Download Explaining Media and Congressional Attention to Global Climate

2009
Institute for Science, Technology and Public Policy Explaining Media and
Congressional Attention to
Global Climate Change,
1969-2005:
An Empirical Test of Agenda Setting
Theory
Xinsheng Liu, Eric Lindquist, & Arnold Vedlitz
Forthcoming in Political Research Quarterly
No part of this paper may be copied, downloaded, stored, further transmitted, transferred, distributed, altered,
or otherwise used in any form or by any means, except: (1) one stored copy for personal, non-commercial use,
or (2) prior written consent. No alteration of the paper or removal of copyright notice is permitted.
Explaining Media and Congressional Attention to
Global Climate Change, 1969-2005
An Empirical Test of Agenda Setting Theory
Xinsheng Liu
Corresponding Author
Associate Research Scientist,
Institute for Science, Technology and Public Policy
[email protected]
Eric Lindquist
Associate Research Scientist
Institute for Science, Technology and Public Policy
Arnold Vedlitz
Bob Bullock Chair in Government and Public Policy,
Bush School of Government and Public Service
Director, Institute for Science, Technology and Public Policy
Institute for Science, Technology and Public Policy
Bush School of Government and Public Services
Texas A&M University
4350 TAMU
College Station, TX 77843-4350
979.862.8855
Authors’ Note: The data utilized in this study is based upon research supported under Award No. NA04OAR4600172
from the National Oceanic and Atmospheric Administration (NOAA), U.S. Department of Commerce. The statements,
findings, conclusions, and recommendations are those of the authors and do not necessarily reflect the views of the
National Oceanic and Atmospheric Administration or the Department of Commerce. An earlier version of this paper
th
was presented at the 77 Annual Conference of the Southern Political Science Association, January 5-7, 2006, Atlanta.
We thank our graduate research assistant Charles Lindsey for providing statistical assistance to this project. We also
thank four anonymous reviewers for their thoughtful comments and excellent suggestions.
Institute for Science, Technology and Public Policy
Bush School of Government and Public Service, Texas A&M University
Explaining Media and Congressional Attention to
Global Climate Change, 1969-2005
An Empirical Test of Agenda Setting Theory
Abstract
Agenda theories suggest that problem indicator, focusing event, and information feedback enhance
issue attention. However, few studies have systematically tested this. This study, using time series data
and vector autoregression (VAR), examines how climate problem indicator, high-profile international
event, and climate science feedback influence media and congressional attention to global warming and
climate change. Our findings confirm that these attention-grabbing factors indeed generally promote issue
salience, but these factors may work differently across agenda venues. Attention inertia, inter-agenda
interaction, and partisan advantage on agenda setting are also included and analyzed in the VAR
modeling. Implications of the study and recommendations for future research are discussed in conclusion.
Keywords: news media; congress; issue attention; agenda setting; global warming; climate change
1. Introduction
The question of why and how some issues are placed on the agenda has long been of interest to
political scientists and other social science scholars. At any given time, policy makers are confronted with
many complex public issues. For a public issue to be seriously considered and handled in the
policymaking process, a necessary condition is that the issue must capture the attention of policy elites
(Cohen 1963; Cobb and Elder 1983; Baumgartner and Jones 1993; Jones 1994; Rochefort and Cobb
1994; Kingdon 1995; Jones and Baumgartner 2005).
One important question for agenda setting scholars is what factors may contribute to higher levels of
attention paid to a public issue. In Agendas, Alternatives, and Public Policies (1995, 90-103), Kingdon
argues that 'problem indicators,' 'focusing events,' and 'feedback' can facilitate bringing a public issue to
decision makers' attention and help that issue achieve higher status on the agenda. In a series of
1
important works on agenda setting, Jones and Baumgartner argue, like Kingdon, that relevant
information surrounding an issue (including problem indicators and information feedback) is important to
decision makers' attention but, so too, is the significant event that provides sudden information shock to
the policy system (Jones and Baumgartner 2005; Jones 1994, 2001; Baumgartner and Jones 1993).
Kingdon and Jones and Baumgartner may differ in the terms they use to describe what factors may
promote issue attention, but they all agree that the dynamics of incoming information sources and flows,
reflected in changing problem indicators, occurrence of focusing events, and information feedback are
essential for understanding how issues move from obscurity to visibility on policy agendas.
In this study, we draw upon the common theoretical elements in agenda theories and apply them to
the issue of global climate change. We are particularly interested in examining whether and how problem
indicators, focusing events, and scientific feedback induced the US news media and Congress to be
attentive to global climate change during the period from 1969 through 2005. First, we briefly review
existing research on what factors contribute to capturing the attention of policy elites. Second, we develop
our hypotheses and describe how we measure the dependent and independent variables as well as how
we collect our data. Third, we employ vector autoregressive (VAR) method for hypothesis-testing and
present the results of analysis. Finally, we summarize our main findings, discuss some implications of our
study, and make a few recommendations for future research. These empirical tests of attention driving
forces of agenda theory are among the first to systematically examine and test the key underlying
relational elements of the theory.
2. Attention-Grabbing Factors in Agenda Setting
Many scholars have examined various factors that contribute to a higher level of attention to a
particular issue. Earlier studies find that issue attention levels are associated with expansion of political
conflicts (Schattschneider 1960), issue-attention cycles (Downs 1972), specific issue characteristics such
Institute for Science, Technology and Public Policy
Bush School of Government and Public Service, Texas A&M University
1
as social significance and temporal relevance (Cobb and Elder 1983), policy elites' interests and
ideologies (Vertzberger 1990), and various issue framing effects (Fiske and Taylor 1984; Hilgartner and
Bosk 1988; Iyengar and Kinder 1987; Iyengar 1990). In Kingdon's Multiple Streams Theory (1995) and
Jones and Baumgartner's Punctuated Equilibrium Theory (2005; see also Baumgartner and Jones 1993;
Jones 1994, 2001), issue attention dynamics is systematically addressed from an agenda system
perspective. Both theories identify several major system factors that can bring a public problem to the
attention of decision makers. In Kingdon's theory, attention attractors in agenda setting include changing
problem indicators, focusing events, and information feedback. In Jones and Baumgartner's theory, issue
attention is enhanced (sometimes suddenly enhanced) by the intrusion of new (or previously overlooked)
information into the policy agenda setting process, and the intrusion of new information is usually
associated with changing social conditions and problem indicators, significant events, and information
feedback that is looped back into the policy system.
2.1. Problem Indicators
For both Kingdon and Jones and Baumgartner, one factor that may increase attention to a problem is
the changing factual indicators surrounding the problem. Problem indicators may come in various shapes,
sizes, formats, and from multiple sources. Some indicators may come from direct experience (paying
more at the gas station indicates certain conditions within the energy sector), but most indicators (such as
unemployment rates, economic growth rates, crime rates, mortality rates, high school drop-out rates, or
highway death rates) are presented to the public and policy elites in an abstract, index-type, quantitative
format from second-hand data sources.
For the issue of global warming and climate change, there are many kinds of factual and baseline
information indicators, ranging from greenhouse gas emissions to precipitation patterns, from arctic ice
sheet melting to sea level rise, from complicated long-term climate change projections to short-term
extreme weather conditions, and from climate-related economic losses, and infectious diseases to landuse practices. Several studies have explored how short-term, weather-related, extreme conditions (such
as unusually hot temperatures and droughts), amplified by interested constituencies, may contribute to
increase attention to the climate change issue (Ungar 1992; Shanahan and Good 2000), but no previous
study has examined the relationship between long-term, system-level, climate change indicators and
national media and congressional attention. In this study, we use two systematic time series of climate
change indicators—the atmospheric carbon dioxide (CO2) concentration levels and the US Climate
Extremes Index (CEI)—to examine the effect of problem indicators on US policy elites' attention to climate
change. We shall return to these two indictors in the variable measurement section.
