The Washington Post in its editions of Feb. 21 incorrectly described the difference between blacks and whites who took nationwide math and verbal tests sponsored by the Pentagon. The error occurred because "percentile" ratings were confused with "percentage" scores. Overall, blacks aged 18 to 23 scored in the 24th percentile in the tests in question. Whites in the same age group, overall, placed in the 56th percentile. Although the 56th percentile is a better ranking than the 24th, the relationship is not directly proportional. Through a misunderstanding, the Post treated those average percentile rankings as percent scores, and drew the conclusion that blacks had done more than twice as poorly as whites. Further checking with the National Opinion Research Corp., which administered the testing for the Pentagon, revealed that the average (mean) score for blacks in percentage terms was 49, and that for whites was 72. Because these were the first tests of their kind, the results do not show whether the gap between blacks and whites is declining or increasing. Historically, blacks have scored worse than whites in math and verbal tests, but recent studies of younger age blacks and whites suggest that the differences are becoming smaller.

Young black men and women did less than half as well as whites in math and verbal tests given by the U.S. government to a representative national sample of the population aged 18 through 23, the Pentagon said yesterday.

The Pentagon commissioned this ground-breaking, $3.5 million test to determine whether the military services were filling the all-volunteer force with below-average people or ones representative of the nation's youth. A scientifically selected sample of nearly 12,000 young people from all strata of society was tested.

While Pentagon officials say the test results show that the services are not taking in below-average young people, as some critics have charged, they predicted that the findings will set off political explosions around the country.

The most controversial finding in the tests, which were commissioned in 1979 and finished recently, is that black men and women scored an average of 24 percent on tests of "developed ability" in math and reading compared to an average score of 56 percent for whites. Hispanics scored 31 percent.

Pentagon officials stressed that the tests do not measure natural intelligence or learning potential. Instead, they are designed to assess what a young person has learned and his or her capability to be trained as a soldier.

One official said the results raise serious questions about the quality of education in predominantly black schools. He cited the gap between blacks and whites in demonstrating ability in mathematical reasoning, clerical skills, reading comprehension and vocabulary.

The national director of the League of United Latin American Citizens said the results would have been a fairer indicator of Hispanics' capabilities for military service if the tests had been given to them in Spanish.

The Reagan administration has been treating the test results like political land mines, trying to keep them from going off by meeting privately over the last several days with representatives of such groups as the National Urban League, the National Association for the Advancement of Colored People and the Mexican American Legal Defense and Educational Fund. The administration had planned to hold off releasing the report until March 1.

However, after learning that The Washington Post had found out the test results, Pentagon officials agreed to discuss them Friday and yesterday. They also called in specialists over the weekend to whip up a summary report for earlier public release, perhaps Monday.

Pentagon officials said yesterday that the testing of 11,914 men and women aged 18 through 23 revealed the following:

* Race: The military services last year signed up young men and women whose test scores were better than the average for their race in the general youth population.

Emphasizing that the scores represent the math and verbal abilities developed by the young people as of the day of the test, the military recruited whites who scored an average of 58 percent compared to the newly computed norm for whites of 56 percent; blacks who scored an average of 33 percent compared to the national average of 24 percent in the black youth population as a whole; and Hispanics who scored 41 percent compared to the average of 31 percent for their ethnic group.

One in three Army recruits in fiscal 1981 was black, even though blacks make up only 14 percent of the population aged 18 to 23.

Pentagon testing specialists said how well people do on the tests depends largely on how much they learned in school and on their own, the cultural influences imposed by family and friends, and personal motivation.

* Region: Young people from New England--Connecticut, Maine, Massachusetts, New Hampshire, Rhode Island and Vermont--averaged the highest score of any part of the United States, 60 percent, while those from the East South Central section--Alabama, Kentucky, Mississippi and Tennesee--the lowest, 42 percent.

* Education: High school graduates scored more than twice as high as non-high school graduates, 57 percent compared to 27 percent.

* Sex: Men and women scored about the same.

The nearly 12,000 young men and women in the sample took the same test the military gives to determine whether a volunteer can be trained for the armed forces--the Armed Forces Qualification Test, or AFQT. How youths in the general population fared on the tests gives the Pentagon a yardstick for measuring the quality of the people volunteering for the military services.

Asked about the signficance to the military of the new findings, Lawrence J. Korb, Pentagon manpower chief, said yesterday, "The study shows that since the Reagan administration took office the all-volunteer force is working well."

Korb said that in fiscal 1981, 80 percent of the people enlisted were either average or above average "in trainability" when measured against the new yardsticks. Also, he said, the military took in a higher average of high school graduates than the proportion in the general youth population, 85 percent compared to 74 percent.

The United States shifted from the draft to the all-volunteer experiment in 1973, right after the Vietnam war. The quality of the volunteer military, especially the Army, has been a bone of contention ever since, with several powerful lawmakers contending that the draft must be reinstituted to provide the United States with a quality fighting force representative of the general population.

Doubts about the quality of the all-volunteer force hardened after the Pentagon disclosed that the aptitude tests it had given enlisted people from 1976 through 1980 were "miscalibrated," resulting in inflating the entrance test scores for enlistees. Congress forced higher quality standards on the Pentagon by limiting the enlistment of below-average volunteers and insisting that not more than 35 percent of those taken into the services could be non-high school graduates.

It was against this backdrop of doubts about the all-volunteer force that the Carter administration in 1979 commissioned the first testing of all strata of the population aged 18 through 23. The Labor Department was the starting point for obtaining this "longitudinal sample."

The National Opinion Research Center at the University of Chicago administered the tests in the summer of 1980, and the data have been in the process of distillation ever since. The results are especially revealing because, to quote the Pentagon, this was "the first time a vocational aptitude battery was given to a nationally representative sample of American youth."