PDA

View Full Version : Can troops be ID’d from survey?



thedrifter
05-18-07, 07:32 PM
Can troops be ID’d from survey?
By Kelly Kennedy - Staff writer
Posted : Friday May 18, 2007 17:29:38 EDT

A survey cataloging whether soldiers or Marines have ever kicked or hit an innocent Iraqichanged rules of engagement, or believe torture should be allowed leaves research experts concerned that those service members could be identified.

But the Mental Health Advisory Team members who conducted the survey said there’s no way anyone could be identified “with any certainty.”

The MHAT IV mental health survey released earlier this month provided troubling, but important, statistics: One out of 10 of those surveyed said they have hit or kicked a non-combatant or damaged his property. Only half said they would report a colleague for hurting or killing an Iraqi civilian. And one-third said they would condone torture if it produced valuable information.

The report also found suicides are up in the war zones, mental health issues are worse for those who have been deployed longer and more often, and those who regularly go off base are more likely to suffer stress disorders — issues Defense Department officials said they are addressing immediately.

When the service members took the survey, they were told it was anonymous.

But on the first two pages of the survey, 1,320 soldiers and 447 Marines filled out circles indicating their age range, gender, race, rank, branch of service, job skill, location by base, years in the military, highest level of education, number of times and where they had deployed, month they arrived in theater, and unit of assignment down to the company or battery level.

According to a report released by the team, 30 of those Marines were women; 69 were married with children; three were 40 or older; 18 were Asian; three were senior enlisted; and two had a master’s degree or doctorate.

For the Army, 188 were women; 43 were ages 18 or 19; 59 were Asian; eight were senior enlisted; and 14 had a master’s degree or doctorate.

“The service members were told the survey was completely anonymous,” said a researcher who used to work with the Defense Department but asked not to be named for fear of retribution. “But if you look at the survey itself, that is … a blatant misrepresentation.”

Though the survey doesn’t ask for a name or Social Security number, the source said anyone with a list of company demographics could potentially figure out who said what. Because some of the reported activity falls under violations of the Uniform Code of Military Justice, the source worried a court could subpoena those surveys.

“All this stuff is sitting in … headquarters,” the source said. “As long as these data exist, they’re a threat. What if there’s some future Abu Ghraib? It will be, ‘Did they take this survey? What did they say?’ ”

Col. Carl Castro, who led the team, said ethics questions about anonymity are “fair and legitimate” and he isn’t surprised someone would ask them.

“Methodology is very, very important,” he said. “We always give the soldiers an envelope so they can put it in themselves and seal it, and we tell them, ‘Skip anything you’re uncomfortable with.’ We try to make them as anonymous as possible.”

He said there’s no way a person could be absolutely identified.

For example, he said, some service members could say they’re female when they’re male, or give false unit information.

“You can’t with any certainty say, ‘This is the person,’ ” Castro said. He was unsure if the records could be subpoenaed.

Peter Finn, chair of the institutional review board at Indiana University, said the service members should have been told the surveys were “confidential,” and that the identifiers should be removed from the rest of the survey to guarantee its confidentiality. The surveys are at the Walter Reed Army Institute of Research.

Institutional review boards look at research on humans to make sure no harm is done. The Army has such a board, but it did not look at this survey because it was done in response to a question from Army Gen. George Casey Jr., former senior U.S. commander in Iraq and now Army chief of staff, rather than as “research.” This is proper procedure, according to the researcher who formerly worked with the Defense Department.

“The MHAT was conducted as a unit-needs assessment,” said Col. Charles Hoge, director of the division of psychiatry and neuroscience at the Walter Reed Army Institute of Research. “Results were only presented in aggregate with no individual information presented.”

But Finn said this survey would never have made it through a review board.

“You could imagine there could be legal action because of it,” Finn said. “It could represent a real risk to some of these individuals.”

Castro disagreed. “To me, confidential means if I did a survey, you could figure out it was me,” he said, but there are no names or Social Security numbers that could make identification certain.

Finn also worried that service members could have been coerced — or “voluntold” — to fill out the survey, which is another violation of research practice. But Hoge and Castro said troops were given the option not to take the survey, and also could have opted to seal a blank survey in an envelope.

The surveys were conducted by Combat Stress Control team members who were not assigned to the surveyed units. Some of them were hand-delivered to the MHAT team, some were mailed, and some were sent via FedEx, but Castro said he personally inspected each sealed envelope when they arrived at the MHAT.

“Our senior leaders really do care,” Castro said. “I always tell them not to ask questions they don’t want to know the answers to, and these were difficult questions. You don’t do that if you’re trying to hide something.”

The researcher who worked with the Defense Department agreed the information gained is important, but added a word of advice for service members taking such surveys:

“Don’t fill out the first two pages.”

Ellie

Zulu 36
05-18-07, 07:54 PM
Wow. As a trained researcher, thats a lot of info to ask on an anonymous survey. First it makes for a long survey, second it does make it easier to ID respondents.

It is possible to ID WITH certainty some respondents if their demographic and personal data is unique and their MOS or some other categories are small. It might be possible to ID down to five or six people in larger MOS'. Of course, all that presupposes truthful responses.

I agree with the "experts." Don't fill out the first two pages. I also agree that it probably would not make it through a proper academic IRB (Internal Review Board). As a researcher, all of that demographic info is nice to have, you get to play all sorts of games with the data, but most of it is not critical to drawing reasonably accurate conclusions on the core reasons for the survey.