retaliation by the Ann Arbor police against women for staying in an abuse situation. n115 It is interesting that there are so few prior calls in Ypsilanti. Either people are not calling the police, or once the police are called, they rarely get repeat calls; it is unclear whether this is because the problem is solved, or calling the police does more harm than good.

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n115. For example, Detroit Police Commander James Bannon, addressing the American Bar Association in 1975, stated that the police screened calls and only responded to "serious" injuries, and if there were only minor injuries to the woman when they arrived, the police would be slower to respond the next time. See Zorza, Criminal Law, supra note 63, at 48.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

[SEE CHARTS IN ORIGINAL]

Looking at the cases where only a man is arrested (i.e. excluding dual arrests), whether police have been called before is not significant.

[SEE CHART IN ORIGINAL]

It seems that if a woman says she has been hit before, the police are more likely to arrest the man, but if she has actually called the police before, they are not. Again, when there are prior calls to the police, the police appear less likely to believe the woman.

3. Compared by City

The statistics also show that whether the police have been called before significantly depends on the city.

[SEE CHART IN ORIGINAL]

The difference between cities may be in the nature of the abusive relationships (i.e., Ypsilanti the abuse is usually first time), the likelihood of people in that city to call the police, or in record-keeping procedures. n116

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n116. The chi-square test does not tell us which way the variables affect each other.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

When women are arrested or only men are arrested, there is a significant correlation between prior calls and arrest.

[SEE CHARTS IN ORIGINAL]

Whether or not history is reported when women are arrested depends on the city. This may be because of better questioning techniques in Ann Arbor, or because there really is more prior abuse in Ann Arbor.

[SEE CHART IN ORIGINAL]

The result still holds when men and women are combined.

[SEE CHART IN ORIGINAL]

There is also a correlation between whether or not an abuse history is reported and the offender's gender. This chart includes all calls, not just those where someone was arrested. n117

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n117. However, when the cities are separated, the results are not significant. This may be due to a sample size issue, not to an effect in the cities themselves.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

History by Gender for Offenders

[SEE CHART IN ORIGINAL]

4. Mixed Race Couples

Examining the effects of race in more detail, we find that a woman's arrest is not significantly predicted by the fact that the woman is in a mixed race couple.

[SEE CHART IN ORIGINAL]

However, being in a mixed race couple is a significant predictor of whether a man is arrested. The difference from the result for women may be a sample size problem (only three of the arrested women were in a mixed race couple), or it may be showing the intersection of racism and sexism.

[SEE CHART IN ORIGINAL]

Further, the arrests of men were not significantly affected by their own race alone.

[SEE CHARTS IN ORIGINAL]

The city is likewise not a significant predictor of whether a man will be arrested.

[SEE CHART IN ORIGINAL]

5. Harms

Following are breakdowns of the harm caused compared to arrest patterns. I have excluded harm levels not represented in the cases. The first table shows the results for women arrested; the only harm level that is significant for predicting arrest of women is hitting with an object, i.e., lamp, chair, glass, etc. The second shows the corresponding results for men; the category of slapping, hitting with fists, kicking, etc. is the only significant predictor for arrests of men.

[SEE CHART IN ORIGINAL]

[SEE CHART IN ORIGINAL]

These results seem counter-intuitive, but I cannot say with certainty that certain harm levels were significant predictors while others were not, since each group (harm level) contains so few cases.

B. Logistic Regression - Arrest Patterns

The last part of the analysis involved several logistic regressions. In running the logistic regressions, only variables that the previous chi-square tests indicated were significant were included. The chi-square tests give an idea of what variables should be included in the logistic regression, and the logistic regression does a more sophisticated analysis, determining how the factor affects arrests.

1. Women Arrested Overall

The first test I ran examined the factors predicting whether a woman would be arrested. Below are the tables the computer returned.

[SEE CHART IN ORIGINAL]

These results demonstrate that prior calls and prior history are the variables that matter in explaining why a woman is arrested. Prior calls is negatively associated. Thus, if someone has called the police before, the woman is more likely to get arrested. Prior history is positively associated, meaning if the police find out about a prior abuse history, they are less likely to arrest the woman. n120 These results are interesting and can perhaps be attributed to the way police in the two cities are trained, whether the prosecutors pursue the arrests, or a combination of both.

