The Labor Department announced yesterday it is abandoning a controversial test designed by the federal government to help state employment agencies boost the job chances of minorities.

The test, or variations of it, is a major tool used by state employment services to determine aptitude for blue-collar jobs.

Blacks and Hispanics typically scored lower than whites on the exam, known as the General Aptitude Test Battery (GATB). To minimize the disparity, the Labor Department scored the minorities separately and the top candidates in each group were referred to private-sector employers for interviews.

The test is used by 35 states. The Labor Department said it is used extensively by the Virginia Employment Service, but is not used in Maryland or the District. The state services pay unemployment compensation and refer job applicants to employers. The test scores are used to determine who gets referred to job interviews.

Yesterday's announcement was the latest chapter in a long battle over the effectiveness of the test in predicting who would be a good worker. In 1986, the Justice Department said it was concerned that the way the test was scored amounted a system of reverse discrimination.

Labor Secretary Elizabeth Hanford Dole said the department is proposing suspension of the test because of concerns it does not adequately serve all individuals seeking help from the employment services. "We must ensure that GATB is scientifically and legally beyond reproach," Dole said.

Because of federal regulatory requirements, the Labor Department must wait until near the end of the year before it can order the states to stop using the test for referrals. The department first must submit its proposal for a 30-day comment period and then wait an additional 90 days before it can act.

After that, states that insist on using the test risk losing federal funding for their employment services. Most states depend entirely on the federal government to finance their employment services.

Robert Litman, a Labor Department employment specialist, said that in the absence of the test, minority job applicants would still be protected by federal anti-discrimination laws. He said that under yesterday's proposal, states could use the test for determining the aptitude of a job applicant, but could no longer use the test for ranking applicants for job referrals.

The test, which has been in use for many years, was changed in the 1980s to allow separate scoring for minorities. The scoring system became particularly popular with so-called Fortune 500 companies, the nation's largest employers.

As an example of how the system worked, a black applicant seeking a semi-skilled job might score a 276 on the test to be rated at the 50th percentile, halfway down the list of all blacks taking the test. A Hispanic might have to score 295 to be rated at the 50th percentile among Hispanics and a white might have to score 308 to be classified at the same level.

The Justice Department agreed not to take any legal action against the test until it could be reviewed by the National Research Council of the National Academy of Sciences. That review, completed in May 1989, concluded that the separate weighting of minority test scores was essentially fair, but that the test was not an accurate predictor of who would be a good work- er.

The researchers said they were unable to find an "appropriate alternative."

The Labor Department intends to conduct a two-year study to determine if it can sharpen aptitude measurement. In the meantime, job referrals will be based on more traditional methods, such as interviews and job experience.

Litman said the problems with the testing were not confined to minorities. "The study pointed out that there was a chilling effect, not just on minorities, as a result of the test," he said.