In 1974 the government laid down a rule to prevent social scientists with federal grants from injuring the people they study.

It all sounded fine in theory. But in practice, many social scientists say, it has led to a new phenomenon in American academic life: a kind of federally supported censorship.

Consider the case of the University of Minnesota psychiatrist and the impoverished Asian refugees.

Dr. Joseph J. Westermeyer wanted to study the way the refugees were adapting to American life. So he went to the university's institutional review board, the organization that checks projects to make sure they conform to the government rule.

Fine, said the board. Just list all the questions you plan to ask and get the refugees to sign a form agreeing to answer.

Westermeyer thought that was "absolutely crazy." His subjects were all illiterate. His questions would depend on individual problems. But he wanted to do the research, so he agreed, "just to satisfy the bureaucracy."

However, when he took his application to the National Institute of Mental Health for funding, officials there decided that anyone who would approach illiterate subjects that way must be incompetent. Grant denied.

Institutional review boards (IRBs) like the one at the University of Minnesota have been giving social and behavioral scientists heartburn for years. Set up under a 1974 regulation of the Department of Health, Education and Welfare, the boards were designed to protect human subjects involved in any government-sponsored research.

But they evolved, critics say, into bureaucracies that impose their own requirements -- ranging from the merely burdensome to the downright absurd -- on researchers. And in some cases they have become battlegrounds for institutional rivalries in which their members use their position to kill off research that is unpopular or that competes with their own.

The University of California medical school IRB wanted a researcher to obtain a consent form from husbands before it would allow the researcher to ask wives some questions about them. A University of Michigan student was denied permission to study the public behavior of organized crime figures when she refused to seek release statements from all of them.

Trying to get rid of some of the silliness, the National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research took a long look at IRBs recently. It made a number of recommendations that are soon to emerge as HEW rules governing all research on human subjects at institutions that receive federal money.

Many researchers think the new rules will probably make things worse.

"They call for more detail, more committees, more hassle," complained Ernest Hartmann, a psychiatry professor at the Tufts University School of Medicine. "The balance has shifted so far in favor of protection, protection, protection of human subjects that it's very hard to get any decent work done."

The initial idea was that IRBs would make sure new drugs, surgery tricks and replacement parts were not necessarily tested on prisoners, children, the poor, the mentally deficient or even on ordinary folks who didn't know what they were becoming involved in.

To that end, the 1974 regulation required the informed consent of research subjects, in writing. It said personal information should be kept confidential and people should be told the truth about what was going to be done to them. The IRBs could veto projects they thought might be too risky.

However, the rules didn't define "risk."

"I had a graduate student talking to people in wheelchairs who were trying to get out of their nursing homes," said Edward Rose, professor emeritus of sociology at the University of Colorado in Boulder. "Our committee wanted to know the questions, the object, the entire research procedure." He said the IRB involvement was "close to infringement of first amendment rights" of freedom of speech, when the survey "was ethnography, talking to people, that's all."

The rules allowed the IRBs to demand disclosure of purpose even where it interfered with the research. No one could study whether income supplements kept people from seeking better jobs if the people had to be told in advance that they were getting the extra money only so that they could be studied.

Alcoholics and drug addicts and criminals didn't mind talking anonymously, but they didn't want to sign any forms. Sociologists and anthropologists rebelled at the idea of submitting their interview subjects a list of questions in advance.

Some IRBs were accused of playing politics. Researchers in Boston and New York complained they could not get approval for studies of compatibility among children being bused for racial integration purposes because preliminary results showed that the youngsters wound up disliking each other. "The results were not what the university wanted to find," said an HEW official close to the project.

Researchers, not to put too fine a point on it, began to lie. "The students get around the board by talking about 'human beings' instead of 'human subjects,' and that is really weird," said Rose.

"The agencies generally are responding to the regulation by ignoring it," said consultant Richard Tropp in documents supporting the commission study. The overall report said that although 55 percent of the behavioral and social scientists it queried thought that IRBs were a good general idea, 54 percent of them thought the boards had "impeded progress... to some extent" in their fields.

The commission's new recommendations have tried to get around some of these problems but have not always succeeded. Seeking to cut the number of silly rejections, for example, the commission would allow "expedited review" in certain categories of projects, including "collection (in a non-disfiguring manner) of hair, nail clippings and deciduous teeth." Researchers might also collect sweat, feces, teeth plaque or saliva without IRB attention, the report suggests. It is not that specific, however, in the grayer areas of social research, where most disputes have occurred.

While noting that many IRB actions drag on for months, the commission makes no specific recommendation for a time limit. A graduate student of Hartmann's, trying to study psychological changes in alcoholics on a certain medication, tried for six months to get the idea through her board, and finally gave up from sheer lack of time.

"I'm here forever. I have a secretary to fill out the forms and push it through," Hartmann said. "It's the smaller projects that will be big ones in five years that are getting hurt now."

The recommendations would centralize monitoring of IRBs in HEW, thus ensuring their uniformity. But an HEW official wondered whether the Defense Department, say would be happy to have HEW inspectors telling it that one of its projects had to be changed.

Much-needed continuing review of research is recommended to make sure the scientists are doing what they said they would, but the suggestions aren't clear and have no teeth, said Vernon E. Wilson, vice president for medical affairs and chairman of Vanderbilt University's IRB "We've read it and reread it and it sure needs clarification," he said.

Commission staff director Michael Yesley insisted the new proposals will improve the situation. "People aren't reading them right. They're an attempt to alleviate the problems... the IRBs should do more, not less."

The commission itself dissolved at the end of the year and was replaced by HEW's Ethics Advisory Board, which has so far concentrated on the issues involved in research work with human embryos.The new regulations are expected to be issued in late spring.

"We have to make sure no one is hurt by federal activity," said Barbara Mishkin, deputy staff director of the ethics board. "If you don't want to do that, no one requires that you take HEW money in the first place."