AbstractThere is a consensus regarding the fact that students encounter difficulties in understanding scientific concepts, such as the particulate nature of matter, the mole, and the interpretation of chemical symbols. Researchers and practitioners have been looking for teaching methods to improve students' understanding of quantitative chemistry and their ability to solve related problems. This study describes the Multidimensional Analysis System (MAS), an approach to constructing, classifying, and analyzing quantitative chemistry problems. MAS enables classification based on complexity and transformation levels of a quantitative problem. We define three transformation levels: symbol ↔ macro, symbol ↔ micro, and symbol ↔ process. Applying this framework to teaching and research, we investigated the relationships between MAS‐classified chemistry problems and student achievement in solving these problems. The research population, 241 high school chemistry students, studied problem solving according to MAS for 9 weeks; the control group studied the same topic for the same duration in the traditional way. Student achievement was sorted by mathematics level and gender. We found that the success rate of the entire student population in solving these problems decreased as the problem difficulty increased. Experimental group students scored significantly higher than their control group peers. The improvement in student achievement was significantly dependent on the pretest score and the mathematics level, and independent of gender. Students who studied mathematics in the basic level benefited significantly more from MAS‐based teaching than their peers, whose mathematics level was advanced. Based on the research findings, we recommend applying the multidimensional analysis approach while teaching quantitative problems in chemistry. © 2003 Wiley Periodicals, Inc. J Res Sci Teach 40: 278–302, 2003