Methodological convergence as an issue within environmental cognition research

Two perspectives in the analysis of pointing and mapping tasks as the measure of representations of the large-scale environment are examined. These two perspectives are: (1) an individual difference approach; and (2) a cognitive representational approach. Convergence between methods assessing the same geographical/spatial knowledge is necessary as evidence for the existence of unified cognitive-spatial representations of the environment.Three sets of analyses interrelate performance on pointing and mapping tasks. In the first analysis, a confirmatory factor analytic model is applied to short tests of pointing and mapping accuracy to determine whether one or two factors are needed to account for covariation between the tests. In the second analysis, covariation among errors in pointing and mapping of specific locations is partitioned into general and specific method factors using the Schmid-Leiman procedure. In the third analysis, pointing errors for identical locations within the mapping and pointing tasks are directly compared on the basis of directional errors.The three analyses indicate that: (1) tests of pointing and mapping measure highly related abilities; (2) the targets used in pointing and mapping tasks are of differential importance in identifying general and specific method factors; and (3) there is little or no direct correspondence between directional errors made in pointing tasks and those occurring in mapping tasks for the same locations.When results of the three analyses are examined in relationship to criteria for convergence of pointing and mapping tasks, little evidence is found to suggest that directional errors in these tasks arise from a unified mental representation of the geographical environment. However, substantial predictable individual differences are apparent for both tasks.