Lookup NU author(s): Professor Peter Wright,
Professor Michael Harrison
Full text for this publication is not currently held within this repository. Alternative links are provided below where available.
Organizational context is now accepted as a central concept in attempts to understand error in human–machine systems. However, accounts which emphasize the processes of everyday organizing, such as accountability and work activity, are needed in order to establish organizational requirements for design. In this article, we provide a framework for the consideration of organizational contexts of human error in high-consequence work systems, with a view to integrating empirical insights and supporting practical design work. We draw on computer-supported cooperative work conceptualizations of the process of everyday organizing, particularly the notion of “accountability for work activity” which is pivotal to our organizational account of error. The conceptual framework is characterized here as a set of dimensions which are expressive concerning the relationship between accountability and work activity in different contexts: (1) explicit–implicit; (2) global–local; (3) stable–transient and (4) dependent–independent. The framework is demonstrated with respect to everyday work practices in a radiology department and its analytical utility validated with respect to two documented aviation system failures. Applying the framework has enabled us to identify and define, in terms of the dimensions, a number of contexts for vulnerability in high-consequence systems: contexts for collusion, violation, deference, loss of control, buck passing and complacency. These are discussed in terms of requirements for error-tolerant design. In the final section of the article, links between the various contexts for vulnerability and the design process are explored.
Author(s): McCarthy JC, Healey PGT, Wright PC, Harrison MD
Publication type: Article
Publication status: Published
Journal: International Journal of Human-Computer Studies
ISSN (print): 1071-5819
ISSN (electronic): 1095-9300
Publisher: Academic Press
Altmetrics provided by Altmetric