The National Institutes of Health promised a series of steps to help black researchers win scientific study grants in light of a new study showing a persistent racial gap in its review processes. The study, carried out with the cooperation of the NIH, found that applications from black scientists were 13 percentage points less likely to win grants than were applications from white investigators.Given that the NIH approves only about 30 percent of its grant applications, the gap means that whites are about twice as likely as blacks to win approval, the study found.

“This situation is not acceptable,” Director Francis S. Collins of the NIH said in response to the study published Thursday in the journal Science. “This data is deeply troubling.” The NIH, with an annual budget of about $31-billion, is the world’s largest source of financing for medical research. Yet despite decades of trying to reduce racial disparities in its distribution of that money, the study showed the agency that it still has “a long way to go,” Dr. Collins said.

The study, led by Donna K. Ginther, a professor of economics at the University of Kansas at Lawrence, was based on a review of more than 80,000 grant applications over seven years. The study also found disparities in approval rates involving Asian and Hispanic applicants, though not to the same degree as those involving blacks.

Ms. Ginther and her co-authors said they tried, with little success, to identify key causes. Even after adjusting their figures for a host of possible factors that could contribute to the black-white disparity—including the applicant’s educational background, country of origin, training, and professional publication record—the authors still found black applicants facing a 10-percentage-point gap as compared with whites.

“That’s the frustrating thing about this paper—in most cases, you can come up with a reasonable explanation looking at the observable characteristics, and we haven’t been able to,” Ms. Ginther said. The only conclusion is that the remaining racial gap is due to either differences in the quality of the applications, or some type of racial bias, or both, she said.

Dr. Collins, seeking to demonstrate the seriousness with which the NIH takes the findings, wrote his own response for publication in Science alongside the Ginther study, and spoke with reporters in advance of publication to elaborate. His promised steps included bolstering existing NIH programs that help younger researchers, especially those from disadvantaged backgrounds, gain hands-on experience in writing and assessing NIH grant applications.

The director also pledged new studies of the NIH’s processes for reviewing its grant applications, including tests of whether approval rates might change if officials made a deliberate effort to remove all indications of an applicant’s race from applications. Applicants do indicate their race when submitting grant requests to the NIH, but that information is removed from the paperwork that is seen by the review panels. Other possible clues to an applicant’s race, such as his or her name and educational background, still remain on the paperwork.

The NIH will also try new methods for training participants in review panels to overcome conscious or subconscious racial bias, Dr. Collins said. Part of the problem, he acknowledged, may be the racial mix of the scientific experts chosen to serve on NIH grant-review panels. But with blacks representing only about 2.9 percent of full-time medical-school faculty, and only about 1.2 percent of the principal investigators on NIH grants, the agency has limits on some of its short-term options for choosing review-panel participants, he said.

That current shortfall in black scientists points out the circular nature of the problem, said one of Ms. Ginther’s co-authors, Raynard S. Kington, president of Grinnell College. Often the few black scientists at a university are called upon to serve on committees and help mentor minority students, further reducing the amount of time they have to focus on their research, he said.

“It may be a complicated, push-pull type of thing,” said Dr. Kington, who served as interim NIH director before Dr. Collins arrived two years ago. Grinnell has one possible solution, he said, in the form of the Grinnell Science Project, formed in 1992 as a means for preventing students from traditionally underrepresented groups from dropping out of science courses.

For the NIH, fixing the problem may not be complicated, but could entail hard work on a more personal level, said Michelle Asha Cooper, president of the Institute for Higher Education Policy, which works to help minority and low-income students.

It may require agency officials and universities to actively seek out minorities, in a variety of settings, and invite and encourage them to get more deeply involved in the NIH and its processes, Ms. Cooper said. “You can’t continue to rely upon that established core of people who are part of the network,” she said. “You’ve got to go make some new friends and branch out and extend your network.”