By year's end, researchers hope to have developed a more effective pilot training method.

It won't be a "one size fits all approach," but instead will be designed to be a useful tool for pilots of turboprops, regional jets and jumbo jets.

"Right know most airlines are teaching procedures, but not much knowledge about automated systems," said Deborah Boehm-Davis, a psychology professor at George Mason University. "We know the airlines are not going to add a lot of time to their training programs. The question is what can the airlines take out, and can we find procedures that work better than others?"

Boehm-Davis and Florian Jentsch, director of the Team Performance Laboratory at the University of Central Florida, have been researching effective pilot training techniques for a number of years with Federal Aviation Administration (FAA) grant money.

The pair has developed prototype training interventions that are being tested by regional and network carriers. Jentsch said they have developed a computerized evaluation tool that bridges the gap between multiple-choice tests and airmen checks on simulators that the carrier can use to determine the effectiveness of the instructions.

Neither researcher would disclose the airlines that are participating in the research.

"So far we have validated our tools against multiple choice tests, but we have not had the opportunity to observe it in a simulator," he said. The test will be to determine if those pilots who are trained using Jentsch and Boehm-Davis's theories react "better" performing difficult tasks in unusual flight situations that their peers who have been taught the industry's tried-but-true methods.

By year's end, Boehm-Davis said they hope to a prototype manual ready and will then spend another two years evaluating the effectiveness of their techniques.

The industry is vulnerable right now, said Beth Lyall, president of Research Integrations, who has also spent years determining the best way to teach pilots to fly automated aircraft. "We are vulnerable in not training them as effectively as we could. There is a vulnerability in that we have not defined how the programs are not effective."

Based on their research, Jentsch said pilots need to be taught to "follow the information," instead of procedural skills.

"Current training is largely focused on giving pilots procedural skills. Here is sequence of buttons to push to achieve a particular goal. While this is helpful to teach people quickly what to do, it doesn't tell them why they have to do it or when it would be a good idea to use a different strategy."

In their approach, which they have dubbed "follow the information," the pilot is taught where a particular piece of information comes from and how it is used by a plane's automated systems. If an error is made in entering data or pushing a sequence of buttons, this theory explains just what will happen, Jentsch said.

"The current training is effective for the average situation that they may encounter. The current training needs improvements for situations outside the normal flight envelope. Pilots don't understand why something is happening so they cannot project forward what might happen in extraordinary situations," Jentsch said.

During the research phase at a major carrier, he said, they discovered that a whole class of pilots did not understand how much fuel was still available after flying a holding pattern because they did not understand how the plane calculated the amount of fuel that remained. He noted that the airline immediately changed its instructions so that the pilots could better understand their fuel situation.

"People tend to attribute more intelligence to an automated system that it has," he said. "Too many believe that automation has the right answer."

There continues to be a debate in academic circles on the best instructional method, Lyall said. Prior to automation, pilots were taught how to build an airplane and how to fix it when it broke down. In the automated aircraft, the instruction has simply been how to use the electronic and digital devices. "The philosophies are again should we teach them how to build the plane so they can predict how it will react or just teach them how to use it. Where is the middle ground? That has not been determined," Lyall said.

Lyall's research has indicated that the methods that may be most effective with the network carriers may not "generalize" to the regional carriers. The researchers need to take into account that most regional pilots don't have a lot experience with automated aircraft. This lack of experience, she said, can change the training objective and methods.

Jentsch noted that in work with entry-level first officers, they discovered these pilots overestimate the risk and underestimate the time pressure. "They agonize over what is the right thing to do, thinking they have a lot of time to make a decision, but they don't have a lot of time. At the same time, if they made a snap decision, it would probably be OK because the consequences are not as bad as they think. Their sophisticated aircraft can deal with these situations that perhaps a single-engine aircraft cannot," Jentsch said.

>>Contacts: Deborah Boehm-Davis, George Mason University, dbdavis@gmu.edu; Florian Jentsch, Central Florida University, (407) 384- 2091, fjentsch@ucf.edu; Beth Lyall, Research Integrations, (480) 893-1600.<<