Content-Type: text/html
Swamped in Politics: News Coverage of Wetlands
in Three Presidential Administrations
Bonnie Parnell Riechert, Ph.D.
Coordinator of Communications
Agricultural Experiment Station
The University of Tennessee
Knoxville, TN 37996-4500 USA
Phone: (423) 974-7362
e-mail: [log in to unmask]
M. Mark Miller, Ph.D.
Professor of Journalism
College of Communications
The University of Tennessee, Knoxville
Knoxville, TN 37996-0330
Phone: (423) 974-4452
e-mail: [log in to unmask]
Paper submitted for presentation to
Association for Education in Journalism and Mass Communication,
Communication Theory and Methodology Division,
August 1997, Chicago.
Swamped in Politics: News Coverage of Wetlands
in Three Presidential Administrations
Bonnie Parnell Riechert, The University of Tennessee;
and M. Mark Miller, The University of Tennessee, Knoxville
Abstract
Several studies suggest that U.S. Presidents dominate the media agenda.
This computer-assisted content analysis of Associated Press coverage of wetlands
over a 12-year period
compares the administrations of Reagan, Bush, and Clinton. Concept mapping
procedures identify
four major themes in 1,756 wetlands news articles from 1984-96: Habitat
Restoration,
Jobs/Development, Regulation Legislation, and Agriculture/Spending. Shifts in
relative
prominence of major themes suggest an active, but not necessarily dominant,
influence of U.S.
presidents on the media agenda.
Swamped in Politics: News Coverage of Wetlands
in Three Presidential Administrations
The question of who sets the media's agenda is one of critical interest to
mass
communication scholars. This study explores the influence of U.S. presidents in
shaping the
content of news media coverage.
The role of the news media in society is as much a manifestation or product
of culture as it
is a source of culture, Shoemaker and Reese (1991) have suggested. Agrees
Kennemar
(1992/1994), the news media "are as much the target of agenda-setting as they
are the source"
(p. 9). The focus of this paper is on this recycling, recursive role of the mass
media in society
(borrowing from the language of Derksen & Gartrell, 1993).
The importance of studying media content "the message itself as a dependent
variable" is emphasized by Shoemaker and Reese (1991, p. 8) in their pioneering
volume,
Mediating the Message: Theories of Influence on Mass Media Content. They argue
that rather
than viewing media content as a given, instead scholars should examine the
factors inside and
outside media organizations that affect media content (p. 1). It is in the
spirit of this approach that
this analysis is conducted.
The term media sociology often is applied to studies with this focus: "A
number of
researchers who previously studied media effects including ourselves now find
themselves
asking why such effect-producing content exists to begin with" ((Shoemaker &
Reese, 1991,
p. 3).
An intuitively-appealing hierarchical model of influences on media content
is offered by
Shoemaker and Reese (1991). The model divides influences on media content into
five levels of
widening scope: the individual level, the media routines level, the organization
level, the
extramedia level, and the ideological level.
Dearing and Rogers (1996) explicate the agenda-setting process as "composed
of the
media agenda, the public agenda, and the policy agenda, and the
interrelationships among these
three elements . . . So, the agenda-setting process is an ongoing competition
among the
proponents of a set of issues to gain the attention of media professionals, the
public, and policy
elites" (pp. 5-6). They call for increased efforts of scholarly studies that
focus on single issues to
provide better understanding of these dynamic processes.
Useful analyses of single issues include Lang and Lang's (1983) examination
of the role of
the media in the Watergate scandal, Gitlin's (1980) provocative study of
coverage of the Students
for a Democratic Society during the Vietnam War era, and Entman's (1991)
analysis of coverage
of the Korean Air Lines and Iranian air incidents. Additional in-depth
investigations on single
issues are needed, according to Dearing and Rogers (1996), to address such
questions as who
puts an issue on the national agenda: "What institutions in addition to the New
York Times and the
White House put an issue on the national agenda?" (p. 95).
Agenda studies are critically important in the field of communication
research, although
they have been plagued by problems of terminology. Berkowitz (1992/1994)
provides a helpful
organization of the linguistic hodgepodge of terminology the semantic
schizophrenia weighing
down scholarly discussion of the mutual influences of policymakers and
journalists. The terms
agenda-setting, agenda-building, and information subsidy have "added both
understanding and
confusion to the study of journalists and news sources" (p. 82). Sadly but
truly, the absence of
consensus among communication scholars on the application of these various
terminologies
effectively mires the field in a bog of confusion and lack of clarity.
Berkowitz (1992/1994) recommends the use of the term agenda-building or
media
agenda-building in discussion of the "interactive, give-and-take process of
shaping media
content" (p. 87). Not viewed simply as a transferral of priorities, this notion
of the media agenda
acknowledges the dynamic nature of the continuing process. As Berkowitz
(1992/1994) notes,
"Beyond journalism itself, the shared culture of journalists and policymakers
leads toward a
dynamic relationship that is sometimes more adversarial and sometimes more of an
exchange"
(p. 102).
The needs for investigation of news media content as dependent variable,
noted by
Shoemaker and Reese (1991), as well as for in-depth analysis of single
agenda-related issues,
recognized by Dearing and Rogers (1996), point toward studies such as this one
on the content of
news media coverage of wetlands. Moreover, Shoemaker and Reese (1991) call for
improved
precision in methods for investigating media content, and this study employs
rigorously precise
measurement and comparison techniques made possible by developments during this
decade in
computer-assisted content analysis.
Recent investigations demonstrate that the White House is highly
influential "in putting an
issue on the U.S. media agenda" (Dearing & Rogers, 1996, p. 17). The
"notoriously" powerful
role of U.S. presidents in influencing news coverage is observed by Shoemaker
and Reese (1991,
p. 153). Cohen (1989, cited in Shoemaker & Reese, 1991, p. 153) makes a similar
observation:
The president has a natural advantage over the media, giving him a
commanding
lead in setting the terms of public-political discourse. The president
is invariably
news . . . The White House can therefore create the news necessary to
give the
president's agenda wide publicity and even dominance in the political
marketplace.
(p. 2-3)
The U.S. president "can put an issue on the national agenda just by giving
a talk about it"
(Dearing & Rogers, 1996, p. 75). In fact, Kingdon (1984) posits that "No other
single actor in the
political system has quite the capability of the president to set agendas in
given policy areas for all
those who deal with these policies" (p. 17).
Bosso (1987) further suggests the U.S. president "is the political system's
thermostat,
capable of heating up or cooling down the politics of any single issue or of an
entire platter of
issues" (p. 261). Other influences such as the prestige media, scientific
journals and the results of
public opinion surveys also are powerful. Recent studies also provide evidence
for the influence of
Congress on setting the media agenda (Goodman, 1994).
An earlier investigation of news coverage of wetlands by Associated Press
demonstrated
systematic shifts over time in relative prominence of the competing conservation
and property
owner concerns (Riechert, 1996). The study covered an 11-year period from
1984-95 and
identified shifts in coverage content across years. The relative prominence of
competing
stakeholder concerns was shown to be related to news sources either mentioned or
quoted in
news articles. Some 1,456 articles were included in the computer-assisted
content analysis. The
study did not investigate systematic shifts in coverage across four-year periods
corresponding to
U.S. presidential administrations.
