A national commission yesterday challenged the accuracy of state and local unemployment data, used to determine how $10 billion worth of federal money is divided up for public service jobs and other programs during fiscal 1979.

The National Commission on Employment and Unemployment Statistics, chaired by economist Sar A. Levitan, issued a draft report on how the nation counts its unemployed, the first major study of this subject since another such commission in 1962 cleared the labor data of charges of political bias.

The commission was set up by Congress in response to widespread agreement that the unemployment and employment data needed a facelift. Many critics felt that the unemployment number -- perhpas the best-advercised of all government statistics -- was being overstated. Others thought that inadequate statistical methods understated the number. (Yesterday's report gave the national jobless number "high marks for accuracy.")

But as Levitan said yesterday, the main impetus behind the chartering of the commission was the belief by many, especially congressmen, that their own jurisdictions received too little federal aid, while others were getting a lion's share.

The state and local data, prepared independently from the national labor force data, are the basic yardstick for determining how much federal money goes to 6,100 different geographical areas for economic development, public service jobs, and other programs.

The Levitan study, in the works for more than a year, proposes major revisions in the definitions and concepts of the widely published national, monthly labor force data, including a proposal to count members of the armed forces along with civilians, as job-holders.

As published yesterday, the commission's report is only tentative, and designed to elicit comments from public and private agencies prior to April 2. A final vote of the nine commissioners has been deferred on all proposals until then. Recommendations will go to Congress and the president in the fall.

A key recommendation of the draft report, reflecting a lifetime objective of Levitan's, is the need to include some measure of "economic hardship" along with the regular jobs and unemployment data. "The commission finds existing statistics deficient in that they measure whether individuals are working or not, but tell nothing about how well or badly they are doing in ability to meet their economic requirements and those of their families," the report said.

This has been a bitterly contested issue among manpower experts. The late Commissioner of Labor Statistics Julius Shiskin (who helped get the Levitan Commission under way) felt that a hardship index was subjective and likely to have political overtones. The labor unions in the past have shied away from the economic hardship concept, because it blurs the existing sharp focus on the unemployment number. But Levitan said yesterday he expected the union to go along with the recommendation.

The commission, which includes representation from labor as well as business and academic groups, did not settle on any single approach to hardship, but listed a variety of ways of constructing a hardship index by relating the jobs data to earnings and incme.

Other key recommendations:

The commission flatly rejected proposals that older workers be cut out of the labor force count. But it has not yet decided whether to recommend that the count begin at the age of 18, instead of 16, as at present.

It supported regular publication by the BLS of an employment-population ratio, measuring the percentage o fthe population of working age that holds jobs. But this is not "an adequate replacement for the present unemployment index,"

The commission said that a rise of 1 percent in the unemployment rate for a major metropolitian area like Chicago would bring it a 16 percent jump or $18 million, in its public service (CETA) jobs program grant for 1979.

In a telephone interview, Levitan said that of all the potential recommendations in the detailed, 347-page report, nothing was more important than bringing greater accuracy into the state-and-local data, especially if the nation faces an economic downturn in the next year as forecast by some.

In the post 1974-75 recession period, Levitan estimated, as much as $17 billion in federal disbursements depended on state and local unemployment numbers. "What you've got," Levitan said yesterday, "is a multi-billion sweepstakes that depend on these numbers."

The main corrective measure proposed by the commission is a broadening of the size of the Current Population Survey so as to provide better data for jurisdictions of 1 million or more, with particular emphasis on blacks and Hispanics.

The cost was put at from $5 million to $20 million, or up to, double the current cost.To supply completely trustworthy data, the commission said, would cost an unacceptable $2 billion a year.

Rationale for inclusion of military personnel in the labor force is that with the evolution of a volunteer armed service, the similarities between civilian and military employment outweigh the differences. But military personnel would not be counted in local job statistics because there is "no flow between the local labor force and the military job slots."