In the Head Start Public School Transition Demonstration Project, 32 demonstration sites across the country have extended Head Start-like services to help low-income children transition smoothly into the elementary grades. Preliminary evaluation of a demonstration program in two West Virginia counties suggested that Head Start students receiving transition services did not gain academically. In 12 schools, standardized tests of verbal achievement and basic math problem solving were administered to 290 students entering kindergarten and 243 students at the end of second grade. Students had received no Head Start services, Head Start only, or Head Start plus transitional services. Head Start participation had no effect on kindergarten results, and neither Head Start nor transitional services affected second-grade results. Ethnographic data collected by participant-observers during the program indicate that the project encouraged teaching practices that may be inconsistent with the standardized tests used to evaluate the program. Local staff development meetings (four to seven per year) focused on practices that are child-centered, developmentally oriented, and based on constructivist principles and teacher reflection. Although teachers were given few concrete recommendations about how to implement such practices, some shifted away from their former practice of "drilling skills" and teaching to standardized tests and became more passive in guiding children toward specific learning goals. The results suggest that it may be harmful to recommend developmental, constructivist education for low-income children with limited means of acquiring the dominant cultural knowledge that comprises standardized tests. (SV)