The shift from a centrally-controlled to a market-oriented economy in the early 1990s put pressure on the Romanian education system to reform both the curriculum and methods of instruction employed in primary and secondary education. At least part of the pressure to implement educational reforms came from the relatively poor performance of Romania's 8th grade students in the 1995 Third International Mathematics and Science Study (TIMSS), where Romania came 34th in mathematics and 31st in science out of 41 countries, far behind its Central and Eastern European neighbors Bulgaria, the Czech Republic and Hungary. While Romanian education has in the past served some students well, as evidenced by successful performance in math and science Olympiads, the TIMSS results suggested the need to specifically address the mathematics and science education of the broader school population. For Romania to effectively compete in an increasingly global economy, it must improve the science and technology skills of its entire workforce, not just the skills of the elite. To achieve this goal, policy makers need to know what contributes to current gaps in the achievement of Romanian children, as well as what school level-factors can be leveraged to improve the achievement of all children. To contribute to this discussion, this article assesses the size of current gaps in science achievement near the end of lower secondary education using Romanian data from the 2003 TIMSS. Variation in the socio-economic background of students, teachers' subject matter knowledge, experience, and instructional practices, as well as the academic climate of the schools are examined as possible sources for variation in students' science proficiency and the urban/rural gap in achievement. (Contains 5 tables.)