Blog By Barry

January 27, 2009

Response to Christina Hoff Sommers, part 3: Truths and Lies

In a speech, self-described “conservative feminist” Christina Hoff Sommers said:

Let me turn to my second major objection to contemporary feminism: its reckless disregard for the truth. In doing research for my books, I looked carefully at some standard feminist claims about women and violence, depression, eating disorders, pay equity and education. What I found is that most –- not all —- but most of the victim statistics are, at best, misleading –- at worst, completely inaccurate. […]

I partly agree with Sommers: Too many feminists — including those we rely on to get facts right (such as academics and published writers) — have been careless about fact-checking their claims. Critiquing a textbook on domestic violence, Sommers writes:

Zorza also informs readers that “Between 20 and 35 percent of women seeking medical care in emergency room in America are there because of domestic violence.” This claim is ubiquitous in the feminist canon. But is it false. There are two legitimate studies on emergency room admissions: one by the Bureau of Justice Statistics and another by the Centers for Disease Control. The results of both indicate that domestic violence is a serious problem, but that it is far down on the list of reasons women go to emergency rooms. Approximately one half of one percent of women in emergency rooms are there seeking treatment for injuries from domestic violence.

Sommers cites a second recent textbook, The Penguin Atlas Of Women In The World, which repeats the same error. And she’s right — it is an error. (Although, as I’ll show in a future post, Sommers’ counter-claims are just as false.)

I think this is the strongest of Sommers’ claims. Sommers is right to say that “false assertions, hyperbole and crying wolf undermine the credibility and effectiveness of feminism in general.” And many errors could easily be avoided if authors just checked primary sources — something that professional writers and academics should do routinely.

Within feminism, there’s sometimes too little skepticism regarding statistics and news stories which emphasize harms against women. We’ve created a culture which does a rotten job of self-correction.

But although she has a point, Sommers is still substantially wrong, for two reasons. First, Sommers conflates unambiguous errors of fact — which will inevitably happen sometimes, especially in a movement the size of modern-day feminism — with well-reasoned disagreements. And secondly, Sommers’ own work is full of errors, and at times actually deceptive.

In her lecture, Sommers writes:

Some of you are probably thinking –- the literature on feminism is vast and complex –- there are bound to be some mistakes. So what? But I and other investigators have not found “some mistakes.” What we have found is a large body of blatantly false information. The Domestic Violence Law textbook and the Penguin Atlas of Women in the World are not the exception. They are the rule.

So here’s Sommers’ argument:

1) Feminist writers sometimes repeat “blatantly false information.”

2) This errors are the rule, not the exception. This is documented in the works of Christina Hoff Sommers and “other investigators.”

3) Therefore, feminism, as a rule, consists of “a large body of blatantly false information.”

The trick here is in point 2. Sommers wants us to believe that her critiques of feminism, as well as those by “other investigators,” are filled with examples of feminists making unambiguous factual errors. But that’s not true. In Sommer’s book Who Stole Feminism?, Sommers does catch feminists making some unambiguous errors, but most of the book is taken up by subjective political disagreements, not by fact-checking.

In order to accept that Sommers’ work demonstrates that a “reckless disregard for truth” is the “rule,” “not the exception,” we’d have to accept that anytime a feminist disagrees with Christina Hoff Sommers — because such disagreements take up most of Sommers’ work — that is evidence of a reckless disregard for the truth. But of course, it’s no such thing.

So what do I mean when I say that most of work consists of subjective political disagreements? By “subjective political disagreements,” I mean issues that reasonable, honest people, basing their opinion on well-founded evidence, can disagree with Christina Hoff Sommers on.

