Missile Dispute Enters 7th Year As Air Force Takes Over Inquiry
By Keith J. Winstein
Last Friday, MIT turned to a new chapter in the seven-year-old dispute over whether Lincoln Laboratory researchers fraudulently issued a passing grade to a 1997 prototype from the military’s missile defense system.
The Lincoln-led study gave a generally positive review to claims of the defense contractor TRW, now part of Northrop Grumman, that its software was able to tell the difference between an enemy warhead and a decoy balloon in a 1997 missile test.
For three years, ever since a preliminary inquiry found that the matter warranted a full-blown examination, the Institute’s policies have called for a “prompt” investigation into the allegations of scientific fraud levied in 2001 by Professor Theodore A. Postol ’67.
That hasn’t happened, because the Missile Defense Agency forbade MIT to investigate, saying the risk of disclosure of classified information was too great.
But last week, MIT and the Pentagon agreed on an end to the standoff: MIT will diverge somewhat from its written rules and disband its panel of outside investigators, but the Defense Department will conduct its own investigation, by a senior Air Force scientist and administrator, Brendan B. Godfrey.
At the Institute’s insistence, Norman R. Augustine, the retired longtime chief executive of Lockheed Martin Corp., will serve as “advisor and consultant to the investigator, to help assure an impartial and thorough investigation,” Provost L. Rafael Reif wrote in an e-mail to the faculty.
MIT officials clearly regard the compromise as a welcome step toward reaching some resolution of the allegations, even if the MIT rules are not being strictly adhered to.
“Would we prefer to be overseeing the investigation? Yes. That’s why we tried to do that,” said Claude R. Canizares, the associate provost. “Given the limitations that we have not been able to get around and that are the law of the land, this is the best we can do.”
The entire matter is shrouded in secrecy, both because of MIT’s requirement to keep misconduct matters confidential as well as the presence of classified information. As a result, questions raised about the compromise may prove difficult for MIT to answer:
Did MIT submit to so many restrictions on Augustine’s ability to “advise” the investigation that he won’t be able to do anything?
As part of the MIT-Defense agreement, Augustine, who is volunteering his time, is to “have full access to all classified and unclassified documents,” except those on a special list of documents withheld in 2003 from a civil lawsuit that also alleged fraud in the missile defense system.
Unfortunately, the list of withheld classified documents is itself classified, and no MIT official appears to have had access to the list before agreeing to this condition.
That could put Augustine in a difficult position, because the disputed Lincoln Lab-led report itself is on the list, according to Nira Schwartz, the dismissed TRW engineer who filed the civil lawsuit.
Also on the list is the MIT document that lists the six issues that Godfrey will be investigating, according to Reif.
If Augustine can’t read the disputed report or the questions that have been raised, how can he advise Godfrey’s investigation?
“That’s a good question,” Reif said. “Augustine may be able to say at the end, ‘Look, I don’t subscribe to this conclusion, because it was important for me to see this, and unfortunately I wasn’t allowed to see it.”
“If he’s going to say that, I don’t think DOD’s going to like that,” Reif said. “Let’s see how the process evolves.”
How independent will the investigation be?
The choice of investigator and advisor illustrates the difficulty in finding independent experts in the small, classified world of missile defense.
The Defense Department is both the principal beneficiary of the disputed report — the Lincoln-led team’s positive findings helped the Missile Defense Agency end a criminal investigation into fraud in the system — and, now, the entity investigating whether the report’s findings were wrong.
Lockheed Martin is both the home of Augustine, the advisor — he was chairman and chief executive when the disputed test took place and still maintains an office at the company — and a participant in the very analysis now at issue. (Official documents refer to “Lincoln Laboratories/Lockheed Martin” providing data to TRW in the test.)
A 2002 report by the Government Accountability Office blamed some of the problems in distinguishing warheads from decoy balloons on “software that did not function as designed” because of the poor performance of a Lockheed rocket engine used in the test.
“I don’t see how Mr. Augustine can pass the laugh test of being independent and impartial,” Postol said.
MIT officials described the situation as unavoidable and said there is no conflict of interest.
“It’s hard for me to understand how either an employee of the DOD or Norm Augustine would get a benefit of shading that either way,” said Canizares.
“If we want to identify somebody that has the right credentials, who has the right integrity, who has the right knowledge, and is completely uninvolved in anything that has ever happened, it would be almost impossible to find,” Reif said.
Augustine has “certainly shown himself to be an independent-minded person,” said Philip E. Coyle III, a missile-defense critic who was President Clinton’s assistant secretary of defense for test and evaluation from 1994 to 2001. “He has a wonderful reputation.”
Did MIT’s researchers end up giving the government exactly what it asked for?
A large part of the dispute here is about the responsibilities of scientists who agree, under carefully-defined parameters, to prepare reports for the government. Should they bend over backward to report possible errors or flaws in the data they used to draw their conclusions, or may they ethically stick to a narrower scope?
In this case, the disputed report was commissioned under a charter whose narrowness was virtually unprecedented in the history of the relationship, known as POET, between the Missile Defense Agency and its federally-funded research partners.
“Unlike almost any other POET report,” said Edward F. Crawley, the professor who conducted MIT’s preliminary inquiry, “they were tightly constrained, or they were told to be tightly constrained.” (Crawley made his remarks in a 2002 conversation with Postol, according to a transcript of the conversation.)
Postol takes a broader view: “When you write a scientific paper with results, and you present the results based on data that was taken, you explain how you corrected the data for problems in the measurement. That’s part of the procedure.”
In this case, “they simply said that there’s no problem, and they didn’t describe the problems with the data acquisition. That is fraud, period.”
Canizares, the associate provost, said that the Lincoln-led team simply was not supposed to discuss problems “with the data acquisition,” whether they were there or not.
“They said that the software basically worked the way TRW said it worked,” he said. “It wasn’t a statement that IFT-1A worked,” he said, using the abbreviation for Integrated Flight Test. “That’s not what they were asked to look at. That’s not what they looked at. They had a very narrow charge.”
Timeline of events related to this article: