A quasi-experimental study was undertaken to extend first-world research on physics problem solving into disadvantaged South African classrooms. Sixteen urban high schools, involving 189 learners, participated in the study, investigating the effect of a structured problem solving strategy on performance and conceptual understanding. This article focuses on the enhancement of problem solving performance in classroom tests and the midyear examination. The treatment group outperformed the control group by 8% in the midyear examination. Using ANOVA, this increased average score was statistically significant at the .001 level, indicating enhanced problem solving skills. Furthermore, it was demonstrated that the strategy was not transferred successfully to topics studied prior to implementation of the problem solving strategy. A theory is presented to explain the results in terms of the co-development of conceptual understanding and problem solving skill.