With affirmative action in college admissions facing legal challenges, Yale psychology professor Robert Sternberg is developing a test he hopes will help schools gain the diversity they desire without using affirmative action.

Sternberg is heading the creation of a test that would supplement the SAT and evaluate students based on nontraditional skills. Unlike the SAT, which has been criticized for not following a learning theory and producing a discrepancy in performance between white and Asian students and other minority students, Sternberg’s test is based on his own theory of successful intelligence. The theory argues that traditional methods of intelligence are not completely indicative of a person’s performance in school and life.

The exam, which includes exercises such as caption writing for cartoons and telling stories based on pictures, is now entering its second phase of testing, with a sample group of 5,000-10,000 high school and college students.

Sternberg’s study, called the Rainbow Project, is funded by the College Board. Sternberg said the test could be useful in college admissions because it shows student’s creative and practical skills, which are not tested on the SAT.

“It improves prediction of freshman year performance and it helps to create diversity,” he said. “The differences between ethnic groups are smaller than they are on the SAT.”

Dean of Admissions Richard Shaw said the test could be useful to Yale and other schools in quantitatively examining nontraditional intelligence.

“I think sometimes people develop different strengths and they’re not always the traditional ones,” he said. “We’re certainly open and will be interested in seeing what these researchers come up with.”

Shaw said Yale examines applicants on a more individual basis, but he was interested in learning more about the test after the study progresses further.

“I think it also would be used in concert with what we already do, which is to take a look at the whole person,” he said. “So we kind of –Êin a much less scientific way — do this kind of work when they apply to Yale, but certainly it could be helpful if it has a national database and a strong substantive amount of research behind it.”

In the first phase of testing last year, 1,007 high school and college students in Connecticut, Florida, New Jersey, Utah and Virginia took sample exams. The third phase will study the commercial viability of the test.

“If phases one through three work, then it could become a commercial product, meaning that the College Board could use it at least as an option for the SAT,” Sternberg said. “The data from phase one are really good.”