Studying the Use of Digital Documents -
Towards a Socio-Cognitive View of Interaction
Andrew P. Dillon
School of Library and Information Science
Indiana University
Bloomington, IN 47405
adillon@indiana.edu
One of the major difficulties in assessing human use of interactive technologies is adequately capturing the process rather than the outcomes of use. The latter measures have dominated research to date and we know that reading electronically can prove slower, less accurate and more fatiguing (Dillon 1994). Although such findings are important, overcoming them cannot be managed without greater insight into the behavior and cognition from which they derive.
But capturing and interpreting the process of interaction with digital documents is far from simple. Much of the activity is silent, rapid and highly skill intensive. Furthermore, the resources of the users (or readers) are largely engaged in the act of information usage, rendering metanalysis or self-reporting intrusive.
While these are old research problems in reading research, the problems are multiplied in complexity when we enter the digital domain. Not only can the immediately available information space expand significantly, but the elements within this space may take multiple forms (text, photographs, diagrams, sound etc.) and incorporate animation and reconfiguration. Compounding these issues further is the fact that the human response to information technology can shift over time - and what appears difficult to use initially might prove empowering when mastered.
In this workshop I tried to outline how HCI research traditionally can be seen as existing at three levels. At level 1, people measure user performance in terms of efficiency, effectiveness and satisfaction, to determine what features of an interface are most usable, and to try and ensure that any design will satisfy user requirements. This is perfectly good practice but is theoretically weak as it rarely explains why certain interface characteristics lead to the measures that are taken.
At level two, theory enters the picture as various schools of thought seek to explain and predict user responses. To date the field has been dominated by formal model work drawing from the field of cognitive psychology, various anthropological approaches to work and task performance, and European activity theories from work study. None has produced a unifying framework, and most can be applied interchangeably depending on the theorist's persuasion to yield similarly imprecise predictions - with the exception of the formal modeling approach which has strong predictive power but for a very restricted range of user/task scenarios.
At level three, we seek theories that are less concerned with the explaining and predicting level 1 issues of effectiveness, but which seek to support the derivation of technologies that will augment human activities. So far, we have no such theories, even though the call for augmentation dates back to Engelbart in the 1960s.
In order to gain a handle on these problems and to move HCI toward level 3 issues, I proposed conceptualizing information usage over the short and long term in an attempt to bridge various research perspectives on human use and to develop better theory for design. In essence, I am seeking to extend traditional cognitive approaches to include social dynamics, what we may term 'socio-cognitive' systems design. In this approach we seek to understand the processes that shape behavior at the interface over both short and long-term interactions.
Three specific examples were presented. In the first, the relationship between cultural rules of discourse emergent in a community of speakers or writers/readers, and the perception of order in electronically presented documents - the emergence of genres. Here, long term effects of training and shared expectations lead people to significantly outperform non-members in tasks involving the determination of location of previously unseen information. None of the immediate perceptual variables of readability, cue presence/absence, priming etc. could overcome this single source of variance. Specifically, this has implications for the design of digital document sets in which users must navigate. From the perspective outlined here, the issue is less to study navigation, than to determine the perceived or expected shape characteristics of information.
In the second example, I showed how interface evaluations performed by users after exposures of several minutes to the interface can poorly predict how users will view a system hours and eventually days later. By developing two databases that differ in terms of usability, extended exposures by users over 12 days repeated use indicated that large significant differences between users on days 1 and 2 became unnoticeable in behavioral terms by days 9 and 10. However, despite the absence of behavioral measure distinctions, cognitively, these users had evolved different models and were differentially capable of generalizing their skills. Again, this work suggests typical usability evaluations are likely to be limited if reliant on limited exposures, and if they emphasize exclusively observable performance. The process of knowledge acquisition is not directly correlated with behavioral measures in interaction, yet interface variables produce strong cognitive effects even when users appear to be indistinguishable from each other.
Finally I outlined findings from the socio-cognitive analysis of acceptance, where we seek to model the determinants of acceptance or rejection by users at the earliest stage possible. Modifying existing theories of acceptance, it is proposed that early estimates by users of an artifact's usability and utility offer reliable indices that can predict subsequent discretionary usage.
In sum, this approach seeks to de-emphasize outcome measures and to examine directly the underlying socio-cognitive processes of human-computer interaction so that we might build a more complete picture of the dynamics of interaction with and acceptance of digital libraries.
Dillon, A. (1994). Designing usable electronic text: Ergonomic aspects of human information usage. London: Taylor & Francis.
Last Updated: Feb. 17, 1997