In the old days politicians often took their differences, and pistols, to the dueling grounds of Anacostia and Bladensburg. Once there, they usually missed. But points were awarded for grace under fire.

In 1985, the weapons are briefing books chock full of numbers. Statistics by the bucket are what political combatants sling at one another, but only after making sure their numbers prove what they want.

One of the hottest battles around -- at least to the 2.6 million folks who work for Uncle Sam -- is the plan to cut federal pay 5 percent, then freeze it until civil servants start quitting at a rate that is the same as in the private sector.

If you want to take it personally it means that you, as a federal scientist, computer programmer or whatever, would not get another raise until enough of your colleagues sent Uncle Sam a message by leaving government for greener pastures.

The Office of Personnel Management said job turnover in private industry is four times higher than in government. What that means, OPM says, is that many feds are overpaid and realize there is no better place to go in the private sector. OPM says that even if it counted in-house job transfers (which it doesn't do), the federal quit rate is still much lower than the rate in industry.

Not everybody agrees with OPM's conclusions. Many federal workers think the concept of setting pay according to the so-called quit rate is harebrained (at best) or an intentional plan to wreck the career civil service (at worst).

Shortly after OPM announced its plan to link pay and quit rates -- complete with backup statistical data -- the Bureau of Labor Statistics sent a memo to its boss. The memo -- reported here Thursday -- said the bureau doesn't understand OPM's math. It questioned, among other things, the wisdom of comparing stale industry data (the government stopped collecting it, because of budget cuts, in 1981) with newer data on government quit rates compiled by the OPM.

The bureau's professional staff also said it is wrong to compare job turnover in private firms -- the average with 60 workers -- against turnover in the federal government, which is the nation's biggest single company.

OPM said it stands by its data and its conclusions.

Now come two members of Congress -- from districts with sizable federal worker populations -- who express the hope that OPM has shot itself in the foot with its data.

Rep. Steny Hoyer (D-Md.) said, "I find it ironic that OPM wants to compare 1983 federal quit-rate data to a Labor Department (BLS) survey of the private sector that this administration discontinued in 1981 . . . I don't know too many analysts who feel comfortable comparing 1981 data for one group with the experience of another dissimilar group of employes two years later."

California Democrat Vic Fazio said a plan "as radical as OPM's quit-rate proposal with such serious implications for the government ought to be based upon credible data." He said the numbers in the backup data are "so old . . . that when it was released in Washington we cringed at the smell in my district." Fazio said the private sector envies the "stability" of the federal work force. "We never dreamed that OPM would try to sell federal workers down the river with a report that just doesn't hold water," he added.