Abused statistics: domestic violence; like hydra heads or spreading kudzu, the false statistics keep proliferating.

IT DID not take long for advocacy groups and some commentators to claim that the O. J. Simpson case could do for domestic violence what Anita Hill did for sexual harassment. If the Anita Hill analogy refers to gender politics eclipsing truth, common sense, and journalistic skepticism, then that is exactly what's happening.

We are barraged with horrendous figures. CNN's Sheryl Potts said at the end of a 3.5-minute segment on wife-abuse, While you were watching this, 13 women were severely beaten by someone who claims to love them' (i.e., one every 15 seconds). According to the Associated Press, "4 million to 6 million women are beaten [each year]. That means once every 5 seconds ... a woman is punched or kicked ... or held down and pummeled." On Crossfire, Miami radio talk-show host Pat Stevens upped the count to 60 million by reasoning that "there are 6 million reported cases," and "the FBI estimates" that only 1 in 10 is reported. This went unchallenged.

One AP report ran somewhat counter to the general tone, noting that murders of women by male partners had dropped 18 per cent since the late 1970s; but it went on to temper this unduly optimistic message with the assertion that non-fatal violence was up: "From 1980 to 1990, federal figures show, reports of domestic assault ... rose from 2 million to 4 million, according to Rita Smith, coordinator for the National Coalition Against Domestic Violence."

Where do the numbers come from? One source for the 4-to-6-million figure is a 1993 Commonwealth Fund study, which included such acts as shoving, slapping, and throwing things in its definition of battering. The statistic of a woman beaten every 15 seconds is derived from the 1985 National Family Violence Survey by University of New Hampshire researchers Murray Straus and Richard Gelles, which estimated that about 1.8 million American women each year suffer at least one incident of "severe violence" by a partner--a punch, a kick, an assault with an object. But only 7 per cent of the victims required medical care. A study published in Archives of Internal Medicine in 1992 found that, based on reports by wives in marital therapy, 48 per cent of "severe marital aggression" by husbands caused no injury, and 31 per cent caused only a "superficial bruise." While these are still reprehensible acts, most people visualize something very different when they think of "severe violence."

And the reports of domestic assault rising from 2 to 4 million a year over a decade? The figure, said Rita Smith, came from the Justice Department's Bureau of Justice Statistics. Were these assaults reported to police? "Either that, or to medical personnel," said Miss Smith. "I'm not exactly sure how they gathered it, but it was one of the statistics they put out in some publication." What publication? She didn't recall.

An information specialist at the BJS knew nothing of such figures. Most of the Bureau's publications are based on the National Crime Victimization Survey, which puts the number of female victims of assaults by partners at about 470,000 a year.

Another shocking claim was made in a MacNeil/Lehrer NewsHour segment by Connie McCall of Rainbow Services, a battered-women's assistance group in Los Angeles: "Over 50 per cent of women admitted to emergency rooms are admitted for an injury caused by a partner." In a pre--O. J. Simpson speech last March, Health and Human Services Secretary Donna Shalala gave a more modest estimate: "In our hospital emergency rooms, some 20 to 30 per cent of women arrive because of physical abuse by their partner."

These numbers (whose implausibility ought to be evident to anyone who has ever been to a hospital) come in part from studies in which medical charts were reviewed and most unexplained or inadequately explained injuries to women were reclassified as related to abuse, and in part from a 1984 study based on questionnaires distributed to emergency-room patients in a Detroit hospital. About 25 per cent of the female patients answered "yes" to the statement, "At some time my boyfriend/husband or girlfriend/wife has pushed me around, hit me, locked me, or hurt me"; so did about 20 per cent of the men. Most of the abuse did not seem to be directly related to the emergency-room visit. Moreover, a high percentage of the subjects were poor, unemployed, and cohabiting without marriage-factors strongly associated with the risk of domestic violence.

The same fallacy of projecting data obtained by studying the urban poor to the population at large is responsible for another widely reported figure--cited, for instance, in Newsweek's July 4 cover story on battered women: "[I]n 1992 the U.S. Surgeon General ranked abuse by husbands and partners as the leading cause of injuries to women aged 15 to 44." In fact, in a 1992 column in the Journal of the American Medical Association on the medical response to domestic abuse, then-Surgeon General Antonia Novello mentioned that "One study found violence to be ... the leading cause of injuries to women ages 15 through 44." This refers to all violence, not just violence by male partners. The article in which the finding was reported, titled "A Population-based Study of Injuries in Inner City Women," examines emergency-room visits by women "in a poor, urban, black community in western Philadelphia."

