The National Curriculum and Textbook Board in Bangladesh specified 50 terminal competencies which are supposed to be achieved by the students throughout the cycle of primary education. The competency-based test instrument which was developed for Education Watch 2000 was administered on a sample of BRAC school graduates of each year during 2000-2010. This monograph is an outcome of analyses of all these test results. Although the rural BRAC non-formal schools were common in each year, samples from BRAC`s non-formal urban, formal and community schools and the schools for ethnic minorities were taken in some years. Sometimes, the government primary schools were also included. Taking all these test data into account, this monograph presents a repeated cross-sectional analysis of performance of BRAC schools. Comparison of various types of BRAC schools, rural and urban schools, and with government schools are also provided. Separate analyses for boys and girls were done as a cross-cutting issue. The findings reveal that the performance of BRAC non-formal school students increased over time. The students did best in environmental studies (science and social studies) followed by Bangla, mathematics and English. Urban students did better than the rural students and the Bangalis than the ethnic minorities. Best performance was observed in BRAC formal schools followed by community schools and non-formal schools. Variation among schools and students persisted although reduced over time. Three major predictors of both types of variations were quality of teachers assessed by the area managers, proportion of contents taught in the classrooms and religion of the students. The former two positively influenced the achievement of competencies and the Muslim students performed better than the non-Muslims. If the programme organizers were frequently changed or they were given more schools to supervise, the students performed poorly. Contribution of the school and teacher-related factors in predicting learning achievement was more than that of the supervisors and students background characteristics. The BRAC school students were more likely to do well in the test compared to those in the government schools. Magnitude of difference was more among the girls than the boys. However, in both the cases, boys outperformed the girls.