Information was gathered on the high school entered at ninth grade, gender, race, entry age, eight grade reading score, and leave status (still enrolled, transferred, graduated, or dropped out). Regression analysis was used to eliminate the effects of student characteristics. Results indicated that community or neighborhood support for schools is an important variable. Other possible variables include school safety, school discipline, course grading policies, amount of homework assigned, and type and degree of support for less able students. Some of the schools studied had low dropout rates and poor performance rankings. The differences between dropout rank and performance rank could be used to select schools for comparative study aimed at identifying characteristics that make a school effective. Tables and 3 references.