The present study continues this strand of research by examining shifts in
news coverage
from 1984-96, across presidential administrations, involving those of Ronald
Reagan, George
Bush, and Bill Clinton (his first term). Some 1,756 articles are analyzed in
this study.
The research strategy of this study is to examine news coverage of the
particular issue of
wetlands over time, comparing news coverage during the administrations of three
different U.S.
presidents and examining whether and how differences in coverage are manifest
during those
administrations. The topic of wetlands is appropriate for a study such as this.
Wetlands are the
subject of continuing news coverage in recent years as scientists debate their
definition,
conservationists promote their preservation, property owners and developers
lobby for loosened
protection for them, Congress debates their regulation, and courts address cases
involving such
conflicts. Wetlands have received continued and increasing news coverage over
the past few
years, becoming the focus of national attention in 1988 after Presidential
candidate George Bush
adopted the slogan, "No net loss of wetlands," as a campaign promise.
Method
In this study computer-assisted content analysis methods are used to
examine news articles
about wetlands over a 12-year period involving the presidential administrations
of Reagan, Bush,
and Clinton. Key themes in the wetlands news coverage will be identified using a
set of
multidimensional scaling (MDS) techniques referred to as concept mapping
(Miller, 1990/1995).
The method uses mathematical procedures to identify term co-occurrence patterns
within text.
Concept mapping was developed by Miller and relies on the use of the computer
programs VBPro
and VBMap, which he developed for qualitative and quantitative analysis of
verbatim text.
The term co-occurrence patterns will be examined using both cluster
analysis
(hierarchical) and MDS techniques. While the methods often are used in
conjunction by scientists
(Corter, 1996), their application to analysis of term co-occurrence patterns
represents a
methodological innovation of concept mapping. Hierarchical cluster analysis
(1996) and
extraction of principal components in MDS (Coxon, 1982) will be used to produce
ultrametric
tree dendograms and multidimensional scaling configurations or maps to indicate
co-occurrence
similarities of terms and theme (as terms occurring frequently together).
Once the major themes of wetlands news coverage are identified, their
relative prominence
will be compared among the periods corresponding to the presidential
administrations of Reagan,
Bush, and Clinton. These comparisons will be made with one-way analysis of
variance tests
applied to the mean occurrences of theme terms in articles from the three time
periods. The
conservative Scheff test will be applied as a post hoc test to identify the
differences among the
presidential administrations.
Significant differences in theme prominence across the presidential
administrations will
provide evidence for the systematic shifts in news media coverage over time and
will help
illuminate the role of these U.S. presidents relating to the media agenda
involving wetlands.
Findings from this study will contribute to improved understanding of the
complex extramedia
influences on the content of news media coverage. They should suggest avenues
for further study
relating to media content, media sociology, and media agenda setting.
Concept Mapping
Miller and Riechert (1994) employed concept mapping techniques in
identifying major
themes in coverage of pesticides in four magazines including a general news
magazine (Time) and
three diverse specialized magazines (Audubon, Nutrition Today, and Successful
Farming). Key
themes included such topics as the environment, farming issues, and food risk.
Differences in
relative prominence of the themes were identified across the magazines, and were
interpretable by
the different editorial slants of the specialized magazines. These differences
would be expected in
pesticides coverage among the magazines. The distinctions were identified, and
the comparisons
were accomplished, with quantitative precision through computer-assisted content
analysis
procedures. The study illustrates the usefulness of the concept mapping methods
for identifying
themes in text and for making comparisons related to these themes.
The concept mapping procedures have been employed in numerous studies since
1993 to
investigate a number of research questions with computer-assisted content
analysis. These studies
include investigations of national election issues (Miller, Andsager, &
Riechert, 1996); media
framing of hate speech (Miller & Andsager [in press]); trends in scholarly
journal content (Miller,
1994); coverage of women's health issues such as the French abortion pill RU486
and silicone
breast implants (Andsager & Miller, 1994; Smiley & Andsager, 1996); advocacy
group and news
framing of public policy issues (Riechert, 1996); public issues (Miller &
Riechert, 1997); media
coverage of wetlands (Riechert & Miller, 1997); magazine coverage of pesticides
(Miller &
Riechert, 1994; Riechert & Miller, 1994) and several studies on source
credibility (Chew &
Miller, 1993; Chew, Mehta, & Oldfather, 1993; Chew & Kim, 1994). The procedures
also are
described generally in Miller & Riechert (in press).
Analysis
For this analysis all articles mentioning "wetlands" that were distributed
by Associated
Press from July 1984 until June 1996 were obtained in full text from the AP
Newswire database
through the DIALOG information service. This provided a data set of 12 full
years of news
coverage by AP (the full text of AP articles was available beginning with July
1984). Since this
investigation was directed at news coverage and U.S. presidential
administrations, articles
discussing wetlands outside the United States were omitted from the this
analysis. Also eliminated
were editorial roundups, obituary-type articles (reporting the death of
prominent wetlands
advocates) and one article mentioning a college athlete whose last name was
"Wetlands."
Removal of these groups of articles left a total of 1,756 remaining articles on
which this
investigation focused.
These strategies for retrieving and retaining news articles yielded a large
set of 1,756
articles for the analysis in this study. The volume of coverage is adequate to
support examination
of a number of questions relating to the media content. The distribution of the
articles by year is
shown in Figure 1. The set of articles include 43 from July 1984 - July 1985; 74
from 1985-86; 71
from 1986-87; 81 from 1987-88; 144 from 1988-89; 168 from 1989-90; 162 from
1990-91; 220
from 1991-92; 134 from 1992-93; 199 from 1993-94; 169 from 1994-95; and 251 from
1995-96.
The steady increase in news articles about wetlands over this period indicates
continued and
increasing coverage of wetlands in the news media. The distribution of the
wetlands news articles
by presidential terms also shows this pattern:
Reagan 354
Bush 686
Clinton 716
The 1984-96 articles covered the time periods of Reagan's entire 1985-88
second
presidential term as well as the last six months of his first term; of Bush's
entire 1989-92
presidential term; and of Clinton's 1993-96 term except the last six months.
Themes in the wetlands articles were identified using the key word
frequency ranking
procedures and concept mapping techniques made possible by the VBPro set of
computer
programs, which include VBMap (Miller, 1990/1995a and 1990/1995b ). The programs
were
designed specifically for content analysis of natural language text and for
concept mapping, or
multidimensional scaling analysis of co-occurring terms and themes in text.
Entman (1993)
elaborates the manifestation of themes in key words.
Case numbers were added in the text file containing the news articles, and
the file was
formatted in VBPro (Miller, 1990/1995b). The file then was submitted to the
program's
procedures for creating (1) an alphabetized list of all terms used in the
articles (also indicating
frequency of occurrence) and (2) a frequency-ranked list of all the terms,
ordered by raw
frequency of occurrence. The first file listed terms from a, occurring some
21,070 times, to of,
occurring 29,224 times, to water, occurring 3,076 times, to wetlands, occurring
4,566 times, to
Zwannendael, occurring 2 times. From the list of terms in the second file,
ranked by frequency,
were selected more than 100 terms that (1) occurred frequently in the text, and
(2) that are
substantively meaningful and unambiguous. Grammar function words such as a and
of, though
frequently-occurring, are omitted from analysis, as are terms that might have
more than one
meaning, such as act or states. The multidimensional scaling program for term
co-occurrence
accommodates a large number of terms, up to 125 words or groups of words.