2.2. Focusing Events and Information Shocks
According to agenda theorists, the availability of problem indicators and their amplification by
motivated stakeholders may not be enough to cause a condition to rise on the policy agenda as a
problem. Something must happen to push that concern above the noise threshold of other issues. Such
factors are referred to as focusing or triggering events (Kingdon 1995; Birkland 1997, 1998; see also
Cobb and Elder 1983; Dearing and Rogers 1996), or in Jones and Baumgartner's model as information
shocks. Much of the existing work on focusing events/information shocks has concentrated on natural or
man-made crises and disasters. Birkland, in particular, has advanced our understanding of the "politics of
disasters," or the reaction to a focusing event by various stakeholders including the media, decision
makers, and interest groups (Birkland 1996, 1997, 1998, 2004; see also Birkland and Nath 2000;
Lawrence and Birkland 2004).
For agenda scholars, focusing events/information shocks, with the cumulative effect of raising
awareness of the problem, can reinforce existing indicators and become a principal lever for getting an
issue moved to the front of the line (Kingdon 1995; Baumgartner and Jones 1993; Jones 1994; Birkland
1997, 1998). Prominent examples of focusing events include the Three Mile Island incident (Rankin,
Nealey and Melber 1984; Baumgartner and Jones 1993), the September 11, 2001 terrorist attacks
(Birkland, 2004), and Hurricanes Katrina and Rita in 2005 (Liu, Vedlitz and Alston 2008). Various effects
of focusing events or information shocks on issue attention, definition, and access to policy agendas have
been examined by scholars. These studies find that focusing events can bring visibility to hidden issues
(Kingdon 1995; Baumgartner and Jones 1993), introduce new dimensions and new policy alternatives
into policy debates (Hilgartner and Bosk 1988; Baumgartner and Jones 1993; Jones 1994; Birkland 2004),
mobilize interest groups (Birkland 1998), expand issue conflict (Schattschneider 1960), and re-organize
Institute for Science, Technology and Public Policy
Bush School of Government and Public Service, Texas A&M University
2
issue networks and allies (Baumgartner and Jones 1993). Whether such events reinforce existing
indicators or serve as new catalysts to agenda mobilization, the important role of focusing events and
resulting information shocks in sparking and enhancing issue attention is generally shared by agenda
scholars. Following this thinking, this study will examine how focusing events in the climate change
domain generate issue attention, with a particular focus on how certain high-profile international events
may influence the status of the climate change issue on US congressional and news media agendas.
2.3. Feedback
The third factor that may grab policy elites' attention and promote the prominence of an issue is what
Kingdon calls "feedback" and Jones and Baumgartner call "information feedback." Both theories, in line
with the systems analysis literature (Easton 1965; Richardson 1991), maintain that in a democratic polity,
policy elites constantly receive feedback about the operation of existing governmental programs and hear
complaints about public issues being overlooked or neglected.
Feedback sometimes comes internally from governmental officials' routine activities and daily
experience in administering or monitoring existing programs. More often, feedback comes to
governmental officials from external, non-governmental, venues. Public opinion polls, citizen complaints,
interest group pressures, and opinion leaders' criticisms all provide feedback that can amplify the
attention given to an issue. Among various non-governmental venues, the scientific community
sometimes plays a key role in providing information feedback to policy systems, particularly for those
issues involving scientific and technological uncertainty and complexity (May 1990; Birkland 2001).
Through scientific research activities and products (surveys, reports, journal articles, scholarly books), the
science community provides data on natural and social conditions, assesses policy programs, and
identifies problems. In this study, we will attempt to measure the flow of information feedback from the
climate science community and empirically examine the influence of the information provided by this
community on the issue of climate change on US media and congressional agendas.
3. Hypotheses on Media and Congressional Attention Dynamics
Climate change as an issue has appeared in the US news media and in the congressional hearings
since the late 1960s and early 1970s, but the status of the issue on both media and congressional
agendas has fluctuated over the last four decades (see next section for a brief review of the issue status
in both agendas). What explains the rise and fall of media and congressional attention to the climate
change issue? Following the basic arguments of our agenda theorists, that problem indicators, focusing
events, and information feedback are the key factors for an issue to gain attention and agenda status, we
propose the following hypotheses:
 H1a/H1b: News media/congressional attention to climate change is affected by the
indictors of the problem. More specifically, the more severe the problem (reflected by
climate change indictors), the more attention is paid by the news media/Congress.
 H2a/H2b: News media/congressional attention to climate change is positively influenced
by the occurrence of focusing events in the climate change domain.
 H3a/H3b: News media/congressional attention to climate change is positively associated
with the amount of scientific information feedback regarding climate change. More
specifically, the stronger the feedback is, the more attention is paid to the issue by the
news media/Congress.
In addition to the three attention-grabbing factors discussed by agenda theorists, several other forces
may also affect media and congressional attention. First, many scholars have noted that there is a strong
tendency toward inertia in policy elites' attention dynamics – that is, attention paid to a social issue at a
time is strongly associated with the amount of attention to that issue at previous times (Baumgartner and
Jones 1993; Jones 1994; Wood and Peake 1998; True 1999; Soroka 2002; Liu 2006). Second, previous
studies suggest there is often a spillover effect among different policy arenas and institutions (Cobb and
Elder 1983; Hilgartner and Bosk 1988; Baumgartner and Jones 1993; Kingdon 1995, 190-194). As
Hilgartner and Bosk (1988, 67) state: "Through a complex set of linkages, activities in each arena
propagate throughout the others. If a social problem rises in one institution, it is likely to spread into
others." Nevertheless, several recent studies demonstrate that inter-agenda interactions could be far
more complicated than a simple spillover or convergence, suggesting that sometimes considerable
disconnect or divergence could be found across different arenas (Lawrence and Birkland 2004; see also
Jones and Baumgartner 2005). Third, some studies indicate that majority party in Congress may exert
Institute for Science, Technology and Public Policy
Bush School of Government and Public Service, Texas A&M University
3
significant influence on issue agenda setting. More specifically, Republican party as a whole, with its
unfavorable position toward environmentalism and openly hostile stance on global warming, may inhibit
climate change issue in the agenda-setting process (Dunlap, Xiao and McCright 2001; McCright and
Dunlap 2003; Mooney 2005; see also Jones and Baumgartner 2005, 1-2).
Based on these observations, three additional hypotheses – attention inertia, inter-agenda
spillover, and partisan advantage – are proposed here:
 H4a/H4b: News media/congressional attention to climate change at a given time is
positively influenced by news media/congressional attention at previous times.
 H5a/H5b: There is a positive inter-agenda interaction between media and congressional
attention to climate change. In other words, media attention and congressional attention
positively influence each other.

H6a/H6b: There is a negative relationship between Republican's control in Congress and
media/congressional attention to climate change. In other words, the stronger the Republican
advantage, the more restriction of issue access and thereby lower news/congressional attention
to climate change.
4. Dependent Variables and Attention Measurement
In this study, we use proxy measures to trace the rise and fall of US media and congressional
attention to climate change over the time period from 1969 to 2005.
4.1. Media Attention to Climate Change
To measure US media attention to global climate change, we calculated the annual number of climate
change articles published in the New York Times. This method of using the annual number of articles to
measure the media attention to a particular public issue has been widely used in news media agenda
setting studies (see, for examples, Baumgartner and Jones 1993; Soroka 2002; Jones and Baumgartner
2
2005).
