Table 5 shows the correlation coefficients between Chinese and U.S. students’ representational scores and their
performance on the three types of tasks. For both samples, students’ scores on symbolic representations are highly
correlated with their performance on each type of the assessment problems (P < .01). This means that those students
who use symbolic representations generally had better scores on the tasks. This finding supports the hypothesis that
students who used symbolic representations have higher mean scores than those who used other representations, such as visual representations. For the U.S. sample, students’ scores on pictorial representations are highly correlated with
their performance on both computation tasks (P < .01) and non-routine problem solving (P < .01), but this was not the
case for the Chinese sample. For both samples, students’ scores on verbal representations were not correlated with
their performance on any of the types of problems.