The frequently-occurring terms selected for the concept mapping procedure
are listed in
Table 1 along with their frequency of occurrence and additional information that
will be
referenced later. These terms were used to create a search file (Miller,
1990/1995b), with which,
and for which the text file was coded, using the VBPro coding procedures. Terms
that occurred
frequently in singular and plural form were combined in the search file together
for coding
purposes (farm/farms, and so on). The numeric coded file created by the
procedures then was
submitted to the VBMap program, which applies multidimensional scaling
techniques to identify
the term oc-occurrence patterns in the text. Three dimensions or coordinates
were specified,
because three dimensions are easily represented visually in maps or
configurations that may be
constructed using the program's output of coordinate values for each specified
term.
The coordinate values output by VBMap for each specified term also are
listed in Table 1
under the column heading, "Coordinates 1, 2, and 3." These values are useful in
constructing
multidimensional configurations or maps showing the relative location of the
terms in a
multidimensional space. An example of this type of map will be presented and
discussed later in
this paper.
The output of coordinate values for the terms also may be submitted
directly into
statistical programs such as SPSS for hierarchical cluster analysis, using the
agglomerative
schedule and cosine similarity measurement method, to ascertain which terms
occur most often
together. The procedure was followed for the 120+ terms discussed previously and
listed in Table
1. The resulting clustering schedule, shown in a tree-type dendogram,
immediately linked several
groups of terms to reveal that they occur together frequently in the text.
The following steps then were taken to assure the stability of the
clusters. Before the
analysis proceeded further, an additional search file as created with the same
terms plus three
added source identifier terms (to indicate the presidential terms of Clinton,
Bush, and Reagan).
The constructed code terms used were $CLIN, $BUSH, and $REAG meeting the
criteria of
being unique, easily recognizable, and not occurring naturally in the text. The
text file of the news
articles was opened, and within each news article was added one of the
identifier code terms,
depending on the date of the article and the corresponding presidential term.
Thus, the identifier
code term $CLIN was added in articles from 1993-'96; $BUSH was added within
articles from
1989-'92; and $REAG was added within articles from 1984-'88. The text file then
was
reformatted in VBPro before the coding procedure was repeated for the 120+
frequently-occurring terms and the three inserted identifier code terms.
The new numeric coded file then was submitted to the VBMap program for
generation of
coordinate values for three coordinates or dimensions, and the resulting data
file was submitted to
SPSS for hierarchical cluster analysis. (The procedures are followed first
without, then with,
the inserted code terms to confirm the stability of the clusters.)
The term clustering patterns revealed by these two clustering analysis
procedures are
shown in Columns 1 and 2 of Table 2. As indicated in the table, several terms
clustered
immediately with the term agriculture: acreage, crop, growers, producers,
production, subsidies,
benefits, and rural. A second tight cluster was formed by the linking of the
terms conservation,
soil, erosion, farming, USDA, erodible, farmers, and agricultural. The table
lists the terms
connected in 23 tightly-formed clusters of co-occurring terms from among the
120+ terms
included in the cluster analysis.
Examining the co-occurrence patterns of the individual terms, as listed in
Table 2, might
be compared to looking at a grove of trees from an arm's length we see the
individual leaves
clearly, but a different perspective and additional understanding is provided by
looking from
several paces back. (This avoids the well-known "can't see the forest for the
trees" problem.)
Thus, another stage of analysis is undertaken to examine the co-occurrence
patterns of the key
terms from the perspective of stepping a few paces back, combining the
clustering schedules into
fewer and more easily comprehendible categories.
Another way to describe the comparison of relative proximities of
clustering groups of
terms is to think of looking at constellations of terms. The following
procedures offer this type of
examination.
A new search file was created by combining these terms into 23 groups (as
suggested in
the cluster analysis) plus the three identifier terms indicating presidential
administration. This
search file was submitted to the VBPro coding procedure, along with the text
file of the news
articles, to generate a new data file coded for frequency, by case, of terms
from each term cluster,
as listed in Table 2. The data file then was submitted to the VBMap program to
obtain coordinate
values for three coordinates or dimensions. This information was output by the
VBMap program
and is listed in Table 2 under the column heading, "Coordinates 1, 2, and 3."
The coordinate
values may range from -1 to 1, with proximity in the three-dimensional space
representing
frequent occurrence of the terms together in the text. Thus, the terms located
farthest apart in the
derived multidimensional space occur together infrequently or never. For
example, the identifier
terms $CLIN, $BUSH, and $REAG necessarily never occur together in an article.
The relative proximity of the groups of terms, now consolidated into
meaningful clusters
or themes (agriculture, conservation, and so forth), is easily visualized in a
multidimensional
configuration or map. This can be constructed using the coordinate values output
by the VBMap
program. The values listed in Table 2, Column 3, were used to construct the
concept map
presented in Figure 2.
Recall that the purpose of multidimensional scaling procedures is to reveal
the hidden
patterns of data structure (Kruskal & Wish, 1978). The map shown in Figure 2
provides a visual
illustration, in a derived multidimensional space, of the data structure of the
term co-occurrence
patterns in the news coverage on wetlands over a 12-year period. It shows us
what terms
occurred together, and which terms did not tend to co-occur. Circles on the map
represent the
relative location of the terms in the text; thus, the points (represented by
circles) that locate
closest to one another represent terms that occur consistently together in the
text. The shaded
circles indicate the relative location of the inserted identifier code terms
representing the
administrations of Clinton, Bush, and Reagan. Location of the shaded circles
does not indicate the
content of what each president said in the media (that remains for a further
study), but are
meaningful in conjunction with their proximate terms to indicate what themes
occurred most
prominently in news coverage during the distinctive administrations.
The two dark lines represent coordinates or dimensions 2 and 3 in the
multidimensional
space, and the vertical lines to which the circles are attached represent
coordinate 1 in the three-dimensional space. The circles attached at the top of
a vertical line indicate positive values in
coordinate 1, while circles attached at the bottom of a vertical line indicate
negative values.
What does the map tell us about news coverage on wetlands during
presidential
administrations of Clinton, Bush, and Reagan? The configuration reveals that
industry, pollution,
wetlands, habitat, the environment, building, and restoration received
preponderant coverage
during the Reagan administration, compared to his immediate two successors. Jobs
and industry
were primary themes during the Bush administration. With regard to coverage of
wetlands,
legislation, government, rights (of property owners), and Congress dominated
coverage during
Clinton's first term in office.
The relative location of Bush compared to Reagan and Clinton is most
notable, revealing
that media coverage of wetlands was much more similar during Bush's
administration to coverage
during Reagan's administration than during the succeeding term of Clinton. In
other words,
wetlands coverage after Bush was dramatically different than coverage before his
administration.
The dramatic shift suggests the possibility that Bush redirected the media
discussion of wetlands,
which could be accomplished through shifting the national policy agenda related
to wetlands.
Studies by Nicodemus (1995) and Riechert (1996) support this interpretation. A
discussion by
Nicholas (1992) is relevant:
President Bush, who once vowed to protect wetlands "no matter how
small" has
responded to this pressure with a well-orchestrated campaign that
could threaten
fully half of the nation's remaining wetlands. The White House defends
its position
by claiming that previous regulations covered many areas that are not
really
wetlands and went too far in blocking development and denying
landowners their
rights to develop their property. Some members of Congress, from both
sides of
the aisle, have also proposed legislation to rein in wetlands
protection. And
growing numbers of landowners, encouraged by a conservative shift in
the judicial
system during the Bush and Reagan years, are pressing their cases in
courts. (p.