We used the LexisNexis online searchable newspaper database (which contains a large volume of
historical and current news stories from various US newspapers, including all New York Times articles
starting from January 1, 1969), and searched relevant articles containing one of the following key terms:
"climate change," "global warming" and "greenhouse gas." There were two reasons to use these three
key terms for our article search. First, according to previous studies on the history of the climate change
issue, these three terms have been frequently and consistently used in media stories, policy debates, and
scientific research over the last several decades (Houghton 2004; Dessler and Parson 2006). Second,
other empirical studies suggest that using these three key terms (or similar ones) will capture most
climate change articles from LexisNexis (McComas and Shannahan 1999; Liu, Vedlitz and Alston 2008).
To test this, we used a number of additional keywords also related to global warming and climate change
(e.g., "sea level rise," "melting ice sheet," "Kyoto protocol," etc.), and conducted additional rounds of
article search in LexisNexis. We found that the additional keywords did not show significant advantage
3
over the three keyword approach.
4
With the validity check of the three keywords, we followed specific search procedures in LexisNexis.
Our LexisNexis search yielded a total of 4,197 articles published in the New York Times from 1969
through 2005. We took additional steps to ensure that these articles fairly represent the US news media's
attention to climate change. First, we verified that the retrieved articles did not contain a lot of "noise," i.e.,
articles that mentioned global warming or climate change yet were mainly about another topic. We drew a
random 10% sample (420 articles) from these 4,197 search results and carefully reviewed each sample
article. Among the 420 sample articles we reviewed, 268 were validated as true climate change stories.
Time series graphs of the 268 true climate change articles and the 4,197 raw search results yielded
almost identical patterns, and the correlation between the annual numbers of sample true stories and the
raw search results is extremely high (0.98, p<.001). Second, we compared the coverage patterns of the
New York Times with other major national newspapers. Using the same key terms, we conducted search
in LexisNexis and collected the annual number of climate change articles from the Los Angeles Times
and the Chicago Tribune. Almost identical patterns and high correlations were found among these three
newspapers. Based on these examinations, we are fairly confident that the annual number of the New
York Times articles is a good proxy measure of the US media attention to climate change. Figure 1
Institute for Science, Technology and Public Policy
Bush School of Government and Public Service, Texas A&M University
4
displays our search results, showing the annual number of climate change articles published in the New
York Times from 1969 through 2005.
The graph in Figure 1 shows an overall trend in the news media: despite some short-term declines in
coverage, climate change has steadily gained more and more attention in the news media over the last
four decades. Prior to the 1980s, climate change was generally a low-salience issue, with very little
coverage in the media (2.9 articles per year in the 1970s). Starting from the early 1980s, climate change
began to gain increasing visibility in the media, and the average annual number of articles increased to
56.4 articles per year in the 1980s. There were some fluctuations in the 1990s, but the overall media
attention continued to grow with an average of 157.7 articles per year. The first six years of the 2000s
continued this trend and media attention reached a new high, with an average 337.3 articles per year
from 2000 through 2005.
Figure 1: Annual Number of Climate Change Articles, New York Times, 1969-2005
450
400
350
300
Number of
250
New York Times
Articles
200
150
100
50
0
Year
4.2. Congressional Attention to Climate Change
To measure congressional attention, we collected the annual number of congressional hearings on
climate change and global warming. Similar methods to measure the attention paid by Congress as a
whole to an issue were used in several previous policy agenda studies (Baumgartner and Jones 1993;
Jones and Baumgartner 2005; see also MacLeod 2002; Hunt 2002; Sheingate 2006).
There are three major online sources available for collecting congressional hearing data: the Policy
Agenda Project, the Thomas website (a legislative information website from the Library of Congress), and
5
LexisNexis Congressional Publications. All three datasets are searchable by key terms. We choose to
use the LexisNexis Congressional Publications, because this database includes all hearings from 1969 to
the present, while the Policy Agenda Project contains hearing data only up to 1999 and the Thomas
database begins with 1989 hearings. Using the same three keywords as in the news media search, the
LexisNexis congressional search retrieved a total of 148 congressional hearings from 1969 to 2005.
Figure 2 shows the annual number of hearings on climate change and global warming from the
LexisNexis congressional search.
Institute for Science, Technology and Public Policy
Bush School of Government and Public Service, Texas A&M University
5
As Figure 2 shows, congressional attention to global warming and climate change fluctuated over the
last several decades. From the late 1960s to the mid-1980s, congressional hearing activity was minimal.
In the late 1980s, congressional attention to climate change rapidly increased—in 1989, twenty one
hearings were held in Congress. In the following years, there were short term declines of interest, but the
overall congressional attention to climate change was sustained at about eight hearings per year during
the period of 1988 through 2005.
5. Independent Variables and Measurement
Our literature review suggests that attention to a particular social issue may be affected by the
following variables: problem indicators, focusing events, information feedback, attention inertia, interagenda spillover and partisan influence. Now, let us turn our discussion to how we measured these
variables.
Figure 2: Annual Number of Congressional Hearings on Climate Change, 1969-2005
25
20
15
Number of
Congressional
Hearings
10
5
0
Year
5.1. Indicators of Global Warming and Climate Change
Among various indicators of global warming and climate change, the most fundamental one is
perhaps the global carbon dioxide (CO2) concentration data – also known as the "Keeling Curve" named
after Charles David Keeling. As a leading authority on atmospheric greenhouse gas accumulation and a
climate science pioneer at the Scripps Institution of Oceanography, UC San Diego, Keeling started
measuring and monitoring the concentration of carbon dioxide in the atmosphere in the late 1950s.
Keeling's data indicate an upward curve that shows a steady increase in global atmospheric
concentrations of carbon dioxide—one of the major greenhouse gases causing the earth's atmosphere to
heat up and contributing to global warming and climate change.
Several studies demonstrates the significance of Keeling's CO2 measurements in raising the alarm
about global warming and setting the scene for the debate over climate change problems (Bodansky
2001; Weart 2004; Hansen 2005; Briggs 2007; Nisbet 2007), but whether the news media and Congress
systematically respond to this problem indicator over a relatively long time period is an empirical question.
Institute for Science, Technology and Public Policy
Bush School of Government and Public Service, Texas A&M University
6
6
To measure the increasing CO2 levels, we calculated the first order difference of Keeling data —Net
Keeling Level (NKL), which is simply the net change between the levels of carbon dioxide concentration
at the current year t and the previous year t-1.
If the net change of Keeling's CO2 concentration level (NKL) represents a general indicator of the
climate change problem at the global level, the US Climate Extreme Index (CEI) may constitute a specific
indicator of the problem facing the United States. The CEI was developed by the National Climatic Data
Center (NCDC) of the National Oceanic and Atmospheric Administration (NOAA) as a way to quantify
observed abnormal changes in the climate system within the contiguous Unites States for each calendar
year since 1910. It aims to summarize and present "a complex set of multivariate and multidimensional
climate changes in the United States so that the results could be easily understood and used in policy
decisions made by nonspecialists in the field" (NCDC 2006). The CEI is calculated based on a set of
recorded climate information, including maximum and minimum temperatures, extreme precipitation
events, drought severity, and landfalling tropical storms and hurricane wind velocity. The value of the CEI
ranges from 0 to 100, with 0 indicating no climate extremes were recorded for that year and 100 meaning
the entire contiguous United States experienced extreme climate conditions for all the indicators
7
throughout the year.
In this study, we use both NKL and CEI as objective indicators of the climate change problem facing
the United States from 1969 to 2005. Figure 3 displays the two data series. Note that the two series have
different scales in Figure 3 – NKL is scaled to the left y-axis whereas the CEI is scaled to the right y-axis.
Both series show an overall upward trend, suggesting the conditions associated with global warming and
climate change have been getting worse over the last several decades.
