Content uploaded by Darren Halpin
Author content
All content in this area was uploaded by Darren Halpin on Jan 12, 2018
Content may be subject to copyright.
doi: 10.1111/j.1467-9299.2011.02005.x
EVALUATING THE BREADTH OF POLICY
ENGAGEMENT BY ORGANIZED INTERESTS
DARREN R. HALPIN AND H.F. THOMAS III
This article probes the variation in the breadth of policy engagement among organized interests. The
literature, heavily shaped by large-n US studies of Washington and its lobbying system, suggests
many reasons for organized interests to focus policy engagement relatively narrowly. This claim
of policy specialization has been long repeated in the British public policy literature. The aim of
this article is to empirically test the extent to which expectations of narrowed engagement hold
in a UK context. This article uses a new Scottish dataset that tracks actual engagement by any
organized interest on executive policy consultations over a 25-year period. It tracks over 90,000
‘mobilization events’ by over 18,000 organizations in 1,690 distinct consultation issues across the
entire Scottish policy system. In analysing these data, we concern ourselves with establishing: (1) the
extent of generalized engagement; (2) the type of organized interests that are more or less general
in their engagement; and (3) the extent to which a specialized style of policy engagement is on the
increase over time. In the process, we develop measures that are appropriate for assessing breadth
of engagement using issue-based policy data.
INTRODUCTION
This article addresses one of the more fundamental questions about the policy-related
behaviour of organized interests: how broad is their policy engagement? There is a long
tradition of large-n studies of organized interests and ‘lobbying activities’ in the United
States (US) (for a review, see Baumgartner and Leech 1998). Often these studies have
produced maps of the mix of organized interests present in the Washington policy scene
(see, for instance, Schlozman and Tierney 1984; Walker 1991). This scholarly tradition has,
among other things, generated a serious debate about policy ‘specialization’ among orga-
nized interests. In this context specialized means that groups pursue a (relatively) narrow
issue agenda. This theme is worthy of debate not least because it impinges on questions
of democracy and representation. As has been well discussed, the composition of the
system of organized interests is deemed important to ensure that public policy reflects
diverse ‘voices’ (see Schlozman 2009). It follows, therefore, that group specialization –
sticking to narrow policy areas and avoiding conflict – is likely to undermine the plu-
ralistic competition that scholars see as crucial to the democratic contribution of groups.
Moreover, the existence of generalists is viewed as important in linking together the often
narrow deliberations of specialized policy communities in such a way as to maximize the
prospects of policy being informed by a broader public interest (Browne 1990). While US
in its nature, this set of propositions has been highly influential in shaping international
scholarly expectations of the pattern of organized interest engagement in public policy. In
this article we try to push the research agenda forward – both empirically and in relation
to measurement issues – by examining issue-level organized interest data that can tell us
more about the scope and distribution of organized interest engagement in public policy,
as well as that outside the US.
Darren R. Halpin is Associate Professor in the Department of Political Science, Aarhus University, Denmark, and
Visiting Professor, Robert Gordon University, Aberdeen, Scotland. H.F. Thomas III is a graduate student in the
Department of Government, University of Texas at Austin, USA.
Public Administration Vol. 90, No. 3, 2012 (582–599)
©2011 Blackwell Publishing Ltd, 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden,
MA 02148, USA.
POLICY ENGAGEMENT BY ORGANIZED INTERESTS 583
It is worth noting that the style of data used to address this broad theme strongly shapes
how detailed and convincing our answers can be. The tradition of counting populations
of organized interests has yielded a string of impressive studies that show the contours of
the ‘system’ generally (for example, Schattschneider 1960; Schlozman and Tierney 1984).
Subsequent work assumed that the ‘system’ might look different in different policy areas:
for instance, Walker’s (1991) impressive study of Washington groups could tell us in
which policy domains groups were ‘interested’: agriculture, health, and so on. Yet, this
domain study could not say whether ‘interest’ led to actual mobilization; nor could it
deal with the intensity of group activity (for example, how often were they active?). In
this article, we address the question of breadth of policy engagement, and hence about
policy specialization, via directly tracking the engagement patterns of organized interests
across diverse policy terrain and over time. In line with an emerging new generation
of organized interest studies, this article is concerned with using observed engagement
levels as a basis for measuring actual breadth of policy engagement. This style of our data
enables more detailed answers to these long-standing questions.
While a relatively diffuse literature, we locate three propositions regarding specializa-
tion that we test using new longitudinal data for the UK context. The first proposition is that
the organized interest system will be disproportionately composed of policy specialists
with a small breadth of engagement. This hypothesis reflects the broader discussion about
the need for organized interests – mostly for reasons of survival and maintenance –
to avoid policy competition (see Browne 1990). The second proposition we test is that
institutions will mobilize more broadly than membership-based interest groups. In his
seminal US work, Salisbury (1984) argued that institutions with complex interests, but
without the constraints of being accountable to members, would thus move nimbly across
relevant policy areas. Lastly, we examine a far broader proposition: that increased overall
levels of policy activity by government, and a growing size of the overall organized
interest system (the so-called group ‘explosion’), will be associated with increased levels
of policy specialization by individual organized interests. Our UK data span 25 years, and
thus we are in a position to examine this proposition directly.
The article proceeds as follows. We initially address the expectations in the litera-
ture about breadth of engagement. The subsequent section outlines the methodological
challenges that exist in mapping actual engagement and in measuring levels of policy
specialization/generalization with such data. After presenting the Scottish data, the sub-
sequent section examines our findings against the different hypotheses about breadth of
engagement.
ASSESSING BREADTH OF POLICY ENGAGEMENT: SCHOLARLY
EXPECTATIONS?
While the literature on the breadth of engagement by organized interests is not well
organized, it is nevertheless possible to identify some basic expectations in three areas:
the prevalence of generalized versus specialized policy engagement patterns, the types of
organized interests that are likely to be relatively more general in their engagement, and
the trend in relation to specialisation levels over time.
The numerical dominance of specialists?
Expectations with regard to breadth of policy engagement tend to support the idea
that organized interests would engage in a relatively narrow sliver of the policy world.
Public Administration Vol. 90, No. 3, 2012 (582–599)
©2011 Blackwell Publishing Ltd.
584 DARREN R. HALPIN AND H.F. THOMAS III
The general public policy literature – both in the US and UK – has long utilized terms
like ‘iron triangle’, ‘sub-government’, ‘private interest government’ and ‘closed policy
community’, all of which conjure up the impression of groups engaged in narrow agendas.
The policy world is made up of organized interests which specialize in a focused issue
agenda and as such, form rather well developed routines of engagement with civil
servants (for review, see Richardson 1999).
The general interest group literature also tends to support the idea of narrow agendas.
From the perspective of ‘group identities’, it has been suggested that groups follow
agendas that fit with their projected policy identity (Heaney 2004, 2007). Others suggest
that there is a strong motivation for groups to specialize in order to avoid competition
for scarce political attention and thus enhance survival prospects (Browne 1990). The
‘specialize to survive’ position has been further developed by the influential ‘population
ecology’ literature (see Gray and Lowery 2000[1996]).1While these positions provide
cogent arguments for us to expect narrow engagement, some empirical work suggests
this is not always apparent.