36)
Another type of visual representation of the patterns of term
co-occurrence in the text is
provided by cluster analysis in the tree-type dendogram. In such a diagram all
terms are eventually
linked, but terms that are combined earliest in the clustering schedule are the
ones that occur most
frequently together. The dendogram in Figure 3 also shows the relative
co-occurrence similarity
of these term clusters, reasonably conceptualized as themes in the wetlands
articles. The
dendogram was generated by cluster analysis with the cosine method using the
coordinate values
shown in Table 2, which were output by the VBMap program.
The clustering results displayed in the dendogram in Figure 3 provide
justification for
further consolidation of the term clusters into superclusters, dependably
conceptualized as major
themes in the wetlands news coverage. After all, these clustering terms and
groups of terms
emerge from the data themselves. The dendogram suggests broad classification of
the 23 specific
themes into four major themes, which we label Habitat Restoration,
Jobs/Development,
Regulation Legislation, and Agriculture/Spending. These major themes are listed
in Table 3 along
with the term clusters or narrower themes that they encompass. Reasonable
division of the
coverage into these manifest major categories allows for comparison of relative
prominence
across presidential terms. The results of these comparisons will be discussed
following a review of
the nature of each of these four grand topical divisions.
The Habitat Restoration Theme
The major theme referred to here as Habitat Restoration includes the
narrower themes of
permit, preserve, build, habitat, restoration, wetlands, runoff, and scientists.
Grouping of these
topics into this sensible encompassing theme is based on the results of the
cluster analysis shown
in the dendogram in Figure 3.
News coverage with prominence of the Habitat Restoration Theme involves
discussion of
wetlands preservation and restoration, of wetlands-related permits and projects
in building that
affect wetlands. For example, the AP news articles ranking highest in frequency
of the terms
comprising this theme is a Sept. 5, 1994, article datelined Land O' Lakes, Fla.,
carrying the
headline, "Florida's Water Woes." Excerpts of the article include the following:
Catherine Monsees retired to a house on a lake once stocked with fish
and teeming
with wildlife. That's all gone now. What she has left is a dry lake
bed chocked with
weeds and a tarnished dream, for her golden years. . . .The culprits
are varied: too
much water pumped from natural underground reservoirs, persistent
drought,
rapid growth and virtually no thought for water conservation. . . .
An estimated
17,000 acres of wetlands are damaged. Marshes are arid. Cypress trees
are dead.
Ducks, cranes, turtles and alligators are gone.
The Jobs/Development Theme
The Jobs/Development theme incorporates the narrower themes of industry,
pollution,
development, environment, and jobs. News articles with prominent emphasis on
this major theme
discuss industry and development relevant to wetlands, implications to the
environment, and
potential impacts on jobs.
An example of an article with prominent focus on this major theme is an
Oct. 21, 1992,
article datelined Washington and carrying the headline, "Environmentalists Cool
to Bush, Back
Clinton Despite a Mixed Record." The following excerpts are taken from the news
report:
A comic strip congressman recently lectured his wife, urging her to
"face reality,
the real choice. You can either choose jobs or trees, jobs or air,
jobs or . . ." The
distraught woman bolted the frame a choice voters don't have, even
when the
real-life presidential campaign dialogue starts sounding like the
Washington satire
by Mark Alan Stamaty. President Bush alternatively labels Arkansas
Gov. Bill
Clinton a polluter-coddler, who let poultry companies foul his state's
rivers with
chicken waste, and an environmental extremist who has embraced the
ideas of his
greener-than-thou running mate, Sen. Al Gore. . . . Bush came into
office pledging
to be an environmental president but has gradually given more weight
to business
concerns.
A second example of coverage focusing on this theme is seen in a Sept. 23,
1991, news
report datelined Grand Canyon, Ariz., and carrying the headline, "Bush,
Environment." Excerpts
from the 1,269-word article follow:
Like any other tourist, President Bush stood in awe of the Grand
Canyon's
magnificence. His one-word reaction to the canyon's natural majesty:
"grandeur."
The Grand Canyon provided a breathtaking backdrop for the official
announcement of a plan to block pollution over the national park a
lofty
achievement by any standard. . . . But critic's lambast the
administration's pro-industry bent on other issues, such as oil exploration in
the Arctic National Wildlife
Refuge and resistance to costly regulatory measures. . . .
[Wilderness Society
President George Frampton] charged Bush "has stood with big oil, the
timber
industry and the mining companies." . . . Bush frequently states a
goal of
balancing economic and environmental needs. He used that rationale
last month
when he backed off a campaign promise to allow "no net loss" of
wetlands by
rewriting the definition of a wetland. "A pothole in the backyard is
not a wetland,"
Bush told reporters. . . . Among other actions bemoaned by
administration critics:
Refusal to commit to reduction targets for carbon dioxide emissions, .
. . the
decision to drop millions of acres from wetlands protection . . .
Bush's insistence
on opening the Alaska refuge to oil exploration...resistance to
raising fuel
efficiency requirements for cars . . .
The Regulation Legislation Theme
A third major theme, described as Regulation Legislation, incorporates the
smaller themes
of government, legislation, endangered, rights, and regulation. News articles
emphasizing these
topics focus on wetlands legislation, implications for endangered species and
for the rights of
property owners, and the role of government in regulation of wetlands.
An example of news articles with prominent emphasis on these areas is the
March 3, 1995,
report datelined Washington and headlined as "House Approves
Landowner-Compensation Bill."
Excerpts of the 805-word article follow:
Declaring an assault on arrogant government, the House approved a bill
today
requiring federal agencies to compensate landowners for restrictions
that protect
wetlands and endangered species. The legislation, which was denounced
by
Democrats as a thereat to environmental protection and a potential
drain on the
U.S. treasury, cleared by a vote of 277-148. It now goes to the Senate
. . . Called
the "Private Property Protection Act," the House-passed measure would
broaden
substantially the definition courts traditionally have given to a
"regulatory taking"
of property.
The Agriculture/Spending Theme
A fourth major theme referred to as the Agriculture/Spending Theme
represents the
combination of the narrower themes of agriculture, conservation, economic,
spending, and
Congress, as mentioned previously and as listed in Tables 2 and 3. Articles
emphasizing these
topics focus on wetlands and agriculture, conservation practices relating to
wetlands, and
economic/spending issues relating to agriculture. A news article headlined "1985
Farm Law Will
Help Eliminate Soil Erosion: Conservation Group" and released by Associated
Press on Jan. 30,
1990, emphasizes these topics. Excerpts from the 664-word article, datelined
Washington, follow:
Historic provisions in 1985 farm law will help farmers move rapidly
toward the
elimination of severe soil erosion on much of the nation's cropland by
the mid-1990s, a private conservation group said Tuesday. A study of more than
1,000
farms in selected areas across the nation showed that the conservation
provisions
of the 1985 Food Security Act have been a main force in prodding and
luring
farmers into protecting soil resources. The report, released here by
the Soil and
Water Conservation Society of Ankeny, Iowa, focused on the 1985 law's
provisions governing the use of fragile, highly erodible land, and
wetland areas.
