LOS ANGELES -- The nation's first attempt to count its homeless people missed 60 percent to 70 percent of the transient population of downtown Los Angeles, according to a report prepared under a grant from the Census Bureau.

"We seriously call into question the accuracy or validity of the count, and strongly caution against using the 1990 census data for developing or implementing public policy," wrote the authors of the report by the nonprofit Los Angeles Homeless Health Care Project.

The report, which was issued Friday, is one of five assessment efforts by groups around the country that received Census Bureau grants to monitor the homeless count. The Los Angeles report is the first to be released.

In the first comprehensive effort to count the nation's homeless, census workers fanned out across the nation on the night of March 20, surveying transients in skid row flophouses and rural campsites alike. The effort was criticized almost from the start for missing significant chunks of the transient population. No official result from the count has been released.

For Los Angeles and other big cities, a higher count of homeless people as well as other residents could mean hundreds of millions of dollars more in federal aid. "I don't need to remind you it means megabucks for us," Los Angeles Mayor Tom Bradley (D) said Friday at a news conference.

But Bradley said he thought the Census Bureau was doing a good job overall, noting the bureau's announcement earlier this week that it has received census information from 91 percent of the nation's households and 83 percent of Los Angeles households.

Michael R. Cousineau, executive director of the organization that issued the report, said the March 21 homeless census failed because census takers "didn't look in the right places." He said there were too few census takers and too few of them could communicate with the large number of Latino homeless people.

Cousineau's group stationed 80 observers throughout downtown Los Angeles, which contains the largest concentration of homeless people in Southern California, on the night of the homeless count.

John Reeder, the Census Bureau's regional director for Southern California, indicated he was not ready to accept the report's bleak assessment of the homeless count. "We don't know if the figures they put out are correct," he said.

Reeder repeated what the Census Bureau has said all along about the count. "We said it would miss people, that it would be conservative," he said. This partly was because census takers were told not to walk down dark alleys or go other places where their safety might be at risk, he said.

Meanwhile, a federal judge in Brooklyn ruled that it was premature to compel the Census Bureau to use results of a controversial post-census sampling technique to adjust the national count of poor people, minorities and immigrants. The ruling came in a lawsuit brought by Los Angeles, New York City, Chicago and others.

The survey, which has been tested in inner-city neighborhoods in St. Louis and Los Angeles, involves revisiting 150,000 households that filled out census forms in April and determining whether there were more or fewer people living there than listed at the time of the initial count.

Congressional seats and money from the federal government are apportioned on the basis of population figures provided by the census. Many urban Democrats are convinced that they are being shortchanged by a Republican-controlled census that is undercounting the urban population.