ACCOUNTING AND ORGANIZATIONAL CULTURES 341
dialogue noted. Activities in these meetings were subsequently written up in
abbreviated form. Thirdly, data were collected through casual conversations and
by simply ‘‘being around’’.
As the project progressed, the data were repeatedly analysed. At first, it
seemed that there was a real probability of drowning in the data. Transcripts and
notes were accumulating rapidly, and the material appeared to be incoherent.
This is apparently a common feature of the initial stages of cultural research
(Smircich, 1983b). But different ways of making sense of it all were explored and
gradually a pattern began to emerge. At each stage emerging appreciations were
checked against the next round of data in an attempt to confirm the researcher’s
understanding of the situation. This continued until such time as the subsequent
data became predictable.
It may be useful to indicate the precise way in which the data were found to
give a coherent picture. Ultimately, the analysis hinged on three dimensions: role
(function), and level of hierarchy of the subject, and time. Firstly, the data were
categorized according to content and underlying values. Opinions, sentiments,
interpretations, confusions and so forth in each interview transcript were noted.
This was done without reference to the identity of the person who had been
interviewed. The data collected at any one time seemed to fall naturally into
distinct constellations or clusters (in the sense that groups of people expressed
similar views). Attaching identity to the data, it transpired that these constellations
corresponded broadly to interviewees’ roles and positions in the hierarchy. Among
those performing similar roles (functions) at a similar level in the hierarchy, there
was a marked similarity of perspective. Perspectivesdiffered, however, across
roles and at different levels of the hierarchy. These data were set up on a
two-dimensional space with role on one axis and hierarchy on the other.
This exercise was repeated on data collected at different times, and found
to give similar results, in the sense that the data again fell into role-hierarchy
clusters. The specific content of the data differed over time, however. So, the two-
dimensional spaces were set out in chronological order. In effect, a third dimension
was added to the space, representing elapsed time. Studying the content of the
data as one moved through time, it transpired that the opinions, sentiments and
interpretations of each group were in fact evolving in a systematic way. In this
three-dimensional space was a story of unfolding meanings in the organization.
This was indicative of the existence of different cultures in the organization, and
some systematic underlying trajectory in the emergence of those cultures. In fact,
during the data collection process, it became clear that new interviewees’ views
were predictable, given a knowledge of (a) the role (function) of the participant,
(b) his or her position in the hierarchy, and (c) the time of the interview.
At this point, the data were analysed from a different perspective. Specifically,
the level of analysis shifted from content to process. The data were re-examined
to see if the process through which the new meanings were emerging was
observable. Some key turning points were obvious in the data. There was a series
of events and interactions through which the emergence of the new meaning
structure could be traced. These are documented in following sections. Finally, the