By Deborah Howell
Sunday, May 11, 2008

The Post's approach to corrections is not consistent: Some run promptly, others are resisted, a few go astray. And top editors are not looking for trends in the statistics kept on corrections.

Not all correction requests are valid, and the decisions on whether to publish them can be difficult. But rejected requests nag at the ombudsman to investigate.

An April 3 story by Jose Antonio Vargas on how people "live" on the Internet after death focused on the murder of a gay teenager, Lawrence "Larry" King, in California. The story mentioned two other "reminiscent" murders -- that of Matthew Shepard in 1998 and Eddie Araujo in 2002, "both also gay."

Several readers said Araujo wasn't gay, that he was transgender, and though he was biologically male, he identified as a female who was attracted to men. Vargas said he felt that Araujo fell "under the gay umbrella," and his editor, Lynne Duke, agreed.

A half-dozen experts from the American Psychological Association disagreed. Ritch Savin-Williams, chairman of the Human Development Department at Cornell University, said, "The Post would be inaccurate to necessarily call [Araujo] gay. The crucial issue is that, as a woman, she could be either gay or heterosexual. If she's attracted to men, she's heterosexual."

Michael Hendricks, a District psychologist and researcher, said: "This is a distinct difference between gender and sexual orientation. They are independent of each other. Just because someone is transgender doesn't mean someone is gay."

At least a clarification should have run on this issue.

Officials of a company selling a progesterone product asked for a correction on a March 5 story by science reporter Rob Stein about an ongoing major federal study, the Women's Health Initiative. The story said, "Menopausal women who took estrogen and progesterone faced a small, increased risk of cancer for more than two years after they stopped."

The company -- Hormones Etc., Inc./John Lee M.D. Solutions LLC -- said that the study used medroxyprogesterone acetate (MPA), a drug that mimics progesterone and that is called a progestin. Stein said he used progesterone "because progestin is a form of progesterone, and my thinking was that progesterone is a more familiar term to most readers." Progestin was used later in the story. Stein and science editor Nils Bruzelius turned down a correction, concerned that companies marketing "natural" progesterone would say that is somehow safer. "There's no convincing evidence of that," Stein said.

Marcia Stefanick, a research professor at Stanford University, chairs the Women's Health Initiative steering committee. She said that MPA is a progestin, a class of chemicals that have biological properties similar to progesterone, but that MPA is not identical to progesterone. Thus, while it is true that progestin and progesterone are often "used interchangeably, it's technically wrong and WHI did not use progesterone in the study."

Gerardo Heiss, a medical professor at the University of North Carolina and the study's lead author, said progesterone and progestin are not the same but "the way they are used is the same." He said MPA is most commonly prescribed and "in lay terminology" he wasn't sure it was an error to use both terms.

It would have been better to run a correction or clarification, but a carefully written one to avoid giving marketers room to maneuver.

In another case, Steven Milloy, publisher of junkscience.com, said a caption was misleading on a March 10 Associated Press photo from Bejing that accompanied a story by Juliet Eilperin about new studies on carbon dioxide emissions. The caption said, "A heavy haze could be seen in Beijing in August 2007. Two recent reports call for a heightened global effort to reduce carbon emissions."

Carbon dioxide is a colorless and odorless gas, Milloy said, so "carbon dioxide cannot possibly cause smog." Eilperin, Bruzelius and Michel duCille, assistant managing editor for photos, all agreed on the correction; duCille wrote it.

But it never appeared. Executive Editor Len Downie saw that another photo of smog was running with a correct caption, so he thought a correction was duplicative. But the second photo caption did not correct the problem with the first caption.

A March 30 article by Avis Thomas-Lester about the new Gaylord National Resort and Convention Center mentioned Babe Ruth's "historic 715th home run." Reader Michael Curriden of Alexandria wrote to say that "the number 714 was the most recognizable in all sportsdom." Thomas-Lester said she wrote a correction, but it never appeared; no one was able to find it. After I inquired, another was written and was published on Friday.

Post editors approved 1,129 corrections in 2007, down from 1,292 in 2006. As of May 7, 390 have run this year. While corrections are tracked up to publication, no record exists of whether they actually are published.

For years, monthly corrections statistics were sent around to top editors. Because of a glitch, the statistics haven't been distributed for perhaps a year. Apparently no one missed them until my query.

That's a big problem. Editors should be looking at the statistics every month to discern what mistakes are being made and by whom. Corrections are an essential part of the paper's integrity. Section editors approve corrections; they should also start looking at those that are turned down.

Deborah Howell can be reached at 202-334-7582 or atombudsman@washpost.com.