The computer procedures with which these four major themes were identified
assure their
representativeness of the overall news coverage on wetlands, because the terms
comprising them
emerge from the news articles themselves based on their frequency of occurrence
and their
occurrence together in the text. Identification of these themes, represented in
the text by the
appearance of the relevant terms, allows for comparisons of their relative
prominence in the news
coverage.
The efficiency and precision of computer-assisted content analysis is
appreciated when the
task is to code 1,756 articles about 4,000 pages of text for frequency of
occurrence of 100 or
terms (from Table 2) combined into four major topical themes. This process is
accomplished
easily by constructing four new variables from an existing data file in SPSS .
The existing file,
created by VBPro with search files and coding procedures, listed the frequency
of occurrence of
the terms listed in Table 3 for each case or news article. Four new variables
were created using
the computer commands in SPSS and adding, for each case, the frequencies for
the terms
combined within each major theme category. The values generated in this process
then were used
to conduct one-way analysis of variance tests comparing the prominence of each
major theme in
news coverage among the three presidential administrations. These statistical
tests compare the
mean frequencies of the theme terms in news articles distributed during the
three presidential
administrations. The results of these analyses are presented in the following
section.
Results and Discussion
Comparisons of relative prominence were made among coverage in the three
presidential
administrations for each of the four major themes in the wetlands news coverage.
The tests,
therefore, compared the prominence of the major themes of Habitat Restoration,
Jobs/Development, Regulation Legislation, and Agriculture/Spending across the
administrations
of Reagan, Bush, and Clinton. Prominence of a theme was measured by occurrence
of the relevant
theme terms in the news articles. Comparisons involve the mean occurrences of
the theme terms.
Coverage of The Habitat/Restoration Theme
Comparison of the term frequency means revealed that coverage of the
Habitat
Restoration Theme was more prominent in the Reagan administration than in the
following
administrations. Table 4a summarizes these comparisons. One-way analysis of
variance revealed a
significant difference among the presidential administrations, as summarized in
Table 4b. The
Scheff post hoc test was applied, indicating that the mean occurrence of 13.31
relevant terms in
the Reagan administration was significantly higher than the means of 11.40 and
10.10 during the
Bush and Clinton years, respectively.
The environment was not considered an issue on Reagan's agenda, so the
explanation for
this must be found elsewhere. Reagan entered the oval office "as a time of
especially low
prestige Jimmy Carter had just been vicariously held hostage by Iranian
militants for 444 days"
(Manheim, 1991, p. 83), as the news media had reminded the American public on a
virtually daily
basis. The Grenada incident among others served to positively reinvent Reagan in
the public mind
as a world leader. A point to be made here, Manheim argues:
is that public officials, no less than political candidates and
probably more have
reason to find means to structure the ways they are portrayed in the
news media.
Much of their relationship with journalists and through journalists,
with the
public, is guided by this fact. Making news, as much as making or
implementing
policy, is part and parcel of the business of politics. (p. 83)
Coverage of Jobs/Development Theme
Comparisons of prominence of the Jobs/Development theme in news coverage
also were
made for the three presidential administrations. The results of these tests are
presented in Table 5.
One-way analysis of variance tests revealed significant differences among the
presidential
administrations, for this variable. The Scheff post hoc test showed that the
relative occurrence of
these relevant theme terms was significantly higher in the Bush administration
compared to the
Reagan and the Clinton administrations.
Examples of articles emphasizing this theme, provided in the previous
section, were
somewhat critical of Bush while focusing on his themes of jobs and development.
A potential
explanation is that the media frequently quoted Bush discussing jobs and
development while they
also provided criticism of his environmental policies.
Coverage of Regulation Legislation Theme
Comparisons of relative prominence in media coverage of a third major
theme, Regulation
Legislation, are presented in Table 6a. With a mean of 10.07 relevant theme
terms per case during
the Clinton administration, prominence of the Regulation Legislation theme was
significantly
greater compared to coverage during the terms of his two immediate predecessors
(with means of
3.80 and 4.13). Coverage during the Reagan and Bush terms did not differ
significantly. The
results of the associated one-way analysis of variance test are given in Table
6b.
Coverage of Agriculture/Spending Theme
Similar findings related to coverage of a fourth broad theme of
Agriculture/Spending are
presented in Table 7. As indicated table, coverage of agriculture was
significantly more prominent
in Clinton's administration than in the terms of Reagan and Bush, while this
theme received
similar levels of coverage during the Reagan and Bush administrations. One-way
analysis of
variance findings are shown in Table 7b.
To summarize the results of these comparisons, shown in Tables 4-7,
coverage of Habitat
Restoration was highest during the terms of Reagan, coverage of Jobs/Development
was highest
during the term of Bush, and coverage of Regulation Legislation and
Agriculture/Spending was
highest during Clinton's term. These comparisons offer compelling evidence for
the systematic
shifting of news media content over time. The nature of the shifts, however,
suggest an
inconsistently rather than consistently dominant influence of the U.S. president
on the
American news media agenda.
Review of the news articles and findings of additional studies (Nicodemus,
1995; Riechert,
1996) show that during the Reagan's presidential terms in the 1980s, news
coverage of wetlands
emphasized conservation/environmental issues. Conservation advocate groups were
quoted
frequently in the news media as they discussed the benefits of wetlands and
warned of the
environmental threats of continued loss of natural wetlands. Benefits of
wetlands include
preventing flooding, improving water quality, and providing critical habitat
for threatened and
endangered wildlife species (Wetlands Fact Sheets, 1995). Public awareness of
the environmental
importance of these diverse areas was so high in the public mind, in fact, that
Presidential hopeful
George Bush in 1988 issued the slogan, "No net loss of wetlands" as a campaign
promise.
After taking office, however, Bush instead sought relaxation of regulations
that protect
wetlands and who lobbied for legislative reforms that restricted the definition
of wetlands and that
would provide compensation to property owners affected by restrictions imposed
by wetlands
protection. During his administration scores of wetlands legislative reform
bills were introduced,
and these still were being debated in the national policy and media agendas when
Clinton took
office. The ensuing shift in wetlands news coverage of relative prominence of
conservation to
property-owner concerns is linked largely to the public statements and policy
activities of Bush
(Nicodemus, 1996; Riechert, 1996).
Conclusions
Several conclusions may be drawn from this study. Evidence is provided that
in some
cases (with Bush, in wetlands news coverage), presidents may effectively
energize an issue and
reframe it in the media, and yet may not be able to maintain media nor public
support on the issue.
Continued research in this area is needed.
It would be worthwhile, for example, to study the relationship of the
themes in presidential
assertions in news coverage content, relative to assertions by other actors such
as legislators and
to the prominence of these themes in media coverage. This can be studied in
coverage of wetlands
and any other number of topics.
Notable also is the successful demonstration of these computer-assisted
methods in
identifying major themes in text and using those themes for making meaningful
comparisons.
These methods can be applied to studying media content with reliable precision
and appealing
efficiency. The methods can handle enormous amounts of text, making possible the
investigation
of complex agenda processes over long time periods and in volumes of coverage.
Investigation of
research questions with larger sample sizes will make agenda-related studies
more powerful in
examining news coverage for differences of interest. Such large sample studies
allow trends to
emerge because the statistical law of large numbers becomes operative. It also
makes it possible
to look for patterns between and within subsets of the data, which also contain
large amounts of
information.
