Not being there: a pragmatic approach to workplace studies


Lydia Plowman
School of Cognitive and Computing Sciences
University of Sussex
BRIGHTON BN1 9QH


CSRP 341



It is increasingly the case in the UK that research has to be carried out with
an industrial partner if it is to gain funding from the research agencies. The
government white paper on research funding, Realising our Potential (1993),
stresses wealth creation as a desired research outcome and the ensuing need for
contacts between industry and academia in the fields of science, engineering and
technology. These demands extend to many of the EC funding programmes and are
explicit in the DTI's CSCW Advanced Technology programme, which requires the
main consortium partners to be from industry.

The thwarted ethnographer and industry
Ethnographic methods have been associated with the emergence of the concept of
CSCW because their emphasis on the observation and description of a range of
cultures has an obvious affinity with the need to understand the organisational
context in which the proposed system will be used. As academic/industrial
partnerships become increasingly common, particularly in the field of CSCW,
there is an urgent need to consider alternative approaches to workplace studies
and how they can inform system design. In this paper I will 1) outline some of
the problems which arise from undertaking the empirical study of work settings
in the context of a formal collaboration agreement with an industrial partner,
2) attempt to contain the zeal for ethnography in such contexts, and 3) suggest
ways in which some of the precepts of ethnography can be adapted for use within
the workplace, redefining these methods as 'interpretive'.

I will draw on my experience of several studies in the field of CSCW undertaken
since 1988, all of which had the purpose of providing design guidelines for
systems . They have all started from a human-centred perspective, but have
become progressively less 'ethnographic' in orientation as commercial demands
have become more explicit. During doctoral research we have more freedom to
invest the requisite time and energy in grounded theory and the willingness to
risk an approach which is a gamble in terms of translatable findings. If
anything, doctoral researchers spend too much time driven by a constant quest to
satiate their desire to collect enough data. By the time researchers are working
on industrially sponsored projects, accounts rarely disclose exactly how much
time was spent in the field because it's always less than necessary. Time
constraints mean that projects require findings as soon as possible, so that
designers can get on with implementing a prototype or making decisions about
preferred platforms - to the extent that in the present project these decisions
have been made without waiting for an analysis of the initial case study. This
means that findings can be so general and high level as to be almost useless,
and the only realistic option is to offer designers an interpretation of some of
the issues.

Ethnographic methods are more usually associated with the generation than the
verification of theory and it is questionable whether such an approach is useful
for informing system design given that a statement of design guidelines or
issues is the ultimate goal of such studies. The purpose of workplace studies is
certainly much more constrained by the demands of system design than traditional
ethnographic studies, which have a more nebulous purpose. This should make the
task more straightforward...

Being there
There have now been plentiful statements of what constitutes an ethnographic
approach in the workplace, but the main prerequisite for detailed observation is
'being there'  - a state which can be quite difficult to achieve in many
circumstances. Many of the ethnographic studies so far have been based on
studies in which personal contact has been used to gain access (Rogers,
forthcoming), or are of a relatively closed environment, such as a group of
graphic designers (Murray, 1993), architects (Luff & Heath, 1993), or a city
dealing room (Heath et al, 1993). Others use an educational environment to which
the researcher has uncomplicated access (Plowman, 1991 & 1993), or use
organisations in the public sector, which are less constrained by protecting
commercial confidentiality, such as the London underground (Heath & Luff, 1992),
or an air traffic control centre (Hughes et al, 1993).

The emphasis of ethnographic methods is on natural settings, and identifying
appropriate methods of data collection and analysis in situ, during a period of
observation and, where appropriate, participation in work practices. One problem
with this approach is that it is extremely labour intensive. Many projects
allocate one person to undertake the workplace studies whilst the rest of the
team are more directly involved in system design and implementation  so it is
unrealistic to appropriate methods which rely on total immersion in a culture
over a long period. Because such an approach leads to an overwhelming quantity
of potential data, some assumptions have to be made early on about the
parameters of the study and CSCW now has a number of established candidates for
this focus, such as the use of shared artefacts, mediating representations, and
communication practices. These methods also depend heavily on the cooperation of
the institution and the participants. It is often the case that the industrial
partner is situated at a considerable distance from the researcher (in my case,
two sites, each of which is several hours journey time from base) so visits have
to be pre-arranged, are costly, and infrequent.

