With about half of the votes counted--we have scanned and recorded just over 18,500 ballots--U.S. workers are overwhelmingly rejecting as unfair the Reagan administration plan to raise their retirement age to 65.

By a razor-thin majority, employes say they endorse the concept of basing in-grade pay raises and job security on performance, rather than seniority. But when asked if they think such a system would work in their offices, there is a top-heavy "no" vote.

On April 17 this column asked the area's 346,878 federal employes to fill in and mail a questionnaire on various personnel reforms that the White House is proposing in the way workers are paid, promoted and fired. We asked people to mail in ballots no later than last Monday. As of late Friday afternoon there were more than 30,000 ballots on hand, representing nearly 1 in every 10 workers.

More than 96 percent of those responding have identified themselves as either federal workers or former federal workers. Although most of the nonfederal workers voting said they favored raising the retirement age, their voting pattern on other issues was similar to that of government employes.

Here are the questions we asked and the results so far:

* Regardless of how it affects you personally, do you think it is fair to make government workers work until age 65 to draw full benefits? Yes, 3,594; No, 14,046.

* Would you support or oppose a system that ranks performance over seniority if you thought that performance appraisals would be made fairly? Yes, 9,221; No, 9,145. It should be added that many voters commented that the question was "loaded" by the inclusion of the word "if."

* Do you think such a system would work in your office? Yes, 4,401; No, 14,371.

* Do you feel this administration is better or worse than the Carter administration in its treatment of federal workers? Better, 2,040; Worse, 16,101; About the same, 400. In one case a man who said he worked for a regulatory agency said he thought the administrations were about the same, while his wife, who does not work for the government, said the Reagan administration was worse.

* If you had it to do over again, would you go to work for the government? Yes, 4,061; No, 13,239.

So what does it all prove? The poll wasn't run along scientific sampling lines. And it was hard to construct short, easy questions on very complicated subjects. Even so, that is a lot of mail from a lot of people. And some very important folks have been watching the vote.

Office of Personnel Management Director Donald Devine, here Friday for a luncheon, stopped by and looked at the mail and the counting process. So did Reps. Steny Hoyer (D-Md.) and Frank Wolf (R-Va.). Kenneth Blaylock, president of the American Federation of Government Employees, saw the count. Even though he had some problems with the wording of one of the questions, he said the results speak for themselves.

Representatives of the National Treasury Employes Union and the National Federation of Federal Employes also dropped in for a look.

The results--we hope to have a final count for you next week--may not change anything. Administration officials say they still plan to introduce their retirement reform package. And the administration still hopes to make a rule change effective this October that would make performance rather than seniority the determining factor in who gets in-grade pay raises and who gets RIFfed.

But the results may have some impact. Before the poll, Reagan administration officials said they had not detected much opposition to or interest in their proposed reforms on the part of federal workers.

They can't say that anymore.