These findings of the comparisons in this study, considered together with
reviews of the
news articles emphasizing each major theme, show that the role of the U.S.
president in the media
agenda is more complex than some studies have suggested. While the president may
place an issue
on the public agenda and the media agenda, the president may not always control
its evolution.
Other actors in the process, including legislators, compete for dominant
influence.
It is possible that historically U.S. presidents have enjoyed more
influence on the media
agenda than they have enjoyed in recent years. Investigations by Hughes (1995)
and others
provide evidence that this is the case. Further study is needed to determine
changes in the
president-media relationship over time.
Journalism norms and values prescribe that the president will be covered
prominently in
the news media. The position and the activities of the president hold the news
values of
prominence, impact, and many times conflict, occasionally even the bizarre.
Numerous studies
cited here show that media continually and consistently report the comments of
the president. The
findings of this study, however, along with those of Hughes (1995), Gonzenbach
(1992), and
others, also indicate that the media coverage of the U.S. president includes far
more than what the
president says and wants to be said. These findings suggest that the U.S.
president is powerful
enough to tell the media what to talk about, but is not powerful enough to tell
them what to say. References
Aldenderfer, M. S., & Blashfield, R. K. (1984). Cluster Analysis (Sage
University Paper Series on
Quantitative Applications in the Social Sciences, series no. 07-44). Newbury
Park, CA: Sage.
Andsager, J. L., & Miller, M. M. (1994, November). Exploring patterns of
controversy: News
coverage of RU486. Presented to the Midwest Association for Public Opinion
Research, Chicago.
Berkowitz, D. (1992/1994). Who sets the media agenda? The ability of
policymakers to determine
news decisions. In J. D. Kennamer (Ed.), Public Opinion, the Press, and Public
Policy (pp. 81-102). Westport, Conn.: Praeger Publishers.
Bosso, C. J. (1987). Pesticides and Politics: The Life Cycle of a Public Issue.
Pittsburgh:
University of Pittsburgh Press.
Chew, F., & Kim., S. (1994, August). Using concept mapping to go beyond the
source credibility
model in assessing celebrity-message congruence. Presented to the Association
for Education in
Journalism and Mass Communication, Atlanta.
Chew, F., & Miller, M. M. (1993, November). When the source is the message:
Going beyond the
source-message effectiveness model to assess presidential attitudes congruent
with voter
preference. Presented to the Midwest Association for Public Opinion Research,
Chicago.
Chew, F., Mehta, A., & Oldfather, A. (1993, August). A re-examination of the
source-message
effectiveness model: Analyzing pieces of the celebrity endorsement puzzle via
source models,
cultural meaning transfer, and concept mapping. Presented to the Association for
Education in
Journalism and Mass Communication, Kansas City, MO.
Corter, J. E. (1996). Tree Models of Similarity and Association (Sage University
Paper Series on
Quantitative Applications in the Social Sciences, series no. 07-112). Thousand
Oaks, CA: Sage.
Coxon, A. P. M. (1982). The User's Guide to Multidimensional Scaling. Exeter,
New
Hampshire, Great Britain: Heinemann Educational Books, Inc.
Dearing, J. W., & Rogers, E. M. (1996). Communication Concepts 6:
Agenda-Setting. Thousand
Oaks, CA: Sage.
Derksen, L., & Gartrell, J. (1993). The social context of recycling. American
Sociological
Review, 58, 434-442.
Entman, R. M. (1991). Framing U.S. coverage of international news: Contrasts in
narratives of
the KAL and Iran air incidents. Journal of Communication, 41(4), 6-27.
Entman, R. M. (1993). Framing: Toward clarification of a fractured paradigm.
Journal of
Communication, 43(4), 51-58.
Gitlin, T. (1980). The Whole World is Watching: Mass Media in the Making and
Unmaking of
the New Left. Berkeley, CA: University of California Press.
Goodman, R. (1994, May). Bush administration congressional versus presidential
agenda-setting:
The China most favored nation controversy. Presented to the International
Communication
Association, Albuquerque, NM.
Gonzenbach, W. J. (1992). A time-series analysis of the drug issue, 1985-1990:
The press, the
president, and public opinion. International Journal of Public Opinion Research,
4(2), 126-147.
Hughes, W. J. (1995). The "not-so-genial" conspiracy: The New York Times and six
presidential
"honeymoons," 1953-1993. Journalism and Mass Communication Quarterly, 72(4),
841-850.
Kennamer, J. D. (Ed.). (1992/1994). Public Opinion, the Press, and Public
Policy. Westport,
Conn.: Praeger Publishers.
Kingdon, J. W. (1984). Agenda, Alternatives, and Public Policies. Boston:
Little, Brown.
Kruskal, J. B., & Wish, M. (1978). Multidimensional Scaling (Sage University
Paper Series on
Quantitative Applications in the Social Sciences, series no. 07-11). Newbury
Park, CA: Sage.
Lang, G. E., & Lang, K. (1983). The Battle for Public Opinion: The President,
the Press, and the
Polls During Watergate. New York: Columbia University Press.
Manheim, J. B. (1991). All of the People All of the Time. Armonk, NY: M. E.
Sharpe, Inc.
Miller, M. M. (1990/1995a). User's Guide for VBMap: A Program for
Multidimensional Scaling
of Concepts. Knoxville, TN: The University of Tennessee, Knoxville.
Miller, M. M. (1990/1995b). User's Guide for VBPro: A Program for Analyzing
Verbatim Text.
Knoxville, TN: The University of Tennessee, Knoxville.
Miller, M. M. (1994, November). Patterns and trends in 20 years of Public
Opinion Quarterly: An
application of concept mapping. Presented to the Midwest Association for Public
Opinion
Research, Chicago.
Miller, M. M. (In press). Frame mapping and analysis of contentious issues.
Accepted for
publication in Computers and the Social Sciences.
Miller, M. M., & Andsager, J. L. (In press). Media framing of hate speech: From
campus to the
public sphere. Accepted for publication in Newspaper Research Journal.
Miller, M. M., Andsager, J. L., & Riechert, B. P. (1996, November). Framing the
candidates in
presidential primaries: Issues and Images in press releases and news coverage.
Presented at the
Midwest Association for Public Opinion Research, Chicago.
Miller, M. M., & Riechert, B. P. (1994, August). Identifying themes via concept
mapping: A new
method of content analysis. Presented to the Association for Education in
Journalism and Mass
Communication, Atlanta.
Miller, M. M., & Riechert, B. P. (1997, May). Frame Mapping: A quantitative
method for
investigating issues in the public sphere. Presented at the American Association
for Public Opinion
Research, Norfolk, VA.
Miller, M. M., & Riechert, B. P. (In press). Frame mapping: A new method for
investigating
issues in the public sphere. Accepted for publication in Computer Content
Analysis: Theory,
Methods, Applications (Mark D. West, ed.), Ablex.
Nicholas, S. (1992). The war over wetlands. Issues in Science and Technology,
8(4), 35-41.
Nicodemus, D. (1995, March). The prestige press and environmental communication:
A case
study of The New York Times and the national wetland policy during the Bush
administration.
Presented to the Conference on Communication and the Environment, Chattanooga,
TN.
Riechert, B. . (1996). Advocacy Group and News Media Framing of Public Policy
Issues: Frame
Mapping the Wetlands Debates.