Several US studies have been designed in order to quantify the question of breadth of
‘policy engagement’. And here, the findings point to a broad ‘interest’ (if not engagement)
in political issues. The work of Jack L. Walker (1991) was perhaps the first explicit attempt
to empirically address breadth of engagement. His survey of Washington groups asked
respondents to nominate their degree of interest across a list of ten policy domains.
Counting those Washington groups who indicated they were ‘very interested’ in any
given domain, Walker was able to demonstrate the sheer diversity in policy engagement.
Based on subsequent re-analysis, Frank R. Baumgartner and Beth L. Leech provide an
excellent review of the Walker data (1998, p. 158). The data show that over 10 per cent
of groups were not ‘very interested’ in any domain – suggesting ‘public policy is not
their main concern’. Moreover, just over half the groups surveyed indicated they were
‘very interested’ in just one or two domains. The balance could be considered to have a
somewhat broad interest in policy issues – a generalized level of attention. But only 2 per
cent of the total sample indicated they were very interested in eight or more domains.
However, when the threshold was loosened to ‘somewhat interested’, the vast majority
of groups were active in more than one domain. They conclude that ‘The complications
of policy-making force the majority of them to attend simultaneously to several different
policy areas at once’ (Baumgartner and Leech 1998, p. 160). At least at the level of being
‘interested’ in public policy matters, the evidence from Walker’s study is that attention is
spread quite broadly.
This finding is affirmed by another influential US study. Based upon self-reporting
survey data from groups sampled from Washington directories, Heinz et al. (1993)
found that on average their respondents reported ‘some interest’ in 11 events in their
domain from the possible 20 they were presented with. They found that most groups
engaged well beyond the confines of any single governmental department or agency.
Indeed, they conclude ‘...the average representative spends time in more than half the
subfields within his or her policy domain and, in addition, in some four other major
policy fields’ (Heinz et al. 1993, p. 379). In summing up their research, Heinz et al.
(1993, p. 380) remark that ‘numerous interest groups monitor any given policy question
and consider taking a more active role in the debate’. These studies do not quantify
engagement, but the suggestion that organized interests ‘monitor’ broadly at least qualifies
the assumption that specialization is a default position (see Baumgartner and Leech 1998,
p. 157). As outlined above, while establishing if such empirical patterns are applicable
Public Administration Vol. 90, No. 3, 2012 (582–599)
©2011 Blackwell Publishing Ltd.
POLICY ENGAGEMENT BY ORGANIZED INTERESTS 585
to the UK context is salient, there has been no attempt to utilize mapping methods to
establish if they do indeed exist. We seek to empirically explore these arguments in a
UK context.
H1: The majority of organized interests mobilize narrowly.
Institutions engage more broadly than interest groups?
The studies of the breadth of policy ‘interest’ discussed above utilized survey evidence of
interest groups. A quarter of a century ago, Salisbury (1984) counselled that ‘institutions’
rather than ‘interest groups’ numerically dominated populations of organized interests:
the term ‘organized interests’ covers both interest groups and institutions. When using
the term institutions, Salisbury had in mind ‘individual corporations, state and local gov-
ernments, universities, think tanks, and most other institutions of the private sector’ (1984,
p. 64; in original). In this respect, Salisbury (1984) contends that each type of organization
could be expected to make different types of choices with respect to apportioning policy
attention. For Salisbury, institutions are demarcated from ‘voluntary associations of mem-
bers’ (a.k.a. interest groups) by the fact that they are ‘hierarchical structures which exercise
authority over people within their jurisdiction’ (1984, p. 67). Institutions have interests
that transcend the views or interests of their particular constituent members (Salisbury
1984, pp. 67–8). It is this basic difference that leads Salisbury to reason that institutions are
more easily able to enter the political arena (they do not need to persuade membership)
and that as complex organizations they will have a broader range of political concerns
(or at least governmental activity will encroach more often on their interests) (Salisbury
1984, pp. 68–9).2This claim is salient in the UK context, particularly given the role played
by local authorities – which would come under Salisbury’s label of institution – in the
implementation of central government policies. As discussed below, our approach to data
collection – we map all policy active organizations – provides a unique opportunity to
examine the differential patterns of breadth of engagement by institutions as well as
groups. We put the hunch that institutions are more generalized in their engagement than
groups to the test.
H2: Institutions engage more broadly than interest groups.
The growth in specialization?
As discussed above, there is a growing thread in the interest group literature that discusses
the motives for the pattern of engagement as viewed by organized interests themselves.
And it suggests that there are good reasons for organized interests to pursue policy
specialization to secure their survival. It has been argued that as the group-system grows,
organized interests would seek to find narrow policy niches – to specialize as a survival
strategy (Browne 1990; Gray and Lowery 1996). The argument, crudely put, is that in
the face of competition for scarce resources (specifically the attention of policy-makers)
organized interests react by narrowing their policy focus. Even if one does not accept
the assumption that cohabitation of the same policy space is a proxy for competition,
one could reasonably expect it to be better for an interest organization, in terms of their
survival prospects, to ignore most of what is going on in government (even when it
may be relevant to the ‘interests’ they advocate for) and concentrate their own resources
and attention towards an issue niche (for further discussion, see Baumgartner and Leech
2001). This finding resonates with the broader political science literature suggesting that
the growth of overall group numbers reflects an underlying current towards policy
Public Administration Vol. 90, No. 3, 2012 (582–599)
©2011 Blackwell Publishing Ltd.
586 DARREN R. HALPIN AND H.F. THOMAS III
specialization and the pursuit of narrow issue agendas (for the US, see Skocpol 1999; for
the UK, see Grant 2008). If the specialization thesis holds, we would expect to see the
breadth of engagement of organized interests to fall as the overall size of the organized
interest population increases.
H3: Aggregate specialization levels grow over time.
In this article we focus on these related propositions. That these propositions are
strongly held rules of thumb underlines the importance of our findings for this area
of scholarship.3The sheer effort required to conduct large-scale empirical work on
breadth of policy engagement means that existing findings are limited to a few US
studies. For instance, systematic quantification of the actual mobilization of organized
interests around public policy issues or domains is an infrequent approach in the UK.4
This is despite the UK public policy literature containing long-standing claims about
the specialized pattern of organized interest government engagement (for a review, see
Richardson 1999). It has long been asserted that the ‘consultative’ system involving close
contact between specialized organized interests and sub-sections of the bureaucracy: (1) is
important alongside the parliamentary system (McKenzie 1958; Rose 1984); (2) constitutes
the British (and Scottish) ‘policy style’ (Richardson and Jordan 1979; Cairney 2008); and
(3) is the ‘orthodox’ UK public policy approach (Grant 2001). Others have highlighted the
importance of the civil service in the legislative process (Page 2003). We acknowledge
that the three propositions outlined above have been repeated in the British public policy
literature over many decades (see the discussion above); but here we examine whether
or not these are empirically borne out in systematic mapping data. We are not so much
replicating past studies, but beginning to empirically replace a long-standing assertion.
In so doing, this article seeks to extend what is a US research tradition to a new context,
the public policy process in Scotland.
In this article we scrutinize these expectations in the literature – shared by US and UK
scholars – using a new dataset based on Scottish public policy. Specifically, we utilize
data precisely from this ‘consultative system’ – involving the engagement of organized
interests in Scottish public consultations over a 25-year period – to test several general
propositions in the literature. In previous studies, the data available to measure breadth
of engagement has been based on self-reporting surveys. As such, analysis was naturally
restricted to measures of the extent to which groups had an overlap of ‘interest’ across
provided lists of policy domains (see Walker 1991). In this article we utilize a dataset
that links engagement to particular issue contexts. An overlap of ‘interest’ measure was
perfectly reasonable when one only had survey-based data, but our data require new
approaches: they present unique challenges with respect to measurement. That the style
of data collection we analyse is becoming more and more the norm adds further weight
to why (re)examining measures of breadth of organized interest engagement is timely.