Figure 3: Net Change of CO2 Level (NKL) and Climate Extreme Index (CEI), 1969-2005
NKL: Net Keeling Level of CO2
CEI: Climate Extreme Index (right axis)
4.5
45
4
40
3.5
35
3
30
2.5
25
2
20
1.5
15
1
10
0.5
5
0
0
Year
Institute for Science, Technology and Public Policy
Bush School of Government and Public Service, Texas A&M University
7
5.2. International Focusing Events
Much of the existing work on focusing events suggests that crises or natural disasters spark intense
issue attention. However, in the case of climate change, it is difficult, if not impossible, to directly link any
natural disaster with global climate change in a causal manner. In the climate science community,
scholars continue to debate whether there is linkage between severe natural events and global climate
8
change. Among the general public, a national survey released by the Pew Research Center (2005)
indicates that only one in four Americans assumed any linkage between the severity of Hurricane Katrina
and global climate change, while two-thirds believed it was just the kind of severe weather event that
happens from time to time.
Focusing events include major crises and natural disasters, but "focusing events are not always so
straightforward" (Kingdon 1995, 96). More often, focusing events take other forms: scandals, trials,
protests and demonstrations, international settlements, agreements, or treaties, creation of new
institutions or reconfiguration of existing organizations, scientific discoveries, and even prominent books
(e.g., Rachel Carson's Silent Spring) (Kingdon 1995; Hilgartner and Bosk 1988). In line with this thinking,
we argue that certain high profile, symbolic, landmark, international events, such as the creation of a
large scale international organization or institution on global warming and climate change, may constitute
a focusing event in the climate change domain. In this regard, we expand the traditional definition of
focusing event concept to include other types of events.
There have been many climate change events over the last several decades. In this study, we are
particularly interested in the major events occurring at the international/multinational level. We define an
international focusing event in the climate change field as an historical and highly publicized event that
involves many countries and/or international institutions. More specifically and operationally, an event
must meet at least one of the following four criteria to qualify as an international focusing event (IFE): (a)
creation of an unprecedented international agreement, protocol, or treaty on mitigating global warming
and climate change; (b) establishment of a new international or intergovernmental institution on climate
change; (c) occurrence of a world-wide, high profile, convention or conference on global warming and
climate change; or (d) release of a new, landmark type, scientific assessment on climate change by a
highly reputable international scientific organization.
Based on our definition and operational criteria, we examined a variety of international/multinational
events on global warming and climate change. Three official reference documents with relatively
comprehensive chronicle records on major climate change events were closely reviewed by our
9
researchers. Eight international events met our operational definition and were identified as historical
milestones on global warming and climate change. These events are (1) the Montreal Protocol in 1987; (2)
the Toronto Conference on the Changing Atmosphere and the creation of the Intergovernmental Panel on
Climate Change (IPCC) under the United Nations Environmental Programme and the United Nations
World Meteorological Organization in 1988; (3) the IPCC First Assessment Report on Climate Change
and the World Climate Conference in Geneva in 1990; (4) the creation of the United Nations Framework
Convention on Climate Change (UNFCCC) in 1992; (5) the IPCC Second Assessment Report on Climate
Change and the First Conference of Parties to the UNFCCC (i.e., COP-1, which produced the "Berlin
mandate") in 1995; (6) the Second Conference of Parties to the UNFCCC (COP-2) in Geneva in 1996; (7)
the Third Conference of Parties to UNFCCC (COP-3) in Kyoto and the Kyoto Protocol in 1997; and (8) the
10
IPCC Third Assessment Report on Climate Change in 2001.
All these international focusing events (IFE) were coded as a dummy variable in our dataset.
Focusing events may have both contemporaneous and lagged effects on media and congressional
agendas. We shall return to this point and address how we incorporate both immediate and lagged
effects of focusing events in the VAR modeling section later.
5.3. Scientific Feedback
The third independent variable in our study is the information feedback from the climate science
community. This community constantly provides feedback to journalists and policy makers through a
variety of activities and channels: research projects and funding proposals, scientific discoveries and
research findings, academic publications, public speeches, congressional testimonies, participation in
government consulting and science advisory committees, memberships in academic societies and
associations, presentations at professional conferences and meetings, and popular science articles in
newspapers and magazines. Due to the wide-ranging activities and efforts, it is a rather difficult task to
measure the feedback from the climate science community. In this study, we choose to utilize a proxy
Institute for Science, Technology and Public Policy
Bush School of Government and Public Service, Texas A&M University
8
measure: scientific publications by climate scientists. Using scientific publications to trace the activity
levels of a science community is not new to agenda studies. For instance, Walker (1977), in his study of
the US Senate's problem recognition and selection process, used the number of articles in technical
journals to measure the professional activities of the science community in the areas of traffic safety and
occupational health.
For this study, we used the three same key terms —"climate change, " "global warming," and
"greenhouse gas"— to search climate change related publications in the Science Citation Index
Expanded (SCI) and the Social Sciences Citation Index (SSCI) databases available at the Web of
11
Science. Both SCI and SSCI databases are searchable by key terms, document types (article, abstract
of published item, book review, etc.), languages (English and other languages), author's address (for
example, US author), and publication year. We used the three key terms to search in article abstracts, set
12
the search span from 1969 to 2005, and restricted our search to publications in English language only.
A total of 33,557 climate change publications were found from SCI and SSCI databases for the period
13
from 1969 to 2005. The annual number of these scientific publications was then computed. To more
precisely measure the degree of scientific feedback from climate scientists, we calculated the Net
Scientific Publication (NSP) on climate change each year – the difference between the number of
scientific articles published in the current year and the number of articles published the previous year.
The NSP series is presented in Figure 4.
Figure 4: Net Scientific Publications (NSP) on Climate Change, 1969-2005
700
650
600
550
500
450
400
Net Scientific
350
Publications
300
250
200
150
100
50
0
Year
5.4. Attention Inertia, Inter-Agenda Connection and Partisan Advantage
Previous policy studies have indicated strong attention inertia in many policy issue fields. The
attention inertia variable in both news media and congressional attention models is simply the lagged
attention variable measured by past values of New York Times articles or congressional hearings on
climate change. Specific lag orders will be discussed and determined in the VAR modeling section below.
Since we are also interested in detecting whether there is inter-agenda spillover between the news media
Institute for Science, Technology and Public Policy
Bush School of Government and Public Service, Texas A&M University
9
and Congress, the congressional attention variable is then used as an independent variable in the media
attention model, and the media attention variable is used as an independent variable in the congressional
attention model. Republican advantage (REP) is measured by that party's control in Congress. REP is
coded as 1 for the years when the GOP was the majority party of both the House and Senate, 0 when it
controlled only one chamber, and -1 when both chambers were under a Democratic majority. This
measure will allow exploration of possible Republican advantage on restricting discussion of the climate
change issue.
6. VAR Modeling and Result
The time span of our analysis is from 1969 to 2005. It should be reiterated that the two dependent
variables we attempt to explain are measures of the media's attention and congressional attention to the
climate change issue, respectively, not specific proposals, actions, or policies to mitigate climate change.
Given the longitudinal nature of our data, we employ time-series regression methods to test the
hypotheses and examine the relationships between the dependent and independent variables. We
estimate two general equations to explain the media and congressional attention:
H
I
J
K
L
M
N
MAt = α 1 + ∑ β 1( h ) MAt − h + ∑ β 2(i ) CAt −i + ∑ β 3( j ) REPt − j + ∑ β 4( k ) NKLt − k + ∑ β 5(l ) CEI t −l + ∑ β 6( m ) IFE t − m + ∑ β 7 ( n ) NSPt − n + ε 1t
h =1
i =1
H
I
J
j =1
K
k =1
L
l =1
M
m =1
N
n =1
h =1
i =1
j =1
k =1
l =1
m =1
n =1
CAt = α 2 + ∑ β 8( h ) MAt − h +∑ β 9(i ) CAt −i + ∑ β 10( j ) REPt − j + ∑ β 11( k ) NKLt − k + ∑ β 12(l ) CEI t −l + ∑ β 13( m ) IFE t − m + ∑ β 14( n ) NSPt − n + ε 2t
These two equations treat the media attention (MA) and congressional attention (CA) as a dynamic
system. With the two-equation system, the current values for each of the two endogenous variables (i.e.,
MA and CA) are predicted by the following variables: (1) past values of the endogenous variable itself; (2)
past values of the other endogenous variable; and (3) exogenous variables with or without lags, including
the Republican advantage (REP) and the attention-attracting factors: net change of Keeling's CO2
concentration level (NKL), US Climate Extremes Index (CEI), international focusing event (IFE), and net
14
scientific publication (NSP).