Furthermore, once access has been negotiated would-be ethnographers have a
marginal role and there are pressures on them to take the line of least
resistance (Hammersley, 1983). As a person with privileged access to the site,
it is easy to feel that one shouldn't get in the way or be too demanding in
terms of further access and there is a constant need to be adaptable. This can
be heightened by the widespread view in industry that academics have fewer
constraints on their time and can be more accommodating of last minute demands
or changes.

Not being there
The case study on which I have been working for the last nine months has been
allocated by the company without consultation with the academic partners and
with greatly restricted access to information. Case C is almost completely
atypical and has the great disadvantage for an ethnographer of being over before
'being there' was a possibility.  This means that the researcher is restricted
to retrospective accounts of events, leading to obvious problems in terms of the
status of recall, finding relevant memos or sketches, and key personnel leaving
the company.

One of the main functions of the proposed tool, which will run on laptop PCs, is
for salespeople to elicit the customer's requirements by building up a block
diagram of the equipment required. A coherency checker will ensure the system's
functionality. The opportunity to accompany salespeople would provide insights
into the design of the proposed tool and how it could be used, such as the
extent to which the customer would have a shared view of the screen and what
information should be available. Access to these meetings is not available as
the information is considered commercially sensitive.

Interviews with sales staff have established that each person has their own
methods for selling systems which have been refined as a result of experience.
One view is that if the company represents itself as selling technical
expertise, it is counterproductive to ask the customer interminable questions
and a prescribed sequence could hinder the relationship the sales rep is trying
to establish. There is a general feeling that the customer should see the design
emerging and how the costs accumulate, but management are reluctant even to
allow a rough order of magnitude price before the engineers have costed the
system. Thus findings which represent the users' wishes may be ignored.

Difficulties such as these lead the researcher to adopt a combination of
pragmatism and intuition when working in the field and interpreting data which
isn't necessarily acknowledged in accounts of the research. It is rare for our
research practices to be entirely in harmony with our methodological
objectives.

An interpretive approach
The situation described here is probably more widespread than the literature
would lead us to believe. From this realistic, if jaundiced, view, pragmatic
considerations assume more importance than adherence to a coherent theoretical
stance. Where the researcher is dependent upon the goodwill of the company to
gain access and undertake the research, it is likely that compromises will have
to be made between the parties' desired outcomes.

Erickson (1983) refers dismissively to the romantic 'toothbrush and hunting
knife' school of ethnography, claiming that explicitly framing research
questions and deliberately seeking relevant data enable and empower intuition,
rather than stifle it. He suggests that 'interpretive' is a more useful
description of these research methods as it is more inclusive than ethnography
as a term and points to the key features of such research - elucidation and
exposition by the researcher. The term presupposes a certain level of
observation and description, but it does not have to conform to the definition
of this suggested by ethnography. Although he refers to educational research,
the limitations on conventional ethnography in workplace settings make an
'interpretive' approach particularly applicable in this context.

The emphasis on interpretive skills seems a much more honest description of the
researcher's role. Rather than agonise over the many ways in which research
methods fall short of the requirements of ethnography, it is much more
straightforward to acknowledge that restrictions on access to sites, to people,
and to artefacts mean that the researcher will take a pragmatic approach in
which various methods are used depending on which promises to be most rewarding.
In terms of informing system design, there is usually some guidance on which
areas need to be investigated, from the system developers, the system users, or
the researcher's experience.

Using this approach for case C, I made a couple of reconnaissance visits to get
a feel for the workplace and to meet some of the people I would later need to
contact. This was followed up by tape-recording interviews with the main people
involved, initially using an almost unstructured, conversational interview
technique. Relevant materials, such as diagrams and notes were also collected.
For subsequent visits I had transcribed sections of the interview and identified
some of the issues for further discussion. In the interim there would be
occasional email or other written contact in which I made specific enquiries,
the feedback also becoming a source of material for further analysis. Where a
site visit was not warranted I used telephone interviews, sending the person
concerned a list of possible topics for discussion in advance. Telephone
interviews can be adequate if you have already met the person face-to-face, and
if they do not need deictic communication, such as an engineer referring to the
log book or pointing to information on the computer screen. The interim report
on the case study was circulated to all participants in addition to the project
partners, and feedback on it was solicited.