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n118. "B" is the coefficient that goes in the equation.

n119. "Sig" tells the significance. Again, the variable is significant, statistically speaking, if the value is less than 0.05. All significant variables are italicized.

n120. When those equations are run separately by city, in Ann Arbor, harm caused is the only significant predictor (sig.=.0109), and it is positively correlated. No other variables matter at all.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

[SEE CHART IN ORIGINAL]

So, overall, we know that prior history, mixed race, and the city predict why men are arrested for domestic violence. The correlations tell us that when prior abuse history is reported, a man is more likely to be arrested; when the couple is of different races, the man is less likely to be arrested; and when the city is Ypsilanti, a man is more likely to be arrested. n121

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n121. Once again, when broken down by city, the results change. The only significant predictor in Ann Arbor is prior history (sig.=.0863); no other variables matter.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

3. Final Disposition

I also ran tests for final disposition results. In these tests, I excluded the dual arrest cases to simplify the analysis. I ran six groups of chi-square tests. First, I looked at either whether the case was dropped or not, and then whether the case resulted in some sort of guilty verdict or not. In other words, cases where there was some follow-up, but not a guilty verdict, or the information was not recorded were characterized as not dropped and also not guilty. Within both of those two major groups, I ran tests of cases overall, cases in Ann Arbor, and cases in Ypsilanti.

The factors that showed a significant correlation with a case not being dropped were prior history for male and female arrestees in Ann Arbor and calls for male and female arrestees in Ann Arbor. For guilty verdicts, the significant factors were for both cities combined, prior history for male and female arrestees, prior history for male arrestees, prior calls for male and female arrestees, and prior calls for female arrestees; and for Ann Arbor, prior calls for female arrestees.

[SEE CHARTS IN ORIGINAL]

[SEE CHARTS IN ORIGINAL]

[SEE CHARTS IN ORIGINAL]

[SEE CHARTS IN ORIGINAL]

Thus, overall, for both men and women, the only thing that predicts whether a case will not be dropped is the city. In other words, a case is more likely to be dropped in Ann Arbor.

Conclusion

That day I saw what appeared to me to be a disproportionate number of women in misdemeanor court charged with domestic violence turned out to be less of a pattern than I originally surmised, although the overall rate of women being arrested, twelve percent, is much higher than one would expect, given who is harmed most by domestic violence. It is also interesting that the items of statistical significance relate to location of prior calls and to race more than to gender.

The Ann Arbor Police Department is more likely to drop cases with its "must arrest" policy, while cases tried in Ann Arbor have a statistically significant correlation with acquittals. It may be that the police exercise discretion better in Ypsilanti, or it may be that there is more post-arrest prosecutorial discretion in Ann Arbor.

Many saw these "must arrest" and "should arrest" statutes as a major answer to the problem of domestic violence, but clearly they are not; domestic violence is still endemic in society. Between twenty-two and thirty-five percent of all women who seek treatment in hospital emergency rooms are there because of abuse by a man with whom they either are, or were, intimately involved. Forty percent of the injuries come from deliberate assaults by partners, and nineteen percent of those women have a previous history of abuse-caused injury. n122

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n122. See supra notes 38&ndash;39 and accompanying text.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

Nonetheless, the high incidence of women arrested, as well as the negative response of the police to women when they have previously called the police, suggests some need for reevaluation of policy. My suggestions are twofold. First, the "should arrest" policy is the better policy. If the police feel that they are not trusted to determine when an arrest should be made, they will inevitably become resentful and may end up "retaliating" by arresting a battered woman who returns to her batterer and once again needs help.

Second, I propose that after there has been an arrest, but before there is a prosecution, that the prosecuting authority set up a review team to look at each case to discern whether the "right" person has been arrested, n123 to determine if there is a history of violence in this relationship, and to monitor what the police are doing. In this way, the prosecutorial and law enforcement agencies can learn from each other and can see to it that those who should be prosecuted are prosecuted.

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n123. For example, it is not unusual for a woman to tell the police the truth when asked if she hit her husband by saying "yes," even if she hit him trying to defend herself, and for her husband to lie and say he did not hit her. Under these circumstances, the police might actually arrest the "wrong" person.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -