Models For Research into Decision‐MakingProcesses: On Phases, Streams and Decision‐Making Rounds
| Author | Geert R. Teisman |
| Published date | 01 December 2000 |
| Date | 01 December 2000 |
| DOI | http://doi.org/10.1111/1467-9299.00238 |
MODELS FOR RESEARCH INTO DECISION-
MAKING PROCESSES: ON PHASES, STREAMS
AND DECISION-MAKING ROUNDS
GEERT R. TEISMAN
This article elaborates on the question of how complex decision making can be
analysed. Three conceptual models are compared: the phase model, the stream
model and the rounds model. Each model is based on specific assumptions about
what decision making is and how it should be analysed. The phase model focuses
on successive and distinctive stages in a process, i.e. defining a problem, searching
for, choosing and implementing solutions. The stream model emphasizes concur-
rent streams of participants, problems and solutions, defining decision making as
the connection between these streams. The rounds model combines elements of the
other two models, in assuming that several actors introduce combinations of prob-
lems and solutions, and create progress through interaction. Each model generates
specific insights, as is shown from the example of the ‘Betuwe line’, a railway line
intended for the transport of cargo, in the Netherlands. The phase model concen-
trates on decisions taken by a focal actor; the stream model focuses on the coinciden-
tal links between problems, solutions and actors; and the rounds model on the inter-
action between actors.
1 MODELS FOR THE RECONSTRUCTION OF DECISION MAKING
Public administrationists agree that decision making has become more com-
plex. Several reasons can be identified for this increased complexity. Two
important ones are: increased uncertainty about the global economy and
the rise of the power-sharing world or ‘network society’, where nobody is
in charge (Bryson and Crosby 1992; Kickert et al.1997). This article is based
on the assumption of increased complexity. Complexity raises the question
of how researchers should handle this problem (Butler 1991; Mintzberg
1973; Teisman 1992, 1995, 1998). Furthermore, the question of complexity
is part of the wider discussion on governance in networks (for example: in
Germany, Marin and Mayntz 1991 and Scharpf 1997; in France Crozier and
Friedberg 1980; in Great Britain, Rhodes 1996a and in the USA, E. Ostrom
1990; Smith 1998)
This article focuses on the question of how to depict decision making in
societies that are confronted with network structures. I will concentrate on
the characterization of successive decision-making activities and of concur-
rent decision-making activities. I will also discuss the assumed relations
Geert R. Teisman is Professsor in the Department of Public Administration at Erasmus University
Rotterdam.
Public Administration Vol. 78 No. 4, 2000 (937–956)
Blackwell Publishers Ltd. 2000, 108 Cowley Road, Oxford OX4 1JF, UK and 350 Main Street,
Malden, MA 02148, USA.
938 GEERT R. TEISMAN
between the activities, in sequential as well as parallel combinations. The
following research questions can be formulated:
(1) Which assumptions are made with respect to chains of activities in
decision-making processes? I will analyse different criteria to dis-
tinguish between: strands of activities, typologies of decision making
as a whole and elements of this process, and signs of progress in
decision making.
(2) What sorts of assumptions are made with respect to simultaneous
activities in decision-making processes? Specifically, I will distinguish
here between (more or less relevant) actors, the relationship between
problems and solutions, and the content of decision making.
To analyse decision making, the researcher needs to make a reconstruc-
tion of the study object. Such a reconstruction will be selective in nature.
Observation is not simply an effort to learn what is going on. Rather it is
a process where observations are made to conform to sets of assumptions
(Edelman 1971). The gathering and classification of empirical observations
into meaningful information is based on the a priori images of decision
making used. We cannot depict decision making without making assump-
tions about its appearance. Various terms are used to describe such a set
of assumptions: model, image, metaphor, referential framework or method-
ology. In this article, the term ‘model’ will be used. Models help us to
understand decision making in distinctive yet partial ways (Morgan 1997,
p. 4).
Three models will be discussed in this paper. Two of these are generally
accepted and respected, i.e. the so-called phase model and the stream model.
The phase model is the most common approach, both in science (Anderson
1979; Bryson and Crosby 1992) and in policy practice (procedures are often
based on the concept of phasing). Decision making is represented in terms
of a number of distinct stages (Mintzberg 1976). Phase models distinguish
between (at least) policy formation, policy adoption and policy implemen-
tation. Each phase has its specific characteristics and participants. Minis-
tries, for instance, are often divided into departments that are responsible
for policy formation and others responsible for implementation.
The stream model depicts decision making as a combination of three
separate concurrent streams (Kingdon 1984). One stream consists of prob-
lems, another of policies/solutions, and a third one of politics/participants.
Like the phases, streams have their own characteristics, but they exist side
by side. A decision becomes the coincidence of streams.
In this article I will emphasize a third conceptual approach to decision
making, the so-called rounds model. In this model, decision making is
assumed to consist of different decision-making rounds. In all sets of
rounds, the interaction between different actors results in one or more
definitions of problems and solutions. All participants can score points in
each round, in terms of a leading definition of the problem and the
Blackwell Publishers Ltd. 2000
To continue reading
Request your trial