When cognitive fit outweighs cognitive load: Redundant data labels in charts increase accuracy and speed of information extraction

Abstract Organizations are increasingly flooded with large amounts of data on which they base their business decisions. It is thus crucial to visualize relevant data efficiently so that employees are able to extract relevant information most accurately and quickly. However, at present, information visualization research lacks a coherent and evidence based theoretical framework providing clear guidelines on how to design efficient visualizations. When it comes to the use of redundant elements in charts like commonly used data labels, cognitive fit theory and cognitive load theory make different predictions. Whereas the latter assumes that redundant labels generally reduce visualizations’ efficiency due to increased cognitive load, cognitive fit theory predicts increased efficiency for some task types that shall be solved by means of the visualization. In the online experiment presented in this paper, we investigated the effect of data labels in line and bar charts on users’ accuracy and speed in solving chart-related business tasks dependent on different task types. Our results reveal that users perceive charts with redundant labels as significantly more efficient and answer related questions significantly more accurate and faster, which we explain by the help of cognitive fit theory. We provide valuable insights into cognitive processing of charts and encourage graph designers to consider redundant elements as a possible means to increase efficiency under particular circumstances.