ASSESSING POLICY ‘INTEREST’ VERSUS ‘ENGAGEMENT’: DO NEW DATA
NEED NEW MEASURES?
In the discussion above we have established that the extent to which organized interests
spread their attention across policy domains or pursue a more focused pattern of engage-
ment is a salient question for political scientists to consider. But the empirical study of
breadth of engagement is plagued by the dual difficulties of sampling/data collection and
measurement. Past studies gather survey evidence of policy interest from interest groups.
Public Administration Vol. 90, No. 3, 2012 (582–599)
©2011 Blackwell Publishing Ltd.
POLICY ENGAGEMENT BY ORGANIZED INTERESTS 587
Here, in step with a new generation of studies on organized interest engagement in
political life, we collect data from observations of actual policy engagement by organized
interests. In this section we outline the type of data we have collected, and the implications
this has for how we can go about measuring breadth of engagement.
Data source
In this article we utilize a large dataset compiled from records of policy consultations
launched by the Scottish Office (1982–1999) and Scottish Government (1999–2007).5Our
decision was based on the nature of the British – and Scottish – political system, which
means the bureaucratic avenue is likely to be the most frequently deployed, the most open
and accessible, and the most productive for organized interests (see Jordan and Maloney
2001, p. 44).6Thus the consultation process provides an important window into policy
mobilization by groups. We are concerned with the question of engagement, and not with
influence.
In the UK, public policy consultations are routinely conducted on a broad range of
issues, which may include calls for comments on draft bills, initial agendas for discussion,
proposals for amendments to regulations, the details of implementation of EU directives
or similar. Indeed, as mentioned above, such processes have been referred to as both
the British and Scottish ‘policy style’. Precisely how decisive these exercises are for the
final policy outcome is not clear. Evidence from several interviews with civil servants in
Scotland suggests that new insights and information arise from such exercises: there are
sometimes surprises (see Halpin 2011). Moreover, a survey of groups mobilizing suggests
that 76.7 per cent of groups (n =360) engage in consultations very or fairly often.7Of
course, there remains some scepticism. Indeed, participants interviewed by the authors
openly admit that they are not always clear on how important their participation is, or if it
always matters (see also Beales et al. forthcoming) . Yet, resolving this particular question
is not critical for the present article: it is not a study of influence, and our justification
for using these data is simply that they provide a convenient window for assessing the
breadth of policy engagement.
The process of UK devolution, which in 1999 granted Scotland its own parliament and
allocated it direct responsibility for so-called ‘devolved’ policy matters, has not changed
the practice of policy consultations. But, given that Scotland does not have responsibility
for all policy matters (the so-called reserved matters) our data do not cover policy issues
in areas such as defence, national security, international trade, or foreign affairs. It is
standard UK-wide practice that consultations are launched by a team within a rele-
vant government department, with invitations being sent to stakeholder lists and (more
recently) invitations made on the government web site (in practice, for consultations, the
access barriers are extremely low).
The consultations dataset is able to plot levels of mobilization over time by all types
of actors (individuals, institutions, interest groups, government departments/agencies,
and ministers or legislators), starting at the issue level and working upwards to domain
and then system levels. In total, the dataset includes 1,690 consultations, which received
over 180,000 responses from more than 18,000 discrete policy actors. This dataset was
compiled largely using paper-based records held in the Scottish government library and
its document storage facility in Edinburgh, but with the addition of some more recent
documentation only available electronically on the publications pages of the Scottish
government web site. No definitive list of consultations conducted by the Scottish
government exists;8we can therefore definitively say that we mapped each and every
Public Administration Vol. 90, No. 3, 2012 (582–599)
©2011 Blackwell Publishing Ltd.
588 DARREN R. HALPIN AND H.F. THOMAS III
consultation where data are available in the public domain, and our dataset is as
comprehensive as it can ever be.
Most importantly, these data quantify actual policy mobilization: we do not have to
decide objectively who is or is not a relevant actor in a policy domain since we simply let
the actors decide for themselves.9In that respect, our data are substantially different from
those used by past studies discussing breadth of ‘attention’ by interest groups. However,
they are almost identical in design to a new generation of datasets being collected in the
US and beyond (see those used by Gray and Lowery 1996; Baumgartner and Leech 2001;
Rasmussen 2010). As such, the discussion below dwells both on measures of breadth of
engagement and on substantive findings about organized interest behaviour.
New measures: from ‘interest’ to ‘engagement’
As discussed above, the state of the art when it comes to measures of breadth of
engagement involve measures of ‘interest’ in policy areas (Walker 1991; Heinz et al. 1993).
Typically scholars have measured the extent to which groups with a policy ‘home’ in,
say, agriculture self-reported an interest in other domains. Because our data are based
on actual policy activity we have some new possibilities with respect to measurement,
but also some novel challenges. In this article we adopt three different approaches to
measuring breadth of policy engagement.
Domain count
The first is the number of domains that an organization is active in.10 To measure this
we take a count of the number of domains a group has been active in over the entire
25-year period. In order to identify which domain a consultation belongs to, based on the
subject matter of each consultation, we applied a Policy Agendas major topic code.11 We
considered each major topic code – agriculture, health, education, environment, and so
on – as constituting a unique policy domain. The advantages of this coding process are
threefold. First, we are using an established and respected method for coding government
policy. Second, the use of this codeframe makes this and other policy agendas projects
data sources easily comparable (for a review, see John 2006). Thirdly, it overcomes a major
problem with the only other viable coding alternative, departmental structures; namely
that such structures frequently change boundaries over time. While a policy agendas frame
has been applied to governmental agendas, we see our approach as a logical extension
since organized interests also face finite resources and attention which necessitates
prioritization. Table 1 reports the frequency of consultations across the domains. As is
evident, consultations were conducted in 15 domain areas, with far more concentrated
in the areas of agriculture, health, education, housing and crime.12 This broadly reflects
those policy areas in which the Scottish administration has primary legislative powers.
Issue entropy
The above domain measure simply reports the total count of domains in which a single
organization is active without respect to the distribution of this activity. The problem is that
two groups active in the same number of domains may have widely divergent proportions
of activity across them. To capture this variation we utilize a measure of ‘issue entropy’,
namely a Shannon’s H entropy score. This has its roots in information theory (Shannon
and Weaver 1949; Shannon 1950; Chaffee and Wilson 1977; Lasorsa 1991; Culbertson
1992) but is increasingly used in studies of policy agenda diversity (McCombs and Zhu
1995; Baumgartner et al. 2000; Talbert and Potoski 2002; Jones et al. 2005; Sheingate 2006;
Public Administration Vol. 90, No. 3, 2012 (582–599)
©2011 Blackwell Publishing Ltd.