Due to the high possibility of autocorrelated variables in the time series equations, we utilize the
Vector Autoregression (VAR) method for our model testing. One advantage of VAR modeling is that it
provides strong control for autocorrelations in the system. Another advantage of VAR is that it can provide
empirical evidence for Granger causality of two possibly interactive variables (i.e., media attention and
congressional attention in this study), particularly when there is no prior theory to establish direction of
15
causation between these two variables. The notion of Granger causality is an econometric definition of
causality: "Yt is causing Xt if we are better able to predict Xt using all information than if the information
apart from Yt had been used" (Granger 1969, 428). In other words, the test of Granger causality asks, "Is
there variation in one variable that cannot be explained by past values of that variable, but can be
explained by past values of another variable?" (Menard 1991, 57). If the answer is yes, then the second
variable Granger-causes the first.
In these two equations, h, i, j, k, l, m, and n are the number of lags for each variables. One major task
in VAR modeling is to determine the proper number of lags that should be used for these variables. The
strategy to determine lag length in vector autoregression is usually based on either theoretical plausibility
or lag-order selection statistics (Hill 1998; Soroka 2002). In our VAR testing, we chose one lag for each of
the endogenous variables based on both theoretical plausibility and pre-estimation of lag-order selection
tests. First, we think that one year is probably the most theoretically plausible lag number for the
endogenous variable, because it is quite unlikely, if not totally impossible, that the current value of one of
the endogenous attention variables, say congressional attention, would be affected by the past values of
the other endogenous variable (i.e., media attention) more than a year previously. Second, we utilized the
standard lag selection procedure to identify how many lags to use in the VAR. The pre-estimation of lag16
order selection statistic indicated that the optimal lag for the endogenous variables was one.
In the equations, real-world problem indicators (NKL and CEI), international focusing events (IFE),
and scientific feedback (NSP) are exogenous attention-grabbing variables. Since real-world data were
compiled and released on an annual basis and the current year indictors would not be available until
sometime in the following year, we chose one lag for NKL and CEI. We also chose one lag for the
Institute for Science, Technology and Public Policy
Bush School of Government and Public Service, Texas A&M University
10
scientific feedback variable (NSP), because it should take some time for the scientific publications to be
transformed into certain scientific feedback, and then for the media and the Congress to react to the
scientific feedback. For the international focusing events (IFE) variable, we used both current year t-0 and
last year t-1, as these events could have immediate impact and some short-term influence on the media
and the Congress. In addition, we expect the news media to react to international events more quickly
than Congress, and by including both the current and preceding year's events in the system of equations,
we can examine whether and how the international events affect the media and the Congress in terms of
different time frames. For the variable Republican control of Congress (REP), we did not use any lag,
because partisan advantage on issue agenda setting, if there is any, should have no delay.
Based on the discussions above, a reduced form of our VAR model is presented as follows:
MAt = α 1 + β 1 MAt −1 + β 2 CAt −1 + β 3 REPt + β 4 NKLt −1 + β 5 CEI t −1 + β 6 IFEt + β 7 IFEt −1 + β 8 NSPt −1 + ε 1t
CAt = α 2 + β 9 CAt −1 + β 10 MAt −1 + β 11 REPt + β 12 NKLt −1 + β 13 CEI t −1 + β 14 IFEt + β 15 IFEt −1 + β 16 NSPt −1 + ε 2t
We used Stata 10 to run the reduced VAR model. Lagrange-multiplier test indicates no
autocorrelation at lag order 1. Jarque-Bera diagnostics show the data are from normal distribution, and
VAR satisfies Eigenvalue stability condition. There are no multicollinearity issues among the exogenous
variables (all VIFs < 2). Table 1 reports the results of our VAR model of media and congressional
attention dynamics.
Table 1 provides empirical evidence for our hypothesis tests. First, the results show a strong attention
inertia effect. In both the media attention and congressional attention models, current attention is
significantly affected by past attention to the issue of climate change. This finding is consistent with other
policy agenda studies that demonstrated strong effects of attention inertia in various policy issue domains
(Baumgartner and Jones 1993; Jones 1994; Wood and Peake 1998; Soroka 2002; Liu 2006).
Second, there is no evidence to support the inter-agenda spillover hypothesis with these annual data
series using one year lag: past year's media attention does not Granger cause current year's
congressional attention (p=.11), and neither does past-year congressional attention Granger cause
current year's media attention (p=.94) when controlling real-world indicators, international focusing events,
scientific feedback, and Republican control of Congress. This finding seems to be in line with some recent
empirical studies, in which inter-agenda interactions are found to be strong in certain issue fields, but very
weak or non-existent in the domain of environmental issues (see, for example, Soroka 2002). A
comparison of the trend lines in Figures 1 and 2 indicates that congressional attention and media
attention to climate change generally mirror one another until 1999. So, it is possible that inter-agenda
effects may exist for certain time periods. After 1999, media attention generally keeps climbing and
maintains a relatively high level, but congressional attention drops to the lowest level since the late 1980s,
particularly after 2001, when the Bush administration minimized interest in global warming and climate
change. It is also possible that sporadic jumps in congressional activity could spark media attention (or
vice versa) on a daily or weekly basis, however, when using the data series spanning several decades
and aggregated at an annual level, we simply find no long-term sustainable inter-agenda effects between
media attention and congressional attention to climate change.
Third, Republican control of Congress appears to have some effect of inhibiting discussion of climate
change in both the news media and the Congress – the stronger the GOP's control, the less attention is
paid to climate change. However, both models show that the Republican advantage in constraining
climate change issue is not statistically significant.
Institute for Science, Technology and Public Policy
Bush School of Government and Public Service, Texas A&M University
11
Table 1: The Determinants of Media and Congressional Attention
Media Attention
c
(n=36)
Independent Variables
a
Congressional Attention
c
(n=36)
b
Attention Inertia
.717***
(.154)
.573***
(.119)
-.187
(2.386)
-.013
(.008)
-.221
(15.233)
-.549
(.759)
33.131
(20.682)
-.768
(1.224)
1.952*
(1.031)
-.033
(.061)
56.640**
(25.814)
-20.683
(25.325)
-1.534
(1.287)
7.234***
(1.262)
.211*
(.117)
.011**
(.006)
R
0.857
0.777
F
19.395***
11.290***
Media (last year) / Congress (last year)
Inter-Agenda Influence
Congress (last year) / Media (last year)
Partisan Advantage
Republican Control in Congress
Problem Indicators
Net Keeling Level (last year)
Climate Extreme Index (last year)
Focusing Events
International Focusing Event (current year)
International Focusing Event (last year)
Feedback
Net Scientific Publication (last year)
2
* p < 0.1; ** p < 0.05; *** p< 0.01. Cell entries are the regression coefficients with small sample degree-of-freedom adjustment.
a
For the media attention model, inertia is measured by last year media attention, and inter-agenda effect is measured by last year
b
congressional attention. For the congressional attention model, inertia is measured by last year congressional attention, and inter-agenda
c
influence is measured by last year media attention. Small-sample t-statistics are computed due to the relatively small sample size for VAR.
The small sample t-statistics are in parentheses.
