WHEN IT comes to "freezing" the federal work force, as the president-elect has pledged to do, it will hardly be possible to ignore that army of "off-book" workers who aid the government as experts, consultants and advisers, but who do not show up on the budget as regular employees. Presumably, these outside helpers have been increasing in number if only because the size of the permanent bureaucracy has stayed constant while the business of government -- as well as the number of its clients -- has grown enormously.

Good numbers on this are hard to come by. The government directly employs a lot of temporary workers, but most of them do things like help out in the post office at Christmas or take the census. And as a result of Carter administration cuts, the number of on-call experts, consultants and advisory board members is apparently relatively small, about 7,000 -- down from over 16,000 in 1978.

But even here the numbers don't tell you anything like the full story, since most consultants are not hired directly but rather through government grants and contracts. So the trick is knowing which part of such arrangements is dispensable and how to cut it.

There are several ways to cut grants and contracts. One is by "brute force": cutting research budgets across the board and hoping that the bad research is cut in at least reasonable proportion to the good. Unfortunately, this method (which is the one that has mainly been used by the administration and Congress in recent years) doesn't usually work out that way.

A second Method is "strangulation" -- loading down the contracting process with so much red tape that few contracts get let at all, and those that do expend much of their resources on non-productive reporting and compliance activities. This approach has been pursued with such vigor in recent years that some researchers say they spend almost as much time trying to comply with the ever-changing and increasing burden of regulation as they do in actual research.

The third and most sensible way is to cut grants and contracts through what is called "quality control" -- making managers of government research programs directly responsible for the quality and usefulness of the grants and contracts they supervise. This method includes requiring agencies to produce clear statements of what they need to know and then making them find it out. Self-evident as this sounds, it is rarely done. But an approach of this sort could, in itself, cut the amount of contract and grant research, because it requires much more thought and effort on the part of government project directors. And it could also increase the chances that those dollars that are spent are spent well. That is, after all, the point of the exercise, in respect to this aspect of government business and many others. Merely lowering or, for that matter, raising the number of employees or programs or dollars ensures neither savings nor efficiency.