Robert J. Sternberg often writes about a lecture-style psychology course he took as a college freshman in which he got a C. "There is a famous Sternberg in psychology," the professor told him at the time, "and it looks like there won't be another." To Mr. Sternberg, the vignette illustrates that conventional assessments do not measure all the abilities students need to succeed in life. A nationally known psychologist, he has spent much of his career designing new measures that might more accurately capture the full range of students' intellectual potential at the university level. Now, a team of Yale University researchers is using the same ideas to rethink the tests that schools use to identify pupils for gifted and talented programs in elementary schools. The team's Aurora Battery, named for the colorful spectrums created by the northern and southern lights, is being translated and tested with tens of thousands of 9- to 12-year-olds, not only in the United States, but also in England, India, Kuwait, Russia, Saudi Arabia, Spain, and other countries. If the preliminary results from those tests are borne out, its developers say, the new assessment could yield a very different pool of gifted students--one that includes a higher proportion of students from traditionally underrepresented minority groups than is often the case now. The new battery is based on Mr. Sternberg's definition of "successful intelligence," which holds that people who succeed in the real world possess a combination of practical, creative, and analytical skills. Traditional intelligence tests, these researchers say, measure only a narrow subset: memory and analytical skills. Also known as "g" for general intellectual ability, those skills come in handy for comparing and contrasting, analyzing, judging, and classifying, and they are the kinds of abilities that teachers tend to value and emphasize in the classroom.