POLICY ENGAGEMENT BY ORGANIZED INTERESTS 589
TABLE 1 Frequency of policy issues and activity by policy domains
Domain Issues Percentage Activity Percentage
Macroeconomics 46 2.74 1,255 1.40
Civil rights 24 1.43 2,267 2.53
Health 173 10.30 10,510 11.72
Agriculture 395 23.51 7,495 8.36
Labour 12 0.71 542 0.60
Education 147 8.75 16,830 18.76
Environment 211 12.56 9,336 10.41
Energy 21 1.25 1,038 1.16
Transportation 52 3.10 2,927 3.26
Law, crime, and family issues 175 10.42 7,905 8.81
Social welfare 37 2.20 2,952 3.29
Housing and urban development 254 15.12 15,751 17.56
Commerce and banking 32 1.90 1,958 2.18
Government operations 83 4.94 7,337 8.18
Public lands 18 1.07 1,603 1.79
1,680 100.00 89,706 100.00
Boydstun 2008; Wolfe et al. 2009). While capturing the same underlying concept as other
similar measures used in the interest group literature – such as the Herfindahl-Hirschman
Index13 (see use by Gray and Lowery 1996) – Shannon’s H is utilized in this application
as it is believed to be more sensitive to high levels of scope and results in wider variation
(Bevan 2008). The issue entropy scores we calculate are normalized to range from 0.00
to 1.00, and describe the distribution of each organizations domain-level activity. When
organizations participate in only one domain, the entropy score is near zero and indicates
a narrow breadth of engagement. Ultimately, this issue entropy score supports an analysis
of system-level breadth of engagement at the individual organizational level, and can be
compared across all organizations.
Generalization (entropy X activity)
Though issue entropy is used extensively in the analysis to follow, we develop a third
measure of breadth of engagement that incorporates both the distribution of intensity
across domains and absolute levels of activity for each organization. Given that our
calculation of issue entropy utilizes proportions of activity at the organization level,
we multiply this score by each organization’s count of consultation responses (total
activity) to generate a relative generalization measure. Though mathematically simple in
calculation, this score provides a method by which we can compare the distribution of
activity across 15 policy domains given an organization’s total level of participation. Here,
we are implying that despite even distributions of intensity across domains, organizations
with substantial numbers of overall consultation responses are fundamentally different
than those with limited activity levels.
In summary, we develop three measures that enable us to compare the relative levels of
actor participation across 15 policy domains: (1) domains (a simple count of the number
of domains in which an actor participates); (2) issue entropy (a normalized Shannon’s H
score using proportions of actors’ activity); and (3) generalization (the product of issue
entropy and actors’ absolute activity). We are clear that these all provide distinctive
measures, but they are of course correlated with one another.14
Public Administration Vol. 90, No. 3, 2012 (582–599)
©2011 Blackwell Publishing Ltd.
590 DARREN R. HALPIN AND H.F. THOMAS III
FINDINGS
Generalized versus focused policy engagement: what is the mix?
Past studies of breadth of engagement have measured the extent to which groups – based
on survey data – indicate an ‘interest’ in various policy domains. The usual analysis is
to look at the extent to which groups are interested in one or more domains. Here we
map actual policy engagement across policy domains. Table 1 also reports the frequency
with which organizations engaged across the 15 policy domains over the 25-year period.
We counted a single response over this period to a consultation as engagement in a
domain. As such, the ‘domain’ measure offers a very low threshold test of engagement,
and does not capture the intensity of engagement (we come to that in a moment). As
the table shows, the vast majority of our actors, almost three-quarters of the entire pop-
ulation, are active in just a single domain. While we cannot discount that a few may
have changed names in ways we could not detect (which would mean we have over-
counted the number of unique organizations), this does not diminish the overwhelming
finding of very partial political engagement. This analysis does suggest, at face value,
that most organizations in our population are domain specialists: they focus in just one
policy area. Further, it indicates that beyond these organizations there is a rather smaller
minority engaged in a handful of related domains (say, agriculture and environment, or
education and social services). It could, of course, be that the most frequent actors are
simply those that engage more broadly? That is, broad engagement could be a proxy
for a high level of engagement. To rule out this explanation, we looked for any correla-
tions between activity and domain spread. A scatter plot revealed no direct correlation
between the number of domains an actor is active in and their level of overall level of
activity.
It is interesting to compare prior work utilizing a measure of ‘interest’ with our findings
deploying a measure of ‘engagement’. Our results suggest the dominance of a narrow
policy focus. But recall, in their reworking of Walker’s data, Baumgartner and Leech
(1998, p. 158) – based on a similar analysis to table 1 – suggest that two per cent of total
were interested in ‘eight or more of the ten areas’. This different finding suggests that
while breadth of monitoring may be quite broad, for most organized interests actual
engagement is narrowly drawn.
This preliminary conclusion is reinforced when we consider domain overlap (that is,
what proportion of all active organizations are engaged in common policy domains).
When Walker (1991) asked his respondents to indicate whether they were ‘interested’ in
a range of policy areas, he could then examine overlaps of policy interest. In our data,
we are fortunate in that we can look at the proportion of the population who actually
engage (at least once) in one or more policy domains. The comparison is enlightening.
When we tried to replicate his table showing the extent of overlap among actors who
worked in similar policy domains, the highest figure in our data is 4.9 per cent. In
contrast, Walker’s data showed a high in terms of overlap of 13.9 per cent. Part of the
reason for the difference is that we are looking at the overlap among all organizations
while Walker looked only at whether those ‘very interested’ in their ‘home’ domain also
worked in others. Nevertheless, in and of itself, our finding tells us that beyond the well
known political interlocutors (who we ‘know’ are important and invite to respond to
surveys), organized interests engage only sporadically in contemporary policy-making.
Our data on engagement – versus ‘interest’ – suggest most organized interests stick to
their ‘home’ domain.
Public Administration Vol. 90, No. 3, 2012 (582–599)
©2011 Blackwell Publishing Ltd.
POLICY ENGAGEMENT BY ORGANIZED INTERESTS 591
TABLE 2 Grouped entropy across all actors
Entropy Frequency Percentage Cumulative percentage
0.00–0.09 13,906 75.46 75.46
0.10–0.19 347 1.88 77.34
0.20–0.29 2,233 12.12 89.46
0.30–0.39 625 3.39 92.85
0.40–0.49 651 3.53 96.38
0.50–0.59 330 1.79 98.17
0.60–0.69 197 1.07 99.24
0.70–0.79 88 0.48 99.72
0.80–0.89 52 0.28 100.00
18,429 100.00
The finding that most organized interests in our data ‘engage’ rather narrowly is
supported when we deploy our other measures of breadth of engagement. Using the issue
entropy scores we can compare all our organizations over 15 domains in a fashion that
takes available data into consideration (see table 2). As we describe above, the measure
captures the evenness of actors’ distributions of activity across domains. It is salient
here that the vast majority of actors are active only in one domain; that is, they have
an issue entropy score of near zero. It is only when we get to the last 500 or so actors
that we see any organizations that could in any meaningful way be described as real
system-wide generalists. This reflects the broader pattern of policy activity reported in
table 1. Our data show that most organized interests engage very narrowly in public
policy.
Are institutions more broadly engaged than interest groups?
As outlined in the data section above, a by-product of our observational – rather than
survey based – approach is that we analyse the breadth of engagement of all ‘active’
organizations. We are not constrained by a list of ‘important’ or ‘salient’ organizations in
a list of survey respondents. This provides a unique opportunity to evaluate Salisbury’s
(1984) hunch that institutions, and not interest groups, would tend to be more broadly
engaged in public policy. So what did we find?
