Predictors of Success in Diagrammatic Problem Solving

Abstract. We conducted an eye-tracking study of mechanical problem solving from cross-sectional diagrams of devices. Response time, accuracy and eye movement data were collected and analyzed for 72 problem-solving episodes (9 subjects solving 8 problems each). Results indicate that longer response times and visually attending to more components of a device do not necessarily lead to increased accuracy. However, more focus shifts, visually attending to components in the order of causal propagation, and longer durations of visual attention allocated to critical components of the devices appear to be characteristics that separate successful problem solvers from unsuccessful ones. These findings throw light on effective diagrammatic reasoning strategies, provide empirical support to a cognitive model of comprehension, and suggest ideas for the design of information displays that support causal reasoning.

• Type: Conference
• Year: 2004
• Where: DIAGRAMS
• Authors: Daesub Yoon, N. Hari Narayanan
• Keywords: Artificial Intelligence, DIAGRAMS 2004, Effective Diagrammatic Reasoning, Longer Response Times, Response Times