188 7: Decision Th eory
because there is order in loosely coupled decisionmaking processes, but it is not
conventional order.
Th e best-known alternative explanation of order in loosely coupled settings is
garbage can theory. Th e famous Cohen, March, and Olsen description of organi-
zations as decision garbage cans is this: “An organization is a collection of choices
looking for problems, issues and feelings looking for decision situations in which
they might be aired, solutions looking for issues to which they might be the an-
swer, and decision makers looking for work” (1972, 2).
Th is is a distinctly process-oriented description of decision theory. In the “de-
cision soup” there will be institutional competencies and social or political needs
and preferences. Under the right circumstances, competencies and needs will fi nd
each other, bond, and thereby signifi cantly modify or adapt institutional arrange-
ments, preferences, and decision processes. In the conventional decision model,
means are applied to ends. In the garbage can, it is just as likely that ends will be
applied to means. It is important to note the relative unimportance of effi ciency
or rationality in this conception of decision theory. Th e garbage can theory of de-
cisionmaking may not be rational in the traditional means-ends understanding of
rationality, but under certain circumstances it “makes sense.” Such sense making
is retrospective, the sense derived from looking back. “Doing something requires
such an active and immediate engagement with the objects of our attention that
only aft erwards are we able to stop and refl ect on, to ‘see,’ what we have done”
(Harmon and Mayer 1986, 355).
One might assume from this that decisions and actions are accidental, random,
purposeless, and chaotic. Not so, argue decision theorists from the sense-making
perspective. Decisionmaking is less a process of rational choices and more a pro-
cess of the temporal mixing of decisions and actions and the attendant decision
processes of enacting the future (Yanow 1996).
Perhaps the best-known application of garbage can theory in the public sec-
tor is John Kingdon’s Agendas, Alternatives, and Public Policies (1995). He de-
scribes three essentially parallel but independent streams: the political stream,
the policy stream, and the problem stream. Triggers can cause the streams to
fi nd each other in windows of opportunity. Triggers include changes in the col-
lective understandings of problems, changes in political power, possible new
ways of dealing with problems, or a focusing event. Using these streams and
the garbage can metaphor, Kingdon describes a range of policy changes that
resemble the patterns of attraction between decisionmakers, problems, and
solutions, and the eventual “solution” to problems. As March suggests, garbage
can theory is essentially a temporal sorting process under conditions of very
loose coupling. At some point, the attention of decisionmakers may be uniquely
focused on a particular problem; this is especially important because attention
is scarce and carefully rationed. If a problem has achieved attention, the ques-
tions change to defi nitions of the problem and its possible solutions. If, how-
ever, a possible solution is already available and there is agreement regarding