This document is 7th in the Journal for Research in Mathematics Education monograph series. The mathematical performance of (n=250) U.S. 6th-grade students from both private and public schools and (n=425) Chinese 6th-graders from both key and common schools was examined on multiple-choice tasks assessing computation and simple problem solving, and on open-ended tasks assessing complex problem solving. Chinese students performed significantly better than U.S. students on both computation and simple problem solving. These results were about the same for the two samples on complex problem solving. Moreover, when subsets of U.S. and Chinese students were matched on their computational performance, the U.S. students scored significantly higher than comparable Chinese students on the measures of both simple and complex problem solving. The results of this study suggest not only the complexity of examining mathematical performance differences, but also the inadequacy of using a limited range of tasks to measure mathematical performance in cross-national studies. (Author/MKR)