Two weeks ago, the one-minute slot on ABC’s radio news broadcasts — the one that fills in at the hour and half hour of many talk and music stations — included a 30-second story highlighting a medical study that said that 20 percent of pregnant women who died were murdered. When I looked up the story on the Web, I found that it was fully four months old. The original, a long thumbsucker by Bryan Robinson, appeared on the ABC News website April 25, under the headline, “Expectant Victims: Pregnancy Doesn’t Protect Women From Abuse — or Murder.”
What accounted for its recycling? Headlines, of course. Two prominent murders of pregnant women — Peterson, Hacking — provided the hook. The obvious talking points, helpfully highlighted by subheads in the original article, fit the template of post-modern feminist media coverage: “The Need for Control,” “Maybe a More Serious Problem.” That is, men are beasts. Super Bowl wife abuse, anyone? As well, it allowed ample quote space from favored sources, like the National Domestic Violence Hotline, The Sexual Homicide Exchange, and anxious-to-talk academic cud-chewers. “Experts and women’s advocates,” as the story said.
In fact, the study cited by ABC dates back even further. “According to a 2001 study published by the Journal of the American Medical Association, approximately 20 percent of Maryland women who died during pregnancy were murdered. This supported the findings of previous studies in Cook County, Illinois, and New York.” Those “previous studies,” as the ABC article said, “…have found that homicide is the leading cause of death among pregnant women outside of medical complications.”
THAT’S QUITE AN ASSERTION. Homicide the leading cause of death outside of medical complications among pregnant women? Only problem: A fairly easy bit of statistical research shows it’s almost certainly not really provable, or supportable, to say so. “Not true” would be too easy a thing to say, and not provably correct, either. The women are dead; they were murdered; they were pregnant; they do represent a given percentage of a given population. What it means, however, appears to be entirely different, given a little investigation.
First, do peak child-bearing years coincide with peak murder-victim years? Simple answer, yes. According to Murder Victims.com, 44 percent of murder victims in the United States in 1997 were between 25 and 34 years of age. And according to the Center for Disease Control’s Disaster Center statistics, about 9 percent of all people 25-44 who die, die from “homicide and legal intervention” (9,261 per 100,000).
So that 9 percent figure, unfortunately, gives us a norm. The 20 percent of deaths of pregnant women attributed to murder in the JAMA-published study of Maryland would indeed appear to be unusual.
But is it unusual for Maryland or for the Chicago area? (I exclude the named “New York” mentioned by ABC because the story does not specify state, city, or metro area.) Unfortunately again, no. Baltimore and Chicago have violent crime rates much higher than the national average, according to 2001 FBI uniform crime statistics compiled by Sperling’s Best Places. The nationwide average per 100,000 for murder, robbery and assault is 506.1; in Baltimore, it’s 2,469.8; in Chicago, 1,630.6.
As well, Maryland’s population is, according to the Maryland state website, 36 percent black. Cook County, Illinois’s black population, according to The Nationmaster online encyclopedia, is 26 percent black. Both those figures, reckoned as of 2002, come in significantly above the national average of about 12 percent for African-Americans in the U.S.
Again, according to Murder Victims.com, the largest single ethnic group of murder victims in the country in 1997 was black — 49 percent, a figure four times greater than African-American representation in the population as a whole.
The maternal mortality rate for black women follows the same pattern. In Maryland, according to the Center for Disease Control’s Morbidity and Mortality Statistics, from 1987 to 1996, the mortality rate for white women in the state was 6.1 per 100,000; for black women, 15.9.
Actually, the largest single grouping of murder victims was men, at 77 percent. So we must exercise a considerable degree of care before deciding that the JAMA study ought to get us alarmed about race and murder or about pregnancy and murder — or indeed about anything. Because the 20 percent figure comes at the very narrow end of a long reversed statistical telescope. Not many women get murdered, comparatively; not many women, comparatively, die at all during child-bearing years.
SO HOW MANY PREGNANT WOMEN were murdered in a given year in Maryland? The State itself says 77 women died of “pregnancy-related causes” in the years 2000-2001. Causes of death included (in order of frequency), cardiovascular disease, homicide, suicide, substance abuse, and accidents. Finding out how many pregnant women died in a given jurisdiction — of all causes, certainly a higher number — is just about impossible, because almost nobody studies such a thing. Note that Maryland itself listed murder as the second most common cause of “pregnancy-related” deaths.
Does that seem like very few murders of pregnant women? Yes, it does, but it is supported by broader statistics. Maryland had 430 murders in the year 2000, according to the Center for Disease Control’s Disaster Center stats. Twenty-three percent of that figure (representing the proportion of women murdered nationwide for the year 2000) is 98.9. Fewer than 100 women were (likely) murdered in Maryland in the year most likely to be the subject of the 2001 JAMA article.
You can ring all these stats fully round with caveats. But there is enough here, from multiple sources, to call the JAMA-ABC study seriously into question, on the basis of tiny sample size alone, not even including other powerful demographic factors. And the countering statistics all support one another.
These days, the news media publish “studies” almost every week which purport to prove something or other. And we have dozens, if not hundreds, of organizations, anything but disinterested, that will provide such “studies” to lazy reporters, who will simply recycle them without looking at them.
If they fit the template.