They do expect quick service.

National fast-food chains have developed different ways to try to meet that expectation by designing methods to help workers decide how much food to cook and when.

However, a computer system used by one national chain to do that is ineffective, a UB researcher has found.

A study led by Ann Bisantz, assistant professor of industrial engineering, showed that the fault lies with the fact that the system's designers did not fully appreciate what workers deal with on a daily basis, a conflict between workers and systems that exists in many industries.

The study is the first to apply human factors, or ergonomics principles, which describe how to appropriately design jobs and systems to match workers' capabilities, to study decision-making in fast-food environments.

Most human-factors research on helping workers make better decisions has been done in high-risk sectors, such as the military and the airline and utility industries.

Bisantz and fellow researchers observed and videotaped the operations of several restaurants in a national fast-food chain for six months. (As part of the study, the researchers agreed not to disclose the name of the chain.)

She explained that a fast-food restaurant's decision about what to cook and when, is complex.

That's because profit margins are so small and tolerance for error is low, since customers will walk out if they have to wait too long.

"You want to cook the right amount of food at the right time. You don't want people waiting and you don't want to throw food out," she said. "It's not an easy problem."

According to Bisantz, workers are supposed to check the computer and do what it tells them, based on predicted sales, actual sales and the amount of cooked inventory on hand.

"But the technology doesn't mesh with the actual conditions in the restaurant," she said.

Bisantz noted that many human-factors engineers study relationships between workers and technology in the airline industry or in power plants, where a mismatch can be catastrophic.

Her research has proven that many of the same issues are present in the fast-food industry, where it's razor-thin profit margins-not human lives-that are at stake.

"Managers in these stores deal with business risk on a daily basis," she said. "Those costs are not abstract. They feel a lot of pressure to maintain sales versus costs and to do well on those decisions everyday."

According to Bisantz, factors such as unexpected events that make high-risk systems complex also occur in systems with lower risks, like fast-food restaurants.

"For example, all of a sudden, an entire Little League team will walk in and completely drain your store of food," she said.

The computer system Bisantz studied could not respond to these unexpected events.

Also, computer monitors were installed in managers' offices on the assumption-which was incorrect-that that's where managers spend most of their time.

However, "to keep costs down, the manager acts as direct labor so he or she is rarely in the office," said Bisantz. In addition, she noted, most fast-food restaurants are staffed by teen-agers who may require constant supervision.

"It was really interesting to see how the managers did use the computer," said Bisantz.

Periodically, they would count the amount of cooked food they had on hand. They would compare it to the number the computer system said they should have on hand and then correct that number.

"This was a disconnect," she said. "People above you are saying you must use this technology and you can't use it for that, so you adapt."

In reality, the decision about what to cook and when was made by the sales people monitoring the flow of customers coming into the store.

"The (computerized) aid didn't fit with their system because it was based on the assumption that the food preparers made those decisions," she said.

What Bisantz saw happening in the restaurants clearly demonstrated an axiom that originated with John Carroll, a well-known human-factors theorist, that systems are developed on the basis of often-unspoken theories held by the system's designers about how those systems are going to be used.

"In order to design a successful system, you need to get the theories right," she said. "One way to do that is through field studies like these."

Bisantz began the research as a doctoral candidate at Georgia Institute of Technology. She conducted it with Sally Cohen, Michael Gravelle and Karen Wilson, all of NCR Corp. in Duluth, Ga