This article investigates the claim that student poverty is the strongest significant predictor of science teaching and learning, even more than school factors. To investigate this phenomenon, this study focuses on education in South Africa. South Africa is a country and educational system characterized by communities of both extreme poverty and wealth, and using cross-nationally comparative quantitative evidence, one can draw a clear distinction and estimate of the impact of student poverty on teaching and learning for both an international and a South African sample. The cross-national comparison contrasts education in South Africa with approximately 40 other countries by using internationally comparative data from the 2003 Trends in International Mathematics and Science Study eighth-grade science assessments and background questionnaires. Using this data, the author tests the hypothesis that student poverty impact is a stable, strong predictor of science teacher pedagogy and student performance in South Africa compared to educational systems around the world and at every level of national development. These cross-national data are analyzed using hierarchical linear modeling, which is a multilevel regression technique that nests student-level indicators of teaching and learning within school communities, and school communities within national systems. Results support the hypothesis that student poverty is the most significant influence on science teaching and learning, which has strong implications for teacher accountability and the impact of school factors on student learning in high poverty communities worldwide.