"Killer Vet" Story Worthy but Marred by Bad Numbers, Says Public Editor

Public Editor Clark Hoyt's take on the Times' inflammatory story on murderous veterans: A worthy idea marred by shoddy statistics and the eternal journalistic quest to uncover a trend.

Public Editor Clark Hoyt tackled the paper's inflammatory front-page story from January 13 that baselessly suggested returning veterans were in danger of coming home killers, offering worthless statistics to back up the biased assertion. (The series continued Sunday with another front-page "killer vet" anecdote by the same reporters Deborah Sontag and Lizette Alvarez, the story of a veteran convicted of murder aftercoming hometo South Dakota.)

The column's title, "Stories That Speak for Themselves," implied the "killer vet" story was worthwhile, as did Hoyt, eventually. While he offered some sharp criticism of the useless figures compiled by the Times, he basically said that those figures harmed rather than helped a worthy story.

"Today and for the past two Sundays, The Times has devoted front-page play and pages of inside space to a continuing investigative series called 'War Torn,' about veterans of Iraq or Afghanistan who have killed or been charged with killings after returning home.

"It is an important and tragic subject to which an investigative team has devoted more than eight months of reporting, including research into local news reports and court records, and extensive interviews with some of the veterans, their families, victims' families and law enforcement officials often sympathetic toward both the victims and their killers.

"But The Times made some missteps at the beginning of the series, and critics have pounced, accusing it of demonizing veterans and exaggerating the problem even as some mental health professionals have thanked the newspaper."


"The article said the newspaper had found 121 such cases, many of which appeared to involve 'combat trauma and the stress of deployment - along with alcohol abuse, family discord and other attendant problems.'

"The Times was pointing out terrible examples of something the military itself acknowledges: large numbers of veterans are returning from Afghanistan and Iraq with psychological problems. And, as the initial article said, a Pentagon task force found last year that the military mental health system was poorly prepared to deal with this wave of distress.

The Times was immediately accused - in The New York Post and the conservative blogosphere, and by hundreds of messages to the public editor - of portraying all veterans as unstable killers. It did not.

But, the first article used colorfully inflated language - 'trail of death' - for a trend it could not reliably quantify, despite an attempt at statistical analysis using squishy numbers. The article did not make clear what its focus was. Was it about killer vets, or about human tragedies involving a system that sometimes fails to spot and treat troubled souls returning from combat?

"Finally, while many of the 121 cases found by The Times appeared clearly linked to wartime stresses, others seemed questionable. One involved a Navy Seabee accused of arranging her ex-husband's murder during a bitter child custody battle, and another involved a soldier who was acquitted of reckless homicide in a car crash after a jury concluded that his blood alcohol level was below the legal limit and that many other accidents had happened on the same stretch of road.

"Some readers wanted to know how the rate of homicides by veterans compared with the civilian rate. Several bloggers did back-of-the-envelope calculations and said the homicide rate for returning veterans was lower than the rate for the general population. So, what's the problem, they wondered. I asked Martin T. Wells, a professor of statistical sciences at Cornell University, to take a stab at a comparative calculation. The homicide rate for returning combat veterans could be better or worse than the civilian rate, he determined, depending entirely on how many of the 1.6 million military personnel who have been deployed in the Afghanistan and Iraq wars actually saw combat, a number the Pentagon does not have.

"The journalists most responsible for the series - reporters Lizette Alvarez and Deborah Sontag and their editor, Matthew Purdy - argued against trying to make a comparison to civilian homicide rates. The military does not accept people with mental problems or records of serious crimes - the likeliest killers in the civilian population - so its rate is likely to be lower and the comparison irrelevant.

"But they implicitly invited the comparison with a calculation of their own: Based primarily on news reports, their article said the number of homicides involving active-duty military personnel and new veterans was 89 percent higher in the six years since the wars began than in the six years before. And that increase came, the article said, even though 'the American homicide rate has been, on average, lower.'

"It seems analytically shaky to compare admittedly incomplete news reports from two periods and express the difference as a precise 89 percent - especially, as a Pentagon spokesman said in The Times, given that the news media may not have been as sensitive to the military status of accused killers in the period before the wars.

Hoyt quoted Executive Editor Bill Keller's support for the series and the first "Killer Vet" story:

"I believe this series is an important public service that explores in riveting detail the emotional stresses war places on this important community and the problems the military faces in coping with those stresses."

Hoyt concluded:

"But the questionable statistics muddy the message. A handful of killings caused by the stresses of war would be too many and cause for action. Sometimes, trying to turn such stories into data - with implications of statistical proof and that old journalistic convention, the trend - harms rather than helps."