The Equitable Life Assurance Society, the nation's third-largest life insurance firm, which manages $53 billion in assets, had a problem. Its group insurance division was losing money, partly because of the growing labor cost of processing hundreds of thousands of medical claims.

But that was before the people with the stopwatches arrived two years ago to do time studies of the movements of workers such as Sandra Froio, 30, who punches dental claims data into a computer terminal in Syracuse, N.Y., for $231 a week. The Equitable efficiency experts, after carefully observing the finger, hand, arm and eye movements of workers, concluded among other things that a simple claim -- which they dubbed a "work unit" -- should be done in 6.5 minutes, or roughly nine per hour, a standard that would be adopted as a key determinant in setting Froio's pay and her future with the firm.

The use of such "scientific management" techniques to revamp Equitable's offices has turned the money-losing operation into a profitable one, largely by creating a computer-monitored system to measure each worker's performance, determine pay and help Equitable cut its work force while increasing production.

While the company asserts that most of its remaining employes welcome the new system, the change has taken its toll on those like Froio, a five-year veteran, who said, "The whole system, to me, is a total disaster . . . . The work is tiring. You get backaches, headaches, eye aches, and you are sitting at a desk being timed. All I can think of is the pictures in the old history books, where you see people who did piecework. You are timed, and if you're not producing, it shows in your pay."

Equitable's experience illustrates how computer monitoring can greatly increase productivity but at the same time have counterproductive side effects by increasing tension and worker resentment. What management often sees as an extremely useful tool is sometimes viewed by workers as more of a weapon.

Computer monitoring of workers at Equitable and countless other American workplaces is credited with improving productivity because supervisors such as Froio's can use computers to determine automatically a worker's speed and accuracy and take steps quickly to improve them -- either by helping the employe correct problems or, failing that, by disciplining or firing the worker.

But the use of monitoring is assailed as a sometimes counterproductive intrusion that can dramatically increase the stress in stressful jobs, and that can backfire against employers.

The role of the computer, however, is only that of the messenger reporting the message of worker performance to the manager. The key issues, employers and employes agree, are whether the standards used for job performance are fair to the worker and the boss and whether the use of computer monitoring helps or hinders the effort to achieve higher productivity. Different Standards

Performance standards are nearly as varied as the industries in which they are used: the Equitable dental claims approver must process about nine claims an hour; an AT&T Communications telephone operator is expected to terminate calls in an average of roughly 30 seconds; a Northwest Orient Airlines data-entry clerk is required to hit about 9,000 to 16,000 keystrokes per hour on a computer terminal.

Decisions on pay, promotion and termination are sometimes based on the employe's monitored performance, although in most cases the productivity measurement is only one factor, along with attendance, attitude and other measures of the quality of work.

The debate about job performance standards takes place amid dramatic technological changes in the nature of work in many American offices, where computers have enabled supervisors to simplify or "de-skill" jobs, breaking them down into separate, more routine tasks. Secretaries in large offices, for instance, formerly typed, filed, took dictation, answered letters and talked to clients. Now, in many offices, those tasks have been compartmentalized so that groups of "clericals" perform a single repetitive task: word processing, data entry, telephone answering and other functions so specialized that a large insurance company, for example, may have more than 300 job titles for clerical work alone.

The result of this compartmentalization, experts say, has been to increase efficiency substantially for the entire office, but also to heighten the isolation, boredom and stress in many jobs. In some ways, experts say, the so-called "office of the future" comes to resemble factory assembly lines of the past.

The computer can be used as a way to detect problems and improve a worker's performance, but it can turn into a weapon in the hands of insensitive management, according to Harvard Business School's Shoshanna Zuboff, an assistant professor of organizational behavior who has studied and consulted on the use of "computer-mediated" work for five years.

"It isn't that managers are so interested in monitoring people, but monitoring production," Zuboff said in an interview. "And the system makes it so easy to monitor the weak links, and that increases pressure on people.

"What they are doing is managing a factory, managing a production operation, and from their viewpoint, their employes have to accept that and be willing to produce. That means, 'Keep your head to the tube and don't look up, and keep in-putting.' If that induces a certain level of employe turnover, they are willing to accept it," Zuboff said, "because the skill requirements of the job have been reduced anyway" and employes are more easily replaced.

Computerized performance records offer new vistas of management scrutiny, according to Zuboff. In a Harvard Business Review article, Zuboff wrote that one bank supervisor told her, "Instead of going to someone's desk and physically pulling out files, you have the ability to review people's work without their knowledge. So I think it keeps them on their toes."

Regardless of management's style, Zuboff said, the monitoring systems get results: "In a number of cases, I found that despite amazing levels of demoralization and bad feeling, those offices still experienced quite reasonable productivity increases, and this was because the technology makes it possible."