Institute for Science, Technology and Public Policy
Bush School of Government and Public Service, Texas A&M University
12
Fourth, the VAR results provide certain empirical evidence that generally supports the agenda
theorists' proposition on issue attention dynamics. Among the three attention-attracting factors, climate
change problem indictors appear to have only marginal effect on climate change issue attention, as only
one of the two system-level indicators – i.e., the net change of Keeling's CO2 level (NKL), positively
contributes to congressional attention, while the Climate Extreme Index (CEI) does not affect media
coverage and congressional activity. International focusing events (IFE) demonstrate significant attentiongrabbing power for both the news media and the US Congress. The difference is that the news media
seem to respond to international focusing events more quickly than the Congress. As shown in Table 1,
news media attention is significantly influenced by the current year's focusing event while congressional
attention responds to previous year's focusing event. Scientific feedback, as reflected in net scientific
publication (NSP), also exerts positive influence on both news media attention and congressional activity.
Considering both the media and congressional models together, we see that, overall, the three attentiongrabbing factors—problem indicator, focusing event, and scientific feedback—do have more or less
statistically significant power of drawing policy elites' attention, given the marginal effect of problem
indictor on issue attention and the differences found in how these factors drive the two agendas.
7. Conclusion and Discussion
Policy agenda setting is concerned with issue competition for attention. "Attentiveness" is a key
concept in agenda setting theories to understand how new issues get recognized and considered in
various policy venues. Many agenda studies focus on attentiveness itself and use attention to explain
issue mobilization and policy change, but one less explored, yet important, question is where attention
comes from. Our study, primarily guided by agenda setting theories, attempted to examine this question
by exploring several major forces that drive news media and congressional attention to global climate
change issue.
Drawing hypotheses from Kingdon's, Jones and Baumgartner's as well as other agenda scholars'
perspectives, we attempted to empirically assess whether and how problem indicators, focusing
events/information shocks, and information feedback enhance issue attention. We also assessed the
effect of attention inertia, inter-agenda connection between the news media and the Congress, and
possible partisan advantage on controlling agenda access. Our data were collected from various sources,
and VAR modeling was employed to conduct the hypothesis tests while controlling possible
autocorrelations in multiple time series.
Our VAR analyses showed that attention inertia, as found in many other issue fields, was a strong
factor, positively associated with the level of attention paid to global climate change on both the media
and congressional agendas. Contrary to some existing agenda studies, we found no long-term interagenda spillover effect between media attention and congressional attention to climate change during the
17
36-year period covered by our annual data series. Partisan advantage on issue status was also tested.
Republican composition in Congress appeared to have an effect in restricting climate change issue, but
the partisan impact was not statistically significant in the VAR model.
For the attention-driving forces in Kingdon, Jones and Baumgartner, and other agenda setting
theories, we generally found supportive evidence. One of the two real-world problem indicators (i.e., NKL)
marginally contributed to a higher level of congressional activity on climate change. High-profile
international events in the climate change field had very significant attention-grabbing power for both the
media and congressional agendas. As expected, the news media tended to respond to international
events instantaneously while the congressional reaction to these events showed a time lag. Both media
attention and congressional attention to climate change were found to be also positively influenced by
information feedback from the climate science community. Overall, our findings from the VAR modeling
generally support the agenda theorists' argument on issue attention dynamics.
While our test confirms the general story told by agenda theorists, the findings here suggest that the
attention-grabbing factors may work differently for different policy venues. As shown in Table 1, Congress
seems to be responsive to certain climate change conditions (i.e., the net change of CO2 levels in the
atmosphere), while the news media are not sensitive to either of the two objective climate change
indicators. Moreover, when both the news media and Congress are statistically influenced by climate
science feedback, the US news media tend to respond to international focusing events more rapidly than
the US Congress. This finding of varying effects of the attention-grabbing factors for different policy
venues seems to be in line with the political information-processing model constructed by Jones and
Institute for Science, Technology and Public Policy
Bush School of Government and Public Service, Texas A&M University
13
Baumgartner (2005). For Jones and Baumgartner, policy organizations monitor and process incoming
information about a public problem from multiple diverse sources (in our case, problem indicators,
focusing events, and feedback). However, due to their limited attention span, constrained informationprocessing ability, differing organizational incentives and functionalities, and varying levels of sensitivity
and receptivity, different policy venues tend to respond to the incoming information differently and
selectively (Jones and Baumgartner 2005; see also Baumgartner and Jones 1993; Jones 1994; 2001).
While the agenda theorists seem to be correct about the overall impact of attention-grabbing factors, how
these attention-driving mechanisms may differ in different policy venues deserves further examination in
future research.
There are several other aspects that future agenda setting research should further explore. First, our
study provided confirming evidence for attention dynamics within a specific issue domain—i.e., global
warming and climate change, but whether attention mechanisms can be generalized and applied to other
issue areas remains a question. Several existing studies indicate that attention-grabbing mechanisms
may be quite different in different issue domains (Soroka 2002; Lawrence and Birkland 2004; see also
Baumgartner and Jones 1993, 121-124). In their recent work, Jones and Baumgartner (2005, 208-226)
demonstrated that, in certain issue areas such as macroeconomics or crime, the US media and Congress
did respond to objective economic conditions or crime indicators; however, in other issue areas, such as
social welfare, attention did not respond to the objective conditions of the social welfare problem. Future
research should examine issue-specific attention mechanisms before making generalizations across
issue areas. As Jones and Baumgartner pointed out in The Politics of Attention (2005, 226): "[I]n different
issue areas we see different patterns, each of which makes sense on its own, but we detect no single
pattern that characterizes all policy areas. A general model of policymaking will not emerge from the
study of any single policy area."
Next, our study focused on newspaper attention and congressional attention. It is not clear whether
and how these attention-grabbing factors influence other agenda venues (such as the public, the
president, and TV news networks). And it is also not clear how climate change attention in other venues
18
affects the news media and Congress, or vice versa. Moreover, our model attempted to assess the
overall impact of Republican advantage in Congress, but it is not clear how specific ideological
orientations and structural variables internal to congressional committees and subcommittees affect
climate change issue attention. Other literature indicates that conservatism and partisan politics within
and across relevant congressional committees and subcommittees could also play a critical role in
determining congressional agenda setting. In addition, more in-depth case studies and detailed content
analyses are needed to strengthen the assessment of causal effects between attention attractors and
actual policy elites' attention.
The attention-grabbing mechanisms and processes in real-world policy agenda setting are far more
complex and dynamic than presented here. Although our study sheds some light on where attention may
come from and how attention level may be affected by severity of problem indicator, focusing event and
information feedback, it certainly leaves us with even more questions for future agenda setting research.
Institute for Science, Technology and Public Policy
Bush School of Government and Public Service, Texas A&M University
14
NOTES
1. The agenda setting theory of Jones and Baumgartner is based on an information-processing and
political decision making perspective. The major components of the theory can be found in a
series of books written by Bryan Jones and Frank Baumgartner, jointly and separately (see Jones
and Baumgartner 2005; Jones 1994, 2001; Baumgartner and Jones 1993).
2. As news articles may vary greatly in length, an alternative way to measure the media attention to
climate change is to use the length of news articles. In a recent study, Liu, Vedlitz and Alston
(2008) counted the lines of each article and aggregated the total lines of all articles every year to
measure the issue attention over time. They found that the aggregation of the total lines of
climate change articles each year is highly correlated with the simple aggregation of the number
of articles of each year. In this study, we simply use the annual number of climate change articles
in the New York Times to measure US news media attention to the climate change issue.
3. The additional keywords only brought in a very limited number of new articles. In addition, our
review of these new articles from additional keywords search indicated that the search results
contained a lot of 'noises' – i.e., articles irrelevant to climate change.