Riechert, B. P., & Miller, M. M. (1994). Magazine coverage of pesticide issues:
A computer
content analysis. Presented to the Agricultural Communicators in Education,
Moscow, ID, and
Pullman, WA.
Riechert, B. P., & Miller, M. M. (1997, May). Competing Views in the Case of
Wetlands: Frame
Mapping the Public Issue Landscape. Presented at the International Communication
Association,
Chicago.
Shoemaker, P. J., & Reese, S. D. (1991). Mediating the Message: Theories of
Influence on Mass
Media Content. New York: Longman Publishing Group.
Smiley, L., & Andsager, J. L. (1996, November). Communications in conflict: How
media agents
framed the silicone breast implant controversy. Presented to the Midwest
Association for Public
Opinion Research, Chicago.
Wetlands Fact Sheets. (1995). EPA843-F-95-001. Washington, DC: United States
Environmental
Protection Agency.
Figure 1. Wetlands News Articles Distributed by Associated Press by Year,
1984-96 (Excluding
Articles on Wetlands Outside the United States) Table 1. Terms Occurring
Frequently in News Coverage of Wetlands, Specified for Analysis of
Co-Occurrence Patterns, and Eigenvector Values for Multidimensional Space
______________________________________________________________________________
Term Frequency Coordinates 1, 2, and 3*
of Occurrence
______________________________________________________________________________
wetlands 4,566 ( 0.06, -0.06, 0.35)
environmental 3,566 (-0.24, 0.26, 0.04)
water 3,076 (-0.02, -0.17,
0.14)
farmers 1,977 ( 0.61, 0.10, 0.14)
land 1,977 ( 0.21, -0.01,
0.33)
government 1,825 ( 0.07, 0.38, 0.25)
acres 1,749 ( 0.27, -0.17,
0.31)
senate 1,692 ( 0.11, 0.50,
0.03)
conservation 1,623 ( 0.53, -0.03, 0.16)
Bush 1,660 (-0.09, 0.18, -0.34)
wildlife 1,482 (-0.06, -0.31,
0.28)
congress 1,425 (-0.05, 0.43,
-0.13)
oil 1,404 (-0.14, -0.12,
-0.14)
protection 1,303 (-0.29, 0.28,
0.30)
spending 1,250 ( 0.10, 0.50,
-0.24)
agriculture 1,224 ( 0.61, 0.21,
0.05)
budget 1,169 ( 0.02, 0.37,
-0.33)
Clinton 1,090 (-0.14, 0.45, -0.06)
development 948 (-0.08, 0.03, 0.01)
legislation 940 (-0.11, 0.51, 0.16)
property 875 (-0.22, 0.27,
0.56)
species 859 (-0.31, 0.08,
0.43)
environment 812 (-0.19, 0.22, -0.05)
policy 792 ( 0.09, 0.04,
0.12)
republicans 778 (-0.21, 0.52, 0.12)
environmentalists 772 (-0.15, 0.04, 0.06)
natural 718 (-0.09, -0.17, 0.10)
research 706 ( 0.09, 0.12,
-0.23)
protect 702 ( 0.12, 0.08, 0.20)
everglades 701 (-0.06, -0.15, -0.01)
EPA 693 (-0.20, 0.09, 0.01)
American 692 ( 0.01, 0.05, 0.04)
cost 690 ( 0.02, 0.19, -0.08)
clean 644 (-0.21, 0.14, -0.08)
tax 652 ( 0.00, 0.17, -0.27)
Table 1. (Continued)
______________________________________________________________________________
Term Frequency Coordinates 1, 2, and 3*
of Occurrence
______________________________________________________________________________
fish 650 (-0.12, -0.35, 0.20)
crop 632 ( 0.65, 0.13, 0.04)
pollution 631 (-0.15, 0.08, -0.05)
endangered 630 (-0.33, 0.16, 0.45)
industry 619 (-0.14, 0.04, -0.11)
project 611 (-0.06, -0.19,
0.00)
agricultural 602 ( 0.39, 0.05,
0.14)
birds 568 (-0.07, -0.29, 0.14)
habitat 523 (-0.04, -0.26, 0.31)
soil 511 ( 0.43, -0.07,
0.13)
production 494 ( 0.51, 0.05,
-0.02)
laws 490 (-0.25, 0.32, 0.29)
resources 483 (-0.04, -0.08, 0.09)
lawmakers 455 (-0.11, 0.51, 0.13)
acre 448 ( 0.12, -0.21, 0.18)
vote 442 (-0.06, 0.43, 0.06)
flood 432 (-0.01, -0.12, 0.10)
world 426 ( 0.00, -0.02, -0.16)
landowner 420 ( 0.03, -0.00, 0.31)
corps 414 (-0.10, -0.18, 0.12)
health 400 (-0.14, 0.28, -0.13)
subsidies 394 ( 0.49, 0.21, -0.07)
economic 392 (-0.01, 0.08, -0.11)
construction 389 (-0.11, -0.09, -0.04)
home 387 (-0.02, -0.06, -0.09)
costs 375 ( 0.11, 0.10, -0.10)
benefits 372 ( 0.31, 0.14, -0.02)
energy 370 (-0.15, 0.12, -0.25)
plants 369 (-0.08, -0.06, 0.01)
waste 362 (-0.13, 0.06, -0.13)
beach 348 (-0.08, -0.13, -0.03)
rights 326 (-0.19, 0.30, 0.48)
lakes 326 (-0.08, -0.19, 0.05)
court 322 (-0.10, 0.04, 0.11)
growers 318 ( 0.28, 0.06, 0.01)
gulf 310 (-0.08, -0.09, -0.11)
Table 1. (Continued)
______________________________________________________________________________
Term Frequency Coordinates 1, 2, and 3*
of Occurrence
______________________________________________________________________________
marshes 304 (-0.02, -0.20, 0.11)
restore 299 ( 0.03, -0.02, 0.09)
forest 299 (-0.12, 0.07, 0.03)
island 296 (-0.06, -0.12, -0.01)
erosion 294 ( 0.41, -0.05, 0.14)
Reagan 290 (-0.09, 0.06, -0.17)
build 290 (-0.09, -0.12, 0.07)
USDA 287 ( 0.49, 0.01, 0.12)
acreage 287 ( 0.44, 0.11, 0.06)
toxic 285 (-0.23, 0.11, -0.13)
regulation 285 (-0.19, 0.33, 0.30)
dole 285 (-0.06, 0.23, -0.05)
rivers 277 (-0.12, -0.19, 0.11)
duck 275 (-0.00, -0.24, 0.16)
engineers 269 (-0.12, -0.20, 0.16)
jobs 265 (-0.07, 0.08, -0.20)
waterfowl 255 ( 0.03, -0.24, 0.17)
protecting 255 (-0.05, 0.14, 0.22)
building 253 (-0.07, -0.11, 0.02)
population 251 (-0.09, -0.24 , 0.06)
environmentally 244 ( 0.07, 0.03, -0.06)
preserve 243 (-0.03, -0.06, 0.19)
producers 243 ( 0.50, 0.13,
0.02)
farms 240 ( 0.61, 0.16, 0.07)
urban 239 (-0.11, 0.12, -0.02)
funding 238 ( 0.05, 0.18, -0.12)
taking 234 (-0.11, 0.21, 0.26)
debate 234 (-0.03, 0.36, 0.23)
restoration 233 (-0.02, -0.12, 0.03)
drilling 233 (-0.13, -0.00, -0.22)
farming 231 ( 0.27, -0.02, 0.13)
sensitive 230 ( 0.02, 0.07, -0.07)
permit 228 (-0.08, -0.05, 0.10)
logging 228 (-0.16, 0.16, 0.09)
rural 220 ( 0.25, 0.15, -0.03)
developers 217 (-0.06, -0.07, 0.20)
Table 1. (Continued)
______________________________________________________________________________
Term Frequency Coordinates 1, 2, and 3*
of Occurrence
______________________________________________________________________________
erodible 215 ( 0.57, 0.03, 0.09)
runoff 214 (-0.06, -0.11, -0.04)
chemicals 212 ( 0.04, -0.01, -0.03)
threatened 211 (-0.11, -0.10, 0.16)
owners 209 (-0.18, 0.13, 0.36)
residents 204 (-0.09, -0.17, -0.06)
global 202 (-0.09, 0.11, -0.21)
appropriations 200 (-0.01, 0.37, -0.06)
animals 199 (-0.13, -0.17, 0.15)
taxes 196 (-0.04, 0.13, -0.23)
scientists 195 (-0.10, -0.07, 0.01)
______________________________________________________________________________
*Eigenvalues: 5.6608, 4.9291, 3.8051; Percentage Total Variance Per Factor: 5,
5, 4;
Cumulative Percentage Total Variance: 5, 10, 14; Percentage Extracted Variance
Per Factor:
39, 34, 26; Cumulative Percentage Extracted Variance: 39, 74, 100 Table 2.
