Glasgow Interactive Systems Group (GIST),
Department of Computer Science, Glasgow University,
Glasgow, G12 8QQ.
Tel: +44 141 330 6053
Fax: +44 141 330 4913
EMail:johnson@dcs.gla.ac.uk
What do we expect from a PhD in Human Computer Interaction? This brief paper summarises a range of possible responses to this question. The final section goes on to argue that alternative means of assessment must be considered if we are to apply the multi-disciplinary criteria that many people use to define a PhD in HCI.
Keywords: Human Computer Interaction; Education, PhDs.
There have been a variety of attempts by professional organisations, including the BCS and the ACM, to guide the syllabus for undergraduate courses in Human Computer Interaction (Gasen, 1995). Unfortunately, these initiatives have not considered what constitutes a 'good' PhD in HCI. The inter-disciplinary nature of our subject makes it difficult to identify clear guidelines or standards. The criteria imposed in the field of Psychology cannot easily be applied to assess the quality of work in the field of Computer Science or Sociology. In consequence, external and internal examiners from individual disciplines cannot easily judge the quality of work that may draw upon several different fields. This paper, therefore, argues that we must find alternative mechanisms for examining PhD theses in HCI.
Many Universities require that PhD candidates contribute to their field of research. For example, a doctorate in the field of Computing Science might be awarded for improving the efficiency of distributed memory management or increasing the precision and recall of an information retrieval system. Similarly, a PhD in Psychology might be awarded for results that confirm or refute hypotheses generated from previous experimental investigations. Unfortunately, things are not so simple within the field of human computer interaction. Many people challenge the idea that there is a 'theory of HCI' (Dowell and Long, 1989). This view argues that we are part of a craft discipline. Advances come through experience and practical application rather than through the steady development of a field of research. If this is true then it makes little sense for PhDs to contribute to a mass of theory that nobody uses.
A second criteria for a doctorate in HCI is that the candidate shows a proper grounding in experimental techniques. Unfortunately, this raises the question of what techniques are appropriate for a thesis in Human-Computer Interaction? There seems to be little agreement about the precise nature of experimental techniques that are appropriate within the field of HCI. Lab-based techniques remove the contextual factors, such as co-worker interference and commercial pressure, that have a profound impact upon the 'usability' of interactive systems. Contextual techniques and ethnomethodology cannot easily be used to support the focused hypotheses and unambiguous conclusions that are, typically, required in academic theses, especially within Computer Science and Psychology.
A third criteria that might be used to judge a 'good thesis' in HCI is whether the candidate contributes to the design and implementation of interactive systems. The difficulty with this approach is that it often reveals a clear division between academic and commercial views of human-computer interaction. For instance, an academic interpretation of this criteria might centre upon the strengths and weaknesses of the Seeheim model. A commercial view might focus instead upon the cost-effectiveness of VisualBasicTM compared to other implementation platforms. A related point is that PhDs on the design of interactive systems must also face similar problems to those mentioned in the previous section. The toy problems and limited case studies that are used to analyse user interface toolkits and design methodologies bear little relationship to the ill-defined and poorly structured problems that frustrate the commercial development of real-world interfaces. Again, such applications cannot easily be written up for presentation during PhD vivas.
The weakest condition that might be used to assess PhDs in HCI is whether or not they demonstrate an understanding of inter-disciplinary research. This is a dangerous criteria because candidates must, typically, be examined within Departmental or Faculty structures. It can be difficult to recruit well qualified panels that also satisfy the eligibility criteria for Computer Science, Psychology or Sociology. Multi-disciplinary committees create different problems. For example, it seems unreasonable to expect candidates to attain PhD standard in more than one discipline within the three years that most funding bodies will support. This implies that we may pass multi-disciplinary theses that would not be acceptable within a 'parent' discipline.
This brief review has summarised some of the criteria that can be used to assess the quality of PhDs in human computer interaction. Our survey reflects the tensions that exist within the field. For instance, the debate over whether HCI should be regarded as a craft or as an engineering discipline is a research topic in its own right. The critical point here is that these tensions create pragmatic and theoretical problems for the people who must defend their work in PhD. vivas. Experimental work must be defended against accusations that it fails to explain real-world behaviours. Design innovation must be defended against criticisms that few industrial designers use the products of academic Computer Science. Inter-disciplinary research must be defended against technical criticisms drawn from each of the parent subjects.
I would argue that we need to reassess the criteria that are used to judge PhDs in HCI. A particular concern is that other people should be able to use the products of our research. Experimental theses must be backed up by the real-world analysis of situated interaction. Formal and methodological theses must be supported by evidence to show that designers can actually exploit theoretical results. Inter-disciplinary research should demonstrate that practitioners in one area should be able to exploit the products of another. Until such criteria are widely accepted then the best advice that we can offer to our students is that they should prepare for the standard questions that are asked at almost all HCI vivas. Would your results replicate in a real-world setting? Could real-world designers actually use your formal notation? Will your technique actually result in a 'more usable' interface?...
Jean Gasen, Support for HCI Educators: A View from the Trenches. In M. Kirby, A.Dix and J. Finlay (eds.), People and Computers X, pages 15-20. Cambridge University Press, Cambridge, 1995.
J. Dowell and J. Long, Towards A Conception For An Engineering Discipline Of Human Factors, Ergonomics,(32)11:1513-1535, 1989.