I will focus on one example: the rape prevalence research of Mary Koss. Koss’ research is probably the single example that “conservative feminists” have used most often to “prove” feminist dishonesty, ((Think I’m exaggerating? Here is an incomplete list of books which rehash the “conservative feminist” arguments against Koss’ research: The Morning After by Katie Roiphe; The Politically Incorrect Guide to Women, Sex and Feminism by Carrie Lukas; Dead End Feminism by Elisabeth Badinter; Lip Service by Kate Fillion; Tax-funded Politics by James T. Bennett; A Nation of Victims by Charles J. Sykes; Moral Panic: Biopolitics Rising by John Feteke; The New Victorians: A Young Woman’s Challenge to the Old Feminist Order‎ by Rene Denfeld; The Myth of Male Power by Warren Farrell; Does Feminism Discriminate Against Men? by Warren Farrell, Steven Svoboda, & James P. Sterba. It’s likely there are additional books I’m unaware of, not to mention dozens of articles and hundreds of website.)) starting in the early 1990s in books like Sommers’ own Who Stole Feminism?, and continuing to this day (Heather MacDonald published an attack on Koss’ research just last year). According to the Independent Woman’s Forum, ((A Sommers-influenced “conservative feminist” think tank.)) Koss’ research is the “number one feminist myth” in America.

So what was Koss’ rape research? In the 1980s, Koss pioneered a new approach to surveying populations about their past experiences with rape. Where previous surveys measured rape prevalence by asking respondents a single, sometimes hilariously vague question (“Has anybody ever attacked you in any other way?”), Koss asked a series of comparatively specific questions (“Have you had sexual intercourse when you didn’t want to because a man threatened or used some degree of a physical force (twisting your arm, holding you down, etc.) to make you?”) about respondents’ experiences.

Koss’ study of “hidden rape” proved three important facts, which feminists and criminologists had long suspected: that rape happened much more frequently than official numbers indicated; that most rapes aren’t committed by strangers; and that most rapes are never reported to police or other authorities.

Koss’ study, in the decades since, has led two parallel lives. In one life — a life lived in books funded by right-wing foundations, anti-feminist websites, and the like — Koss’ work is an enduring symbol of feminist dishonesty and deception, and is considered a discredited joke, trotted out for rehashed debunkings every couple of years.

In another life, however — a life lived among academic experts — Koss’ work has been amazingly successful. Decades later, her work is respectfully cited in peer-reviewed studies — a few years ago I found that just two of Koss’ articles had been cited over six hundred times. ((In Who Stole Feminism, Sommers claims that Koss’s work is frequently cited by activists but “not so much by established scholars in the field of rape research.” It would in fact be hard to name a scholar of rape prevalence who has been cited more often in the professional literature.))

Although subsequent research has arguably improved on Koss’ 1980s work, her insight — that rape victims are more likely to recount their experiences in response to a series of behaviorally-specific questions — is accepted by virtually all published rape prevalence researchers. And Koss’ central findings (described above) have been replicated in study after study, including two major studies conducted by the Federal government.

By ordinary academic standards, a frequently-cited study which has been replicated multiple times is solid work. That’s not to say that Koss’ study was perfect — no study ever is — but citations plus replication is the gold standard.

Of course, reasonable people can sometimes disagree with professional researchers, and Sommers and other “investigators” are entitled to their opinions. ((To delve into the details of the debate, including detailed responses to the arguments most often brought up by Sommers and other “investigators,” see my past posts about the Koss controversy.)) But Sommers’ position on Koss’ research isn’t that reasonable people can disagree. Instead, she and other “investigators” have repeatedly used Koss’ research as their major example of feminist lying, even though Koss’ results are widely accepted by experts and have been replicated over and over.

This is the central dishonesty of Sommers’ thesis: She claims her work shows that feminists “as a rule” have “reckless disregard for the truth,” but most of her book concerns matters that an honest person could easily disagree with Christina Hoff Sommers about. ((It’s not just rape prevalence research; I could make similar arguments for how Who Stole Feminism? treats topics like domestic violence, education, the wage gap, etc….))

Sommers has to frame all her disagreements with mainstream feminism as feminist lying, because that is the basis of her case against feminism. If she admits that reasonable, honest feminists can disagree with Christina Hoff Sommers, she loses her claim that modern feminism consists of “a large body of blatantly false information… at best, misleading –- at worst, completely inaccurate.”

