Phase 4: Analysis (Months 7-11)
All data were subject to a preliminary coding, in which we identified distinct episodes of
observation, themes in interview data and email trails. A set of categories and hypotheses
was drawn up, and all the data were subjected to a further coding on this basis. Once
these categories were refined, we undertook a final coding, as well as synthesising diary
and memo entries into provisional analytic descriptions of episodes and themes. Particular
attention was paid to “breakdowns” and decision points; these were, as we had predicted
in the proposal, informative, although given the limited time span of the project, they
were not numerous.
The preliminary coding progressed from attaching codes to individual utterances or
events, into pattern coding of a smaller number of themes. When these themes were
finalised, we returned to the corpus of data as a whole, and triangulated our initial
findings with our primary contact D at the firm, and with some other of the engineers. We
also asked the other engineering firm, and one of the engineering academics, to review
our conclusions, and elaborated them on the basis of feedback received. Finally, we were
fortunate to be invited in month 7 to present our provisional findings to a panel of expert
users (see Section 2), providing an additional opportunity for feedback and revision.
4. Results
These results are arranged in sub-sections, which begin with observation and description
and shift in emphasis through to analysis and theoretical findings.
An overview of the structural engineering design process
An engineering design plan for a building — that is, a plan carried out in advance of its
construction — is typically very complex, and created over months or even years. There
are usually three main stages of design (termed ‘concept’, ‘scheme’ and ‘detailed’) which
represent increasing levels of detail and commitment to the design, from a fuzzy initial
idea to a clearer scheme design, where major decisions have been made but many details
remain undecided, to the final design with all the details elaborated. The mathematical
components of the design work vary widely across the stages and the particular roles of
individuals and groups within the design team.
Within each stage of design, we have classified the structural engineers’ work into three
main areas, labelled in Figure 1 (above) as design, analysis and review. The interplay
of design and review in structural engineering practice (in the form of meetings and
discussions both formal and informal) was found to be largely routine, and not
significantly different from practice within other industrial and commercial settings. For
the current study, the distinguishing characteristic was in the area of analysis: that is,
carrying out the calculations for a design (whether by computer or manually), separate
from design itself. It was found that this separation was ubiquitous and explicit: as one
engineer put it, “when I’m analysing I’m not designing, I’m doing the calculations that
justify my design”.
The separation leads to a significant division of mathematical labour in design projects: it
is younger engineers who are performing the majority of the analysis (especially
computer-based), whilst more experienced engineers tend to handle the broader tasks of
design. This division arises because there is a standard career path by which the younger
engineers are apprenticed over a period of years into the practice of engineering design,