4. The Lexis-Nexis searchable news database was accessed at the following website:
http://web.lexis-nexis.com/universe/form/academic/index.html. Our search procedures in
LexisNexis under "Guided News Search" were as follows: (1) under "Step One: Select a news
category," we selected "General News Search;" (2) for "Step Two: Select a News Source," we
selected "Major Papers," and under "Source List," we selected "New York Times;" (3) under "Step
Three: Enter search terms," there are three boxes to enter search terms, where we entered
"climate change" in the first box, "global warming" in the second, and "greenhouse gas*" in the
third while leaving the box reading "Headline, Lead Paragraph(s), Terms" as is, and changing the
relational box "and" to "or." Note that the third key word "greenhouse gas*" is a wild-card term
that also searches "greenhouse gases;" (4) under "Step Four: Narrow to a specific date range,"
we specified the search date range from January 1, 1969 to December 31, 2005. The search
extracted all the articles in the New York Times defined by our search terms and dates.
5. The three online databases on US congressional hearings can be found at the following websites:
(1) The Policy Agenda Project (directed by Bryan Jones of the University of Texas at Austin, John
Wilkerson of the University of Washington and Frank Baumgartner of Pennsylvania State
University): http://www.policyagendas.org; (2) Thomas—Legislative Information on the Internet:
http://thomas.loc.gov; and (3) LexisNexis: http://web.lexis-nexis.com.
6. Keeling's CO2 data are available here: http://scrippsco2.ucsd.edu/data/atmospheric_co2.html.
7. The US CEI data was collected on March 6, 2006 from the NCDC ftp site:
ftp://ftp.ncdc.noaa.gov/pub/data/cei.
8. Some climate scientists suggest that the intensity of hurricane is associated with rising sea
surface temperatures that may be caused by global warming (Emanuel, 2005; Webster, et al.
2005; Hoyos, et al. 2006; Mann and Emanuel, 2006), but others question the accuracy of
predicting more severe tropical cyclones as a result of warmer sea surface temperatures
(Landsea, et al. 2006) or contend that the linkage between global warming and hurricane is
inconclusive (Pielke, et al. 2005).
9. Three major reference documents were used in identifying the international focusing events: (1)
"16 Years of Scientific Assessment in Support of the Climate Convention," Intergovernmental
Panel on Climate Change, 2004; (2) "Milestones in Global Climate Change Policy," Chapter 4 in
Transportation and Global Climate Change: A Review and Analysis of the Literature, US
Department of Transportation, Publication No. DOT-T-97-03, 1998; and (3) "Convention Timeline
and Protocol Timeline," United Nations Framework Convention on Climate Change, 2006.
10. We understand that identifying IFEs is somewhat arbitrary and involves subjective judgment. We
attempted to review all the events listed in the reference documents closely and apply our
operational definition as objectively as possible. Several events seemed to qualify and appeared
in our initial list of IFEs, but after several rounds of additional review and deliberation, these were
dropped from the final list. For example, the Conferences of Parties (COPs) to the UNFCCC
occurs once every year, but we only counted the first three COPs as IFEs because the first three
COPs were relatively "new" and "unprecedented." After the 1997 Kyoto Conference, COPs were
mostly viewed as routine meetings and were not as highly publicized as the first three COPs.
Institute for Science, Technology and Public Policy
Bush School of Government and Public Service, Texas A&M University
15
11. The Web of Science online searchable databases can be found at http://portal.isiknowledge.com.
The SCI Expanded database, starting from 1900 forward, gathers information from more than
5,900 scholarly journals across 150 scientific disciplines. The SSCI database, beginning from
1956 forward, contains information from more than 1,725 scholarly journals across 50 social
sciences disciplines. Both databases include multiple document types (article, abstract of
published item, book review, etc.) in multiple languages.
12. We assumed that non-English language scientific publications, which comprised less than 2% of
all climate change publications in the SCI and SSCI database, would have little influence on the
US media and Congress. Most publications in English were authored or coauthored by US
scientists. We performed our search two different ways – one for US author/coauthor only, and
the other for both US and non-US authors. The two searches produced very similar results in
terms of annual number of scientific publications. We choose to use the annual number of
publications by both US and non-US climate scientists, because it is hardly conceivable that the
US media and Congress would be responsive only to American scientists as important climate
science contributions have been made by scientists from other countries (e.g., United Kingdom).
13. Both SCI and SSCI frequently update their databases when important new journals and other
sources are introduced. The changing scope of SCI and SSCI over time does not comprise the
validity of our measure for scientific publications, as the constant updating of both databases
primarily reflects what have actually occurred over time in the scientific community in terms of the
content and extend of scientific research. In other words, the changing scope of climate science
production, as updated in both SCIS and SSCI databases, is exactly what we want to measure –
the changing scope of scientific feedback.
14. We understand that the annual number of climate science publications by climate scientists may
be, to some extent, a product of past media coverage and congressional activity on climate
change issue: some research efforts may be influenced by what has been "hot" recently in the
media, and some publications may be a direct result of research projects created and funded by
Congress. However, for the purpose of this study, we simply treat the scientific feedback as an
exogenous variable, as we would like to see if scientific feedback affects news media and
congressional attentiveness to climate change.
15. Recall the debates and disagreements among agenda scholars on spillovers/interactions across
policy venues in our earlier discussion.
16. We conducted the pre-estimation of lag-order selection tests for up to 4 lags, and found that only
lag 1 had the optimal AIC, SBIC, FPE and HQIC with significant level < 0.01. Statistics for
selection for any number of specified lags were obtained through the command "varsoc" in Stata
10.1.
17. Because our study focused on long-term trends and our data series were aggregated at an
annual level, we could not exclude the possibility of sporadic interaction between these two
agendas for certain time periods, or the possibility of inter-agenda connection based on shorterterm data aggregations (e.g., data aggregation at monthly level).
18. It is well known that former Vice President Al Gore played an important role in raising awareness
of global warming, while during the administration of George W. Bush the climate change issue
was substantially suppressed by the executive branch. Although it is likely that attention dynamics
in other venues (e.g., the executive branch) may well influence the news media and
congressional agendas, our analysis could not include these variables due to the constrain of
data availability.
Institute for Science, Technology and Public Policy
Bush School of Government and Public Service, Texas A&M University
16
REFERENCES
Baumgartner, F.R., and B.D. Jones. 1993. Agendas and instability in American politics. Chicago:
University of Chicago Press.
Birkland, T. A. 1996. Natural disasters as focusing events: Policy communities and political response.
International Journal of Mass Emergencies and Disasters 14 (2): 221-243.
Birkland, T. A. 1997. After disaster: Agenda setting, public policy and focusing events. Washington, D.C.:
Georgetown University Press.
Birkland, T. A. 1998. Focusing events, mobilization, and agenda setting. Journal of Public Policy 18 (3):
53-74.
Birkland, T. A. 2001. Scientists and coastal hazards: Opportunities for participation and policy change.
Environmental Geosciences 8 (1): 61–67.
Birkland, T. A. 2004. "The world changed today": Agenda setting and policy change in the wake of
September 11 terrorist attacks. Review of Policy Research 2 (2): 179-200.
Birkland, T. A., and R.Nath. 2000. Business and the political dimension in disaster management. Journal
of Public Policy 20 (3): 279-303.
Bodansky, D. 2001. The history of the global climate change regime. In International relations and global
climate change, ed. U. Luterbacher and D. F. Sprinz, 23-42. Cambridge, MA: MIT Press.
Briggs, H. 2007. 50 years on: The Keeling Curve legacy. BBC News, December 2, 2007.
http://news.bbc.co.uk/1/hi/sci/tech/7120770.stm
Cobb, R.W., and C.D. Elder. 1983. Participation in American politics: The dynamics of agenda-building.
2d ed. Baltimore: The Johns Hopkins University Press.
