Barbara Bailar said yesterday that she quit her post as a high-ranking Census Bureau official last month because she believes Republicans in the Commerce Department had political motives in killing a plan to compensate for an expected severe undercount of blacks and Hispanics in the 1990 census.

Adjusting the official census to compensate for blacks and Hispanics missed by counters on April 1 would benefit Democrats in reapportionment and redistricting of the House and the state legislatures, according to many analysts. The reasoning is that most blacks and Hispanics live in areas that usually vote Democratic. An undercount adjustment would increase the official population of those areas.

The Commerce Department's decision against making an under- count adjustment was announced Oct. 30 by Undersecretary of Commerce Robert Ortner.

Bailar, now president of the American Statistical Association, charged that the decision was not based on technical difficulties but was "a political decision."

Ortner called the charge "nonsense."

In 1980, according to Census studies, the official census resulted in an undercount for the whole population of about one to 2 percent, but for blacks and Hispanics it was 5 to 6 percent.

Bailar headed a group at the Census Bureau that worked for more than six years to develop a method to adjust the undercount by taking a 300,000-household national survey soon after the 1990 census, then comparing the results to determine where the original census erred. A panel of the National Academy of Sciences has endorsed the method.

Bailar, who resigned abruptly in mid-December as associate director of the Census Bureau, said Ortner's decision "was dressed up like a technical decision when everyone knew it was a political decision. That kind of hypocrisy I just can't live with."

She said she and other census officials had received inquiries from an Ortner aide and a Republican Party official, only a few days before Ortner's announcement, on how big the black and Hispanic undercount might be.

Ortner called Bailar's charge "unfounded" and "irresponsible . . . . She ought to know better than that."

Ortner said senior officials of the Commerce Department decided against the undercount adjustment because there was no unanimity, even within the Census Bureau, on whether the proposed methodology was adequate, practical and would improve accuracy. "You can't say there was an absolute, airtight case in favor of adjustment," he said.

Fears of GOP losses were "absolutely not" a factor in the decision, Ortner said. "Nobody, on any political basis, requested this decision," he said, adding that there were "no calls from the White House."

Census Bureau Deputy Director C. Louis Kincannon said that although Bailar is "a person of unimpeachable integrity" and has "a brilliant mind," "I don't have any evidence that this was a partisan political decision." He said, "I don't have any concrete reason . . . on which to base a conclusion that the department's decision about proceeding with an adjustment was based on politics, by which I mean a biased partisan decision to favor one party or the other."

He said that in meetings within the bureau on whether to adjust for the undercount, there was substantial opposition to doing so from some bureau officials.

Bailar said the decision announced by Ortner not only ruled out an adjustment of the official figures but "completely uprooted" a procedure on which there was substantial agreement within the bureau on how and when to make a final decision on whether to adjust the figures.

Under that procedure, she said, the bureau was to go ahead with refinement of techniques for using the 300,000-household survey method and was not to make a final decision on whether to do an official adjustment until the 1990 census results began coming in and an their accuracy was assessed.