Kellermann AL and Mercy JA. "Men, Women, and Murder: Gender-specific Differences in Rates of Fatal Violence and Victimization." J Trauma. 1992; 33:1-5.
methodological and conceptual errors:
- most women kill in defense of themselves and their children. In these common circumstances, lawful self-defense by women against their attackers is not "murder" in any jurisdiction
- the authors' discussion focused almost entirely on guns though the data on knives and other weapons are virtually identical
- the authors failed to note that during the study period the domestic homicide rate nearly halved
- provided no primary research, instead provides largely faulty analysis of FBI Uniform Crime Reports data
- though purporting to assess an aspect of risk, the authors failed to analyze the protective uses of guns -- lives saved, injuries prevented, medical costs saved, and property protected -- no true risk-benefit analysis
- ignored data that suggest guns are actually the safest and most efficacious means of resisting assault, rape, and even non-violent crime
- offered no new insights or solutions to the problem of domestic abuse
Though recognizing the risk and physical disadvantage of women, Kellermann and Mercy attempted to draw us to their conclusion that "...the wisdom of promoting firearms to women for self-protection should be seriously questioned."  No effort was made by the authors to assess the protective uses of guns by women, in fact, the authors attempted to portray legitimate self-defense as "murder." Women are abused 2 million to 4 million times per year.  Their children are similarly abused, even fatally.  Almost all the "spouses and domestic partners" killed by women each year are the very same men, well known to the police, often with substance abuse histories, who have been brutalizing their wives, girlfriends and children.   Defense with a gun results in fewer injuries to the defender (17.4%) than resisting with less powerful means (knives, 40.3%; other weapon, 22%; physical force, 50.8%; evasion, 34.9%; etc.) and in fewer injuries than not resisting at all (24.7%).  Guns are the safest and most effective means of protection. This is particularly important to women, children, the elderly, the handicapped, the weak, and the infirm, those who are most vulnerable to vicious male predators. [See Graph 12: "Rates of Crime Completion by Victim's Method of Protection" & Graph 13: "Rates of Victim Injury by Victim's Method of Protection"]
Would it be more "politically correct" if women or children were killed by their attackers -- the common outcome when women do not defend themselves and their children with guns? Should women, children, the elderly, the physically challenged, or anyone rely on riskier or less effective means of self-protection? Or... should innocent victims defend themselves with the safest and most effective means of defense until such time as prevention strategies become significantly more effective?
The article's title notwithstanding, lawful self-defense is not "murder" in any jurisdiction. It has been estimated that as many as 20% of homicides are self-defense or justifiable in the final analysis.  Since the FBI Uniform Crime Reports records "justifiable homicide" based on the preliminary determination of the reporting officer, rather than upon the final determination, the FBI data dramatically under-reports "justifiable homicide." Knowing one another is sufficient to meet the FBI's definition of "acquaintance," so "acquaintance" includes the maniac in one's apartment building and dueling drug dealers, hardly the type of good people most would call "friends." These are predators that Handgun Control Inc. considers "friends and family."
At unconscionable expense this article recapitulated FBI Uniform Crime Reports data that was already available "off the shelf" for $20 from the US Government Printing Office. The data only bolster what we already knew about women's risk at home, but Kellermann and Mercy -- unjustified by the data -- singled out guns for special treatment. "When women killed with a gun, their victim was five times more likely to be their spouse..."  Kellermann and Mercy failed to acknowledge, however, that the FBI data they recounted showed that when women killed with a knife, their victim was also five times more likely to be their spouse -- and when women killed with other means, their victim was over four times more likely to be their spouse.
The most meaningful conclusion from this study, the conclusion missed by Kellermann and Mercy, is the tremendous restraint shown by women, that they kill so few of their contemptible abusers. Interestingly, during the study period of this article, 1976-87, the domestic homicide rate fell from 2.4 to 1.4 per 100,000   and the number of teen and child gun accident fatalities fell from 530 to 280  -- all this while increasing numbers of guns were in the hands of US citizens. It is also worth noting that the highly touted "proliferation of guns" has not been associated with an increase in rates of gun ownership. 
The male authors' patronizing suggestions about gun ownership by women are not justified by available data. Partisan "scientists" who struggle to sculpt their data to fit their a priori conclusions should be ignored or censured. Statistical legerdemain cannot hide what the authors failed to recognize: a woman's or child's life lost because a gun was absent is at least as valuable as a violent predator's life lost because a gun was present. Women are justified in concluding that guns are the most effective and safest tools of self-defense. Catchy ratios and contrived comparisons detract from the public debate and are little consolation to the brutalized victims or their grieving survivors.