A three-year, $210,000 government study released yesterday confirms that persistent discrimination in the labor market is a major reason that women's pay is 40 percent lower then men's, despite decades of struggle for equity. But it recommends against the government setting any nationwide policy for correcting the disparity.

"Since our conclusion is that the differences are deeply embedded in the institutions and traditions of the labor market, we have not been so naive as to come up with a single, all-purpose solution," said Ann R. Miller, a professor of sociology at the University of Pennsylvania. She chaired the committee that produced the report for the National Research Council of the National Academy of Sciences.

The study focused on the sticky issue of equal pay for work of comparable value, or "comparable worth," which has been called "the civil rights issue of the 1980s." In contrast to equal pay for equal work, which is already required by law, the concept compares occupational apples and oranges -- for example, secretarial work usually done by women and maintenance work done by men -- and seeks to eliminate pay differences based on gender rather than the actual value of the work.

Business groups and government officials who oppose the doctrine have claimed there is no workable definition of comparable worth and argue that its implementation could completely upset the U.S. economy, costing employers billions and putting courts and bureaucrats in the business of setting wage rates for American workers.

J. Clay Smith, acting director of the Equal Employment Opportunity Commission, which requested the study, said the 136-page document will assist the commission in determining the scope of the problem and in understanding what to investigate in its efforts to enforce laws against discrimination.

The study also should alert private-sector employers to the issues and "may result in some self-regulation and the avoidance of government involvement," Smith said.

To the horror of employers and others, the commission had considered issuing guidelines on how to evaluate jobs to establish pay rates. The study recommends against this, suggesting instead further study of job evaluation techniques and more of a case-by-case approach which, Smith said, "makes sense."

"We do not believe that the value -- or worth-- of jobs can be determined by scientific methods," the study said. "Hierarchies of job worth are always, at least in part, a reflection of values."

The committee found that the concept of comparable worth "merits consideration" as a basis for anti-discrimination policies -- within individual firms. But even this, "since much of the wage differential arises because women work in low-wage firms and men work in high-wage firms, would not entirely eliminate the differential."

Causes other than discrimination only partially explain the wage differential, the study said. For instance, the theory that women "chose" lower paying jobs in order to balance family responsibilities is only a partial cause of job segregation. (In the study, the term discrimination "does not imply intent but refers only to outcome.")

The study also found that:

Within firms that use job evaluation plans as a tool for setting wages, women's jobs are paid less on the average than men's jobs with the same rating.

Past discriminatory practices, such as paying women and minorities less for jobs equal to those held by white men, have been entwined in wage structures and continue to operate even in the face of conscious attempts to avoid discrimination.

Reaction from women's groups was mixed. "In a time when the Reagan administration seems to be indicating that sex discrimination is no longer a problem in employment," said Karen Nussbaum, of the Boston-based Working Women, "this prestigious study supports what working women have known all along -- that 80 percent of the nation's 43 million employed women are trapped at the lowest-paying, lowest-status jobs in America."

In Chicago, however, Nancy Kreiter of Women Employed dismissed the study. "It doesn't say much that's new." The report is using a "copout" when it says the problem is too complex to tackle, she added.

In the early 1960s, women who worked full time earned less than 60 cents for every dollar earned by men and that figure has remained virtually constant.

The committee focused on male-female wage differentials because, it said, "the situation with respect to minorities is more complex" and little research is available on them.