And how many women are killed by their current or former mates? CNN has cited the figure of 4,000 a year--even though the FBI statistic is 1,300 to 1,400. The higher number comes from activist groups and includes, Miss Smith told me, "a lot of suspicious deaths that aren't [officially] classified as a homicide." The higher the better: on This Week with David Brinkley, Lynne Gold-Bikin, incoming chairman of the American Bar Association's family-law section, claimed that "some 25 per cent of women who are abused ultimately get killed by their husbands." If, as Newsweek says, "1 woman in 4 will be physically assaulted in her lifetime," this would put a woman's risk of being murdered by a partner at 1 in 16. In fact, it's closer to 1 in 1,000.

A widely cited statistic--which appeared in Sheryl Potts's CNN report--is that "women who leave their batterers are at a 75 per cent greater risk of being killed by the batterer than those who stay." This information was attributed to the NCADV; a brochure issued by the organization cites "Barbara Hart, 1988," as the source. When I asked Miss Smith about this, she said a bit hesitantly, "Barbara Hart is an attorney with the Pennsylvania Coalition Against Domestic Violence--and I have recently talked to her and she said, |I didn't say that.' So, the figure--whether or not it's 75 per cent, I don't know. [But] most of the women who are murdered are either separated or in the process of leaving when they die." She assumed that all these women had been battered in the marriage. However, a number of experts such as criminologist Lawrence Sherman, president of the Crime Control Institute and author of Policing Domestic Violence (1992), dispute the "escalation" theory. Sherman's study in Milwaukee over nearly two years revealed that in 33 domestic murders, 32 couples had no prior police record of battery--and there was 1 homicide among the roughly 6,500 couples with such records.

Epidemic Proportions?

IS THERE an epidemic of domestic violence in America? One could say that we have an epidemic of violence in general, and for all the talk about doctors and executives who beat their wives, domestic violence tends to affect primarily the same segments of society as street crime. Men and, to a lesser extent, women guilty of violence in the home are also much more likely than other people to commit violent acts outside the family. Yet the view that domestic violence is largely the doing of a minority of thugs and sociopaths is anathema to the highly politicized battered-women's advocacy movement. An NCADV pamphlet makes this clear: "A feminist-based analysis of battering [shifts] from a focus on the |pathology' of individuals or families to the particular social policies, norms, and practices which tolerate woman-abuse."

The theme of supposed tolerance for woman-battering in our culture is reflected in the infamous tale of the "rule of thumb," repeated by Cokie Roberts on This Week with David Brinkley: "The rule of thumb, the expression we use ... that was the size of the stick that was acceptable to beat your wife with under common law." Yet, as the scholar Christina Hoff Sommers shows in Who Stole Feminism? this origin of the phrase "we of thumb" is a myth. While two American courts in the nineteenth century alluded to such an "ancient law" they never invoked it as a standard. The first legal code enacted by the Massachusetts colonists in 1652 prohibited wife-beating.

Despite attempts to find the roots of wife-battering in the power of physical chastisement once given to husbands by law, few serious researchers believe that domestic violence is primarily a way to maintain patriarchal control. Even feminist scholar Elizabeth Schneider of Brooklyn Law School concedes, in a 1992 New York University Law Review article, that the "traditional model" of battering as a means of maintaining male dominance "is thrown into question when the violent partner is a woman, or the victim is a man." Though Professor Schneider is referring to same-sex couples, a wealth of research (notably the Straus-Gelles studies) has found that wives and girlfriends are the aggressors at least half the time, though they are five to six times less likely to cause physical damage. (On one MacNeil/Lehrer report, an L.A. police officer said that he had been on many calls where the woman was the assailant; but the panel discussion following the segment focused solely on male batterers and female victims.

For every two women killed by a male partner, there is at least one man killed by a wife or girlfriend--not including cases dismissed as self-defense. It is interesting to recall that in 1989, when San Diego housewife Betty Broderick fatally shot her ex-husband and his new wife after harassing them for months, the incident was not seen as emblematic of the epidemic of domestic violence by women. Indeed, Mrs. Broderick got a great deal of sympathy as a wronged wife (she claimed that Dan Broderick's alimony payments of only $16,000 a month amounted to a white-collar version of battery), and her first trial ended with a hung jury. Male violence toward women fits into the fashionable larger theme of oppression; female violence toward men does not.

With remarkable lack of compunction about exploiting a horrible tragedy for political gain, members of the National Organization for Women, with Patricia Schroeder and other congresswomen on hand, rallied on Capitol Hill to urge passage of the Violence Against Women Act--which, among other things, would redefine wife-beating as a gender-based hate crime and mandate sensitivity training for judges. (In Massachusetts, a similar media frenzy has already led to a situation in which, former Massachusetts Bar Association President Elaine Epstein wrote last year, "it has become essentially impossible to effectively represent a man against whom any allegation of domestic violence has been made.")

Beyond these immediate political goals, the advocates' real agenda is to promote their view that American women are routinely terrorized by men who are intent on keeping them subjugated. Why not just say that 5 out of 4 women are battered by men, and be done with it? Given their track records, the media would probably buy that one, too.