In support of Salisbury’s suggestion, table 3 reports the 15 actors with the highest
‘breadth of engagement’ scores (we report all three measures). These actors are the most
broad in their engagement. The second column designates organizational type. As is
evident, these actors are all local government organizations (with the Convention of
Scottish Local Authorities [COSLA] being the collective voice of local authorities). In fact,
29 of the 32 current Scottish local authorities (Councils) were in the top 100. If there is
a set of actors that constitute a core to Scottish public policy it is local government: that
is, central/local government institutions. This resonates with Heinz et al. (1993), in their
seminal US study, who could not count government, but nevertheless speculated it may
be the core. Immediately salient here is that entropy and entropy x activity scores are
not directly reflected in either domain spread or overall volume of activity. For example,
COSLA is the most frequent actor in the table, and is active across 15 domains (as broad
as any actor in the data). Yet it has a lower entropy score than Perth and Kinross Council.
Similarly, while Perth and Kinross Council has a higher entropy score than Aberdeen City
Council, when we factor in activity levels, we see Aberdeen emerge as a more generalized
Public Administration Vol. 90, No. 3, 2012 (582–599)
©2011 Blackwell Publishing Ltd.
592 DARREN R. HALPIN AND H.F. THOMAS III
TABLE 3 Top 15 system-wide generalists
Rank Name No. of domains Activity Entropy Entropy X activity
1 COSLA 15 677 0.856 579.63
2 Glasgow City Council 15 629 0.846 532.01
3 Edinburgh (City of) Council 15 618 0.826 510.74
4 Highland Council 15 590 0.850 501.74
5 Aberdeen City Council 15 565 0.840 474.82
6 Fife Council 15 545 0.857 467.03
7 North Lanarkshire Council 15 548 0.831 455.45
8 South Lanarkshire Council 15 533 0.841 448.33
9 Aberdeenshire Council 15 522 0.847 442.11
10 East Ayrshire Council 15 499 0.850 423.98
11 Angus Council 15 500 0.838 418.98
12 Renfrewshire Council 15 497 0.832 413.75
13 Perth and Kinross Council 15 448 0.869 389.14
14 Dundee City Council 15 463 0.824 381.45
15 West Lothian Council 15 437 0.849 370.85
Key: COSLA =Convention of Scottish Local Authorities.
actor (Entropy x activity score). The newly devised measures are able to separate out actor
generalization levels.
While there is a preponderance of governmental entities among organizations with the
broadest engagement in policy, it is important to note, however, that being a government
entity does not itself designate such a pattern. In fact, the local authorities make up
a small proportion of the overall governmental entity category which also includes
non-departmental public bodies (NDPBs), government agencies and central government
departments. This is borne out in the analysis below.
Table 4 shows the mean entropy x activity scores for cross-domain generalization by
organization type. Calculations of average entropy x activity scores show that institutions
are, on average, more generally engaged that interest groups (score of 2.35 versus 1.73).
However, it depends on the type of institution. Government, by which we refer to
central/local government, is by far the most broad in its engagement. In a Scottish context
TABLE 4 Mean generalization by actor type
Type of actor No. of actors Entropy Entropy S.D. Entropy X
activity
Entropy X
activity S.D.
Institutions
Public institutions 5,369 0.05 0.12 0.95 7.85
Government 2,591 0.18 0.22 8.22 45.10
Businesses 3,746 0.04 0.12 0.30 1.59
Interest groups
Non-profits and citizen groups 3,836 0.10 0.17 1.35 6.86
Trade associations 1,115 0.09 0.16 1.67 11.01
Professional associations 1,025 0.12 0.18 3.20 18.20
Unions 72 0.21 0.22 5.94 19.65
Other 683 0.08 0.16 1.31 12.08
Total/Average 18,437 0.11 0.17 2.87 15.29
Public Administration Vol. 90, No. 3, 2012 (582–599)
©2011 Blackwell Publishing Ltd.
POLICY ENGAGEMENT BY ORGANIZED INTERESTS 593
this makes sense given they are charged with delivering most public services to citizens.15
Moreover, prior to devolution (pre-1999), they were the only form of elected Scottish
government. By contrast, private institutions, such as businesses, and public institutions
(such as hospitals, schools and such like) are among the least broad. There is substantial
variation. For instance, here unions are the second broadest engaged type of organizations
– on average – even though no trade union is in our top 15.
In relation to institutions versus collective organizations, the stand out finding again
is that central/local government are very broad, as Salisbury would expect, but other
institutions are among the most narrowly engaged. There is little support, at least from
our data, that companies and hospital boards, school districts and the like are actually
broadly engaged. By and large they stick to a narrow domain focus. In relation to collective
organizations, unions and professional associations are the broadest in their engagement,
and voluntary associations/citizen groups the most narrow. The fact that unions have
consolidated over time, and that professional groups (doctors, lawyers, accountants) must
protect members’ interests wherever these emerge, seems to make sense with the data
we report. The citizen group finding also gels with the argument that such groups are
by definition typically narrow and focused in their policy activities. Their generalized
engagement mostly happens through umbrella groups – such as the Scottish Council of
Voluntary Organisations (SCVO) and the Scottish Environment Link.
Is focused engagement on the rise?
There is a strong expectation in the group literature that crowded policy space will drive
organized interests to specialize and focus attention on increasingly narrow policy niches.
One of the advantages of our time-series data is that we can make an initial attempt to
examine the development of policy specialization levels over time. To facilitate this type of
analysis we segmented our dataset by six time periods, each correlating roughly with UK
parliamentary terms.
Table 5 reports two measures of ‘breadth of engagement’ (entropy and entropy x
activity) scores, alongside measures of activity by organized interests across time periods.
We analysed the breadth of engagement only for organizations that were engaged
in more than one domain. Here, we test whether, in aggregate, regular participants
become more broadly engaged over time. The initial salient point here is that the
‘system’ is growing in the first two periods (measures of overall activity by organized
interests), and also that the number of unique organizations actually engaged in policy
affairs is rising. This suggests ideal conditions for a growth in actor specialization over
time.16 However, our data show that there is no correlation between a rapid growth
in both consultation activity and the number of groups active over time, and breadth
TABLE 5 Breadth of policy engagement over time, 1982–2007
Period Average entropy Average entropy X activity No. of actors engaged Volume of activity
1982–1986 .313 2.519 275 1,776
1987–1991 .358 5.635 532 6,150
1992–1996 .337 4.977 1,285 12,598
1997–2000 .350 5.665 1,387 13,825
2001–2004 .339 6.655 1,489 17,513
2005–2007 .354 5.142 911 8,221
Public Administration Vol. 90, No. 3, 2012 (582–599)
©2011 Blackwell Publishing Ltd.
594 DARREN R. HALPIN AND H.F. THOMAS III
of engagement levels.17 Of course, we cannot directly assess the competition levels that
any increase in organizational populations may suggest – but then again the population
ecology literature has suggested that indirect competition is the mechanism that regulates
populations (see Gray and Lowery 2000 [1996]). While we cannot categorically say
what mechanism underlies this process, we can conclude that overall population sizes
are not, at an aggregate level, associated with higher or lower levels of specialized
engagement. This finding supports analysis elsewhere that shows the importance of
individual group resource levels, as opposed to intra-group competition, in explaining
the level of generalized engagement in policy (see Baumgartner and Leech 1998).
DISCUSSION AND CONCLUSIONS
The breadth of engagement by organized interests is a salient topic for political science.
