Press Box

The Muddled Maternal Murder Series

A Washington Post investigation loses its way.

Last week Slate “Today’s Papers” columnist Eric Umansky flinched upon encountering the word “many” in this Washington Post headline:How many is many? he demanded. Surely there’s a better way to express quantity in a headline! In this story, as it turns out, “many” was one-fifth, or 20 percent.

My friend Chuck Shepherd calls such imprecise terms and phrases as “many,” “some,” and “he’s not alone,” journalistic crutches. Reporters rely on them not so much out of abstract laziness as to help them maintain their story bias, he says.

The “many” crutch gets another Page One workout in the first installment (Dec. 19) of a three-part series in the Washington Post titled “Pregnancy and Homicide: The Known Toll.” The headline reads: “Many New or Expectant Mothers Die Violent Deaths.”

Again, how many is many? The precise number of “maternal homicides” is unknown, concedes reporter Donna St. George, because police departments don’t routinely log the maternal status of murder victims. A yearlong investigation by the Post documented 1,367 maternal homicides in the United States in the last 14 years, but these numbers are “far from complete,” St. George writes. A better yardstick was found in the fairly comprehensive statistics complied by the state of Maryland. Extrapolating from those numbers, St. George estimates 295 maternal homicides each year in the United States.

Of course, just one maternal homicide is one more than acceptable, so I’m not arguing with the urgency of St. George’s topic. But she ignores available data that might place the horrific numbers she’s collected into relevant context. According to the Department of Justice, total murders of women in the United States peaked in 1993 at 5,550. The number of murders of women by “intimates”—defined by the government as a spouse, ex-spouse, or boyfriend—has also been falling since 1993 (when there were 1,581), reaching its lowest level since 1976 in 2001 and 2002 (which had 1,202 murders each year). These trends are all the more positive when you factor in the dramatic increase in the U.S. population since that time.

Source: The U.S. Department of Justice

St. George briefly alludes to this good news in a sidebar to Part 1: Criminologist Neil Websdale of Northern Arizona University cautions her about overstating the maternal-homicide problem. More than 1,000 women are killed each year in domestic clashes, Websdale tells her, the overwhelming majority of whom are not pregnant. But she promptly drops the subject. Why? Does she fear that these statistics will undermine her thesis?

Another troublesome aspect of the series is its definition of maternal homicide as the murder of any pregnant or “postpartum months” female. (In the sidebar, St. George reports that “experts” have expanded the definition of deaths associated with pregnancy to 12 months postpartum!) By her definition, maternal homicide would logically include random slayings of pregnant women and new mothers. But this class of murder does not appear in the first two installments of the series. Every maternal-homicide case she discusses involves an intimate of the victim. (The one exception: a case being investigated as a murder-for-hire. A search warrant has been served on the home of the ex-husband, and the ex-husband’s lawyer professes his innocence.) The closest St. George gets to acknowledging that her story is solely about the murders of pregnant women by their intimates comes when she writes, “Many women were slain at home—in bedrooms, living rooms, kitchens—usually by men they knew. Husbands. Boyfriends. Lovers.” (Oooh, there’s that “many” word again.)

How frequently does maternal homicide by intimates occur? St. George studied 72 maternal homicides committed in 2002 in 24 states and found that about one-third were caused by “violence that did not seem related to childbearing: drug dealing, robberies, errant gunfire”; about two-thirds “had a strong relation to pregnancy or involved a domestic-violence clash in which pregnancy may have been a factor.” She doesn’t generalize from this finding to say that two-thirds of all maternal homicides are pregnancy-related murders. But if she did, the 1,367 body count she has documented since 1990 would decline to approximately 910 pregnancy-related murders. Still a horrific number, but a smaller horrific number. (Reading St. George one final time, I wondered if all women of child-bearing age are at greater risk of being murdered by jealous husbands, boyfriends, and ex-boyfriends than those too young or too old to have babies. Surely there are statistics on that to consider.) (Dec. 24 addendum: See this follow-up piece for some debunking math.)

The pity of this series is that after latching on to a compelling subject and reporting the murderous hell out of it, the Post lost track of what all its research added up to. It’s no wonder, then, that the editor assigned to headline the first installment chose the strongest and most specific word of summation he could, and that that word was “Many.”


Yes, I know, I’ve judged a three-part series based on its first two parts. But the promo line for the Dec. 21 installment is “The Legacy” and is about the survivors of maternal murder, so I think I’m on safe ground. When the piece goes up on the Post Web site, I’ll post it here for you to judge. Send e-mail to (E-mail may be quoted by name unless the writer stipulates otherwise.)