Wood, S., Cox, R., and Cheng, P. Attention design: Eight issues to consider. Computers in Human Behavior 22 (February 2006), 588–602. [pdf]
This paper describes eight issues that are fundamental when designing attentive systems. First of all there is no consensus on what is attention. Many models are proposed that tend to represent it as a spotlight or as a ‘coherence field’, or a ‘dymanic telephoto system’. Most research tend to assimilate attention with visual search through eye-gaze. However this is not necessarily correct.
Second point attention is difficult to measure. Vision has a selective nature. The implication is that direction of gaze is not necessarily a synonym with focus of attention and therefore studies will need to validate focus of attention through further evidence.
The third point raised in the paper is to understand how graphical displays interact with attention. Larkin and Simon (1987) demonstrated that the way an extrernal representation encodes information is critical to how easy it will be for the user to find relevant information. This is connected to the split-attention effect. Shimojima (1999) claims that ‘free rides’ and ‘derivative meaning’ occurs when constraints on the relation between local and global structure in the representation and the target domain, are satisfied by the semantic conventions of the representation.
A fourth point of interest is what are the potential effects of introducing artificial feedback in systems designed to monitor the user’s attention. It has been demonstrated that the introduction of an an artificial feedback loop can, in some circumstances, cause variables that are ussually highly correlated to become decoupled.