From a pluralist perspective, assessing the extent to which organized interests spread their
activities broadly – or not – helps in probing the diversity of ‘voice’ in a given political
system and, thus, the prospects for governing in the ‘public good’. While these concerns
have US origins, they have been repeated for the UK context and have been central to the
international literature on organized interests and public policy. In this article we have
explored this issue using a new style of dataset on Scottish policy engagement by organized
interests, which affords us three substantive empirical findings. In the process we have
fashioned some new measures for breadth of policy engagement that – so far as this style
of data collection becomes more common – should prove useful for others (especially in
connection with an emerging neo-pluralist group literature, see Lowery and Gray 2004).
Firstly, we find that most organized interests are rather narrow in their breadth of
engagement. They tend to stick to a single domain, or perhaps stretch to related domains.
By contrast, few organizations could be said to spread engagement across anything close
to the entire policy system. This finding is consistent with what the ‘specialize to survive’
position would expect to see by way of patterns of engagement. As such, the finding is
somewhat at odds with the large-scale empirical studies reporting generalized attention
among organized interests. This dissonance underlines a key implication of our study,
that attention is empirically different to engagement. We have no reason to doubt that the
organizations we observe were attentive in a different pattern to which they engaged. It
is important to keep in mind here that the population of organized interests we observe
are taken from the context of policy consultations, which have relatively open access to all
interested parties. Other studies tend to deliberately sample well-known and regular par-
ticipants in politics: our data are different. Our method of quantification means we include
many, many organizations that are unlikely to conceive of themselves as dedicated policy
organizations. This may also partly explain the extent to which we find few organizations
with a relatively broad engagement. Yet, if we are seriously interested in quantifying
public policy ‘as it happens’ then there is surely value in also from time to time casting
a net in a way that reflects the pattern of policy activity in the real policy world. Thus,
a firm implication from our work is that the study of organized interests ought not to
always focus exclusively on the known quantities of the organized interest universe, but
to look also to the fringes where a very large cast of intermittent pressure participants
reside. It follows, that explaining the organized interest populations that emerge on given
issues requires scholars to resort to theories of conflict expansion which pay attention to
the ways in which the initially disinterested or disengaged (policy amateurs) come to take
an interest in policy matters (see Schattschneider 1960; Baumgartner and Leech 2001).
Public Administration Vol. 90, No. 3, 2012 (582–599)
©2011 Blackwell Publishing Ltd.
POLICY ENGAGEMENT BY ORGANIZED INTERESTS 595
Secondly, following Salisbury, we find that institutions are indeed more generally
engaged, on average. However, it depends on the type of institution. We find central and
local government are the most broadly engaged institutions, but businesses are the most
narrowly engaged. This is an important caveat on the Salibsury account; and also a healthy
reminder that much political ‘lobbying’ is government lobbying government. In relation
to groups, variation also occurs. Citizen groups seem, on aggregate, to be most narrow
in their engagement compared to unions or professional groups. The broader implication
here is to how interest group scholarship attends to the question of survival. How two
different types of interests groups can survive with vastly different breadth in policy
agendas points to useful further work, particularly probing the particular comparative
‘ecologies’ of citizen versus business interests.
These two empirical findings are not likely to shock the long-standing observer of
British public administration. After all, dating back to Sammy Finer, there has been the
assertion that groups pursue rather narrow agendas confined to areas of policy expertise
(for a summary, see Richardson 1999). Yet, these findings warranted testing and empirical
scrutiny. While the US literature has tried to ‘test’ this type of proposition with population
data, they sit in the British literature as largely unsubstantiated claims. This article has
not so much attempted to replicate these past studies, but sought to replace assertion
with systematic empirical data. That in broad brush we uphold these fine assertions does
not devalue the findings, but rather gives them added importance. That we can provide
empirical nuance offers a baseline for others to take up and probe the specifics.
Lastly, we examine a far more general question in the literature about change over time.
Conversely to expectations about a rise in issue specialists, we show that an expanding
population of policy active organizations does not lead to higher levels of specialization
among organized interests that are regular participants (those that are engaged frequently
in policy activity). There has been a repeated finding that organized interest populations
are growing over time – and our data support this for the Scottish ‘system’. But, in
contrast to expectations (see, for instance, Browne 1990) this does not drive organizations
to pursue ever narrower policy niches. This does not rule out, as Heaney (2007) has
suggested, that organized interests might utilize other dimensions of their identity to
differentiate themselves from each other (say, representativeness, expertise, and so on).
But, we can at least make the argument that, in aggregate, policy specialization does
not seem to be associated with population expansion. This is highly salient for the long-
standing debate that aggregate growth in the organized interest ‘system’ was synonymous
with the growing compartmentalization of policy-making. In fact, we find that the system
in Scotland is growing over time, but that it does so by adding new organizations (mostly
intermittently) and that the existing general groups remain general in their engagement.
This fosters an image of a growing penumbra of policy amateurs surrounding a core of
broadly orientated policy generalists. On our evidence, if the system is dominated by
policy niche operators now, it is because it probably has been for several decades.
Of course, the argument we have put forward is positioned against expectations in
the UK and US contexts. We suggest that breadth of engagement is a theme that has
salience outside of these cases, although we also admit that the expectations arising
from US and UK evidence might need recalibrating for other cases. Comparative data
– of which there is a paucity – on different systems will help explore the claims made
in this article further. It is noteworthy that these substantive findings are made against
scholarly expectations established largely using data on ‘reported’ levels of interest in
policy domains. Previous studies asked groups for generalized responses as well as about
Public Administration Vol. 90, No. 3, 2012 (582–599)
©2011 Blackwell Publishing Ltd.
596 DARREN R. HALPIN AND H.F. THOMAS III
‘intentions’ rather than ‘actions’. In recent years, there has been a concerted effort amongst
interest group researchers to develop public policy based datasets that reflect observations
of actual group engagement and do so at the issue level over time (for a discussion, see
Baumgartner and Leech 1998, pp. 119, 124 and Halpin and Jordan forthcoming). This article
utilizes data that in many respects heed this call. As such, the true test of our findings
will be against data from this new generation of studies that take a similar focus on
actual policy engagement of organized interests in political affairs. In taking this agenda
forward, our study offers some important lessons. For instance, the need to collect data
based on policy mobilization then raises the subsequent question over which window
on policy engagement to utilize. The initial choice to map policy consultation processes
was based on three principles: pragmatism (the data were available and accessible);
relevance (the consultation process is very open and would likely capture most politically
active groups); and scholarly salience (at least some suggest the bureaucratic arena is a
likely focal point for the engagement of organized interests in British politics). Similar
justifications of approach are welcome, indeed necessary, both to justify stepping away
from established orthodoxies, and (especially) in taking a first tentative step when no
orthodoxies exist to follow. In sum, beyond our substantive findings, we offer up one
novel way to investigate and measure engagement patterns: methods that are necessary
when new datasets that are built using issue-level information on mobilization become
available. In so far as this type of data becomes more commonly collected, the measures
we devise and apply should prove useful for others.
