College students prefer periodic tests or term papers to big final exams and learn more from teachers who give them substantial, immediate feedback, a Harvard University study found. For example, a ``one-minute, low-tech, no-cost'' technique is to ask students at the end of each class for a quick, one-paragraph summary of what they learned, said Richard J. Light, author of the three-year study issued Monday. The study concludes that modest shifts in academic policies can create major improvements in college classrooms. Light, a professor at Harvard's Graduate School of Education and Kennedy School of Government, based the study on in-depth interviews with 360 Harvard undergraduates _ conducted by their fellow students _ and seminars that included faculty and administrators from Harvard and about 20 other institutions. Among the conclusions of ``Explorations with Students and Faculty about Teaching, Learning and Student Life'': _Students said they learned the most in classes with periodic tests, oral presentations and papers that let them know how they were doing ``mid-course,'' rather than a big exam at the end of the course. In particular, students indicated they preferred opportunities to rewrite term papers after ``red-penciling'' by faculty. ``Students said they learned more from doing the fourth draft than by doing the first draft,'' Light said. _Students found small study groups of up to six people meeting outside the classroom particularly effective. Light said teachers should encourage formation of such groups. _Students in outside activities such as athletics or clubs, jobs or volunteer work for up to 35 hours a week ``are by far the happiest students,'' Light said. He said teachers should urge students to ``get involved.'' While only Harvard students were involved in the study, Light said the in-depth questionnaires and the information from the seminars made the results applicable to other institutions, including high school and graduate school. Light said he was surprised by some of the differences found between male and female students. For example, although there is no substantial difference between the grades of male and female students, 39 percent of the females said they were not happy with their academic performance, compared with 19 percent of men, he said. Also, although both groups indicated that study groups were effective, males were more likely to join one or form one, Light said. Twelve undergraduate and graduate students were trained by Light and other colleagues before they conducted interviews with the 360 randomly selected Harvard students. They used a 23-page questionnaire that included many open-ended questions. Follow-up interviews were done five to six months later. R. Eugene Rice, senior fellow at the Carnegie Foundation for the Advancement of Teaching, said the Harvard study ``confirms a lot that we know about learning, in fact, just about human work in general _ the importance of feedback, its regulation and its timing.'' Rice said he hoped that faculty heeds the study's recommendations, but he noted that universities must then ``reward faculty for putting their time into this kind of teaching which take away time for work on research.''