Int. J. Social Research Methodology
Vol. 8, No. 4, October 2005, pp. 345–353
ISSN 1364–5579 (print)/ISSN 1464–5300 (online) © 2005 Taylor & Francis
DOI: 10.1080/1364557042000203116
A Different Kind of Snowball:
Identifying Key Policymakers
Karen Farquharson
Taylor and Francis Ltdtsrm100258.sgm
Received 14 July 2003; Accepted 28 October 2003
10.1080/1364557042000203116International Journal of Social Research Methodology364-5579 (print)Original Article2 0Taylor & Francis Ltd80 0002004K re FarquharsoSchool of So ial and Behav oural SciencesSwinburne University of TechnologyPO Box 218HawthornVictoria 3122Australiakfa quh son@swin.edu. u
Studying policy networks raises challenges in three important areas: identifying members
of the policy network, gaining access to the network, and reporting findings from the study
while maintaining confidentiality. Using the tobacco control policy and health policy
networks in Victoria, Australia as a case study, the article describes how to use a reputa-
tional snowball to identify a policy network. I argue that the reputational snowball not only
presents a useful tool for identifying micro-level network members, but also provides a
means for assessing which members of the policy network are core, and which ones are on
the periphery. Issues around reporting in studies of policy influentials are also discussed.
Introduction
Studying policy networks raises challenges in three important areas: identifying
members of the policy network, gaining access to the network, and reporting findings
from the study while maintaining confidentiality. The first challenge involves identify-
ing the key influentials in a given policy field. This problem is one of identifying the
central, key members of policy networks as well as those members who are on the
periphery. Another issue faced in studies of influentials is that of reporting. In particu-
lar, how can a researcher accurately describe a policy network while maintaining confi-
dentiality, and are promises of confidentiality important for this type of research?
In 2000 my colleagues and I commenced a study on tobacco control and health
policy making in the Australian state of Victoria. The VicHealth Centre for Tobacco
Control, a tobacco control advocacy organization, funded the research. The goal of the
project was to identify the key players in the health policy and tobacco policy areas in
Karen Farquharson is Lecturer in Sociology in the School of Social and Behavioural Sciences at Swinburne
University of Technology. Correspondence to: Karen Farquharson, School of Social and Behavioural Sciences,
Swinburne University of Technology, PO Box 218, Hawthorn, Victoria 3122, Australia. Tel.: +61 3 9214 5889;
Email: kfarquharson@swin.edu.au

346 K. Farquharson
Victoria, and to find out whether tobacco control was on the then-current Victorian
government’s health policy agenda, what else was on the health policy agenda, and
what views policymakers had of tobacco control.
In Australia, tobacco control policy is done at both the Commonwealth (federal)
and state levels of government. At the state level, tobacco control policy is almost exclu-
sively the domain of the health policy area; tobacco control can thus, at the state level,
be seen as a sub-field of the larger health policy field. We were interested in finding out
who the key policymakers were in health policy and tobacco control policy in our state.
We wanted to be able to claim that the group that we interviewed were the most influ-
ential in their policy area in Victoria. We decided to use a different type of snowball to
identify our sample: the reputational snowball.
This article describes the process we undertook to identify the key policymakers, and
the pros and cons of this type of research method. I argue that the reputational snowball
technique that we used to identify key policymakers was both effective and replicable,
providing a useful addition to existing approaches to the study of policy networks.
Identifying Networks of Policy Influentials
Policy networks are groups of connected individuals or institutions who have the
capacity to influence (e.g. initiate, develop, block) policy in a given field. Policy
researchers have written about policy networks as vehicles through which policies get
initiated, developed, and eventually implemented (Considine, 1994; Marsh & Rhodes,
1992; Marsh & Smith, 2000).
Policy networks are not fixed: their composition varies depending on what policy
issue is under consideration (Laumann & Knoke, 1987) and can change dramatically
when the policy setting changes, such as when a new government is elected. For any
given policy field there will be a few key players who will be influential in all areas of
that field, but most sub-fields will have specialists who only become influential when
proposals in their area are being actively considered. In a broad policy field like health
policy, there are many sub-fields that have their own core influentials. This makes these
networks somewhat difficult to identify.
Various techniques have been used to identify elites. For example, Useem used
Who’s Who to identify corporate executives (1995), and Herzog found elected female
politicians on public registers (Herzog, 1995). Laumann and Knoke took a different
approach, identifying influential organizations and interviewing informants from
them (1987). For Laumann and Knoke, the influential organization was the unit of
analysis, not the influential person. Other researchers have used the traditional snow-
ball technique, where informants are asked to suggest other people for the study
(Ostrander, 1995). These techniques have not often been used to study policy
networks.
Social networks have been studied using a variety of techniques including: various
mathematical tools used to measure links between network members, and to describe
networks and the relationships therein (see Scott, 1991), ethnographic studies that
describe a variety of relationships within networks of people including kinship (Geertz,