* * *

Earlier this post, I said that “Sommers’ own work is full of errors, and at times actually deceptive.” In my next post in this series, I’ll back that statement up, using her discussion of emergency room admissions as my example.

This post appears both at “Alas, a Blog” and at “Blog By Barry.” To facilitate intra-feminist dialog, the comments at “Alas” are only open to feminists, while the comments at “Blog By Barry” are open to all.


January 21, 2009

Homicide is not the leading cause of death among pregnant women

Filed under: Rape, intimate violence, & related issues — Ampersand @ 9:04 pm

At the end of an otherwise interesting list of convicted people various feminists would pardon — including Assata Shakur, The Amiraults, and all nonviolent drug users (a suggestion that would save millions of tax dollars) — one feminist wrote:

I would pardon every woman convicted of killing her husband before the self-defense plea was admissible in all 50 states because, after all, it probably was. We live in a country where the biggest risk factor for the death of pregnant women is homicide and the number of women killed by their husbands or partners constitutes 41 percent of all women killed (only 11 percent of men killed are done in by their wives or partners). It’s not a far leap of logic to think that those women were making sure they didn’t become part of that 41 percent statistic.

Virtually all of that is wrong.

I would pardon every woman convicted of killing her husband before the self-defense plea was admissible in all 50 states because, after all, it probably was.

First of all, there has never been a time when pleas of self-defense were inadmissible. My guess is that she means any woman convicted of killing her husband before expert testimony on battering and its effects (what used to be called “Battered Women’s Syndrome”) was admissible in all states.

Second of all, it doesn’t appear that the inclusion of excluded expert testimony on battering often changes the outcome of a trial. To quote from a congressional report:

With respect to the disposition of cases, a review of state court cases found that convictions of battered women were reversed in less than one-third of the cases appealed and that, of those reversals, less than half were due to erroneous exclusion of, limitation of, or failure to present expert testimony on battering and its effects.

These findings suggest that, contrary to popular misconceptions, the introduction of expert testimony on battering and its effects does not equate to acquittal for a battered woman defendant.

Still, I agree that expert testimony on battering should be included in any relevant case, and probably juries and judges aren’t giving it as much weight as they should. So there are certainly some good pardons in there. But let’s face it — there are also women who kill husbands for motives other than self-defense.

We live in a country where the biggest risk factor for the death of pregnant women is homicide…

We really don’t. Pregnant women in the US are about eight times as likely to die of medical causes (such as bleeding during childbirth) than they are of homicide. Car accidents come second, and homicide comes third.

It’s unclear if homicide is any more common among pregnant women than it is among non-pregnant women of a similar age (young women are both more likely to be murdered and more likely to be pregnant than other women). But maybe it is — the reporting system isn’t great, and some scholars say that homicide of pregnant women is badly undercounted. But there’s no way it’s so undercounted that homicide is “the biggest risk factor.”

I’ve seen feminists make this false claim before. It’s too bad, because it obscures the biggest preventable cause of maternal death in the US — which isn’t murder, but inadequate health care. Better prenatal care could save hundreds of women’s lives every year.

…and the number of women killed by their husbands or partners constitutes 41 percent of all women killed (only 11 percent of men killed are done in by their wives or partners).

This is misleading and wrong.

It’s wrong because the real numbers are actually a lot more extreme: only 2.5% of men murdered are victims of intimate homicide, versus about 33% of women murdered.

It’s misleading because a portion of that difference isn’t caused by more women being killed by intimates, but by more men being murdered by strangers. In 2005, 1,181 out of 3,545 women who were murdered, were killed by boyfriends or husbands, while 329 of the 13,122 men who were murdered, were killed by girlfriends or wives. To just report the percentage of intimate homicides, without reporting the difference in the total number of murders, creates a false impression.

Blog at