Observing loosely coupled organisations is not as
straightforward as it may first appear. Weick (1976) (p. 9)
argues “that interaction data – who speaks to whom about
what – are unlikely to reveal loose couplings. . .[as] people
tend to overrationalize their activities and to attribute
greater meaning, predictability, and coupling among them
than in fact they have.” Loose coupling can only be identified
if a methodology is adopted that “preserves rich
detail about context”. “A contextually sensitive methodology
would record both the fact that some people are in
one place generating events and the fact that some people
are thereby absent from some other place” (Weick, 1976,
p. 10).
Therefore, the research approach adopted had a number
of distinct stages, deliberately employing a mixed
methodology “to maintain the researchers’ sensitivity to
context-specific variations in meanings” (Modell, 2009, p.
208). A further justification of this blend of research methods
can be found in Modell (2005).
After reviewing the literature in the area to enable some
prior expectations to be established, a number of scoping
interviews were conducted. A questionnaire was then prepared,
based on both the prior literature and the interviews,
for distribution to all the Trusts in Wales. Throughout
this process guidance and assistance was received from
the Welsh Healthcare Financial Management Association
(WHFMA).7 The members of this committee facilitated
access to individuals, commented on drafts of all documents
produced and encouraged colleagues to respond.
The initial interviews were held on a totally informal
and unstructured basis and not recorded verbatim. This
process was chosen to allow respondents to talk freely and
in an unconstrained manner to ensure that all issues, no
matter how controversial, could be raised. This style of
interview has been carried out in accounting research in