Towards holistic human-computer interaction evaluation research and practice: development and validation of the distributed usability evaluation method

The traditional notion of usability localises usability at the system interface and does not take into account the context in which people use computer systems to do their everyday work. Recently, arguments have been made by a number of researchers that this notion of usability is outdated and inadequate because it fails to address the use-situation. Proposals have been put forward to extend our current thinking about usability to include the usefulness of systems. The usefulness of systems is manifested in the use-situation because usefulness cannot be understood outside the context in which the system is employed to perform real-life activities. To extend the traditional notion of usability to include the usefulness of systems, Spinuzzi (1999) introduced the notion of distributed usability which views usability as a property of humans interaction with a system, rather than a property of the system itself. Spinuzzi (ibid) argues that, instead of being localised at the system interface, usability is distributed across an entire activity that a human engages in using the system. This view of usability has significant implications for our current usability evaluation methods (UEMs) which are focused primarily on assessing the traditional usability of systems. As a result of this focus, the UEMs suffer from a number of problems and limitations, raising questions about their validity and reliability. This thesis aims to develop and validate a UEM based on distributed usability. The UEM has been named the Distributed Usability Evaluation Method (DUEM). It consists of four phases and is focused on assessing the distributed usability or usefulness of computer systems. Distributed usability is operationalised through the principles of Cultural Historical Activity Theory. Activity Theory is a powerful clarifying tool (Nardi, 1996b) for understanding and explaining human activities in context and, as a result, a suitable underlying framework for a UEM that aims to assess systems in this context. The validation of DUEM indicates that it overcomes most of the problems associated with current UEMs, however in the process of doing so it suffers from its own set of limitations.