There is nothing unusual about these methods - they are all part of the
researcher's armoury of techniques, but there is no reason why researchers in
the field of CSCW should apologise for using them simply because they fall short
of a strict interpretation of ethnography. In a sense, this is perpetuated by
using terms such as 'quick and dirty' ethnography, a misnomer for situations
such as the one I have described, in which the researcher is mainly in absentia.
Goguen (1993, p144) refers to using ethnography to get 'prior orientation to the
milieu to be studied' but this seems unrealistic. He recommends a 'zooming'
method of requirements elicitation, whereby the more expensive but detailed
methods are only employed selectively for problems that have been determined by
other methods to be especially important. Whereas this 'zooming' method has a
lot to recommend it, it is mistaken to work on the basis that this is a
pragmatic solution as ethnography is not a method to be used prior to something
else. It is the ethnographic approach which is so time consuming and depends on
'being there'.

The question of what exactly is contributed to system design remains. Where
research is being undertaken with or for an industrial partner it is likely that
some recommendations may, in any case, be ignored. Probably all one can do is
hope to highlight particular issues which may not have been apparent to the
system developers (who very rarely get the opportunity even to visit the sites
of use) or to elicit previously unarticulated explanations of work practices and
preferences. I am not suggesting that recommendations for system design such as
these couldn't be extrapolated from an ethnographic approach, but this would
probably be after a much greater input of researcher time and in a situation
where fewer company imposed constraints existed. Interpretive workplace studies
need to be aware of situatedness and indexicality, for instance, and these
observations are in large part due to previous ethnographic studies of the
workplace. There is still a place for detailed, labour intensive ethnographic
studies to develop our understanding of cooperative work practices and how they
may best be supported, as their findings can inform the interpretive approach
outlined here. But the future direction of research is likely to be in
partnership with industry, and in such contexts an interpretive approach offers
a more pragmatic solution.

Acknowledgements
The author gratefully acknowledges the contribution of participants in the case
study described. The CORECT project is partially funded by UK SERC grant
GR/J/53454. Among the partners on this project are Intelligent Applications
Ltd., and the Universities of Edinburgh and Sussex.

References
Cooper, Geoff, C. Hine, J. Low & S. Woolgar (1993). Ethnography and Human
Computer Interaction. CRICT discussion paper no. 39, Brunel University.

Erickson, Frederick (1983). Qualitative methods in research on teaching. In
Handbook of Research on Teaching (3rd. ed.), MC Wittrock, ed. Macmillan : New
York.

Goguen, Joseph A. (1993). Towards a social theory of information. In Proceedings
of an invitational workshop on social science research, technical systems and
cooperative work. March 8-10, 1993. Paris, France.

Hammersley, M. (1983). Reflexivity and naturalism in ethnography. In The
Ethnography of Schooling: Methodological Issues, ed. M. Hammersley. Nafferton:
Driffield.

Heath, C., M. Jirotka, P. Luff & J. Hindmarsh (1993). Unpacking collaboration:
the international organisation of trading in a city dealing room. In Proceedings
of the Third European Conference on Computer Supported Cooperative Work, Milan,
Italy.

Heath, Christian & Paul Luff (1992). Collaboration and control: crisis
management and multimedia technology in London underground line control rooms.
Journal of CSCW 1 (1-2) 69-94.

Hughes, John, D. Randall & D. Shapiro (1993). From ethnographic record to system
design. Journal of CSCW 1 (3) 123-141.

Luff, Paul & Christian Heath (1993). System use and social organisation.
Observations on human-computer interaction in an architectural practice. In
Technology in Working Order - Studies of Work, Interaction and Technology, ed.
Graham Button. Routledge: London.

Murray, Dianne (1993). An ethnographic study of graphic designers. In
Proceedings of the Third European Conference on Computer Supported Cooperative
Work, Milan, Italy. pp.295-309.

Plowman, Lydia (1993). Tracing the evolution of a co-authored text. Language and
Communication 13 (3) pp.149-161.

Plowman, Lydia (1992). An ethnographic approach to analysing navigation and task
structure in interactive multimedia: some design issues for group use. In People
and Computers VII, eds. A. Monk et al, pp.271-287. CUP: Cambridge.

Rogers, Yvonne (forthcoming). Exploring obstacles: integrating CSCW in evolving
organisations. To appear in Proceedings of CSCW '94, Chapel Hill, North
Carolina.