ACKNOWLEDGEMENT
This research was funded by an award from the UK Economic and Social Research Council
[Award The mobilisation of organised interests in policy making: Access activity and bias
in the ‘group system’ ESRC (RES-000-22-1932 to Halpin 2006-9]. The authors would like
to acknowledge the support of the SE Library and the staff at the SE Edinburgh document
storage facility for facilitating frequent access to the consultation files. Very special thanks
to Graeme Baxter for assistance in collecting and coding the consultation data. Thanks to
Shaun Bevan for compiling the Stata code for calculating the HHI and entropy scores. An
earlier version of this paper was presented at the Southern Political Science Association
Annual Meeting (2009) and we thank participants and the discussant for their comments.
Finally we wish to thank the journal reviewers for their careful reading of the manuscript
and their constructive comments.
NOTES
1It is important to note that, despite the theoretical formulation (that is, ESA), the empirical findings from population ecology
studies of organized interests support the idea that survival-related specialization is achieved through the partitioning of
resources (like constituency support and funding) and not in relation to avoiding crowding in a given policy space.
2In agreeing with Salisbury’s finding of (numerical) institutional dominance, Lowery and Gray (1998) have argued that this
is because of ‘choices’ to lobby separate from collective associations. This may make sense in analysing lobby data in the US
where ‘institutions’ is assumed to be a synonym for businesses, but our data includes institutions of a governmental nature
where collective action is not an alternative.
3It is also important to acknowledge that these theoretical expectations – which are stated for the UK and US cases – may
not hold for other (sub) national cases. For instance, as one reviewer suggested, one could imagine reasons why corporatist
countries in Western Europe might be approached with different expectations. The limited scope for accreditation as a
negotiating partner with the state might be argued to punish a lack of generalized constituency coverage and constrain
variations in policy breadth by key groups. Thus, while the specific expectations might differ for some other systems, the
theme of breadth of engagement remains salient. Arguably, the reported decline of corporatism in many countries gives this
Public Administration Vol. 90, No. 3, 2012 (582–599)
©2011 Blackwell Publishing Ltd.
POLICY ENGAGEMENT BY ORGANIZED INTERESTS 597
theme more weight – see literature on Scandinavian corporatism and rise of ‘lobbyism’ in particular (see Christiansen and
Rommetvedt 1999).
4Scholars who have attempted such quantification exercises – even in small domains – have typically noted the difficulty
of their task and its necessary partiality (see, for instance, the May et al. (1998) study of trade associations). Work is underway
to count associative activity in the UK through coding of time-series data from the UK Database of Associations (see Jordan
and Greenan forthcoming). Nevertheless, the authors are aware of no current UK-wide or Scottish dataset (institutional or
otherwise) which aims to systematically map mobilization in specific issue contexts.
5It is important to recognize that as part of a broader process of UK devolution, Scotland was granted its own Scottish
Parliament and Scottish Executive in 1999 and, with that, responsibility for some issue areas (so-called ‘devolved matters’).
The UK parliament and government retain responsibility for so-called ‘reserved matters’. Devolved matters include, health
and social work; education and training; local government and housing; justice and police; agriculture, forestry and fisheries;
environment; tourism, sport and heritage; economic development and internal transport.
6As one reviewer suggests, our choice of arena might be expected to influence the type of organizations mobilizing, and thus
perhaps underplay the role of citizen groups. In principle this might make sense. However, consultations in the UK are almost
always openly accessible to any individual or organization, making it easy for even the most vociferous outsider to mobilize
in this arena. That being said, we in no way suggest that counting populations using data on media coverage or access to
parliamentary committees would surface precisely the same population. In other countries where access to administrative
arenas are more tightly controlled this might be a valid concern. More broadly, we accept that different arenas (legislative,
media and administrative) in the same national system will provide different pictures: as Salisbury (1984) demonstrated
25 years ago for the Washington scene.
7Respondents were asked if they thought that consultation results are largely disregarded by policy-makers. Opinions were
almost evenly divided, with broadly similar proportions of the sample agreeing with these statements, disagreeing, or
remaining non-committal.
8The Scottish government’s internal Consultation Good Practice Guidance (2008) recommends that departments, on completing
a consultation exercise, should deposit copies of responses with the Scottish Government Library and also post them on
the Scottish government web site. However, this guidance has not always been followed, and therefore not all consultation
documentation has made its way into the public domain.
9Criticism has been made in the past over using consultation invitation lists as data – mostly on the basis that access to lists is
very open (see Cavanagh et al. 1995). In this data we only map actors who actually responded to policy consultations.
10 Given that we engage with what is a largely US literature, we deploy the US term ‘policy domain’. The equivalent European
or UK term is policy sector or policy area.
11 A topic codebook is available (http://www.policyagendas.co.uk).
12 It should be noted that we dropped policy domains where less than 10 consultations occurred over the 25-year period.
Therefore the number of issues covered drops to 1680 and discrete actors to 89706.
13 Note that we did calculate both HHI scores and Shannon’s H entropy scores for all groups (−.977 correlation), and our broad
analysis/findings remain unchanged regardless of the measure used. Of course, some values for individual groups do differ.
For the reasons discussed in the text, we chose to use Shannon’s H for our analysis.
14 The correlation between ‘domains’ and ‘issue entropy’ measures is highest (.88), which makes sense given that a higher
entropy scores in part means an organized interest is engaged in more domains.
15 This finding also underlines and reinforces the need to resist the tendency among scholars to conceive of organized interests
as though they did not include governmental actors.
16 As one reviewer noted, the potential for mobilization ought to be increased by the supply of more possible issues to engage
on. There is an increase in the overall number of consultations over time for which we have data (generally increasing on
average since 1985, but with no obvious break point for devolution). This does not have any impact on entropy itself, but of
course can increase the entropy X activity measure.
17 In our calculations an organization is counted only once per time period, regardless of how many times it is active in that
time period.
REFERENCES
Baumgartner, F.R. and B.I. Leech. 1998. Basic Interests: The Importance of Groups in Politics and Political Science. Princeton, NJ:
Princeton University Press.
Baumgartner, F.R. and B.I. Leech 2001. ‘Interest Niches and Policy Bandwagons: Patterns of Interest Group Involvement in
National Politics’, Journal of Politics, 63, 4, 1191– 213.
Baumgartner, F.R., B.D. Jones and M.C. MacLeod. 2000. ‘The Evolution of Legislative Jurisdictions’, Journal of Politics, 62, 2,
321– 49.
Beales, D., A. Carter, J. Chrispin, et al. Forthcoming. ‘Groups and the Limited Pluralism of the Set-Piece Consultation’, British
Journal of Politics and IR, in press.
Public Administration Vol. 90, No. 3, 2012 (582–599)
©2011 Blackwell Publishing Ltd.
598 DARREN R. HALPIN AND H.F. THOMAS III
Bevan, S. 2008. ‘Effects on the Scope of Agendas: The Reasons Why the Dog Did Not Bark’, Paper Presented at the Annual
Meeting of the Policy Agendas Projects in Barcelona, Spain, 19– 21 June.
Boydstun, A.E. 2008. How Policy Issues Become Front-Page News. PhD dissertation, The Pennsylvania State University, University
Park, PA.
Browne, W.P. 1990. ‘Organized Interests and Their Issue Niche: A Search for Pluralism in a Policy Domain’, Journal of Politics,
52, 2, 477– 509.
Cairney, P. 2008. ‘Has Devolution Changed the ‘‘British Policy Style’’?’, British Politics,1.
Cavanagh, M., M. Marsh and M. Smith. 1995. ‘The Relationship between Policy Networks and Sectoral and Subsectoral Levels:
A Response to Jordan, Maloney and McLaughlin’, Public Administration, 73, 1, 62–79.
