Errors on the job can be reduced

Nation's Business, April, 1988 by William Hoffer

A veteran commercial pilot inadvertently shuts down his jetliner's engines after takeoff, correcting his mistake just in time to prevent a crash.

A law firm employee doing paperwork on a $93 million lien omits three zeros from the figure, recording it instead as $93,000 and leaving the lien-holder vulnerable to a $92,907,000 loss.

A nuclear plant's control-room operators, trying to cope with a mechanical malfunction, turn off a cooling system during a period when it might have reduced the risk of radioactive material escaping from the plant.

Workplace errors such as these in recent years have spurred the efforts of some behavior researchers trying to determine just why capable people make mistakes on the job, and how such mistakes can be reduced. The approach of some of these researchers has been running counter to earlier strategies in dealing with workers' fallibility.

In the past, many businesses have attacked error by cajoling or threatening their employees to pay more attention to their tasks. Such efforts often took the form of a "zero-defects" campaign. But according to a current opinion in error research, that type of strategy is based upon two erroneous assumptions.

The first is the notion that all errors can be eliminated somehow. "Several studies indicate that errors occur randomly in time," says John W. Senders, resident professor of engineering and psychology at the University of Maine in Orono. "There really is no way to stop them from occurring."

Modern error analysis, therefore, seeks not to counteract or eliminate all errors-now seen as impossible-but rather to acknowledge the existence of error, prepare for it and minimize its consequences.

"Our emphasis is not to eliminate error, but to reduce the incidence of critical error," says Senders.

The second mistaken assumption about on-the-job errors is that they are necessarily the fault of the employees. Alan D. Swain, senior scientist at Argonne National Laboratory and a consultant based in Albuquerque, N.M., contends that most critical errors result from faulty workplace design. And since it is management's responsibility to provide the workplace, management bears the responsibility for error, he maintains.

Swain categorizes workplace errors as either "situation-caused" or "human-caused." Analysis shows that situation-caused errors-those related to the design of the work environment-account for about 85 percent of workplace errors, and the rest are human-caused mistakes. Most errors, says Swain, should be seen "as the natural outgrowth of some unfavorable combination of people and work situation."

This was illustrated by what some error researchers informally call the "Gold Box Study." In a certain manufacturing plant, workers handling an expensive and breakable gold-plated component called the "gold box" often dropped the object. Though management accused the workers of carelessness and threatened punishment, the "gold boxes" continued to hit the floor at an alarming rate.

In desperation, management hired a consultant, a specialist in human-factors research, who took a simple if often-overlooked action: He watched the employees work. Management had told the consultant that each "gold box" was handled no more than 100 times during the production process. But the researcher counted at least 1,000 opportunities for each "gold box" to be dropped. Though employees were being careful, they simply had too many chances to drop the component.

The consultant helped the manufacturer to redesign the assembly line so that the opportunities for error were reduced from 1,000 to under 100. The breakage rate dropped dramatically.

It has been known for some time that certain factors such as routine, fatigue, stress and distraction-notably noise--are associated with errors. Chief among these so-called error factors is routine--a finding that holds even for those whose job is to find errors. Swain says that inspectors-men and women who are supposed to spot other people's mistakes-generally fail to find 15 percent of all defective products. Since most products are not defective, the inspector becomes accustomed to seeing no defects.

Encountering this problem at a defense plant, Swain suggested that no one should be assigned to inspect for longer than a half hour at a time. This would assure that each inspector presumably would be more rested and vigilant"This one change," he was enough to reduce the number of defects getting by the inspectors."

Researchers now understand better how the error factors disrupt thought processes. "I believe it foolhardy to attempt to determine the reason or cause of an incident," says Donald A. Norman, director of the Institute for Cognitive Science at the University of California in San Diego. "Many or most incidents are bound to have multiple causes," he says, and often it is found that the absence of any one of the causes would have prevented the incident.

Norman has spent years compiling examples of errors, such as the incident involving two sales clerks in a department store, standing side-by-side and each talking on the phone. In order to grab a sales form, one clerk moved behind the other, so the two women switched positions. When the first clerk completed her call, she placed her receiver on the other clerk's telephone, cutting off the other call and angering the customer. A quick solution: two telephones of different colors.