Cohen, B. 1963. The press and foreign policy. Princeton, NJ: Princeton University Press.
Dearing, J., and E. Rogers. 1996. Agenda-setting. Thousand Oaks, CA: Sage.
Dessler, A. E., and E.A. Parson. 2006. The science and politics of global climate change: A guide to the
debate. Cambridge: Cambridge University Press.
Downs, A. 1972. Up and down with ecology: The issue attention cycle. Public Interest 28: 38-50.
Dunlap, R.E., C. Xiao, and A.M. McCright. 2001. Politics and environment in America: Partisan and
ideological cleavages in public support for environmentalism. Environmental Politics 10 (4): 23–
48.
Easton, D. 1965. A systems analysis of political life. New York: Wiley.
Emanuel, K.A. 2005. Increasing destructiveness of tropical cyclones over the past 30 years. Nature (436):
686-688.
Fiske, S.T., and S.E. Taylor. 1984. Social cognition. Reading, MA: Addison-Wesley.
Granger, C.W. J. 1969. Investigating causal relations by econometric models and cross spectral models.
Econometrica 37: 424-438.
Hansen, J.E. 2005. Is there still time to avoid "dangerous anthropogenic interference" with global climate?
A tribute to Charles David Keeling. Presented at the American Geophysical Union, San Francisco,
CA.
Hilgartner, S., and C. Bosk. 1988. The rise and fall of social problems: A public arenas model. American
Journal of Sociology 94 (1): 53-78.
Hill, K.Q. 1998. The policy agendas of the president and the mass public: A research validation and
extension. American Journal of Political Science 42 (4): 1328-1334.
Houghton, J.T. 2004. Global warming: The complete briefing. Cambridge: Cambridge University Press.
Hoyos, C.D., P.A. Agudelo, P.J. Webster, and J.A. Curry. 2006. Deconvolution of the factors contributing
to the increase in global hurricane intensity. Science 312: 94-97.
Hunt, V.F. 2002. The multiple and changing goals of U.S. immigration reform: A comparison of House
and Senate activity, 1947-1993. In Policy dynamics, ed. F. Baumgartner and B.D. Jones. Chicago:
University of Chicago Press.
Institute for Science, Technology and Public Policy
Bush School of Government and Public Service, Texas A&M University
17
Iyengar, S., and D. Kinder. 1987. News that matters: Television and American opinion. Chicago:
University of Chicago Press.
Iyengar, S. 1990. Shortcuts to political knowledge: The role of selective attention and accessibility. In
Information and democratic process, ed. J.A. Ferejohn and J.H. Kuklinski. Urbana: University of
Illinois Press.
Jones, B.D. 1994. Reconceiving decision-making in democratic politics: Attention, choice, and public
policy. Chicago: University of Chicago Press.
Jones, B.D. 2001. Politics and the architecture of choice: Bounded rationality and governance. Chicago:
University of Chicago Press.
Jones, B.D., and F.R. Baumgartner. 2005. The politics of attention: How government prioritizes problems.
Chicago: University of Chicago Press.
nd
Kingdon, J.W. 1995. Agendas, alternatives, and public policies. 2 ed. New York: Harper Collins.
Landsea, C.W., B.A. Harper, K. Hoarau, and J.A. Knaff. 2006. Can we detect trends in extreme tropical
cyclones? Science 313: 452-454.
Lawrence, R.G., and T.A. Birkland. 2004. Guns, Hollywood, and criminal justice: Defining the school
shootings problem across public arenas. Social Science Quarterly 85 (5): 1193-1207.
Liu, X., A. Vedlitz, and L. Alston. 2008. Regional news portrayals of global warming and climate change.
Environmental Science and Policy 11 (5): 379-393.
Liu, X. 2006. Modeling bilateral international relations. New York: Palgrave-Macmillan.
MacLeod, M.C. 2002. The logic of positive feedback: Telecommunications policy through the creation,
maintenance, and destruction of a regulated monopoly. In Policy dynamics, ed. F. Baumgartner
and B.D. Jones. Chicago: University of Chicago Press.
Mann, M.E., and K. A. Emanuel. 2006. Atlantic hurricane trends linked to climate change. Eos:
Transactions of the American Geophysical Union (87): 233-244.
May, P.J. 1990. Reconsidering policy design: Policies and publics. Journal of Public Policy (11): 187-206.
McComas, K., and J. Shanahan. 1999. Telling stories about global climate change: Measuring the impact
of narratives on issue cycles. Communication Research 26 (1): 30-57.
McCright, A.M., and R.E. Dunlap. 2003. Defeating Kyoto: The conservative movement's impact on U.S.
climate change policy. Social Problems 50 (3): 348-373.
Menard, S. 1991. Longitudinal research. Newbury Park, CA: Sage Publications.
Mooney, C. 2005. The Republican war on science. New York: Basic Books.
National Climatic Data Center (NCDC). 2006. U.S. climate extremes index.
http://www.ncdc.noaa.gov/oa/climate/research/cei/cei.html.
Nisbet, E. 2007. Earth monitoring: Cinderella science. Nature (450):789-790
Pew Research Center. 2005. Two-in-three critical of Bush's relief efforts: Huge racial divide over Katrina
and its consequences. Washington, D.C.: Pew Research Center. http://peoplepress.org/reports/display.php3?ReportID=255.
Pielke, R. A., Jr., C. Landsea, M. Mayfield, J. Laver, and R. Pasch. 2005. Hurricanes and global warming.
Bulletin of the American Meteorological Society (86): 1571-1575.
Rankin, W.L., S.M. Nealey, and B. Descow Melber. 1984. Overview of national attitudes toward nuclear
energy: A longitudinal analysis. In Public reaction to nuclear power: Are there critical masses? ed.
W.R. Freudenburg and E.A. Rosa. Boulder, CO: Westview Press.
Richardson, G.P. 1991. Feedback thought in social science and systems theory. Philadelphia: University
of Pennsylvania Press.
Rochefort, D., and R.Cobb. 1994. The politics of problem definition: Shaping the policy agenda. Lawrence:
University of Kansas Press.
Schattschneider, E. E. 1960. The semi-sovereign people. New York: Holt, Rinehart, and Winston.
Shanahan, J., and J. Good. 2000. Heat and hot air: Influence of local temperature on journalists'
coverage of global warming. Public Understanding of Science 9 (3): 285-295.
Institute for Science, Technology and Public Policy
Bush School of Government and Public Service, Texas A&M University
18
Sheingate, A.D. 2006. Structure and opportunity: Committee jurisdiction and issue attention in Congress.
American Journal of Political Science 50 (4): 844–859.
Soroka, S.N. 2002. Agenda-setting dynamics in Canada. Vancouver: University of British Columbia Press.
True, J.L. 1999. Attention, inertia, and equity in the Social Security program. Journal of Public
Administration Research and Theory 9 (4): 571-596.
Ungar, S. 1992. The rise and (relative) decline of global warming as a social problem. Sociological
Quarterly 33: 483-501.
Vertzberger, Y. 1990. The world in their minds: Information processing, cognition, and perception in
foreign policy decisionmaking. Stanford, CA: Stanford University Press.
Walker, J.L., Jr. 1977. Setting the agenda in the U.S. Senate: A theory of problem selection. British
Journal of Political Science 7: 423-445.
Weart, S.R. 2004. The discovery of global warming. Cambridge. MA: Harvard University Press.
Webster, P. J., G. J. Holland, J. A. Curry, and H. R. Chang. 2005. Changes in tropical cyclone number,
duration, and intensity in warming environment. Science (309): 1844–1846.
Wood, B. D., and J.S. Peake. 1998. The dynamics of foreign policy agenda setting. American Political
Science Review 92 (1): 173-184.
Institute for Science, Technology and Public Policy
Bush School of Government and Public Service, Texas A&M University
19