NASA to Release Disputed Data

WASHINGTON, Oct. 31 — The administrator of NASA told a Congressional hearing today that his agency would soon release data from tens of thousands of interviews with pilots about safety issues, information that NASA previously said could damage the airline industry.

But at the hearing, the administrator, Michael Griffin, and the survey’s designers disagreed so deeply about the purpose of the survey and the usefulness of its information that they barely sounded as if they were talking about the same project.

They could not settle on how many pilots had been interviewed for the project, which has cost NASA more than $11 million. Mr. Griffin said it was 24,000 airline pilots and 5,000 general-aviation pilots. But Jon O. Krosnick, a professor and survey expert at Stanford, said the number was much smaller.

Mr. Griffin testified before the House Committee on Science and Technology that NASA had no obligation to release confidential data that was commercially sensitive. He said a person knowledgeable about aviation could piece together, from the airport names and the equipment types, the identity of the airline involved in some cases.

He said NASA would eliminate entire categories of data before releasing the results and would finish by the end of the year, although that drew skepticism from the committee.

“If it’s a computer program, why can’t you do it today?” said Representative Bart Gordon, Democrat of Tennessee and chairman of the committee.

Mr. Gordon said that he had asked NASA’s lawyers for examples of data that should not be released, and that they had not supplied any.

Preliminary analysis of the results appeared to show that the rate of planes’ hitting birds or planes’ coming too close to one another was double or more that of Federal Aviation Administration estimates.

But Mr. Griffin said that many of the calculations done from the NASA survey were simply wrong.

For example, he said, extrapolating from the survey data to the whole population of airline pilots, it appeared that airliners had to land two or three times a day because of unruly passengers, when this has happened only a handful of times since the attacks of Sept. 11, 2001.

“I have been a pilot for decades,” Mr. Griffin told a sometimes skeptical committee. “Anyone who knows anything about aviation is going to look at this data and have a lot of questions about it. On its face, you can look at it and extract conclusion from it that are not credible.”

Jon O. Krosnick, a professor and survey expert at Stanford, said that the problem was that funding had been cut off before the analysis was done and before the survey-takers had moved on to flight attendants, air traffic controllers and mechanics. Mr. Griffin said the program had always been intended to run for a limited period and was only to develop survey techniques, not to collect actual data.

But Mr. Krosnick countered: “You don’t do 24,000 interviews of pilots to test the feasibility of a method. You do that many interviews after you know the method is feasible and ready for prime time.”

They also disagreed about who had been surveyed. Mr. Griffin said it was 24,000 airline pilots and 5,000 general aviation pilots, but Mr. Krosnick said it was a much smaller number, one that had been interviewed on several occasions, and that the plan had always been to go back to the same group at regular intervals, to track whether problems were frequent.

While the F.A.A., the nation’s air safety regulator, has also raised questions about the usefulness of the data, NASA has drawn the ire of many safety experts in its handling of the project. Jim Hall, the former chairman of the National Transportation Safety Board, testified, “Transparency forms the fundamental basis for any safety program,” and said NASA should release the data. “If we don’t know something is broken, we cannot fix it,” he said.

NASA earlier this month denied a request for the data under the Freedom of Information Act by the Associated Press by saying that release could hurt the airlines commercially. Mr. Griffin said at today’s hearing that NASA had an obligation not to release commercially confidential data, and that a person knowledgeable about aviation could piece together, from the airport names and the equipment types, the identity of the airline involved in some incidents. He said NASA would eliminate entire categories of data before releasing the survey results, and would finish by the end of the year, although that timetable drew skepticism from the committee.

“If it’s a computer program, why can’t you do it today?” said Bart Gordon, the committee chair. He said he had asked NASA’s lawyers for examples of data that should not be released, and they had not supplied any.

Mr. Griffin said that NASA had not done a good job supervising its contractors in the project or supplied a final report in a timely way. A related problem, according to various aviation safety experts, is that NASA is doing less in civil aviation as it concentrates on a mission to the moon and Mars.