Oklahoma is one of three states to offer a free prekindergarten (pre-K) program to all students in participating school districts on a voluntary basis. Fortuitous circumstances in Tulsa, Oklahoma, the state's largest school district, permitted an unusually rigorous evaluation of the pre-K program in Tulsa. Because 4-year-olds beginning pre-K and 5-year-olds beginning kindergarten were administered the same test in September 2001 and because strict eligibility cutoffs were applied (based on date of birth), it was possible to control for selection effects, in addition to gender, race/ethnicity, school lunch eligibility, and precise date of birth. The evaluation showed strong positive effects of the pre-K program on children's language and cognitive test scores, but not on scores for social-emotional or motor skills. Hispanic children benefited most from the program, and Black children also showed sharp gains, especially when they attended full-day programs. In contrast, only those white children enrolled in a half-day pre-K program showed significant gains and only in language skills. A similar pattern of results characterized children who qualified for free lunch, for whom significant benefits derived from full- and part-day programs; children who qualified for a reduced price lunch, for whom only full-day programs produced gains; and children who did not qualify for lunch benefits, where no net gain was apparent. The findings were examined in light of contemporary controversies regarding targeted or universal pre-K programs; full- or part-day programming; public school or multiple delivery sites; strategies for ensuring high quality, effective programs; and the need for rigorous program evaluations. (Contains 32 footnotes). (Author/KB)