The Office of Management and Budget yesterday formally reshuffled the warm bodies in its defunct Statistical Policy Branch and transformed seven of them into paperwork warriors.

Statisticians who used to involve themselves with economic or statistical analysis for the OMB have become desk officers reviewing the reports, information-collection and regulation-writing efforts of various departments and agencies.

The others have been transferred to a new Regulatory and Statistical Analysis Division, which will worry about such things as "statistical policy issues that either require the cooperative effort of several agencies, or affect policy making or program management in several agencies," according to an OMB briefing sheet.

The abolition of the old OMB Statistical Policy Branch has been decried by, among others, Rep. Henry S. Reuss (D-Wis.), chairman of the Joint Ecomonic Committee, as a move that would "confirm those who have dark fears of a massive statistical cover-up in the making."

Reuss' staff declined comment yesterday on the new organizational chart because they had not seen it, a spokesman said.

Other Capitol Hill sources said that the OMB's intentions appear to be good, but that it is impossible to tell whether the new plan will work.

There is general agreement inside and outside the OMB that statisticians are needed in the paperwork and regulatory review efforts.

The other question, however, is whether the OMB's new division will have enough personnel to deal with questions in governmental statistics-gathering that involve more than one agency.

Chrisopher DeMuth, administrator of the OMB's Office of Information and Regulatory Affairs, will be in charge of both efforts.

"This is not a plan to rob Peter to pay Paul in any way," DeMuth said. "If we do not do a better job in all our responsibiltiies, including statistical policy, we'll change it."

The new statistical policy office will concentrate, the OMB handout said, on building uniformity into statistical definitions (for example, what is a standard metropolitan statistical area?); coordinating cooperative efforts when statistical projects involve more than one agency; getting the statistically loudest bang for the statistically smallest buck, and making government statistics readily available to the private sector.