Having established that the four oral narrative tasks used in the study belong to two types since there were no statistically significant differences between the variants for any of the measures, intercorrelations of the output measures for each task were examined. Table 20 presents intercorrelations for Task 1, while Table 21 shows the same intercorrelations for Task 2. Interestingly, there seems to be a cluster of output measures that tended to change together in the case of Task 1: these were accuracy, the quantity of talk, the D index, and the number of narrative clauses (see Table 20). The rest of the output measures seemed to be more or less independent of each other. No such cluster could be detected in the case of Task 2; almost all the output measures seemed to be independent of one another. It was only the quantity of talk that correlated with two other measures: syntactic complexity and the number of narrative clauses (see Table 21). This might indicate that language proficiency probably determined performance on Task 1 to a greater extent, while other variables, like individual differences, probably exerted a greater effect when students performed the cognitively more complex Task 2.