Solving problems using matrix, network, and hierarchy diagrams: the consequences of violating construction conventions.

In order for a diagram to be useful for solving a problem, it must be constructed so that its perceptual features facilitate inferences relevant to that problem. In Experiment 1, we established the construction conventions, which relate to how information is assigned to different parts of the diagram, for three spatial representations-networks, hierarchies, and matrices. In Experiment 2, participants solved problems using diagrams that either followed or violated these conventions. As hypothesized, participants took longer to draw inferences from convention-violating matrix and network diagrams than from their convention-following counterparts, and these inferences were less accurate. Convention adherence did not affect reasoning time or accuracy for hierarchy diagrams. The authors concluded that the construction conventions are related to perceptual features that facilitate certain types of inferences for matrix and network diagrams, and they discussed why this might not have been the case for the hierarchy.