Assistants
During 2007-2008 we carried out an empirical study of the type and nature of work performed by Administrative Assistants in IBM’s research organization. The practical goal of the work is to discover ways of measuring the amount of work carried out, so that assistance can be better allocated. The conceptual goal of the research is to understand the nature of assistance – interestingly enough, although intelligent assistance has been a theme in information technology for years, no one has actually ever looked at how people do it.
The purpose of this work is to summarize the nature and substance of the work of assistants based on studies of IBM assistants and a review of the literature. It turns out that little research has been done by the Information Technology community on administrative assistants, perhaps because their work has been viewed as routine. However, a large body of research in the social sciences shows that seemingly routine tasks turn out to be surprisingly complex, fraught with errors and exceptions that require considerable local and technical knowledge to rectify. The research findings – derived from two lines of work – are consistent with this view.
The first line of research consisted of a series of interviews of assistants and other stakeholders at IBM Research, based on interviews with assistants, and their principals and managers in IBM Research. We described a model of the work of assistants that illustrates the ways in which they act as articulation workers – that is, managing the logistics (who does what, where, when and how) of their principals’ tasks, and the intensive involvement and knowledge necessary to pull this off effectively. Of particular note is the importance of assistants’ relationships with other assistants (e.g., in scheduling meetings), the degree of autonomy that their deep knowledge of their principals’ priorities and tasks enables, and their involvement in the day to day activities of the principal’s office (which is important to facilitating the first two items).
The second line of research consisted of quantitative analyses of communication among assistants gleaned from automated analysis of their email and instant messaging, and of their principals’ calendars. Here we described the methods and tools used, and presented a model of the assistants’ communication network. The work here was aimed at quantifying the amount of articulation work carried out, and providing a partial basis for the allocation of assistance.
For more information
• Erickson, T., Danis, C., Kellogg, W. A., Helander, M. E.. Assistance: The Work Practices of Human Administrative Assistants and their Implications for IT and Organizations. Proc. CSCW ’08. ACM Press, 2008.
• Contact: Tom Erickson
