Relationships between users' and interfaces' task representations

In a previous experiment, we demonstrated that some users seem to significantly transform the instructions for a graph creation task before the y even begin to interact with the interface, and furthermore, that this can create considerable difficulty with an interface that does not require the transformation. In this paper, we desenbe a contrasting experiment, showing that subjects without pre-existing task transformations initially have considerable difficulty with an interface that requires them, but acquire the transformations relatively quickly. Kitajima and Poison’s LICAI model explains these effects as resulting from the problem representation being elaborated with task-specific schemata during the instruction comprehension process.