Demoralization and bad feelings result primarily when computer monitoring is misused "by unenlightened management looking for short-term profit and not long-term performance," according to James O'Brien, executive director of MTM Association Inc., a New Jersey-based management-consulting organization founded by a group of companies in 1951.

MTM, which stands for Methods Time Measurement, has trained about 70,000 industrial engineers and efficiency experts in MTM technique, which O'Brien summarized as "a way of looking at a manual task and determining the correct time it should require" down to milliseconds.

MTM analysts study jobs in such minute detail that they figure how long it takes a typist's brain to refocus his or her moving eyes on a page. MTM's basic unit of measurement is the TMU (time-measurement unit), which equals 0.036 second or 0.00001 hour. The typical typist's brain takes 7.6 TMUs to send that message.

As O'Brien sees the importance of MTM-style analyses: "A company just says 'Give me a fair day's work.' What is a fair day's work? We try to determine what it is . . . whether it's working on an assembly line, changing a tire, whatever."

Standards are crucial for corporations to determine how many people are required to complete a task if the firm is to stay in business, O'Brien said.

O'Brien said he hears frequently that employes regard computer-monitored workplaces as "sweatshops" overseen by a "Big Brother" keeping track of their work. He said he believes such perceptions arise only when companies do not fully explain the need for monitoring productivity and do not properly train their employes.

Stress on workers will intensify if they believe the system is being imposed against their will, O'Brien said. "They need the opportunity for input and participation. If they perceive this as a speed-up, then, most assuredly, they can get that high-stress level. But if management explains that 'we need you; we need your input,' management can develop the feeling that it is not Big Brother.' "

Under Equitable's new system of "work unit" measurement, supervisors armed with daily computer printouts are instructed to conduct weekly review sessions to let each worker know his or her output and rating, according to Brian Williams, an Equitable official who helped design the system.

Williams said the new system is fairer because workers processing lengthy or complicated claims receive credit for more work units rather than just the raw number of claims processed. "People felt they weren't being treated fairly . . . . Because I do 10 claims and you do 10, they shouldn't be considered equal," he said.

Overall ratings, which determine pay and promotions, include quantity of work units plus factors such as quality of work, attendance, punctuality, written and verbal communication, and attitude, such as willingness to work overtime, Williams said.

The system, he said, has resulted in pay raises of more than $100 a week for fast employes. Williams said the "vast majority" of employes are meeting quotas and that many have received raises under the system. Equitable officials would not discuss the specific performance quotas and corresponding pay raises.

"We have many individuals receiving increases in pay. They are performing better than before," Williams said, adding that if financial incentives spur greater output "then we can do it with fewer people who are higher quality. And fewer people is less dollars."

When employes look at the Equitable monitoring system, however, they see it in a vastly different way.

Concern and complaints about the computer-monitoring system were a primary reason why a majority of the 90 workers at the Syracuse processing center voted in 1982 to join District 925 of the Service Employes International Union, according to regional organizer Regina Canuso.

Equitable's attitude "is to weed out the slow ones and reward the survivors, but it doesn't work out that way," said Canuso. "This company really doesn't realize that stress translates to frustration and anger, morale gets low and productivity suffers."

Froio, who has processed claims at Equitable for five years, said that the combination of uncertain standards and close management scrutiny adds tension to the job. "There should be a base salary, and if you go over the quota, you get more money," she said. Instead, she said, "they have built this whole rating system that you can't even describe." System Confusion

The system, which Froio said she and fellow workers generally do not understand, includes "other factors" such as attitude, attendance and telephone communication skills that are figured in numerically. Froio said her cumulative work performance is rated at 10.09, but said she would need to increase it to 11.54 to be eligible for a promotion.

"People feel these 'other factors' are killing them," she said. "They feel they could do really well on the computer, but the other factors will kill you . . . like attitude and attendance. It's like being in kindergarten again . . . and they know when you are off the computer and when you are on the computer."

Karen Nussbaum, national president of District 925, said, "The question to me is how is the monitoring experienced by the worker? People feel they are being violated eight hours a day, but not knowing when they are being watched . . . . The time spent in the bathroom is counted. The time spent to blow your nose is being counted. It creates a sense of living in a police state at work, like having a policeman in your lap instead of a supervisor."

Workers generally respond better to the "personal control" of a human supervisor than the "impersonal control" of the computer, according to Rosalyn L. Feldberg and Evelyn Nakano Glenn, assistant professors of sociology at Boston University, who have studied office automation.

If management is not successful in motivating its employes, then computer monitoring may set off a "vicious cycle" that worsens the problem, they said. They concluded, "The less management is willing to rely on workers' motivation, the more extraneous, regimented and idiotic work has to become. The more extraneous, regimented and idiotic work becomes, the less management can rely on workers' motivation."