Chaffee, S.H. and D.G. Wilson 1977. ‘Media Rich, Media Poor: Two Studies of Diversity in Agenda-Holding’, Journalism
Quarterly, 54, 3, 466– 76.
Christiansen, P.M. and H. Rommetvedt. 1999. ‘From Corporatism to Lobbyism? – Parliaments, Executives, and Organized
Interests in Denmark and Norway’, Scandinavian Political Studies, 22, 3, 195– 220.
Culbertson, H.M. 1992. ‘Measuring Agenda Diversity in an Elastic Medium: Candidate Position Papers’, Journalism Quarterly,
69, 4, 938– 46.
Finer, S. 1958. Anonymous Empire: A Study of the Lobby in Great Britain. London: Pall Mall.
Grant, W. 2008, ‘The Changing Patterns of Group Politics in Britain’, British Politics, 3, 2, 204–22.
Gray, V. and D. Lowery. 2000 [1996]. The Population Ecology of Interest Representation: Lobbying Communities in the American States,
Ann Arbor, MI: University of Michigan Press.
Halpin. D. 2011. ‘Organized Interests and Cascades of Attention: Unpacking Policy Bandwagon Dynamics’, Governance, 24, 2,
205– 30.
Halpin, D. and G. Jordan. Forthcoming. ‘Estimating Group and Associational Populations: Endemic Problems but Encouraging
Progress?’, in D. Halpin and G. Jordan (eds), The Scale of Interest Organization in Democratic Politics. Basingstoke: Palgrave, in
press.
Heaney, M.T. 2004. ‘Outside the Issue Niche: The Multidimensionality of Interest Group Identity’, American Politics Research, 32,
6, November.
Heaney, M.T. 2007. ‘Identity Crisis: How Interest Groups Struggle to Define Themselves in Washington’, in A.J. Cigler and B.A.
Loomis (eds), Interest Group Politics, 7th edn. Washington, DC: Congressional Quarterly Press, pp. 279– 300.
Heinz, J.P., E.O. Laumann, R.L. Nelson and R.H. Salisbury. 1993. The Hollow Core. Cambridge, MA: Harvard University Press.
John, P. 2006. ‘The Policy Agendas Project: a Review’, Journal of European Public Policy, 13, 7, 975– 86.
Jordan, G. and W. Maloney. 2001. ‘Britain: Change and Continuity Within the New Realities of British Politics’, in C. Thomas
(ed.), Political Parties and Interest Groups: Shaping Democratic Governance. Boulder, CO: Lynne Rienner, pp. 27– 44.
Jordan, G. and J. Greenan. Forthcoming. ‘The Changing Contours of British Representation: Pluralism in Practice’, in D. Halpin
and G. Jordan (eds), The Scale of Interest Organization in Democratic Politics. Basingstoke: Palgrave, in press.
Lasorsa, D.L. 1991. ‘Effects of Newspaper Competition on Public Opinion Diversity’, Journalism Quarterly, 68, 1, 38– 47.
Lowery, D. and V. Gray. 1998. ‘The Dominance of Institutions in Interest Representation: A Test of Seven Explanations’, American
Journal of Political Science, 42, 1, 231– 55.
Lowery, D. and V. Gray. 2004. ‘A Neopluralist Perspective on Research on Organized Interests’, Political Research Quarterly, 57,
1, 163– 75.
May, T.C., J. McHugh and T. Taylor. 1998. ‘Business Representation in the UK since 1979: The Case of Trade Associations’,
Political Studies, 46, 2, 260– 75.
McCombs, M. and J.-H. Zhu. 1995. ‘Capacity, Diversity, and Volatility of the Public Agenda: Trends from 1954 to 1994’, Public
Opinion Quarterly, 59, 4, 495– 525.
McKenzie, R.T. 1958. ‘Parties, Pressure Groups and the British Political Process’, Political Quarterly, 29, 1, reprinted in R. Kimber
and J. Richardson (eds), Pressure Groups in Britain. London: Dent [1974], pp. 276– 88.
Page, E.C. 2003. ‘The Civil Servant as Legislator: Law Making in British Administration’, Public Administration, 81, 4, 651– 79.
Peter, F. and C.H. de Vreese. 2003. ‘Agenda-rich, Agenda-poor: A Cross-national Comparative Investigation of Nominal and
Thematic Public Agenda Diversity’, International Journal of Public Opinion Research, 15, 1, 44– 64.
Rasmussen, A. 2010. ‘Diversity in Interest Representation in Denmark, The Netherlands and the United Kingdom’, paper
presented at the annual conference of the Dutch Public Administration Association, 25– 26 November 2010.
Richardson, J.J. and A.G. Jordan. 1979. Governing Under Pressure. Oxford: Martin Robertson.
Rose, R. 1984. Do Parties Make a Difference?, 2nd edn. London: Macmillan.
Richardson, J. 1999. ‘Pressure Groups and Parties: A ‘‘Haze of Common Knowledge’’ or Empirical Advance of the Discipline?’,
in J. Hayward, B. Barry and A. Brown (eds), The British Study of Politics in the Twentieth Century. Oxford: Oxford University
Press.
Salisbury, R.H. 1984. ‘Interest Representation: The Dominance of Interest Groups’, American Political Science Review, 78, 1, 64– 78.
Public Administration Vol. 90, No. 3, 2012 (582–599)
©2011 Blackwell Publishing Ltd.
POLICY ENGAGEMENT BY ORGANIZED INTERESTS 599
Schlozman, K. 2009. ‘Who Sings in the Heavenly Chorus? The Shape of the Organized Interest System’, in J. Berry (ed.), The
Oxford Handbook of American Political Parties and Interest Groups. Oxford: Oxford University Press, pp. 425– 50.
Shannon, C.E. 1950. ‘Prediction and Entropy of Printed English’, The Bell System Technical Journal, 30, 50– 64.
Shannon, C.E. and M. Weaver. 1949. The Mathematical Theory of Communication. Urbana, IL: University of Illinois Press.
Sheingate, A.D. 2006. ‘Structure and Opportunity: Committee Jurisdiction and Issue Attention in Congress’, American Journal of
Political Science, 50, 4, 844– 59.
Scottish Executive. 2004. Consultation Good Practice Guidance. Edinburgh: Scottish Executive.
Skocpol, T. 1999. ‘Advocates Without Members: The Recent Transformation of American Civic Life’, in T. Skocpol and M.P.
Fiorina (eds), Civic Engagement in American Democracy, 2nd edn. Washington, DC: Brookings Institution, pp. 461– 509.
Talbert, J. and M. Potoski. 2002. ‘The Capacity, Complexity, and Volatility of the House Agenda’, in F.R. Baumgartner and B.D.
Jones (eds), Policy Dynamics. Chicago, IL: University of Chicago Press.
Walker, J.L. 1991. Mobilizing Interest Groups in America. Ann Arbor, MI: University of Michigan Press.
Wolfe, M., A.E. Boydstun and F.R. Baumgartner. 2009. ‘Comparing the Topics of Front-Page and Full-Paper Stories in the New
York Times’, Annual Meeting of the Midwest Political Science Association. Chicago, IL, 2– 5 April.
Date received 6 December 2010. Date accepted 1 April 2011.
Public Administration Vol. 90, No. 3, 2012 (582–599)
©2011 Blackwell Publishing Ltd.
