Announced inspections are being incorporated into restaurant inspection programs to support active managerial control; however, their effectiveness is unknown. The study reported here examined the results of 1314 inspections conducted from June 2001 through August 2003 in Minneapolis, Minnesota. Of these, 343 were routine inspections that preceded and 157 were routine inspections that followed an announced inspection, and 501 were routine inspections of restaurants that did not undergo an announced inspection. Significant reductions in frequency of citations for critical violations in two food safety categories--(1) the person-in-charge demonstrates knowledge of foodborne-disease prevention and (2) prevention of cross-contamination--were seen in establishments that had undergone an announced inspection (relative risk [RR] of 0.7, p = .007, and RR of 0.4, p = .001, respectively). The frequency of citation for these critical violations did not decline in establishments that did not undergo an announced inspection. Announced inspections appear to be effective in supporting active managerial control and represent a promising approach to improving food safety in restaurants.