Paper Reading #24: Using Language Complexity to Measure Cognitive Load for Adaptive Interaction Design

Comments

Reference information
Authors: M. Asif Khawaja, Fang Chen, Nadine Marcus
When/Where: IUI’10, February 7–10, 2010, Hong Kong, China

Summary
This paper aims to measure and cognitive load through certain elements of language complexity. The designers identify limited human memory capacity as the root of cognitive load, and that it manifests well in language deficiencies, particularly in speech pauses, when the mind must catch up. Language complexity is categorized by semantic complexity, which observes the use of words, their frequencies, and their lengths, and syntactic complexity, which observes primarily the sentence length, which is considered as the best indicator of text or language complexity.

Bushfire management teams and their speech conversations were studied in incident control rooms. Tasks they had to face were either ‘low load’ (routine tasks and no time pressure), or ‘high load’ (challenging tasks, time constraints, and a lot of unexpected events and breakdowns happening). Their responses in these situations were analyzed according to six complexity measures, including Lexical Density and Complex Word Ratio. The designers found that as task difficulty increases, vocabulary richness decreases and the use of complex words increases. When dealing with easy tasks, people use short, complete sentences, but do not for more difficult tasks. With these results, they hope to be able to improve the interaction design for interfaces or systems that cause higher cognitive load.

Many things can contribute to cognitive load. Source: diamondclear.com

Discussion
Cognitive load is something often researched, and it was interesting to see how it can be quantified from speech patterns, even if the data is specific only to bushfire management teams, as the paper acknowledges, which was also admirable for the authors to address. I believe this line of research can have great impacts on interface design, as well as further research in psychology. I can already see how this could influence designers  to present information, especially text more intuitively such that it is easily and quickly understood. Run-on sentences are a large source of cognitive load, so I imagine they would be the first things to go. Long horizontal texts are another cause of load, so maybe designers would eventually learn to present text in columns as opposed to letting text run from left margin to right.