Themes in Wetlands News Articles
______________________________________________________________________________
Group Terms Coordinates 1, 2, and 3*
______________________________________________________________________________
(1) Agriculture agriculture, acreage, crop, growers,
(0.16, -0.42, 0.50)
producers, production, subsidies, benefits,
rural
(2) Conservation conservation, soil, erosion, farming, USDA, (-0.02,
-0.35, 0.49)
erodible, farmers, agricultural
(3) Wetlands wetlands, restore, landowner (-0.16, 0.11,
0.12)
(4) Economic economic, tax, budget, sensitive ( 0.38,
-0.39, -0.20)
(5) Jobs jobs, energy, Reagan, taxes, Bush, global (
0.16, -0.33, -0.49)
(6) Regulation regulation, laws, protection, logging (
0.32, 0.47, -0.12)
(7) Endangered endangered, species, court ( 0.09,
0.55, 0.09)
(8) Rights rights, property, owners, taking, protecting (
0.26, 0.51, 0.21)
(9) Industry industry, waste ( 0.02, -0.01,
-0.42)
(10) Pollution pollution, clean, toxic ( 0.05,
0.07, -0.46)
(11) Environment environment, urban, environmental ( 0.19,
0.14, -0.35)
(12) Development development, EPA, forest, environmentalists
(-0.01, 0.16, -0.31)
(13) Spending spending, cost, funding ( 0.46, -0.28,
0.02)
(14) Congress Congress, Dole, Clinton, appropriations ( 0.55,
-0.02, -0.06)
(15) Government government, American, debate ( 0.34, 0.10,
0.24)
(16) Legislation legislation, lawmakers, vote, Republicans (
0.60, 0.24, 0.10)
(17) Runoff runoff, residents, beach, Everglades, island,
(-0.33, 0.04, -0.07)
project
(18) Scientists scientists, plants (-0.18, 0.07,
-0.14)
(19) Permit permit, threatened (-0.19, 0.23,
0.06)
(20) Preserve preserve, developers (-0.18,
0.20, 0.03)
(21) habitat habitat, wildlife, water, flood, waterfowl,
(-0.36, 0.13, 0.09)
resources, duck
(22) build build, rivers, natural, Corps, Engineers, animals
(-0.32, 0.26, 0.00)
(23) restoration restoration, fish, birds, marshes, lakes, population
(-0.45, 0.13, 0.05)
$CLIN [identifier code term in 1992-96 cases] ( 0.37,
0.28, 0.22)
$BUSH [identifier code term in 1988-92 cases] (-0.14, -0.25,
-0.14)
$REAG [identifer code term in 1984-88 cases] (-0.30, -0.06,
0.02)
______________________________________________________________________________
Note. The themes are identified by clusters of terms occurring together,
identified in hierarchical
cluster analysis of frequently-occurring terms.
*Eigenvalues: 2.3186, 1.9114, 1.6362; Percentage Total Variance Per Factor:
12, 10, 9;
Cumulative Percentage Total Variance: 12, 22, 31; Percentage Extracted
Variance Per Factor;
40, 33, 28; Cumulative Percentage Extracted Variance: 40, 72, 100
Figure 2. Concept Map of News Coverage of Wetlands, 1984-96.
Figure 3. Dendogram Output of Hierarchical Cluster Analysis of Clustering Terms
from Wetlands
News Articles. Table 3. Major Themes Formed by Cluster Analysis of 23 Themes in
Wetlands News Coverage
(Frequently-Occurring Term Groups)
______________________________________________________________________________
Major Theme* Narrower Themes (Term Groups**) Included
______________________________________________________________________________
Habitat Restoration Permit, Preserve, Build, Habitat, Restoration,
Wetlands, Runoff, Scientists
Jobs/Development Industry, Pollution, Development, Environment,
Jobs
Regulation Legislation Government, Legislation, Endangered, Rights,
Regulation
Agriculture/Spending Agriculture, Conservation, Economic,
Spending, Congress
______________________________________________________________________________
* Suggested by clustering schedule shown in dendogram in Figure 3
**Listed in Table 2
Table 4a. Prominence of Habitat Restoration Theme in Wetlands News Coverage by
Presidential
Administration
______________________________________________________________________________
Reagan Bush Clinton
______________________________________________________________________________
Mean: 13.31a 11.40b 10.10b
SD: 10.77 10.43 10.38
n: 354 686 716
______________________________________________________________________________
Note: Means in the same row that do not share subscripts differ at p < .05 by
the Scheff test.
Table 4b. One-Way Analysis of Variance for Prominence of Habitat Restoration
Theme in
Wetlands News Coverage by Presidential Administration
______________________________________________________________________________
Source df ss ms F Ratio
______________________________________________________________________________
Between Groups 2 2451.17 1225.59 11.15*
Within Groups 1753 192601.55 109.87
Total 1755 195952.72
______________________________________________________________________________
* p < .0001
Table 5a. Prominence of Jobs/Development Theme in Wetlands News Coverage by
Presidential
Administration
______________________________________________________________________________
Reagan Bush Clinton
______________________________________________________________________________
Mean: 6.64a 8.09b 6.26a
SD: 9.50 9.48 6.68
n: 354 686 716
______________________________________________________________________________
Note: Means in the same row that do not share subscripts differ at p < .05 by
the Scheff test.
Table 5b. One-Way Analysis of Variance for Prominence of Jobs/Development Theme
in
Wetlands News Coverage by Presidential Administration
______________________________________________________________________________
Source df ss ms F Ratio
______________________________________________________________________________
Between Groups 2 1246.69 623.35 8.72*
Within Groups 1753 125340.70 71.50
Total 1755 195952.72
______________________________________________________________________________
* p < .001
Table 6a. Prominence of Regulation Legislation Theme in Wetlands News Coverage
by
Presidential Administration
______________________________________________________________________________
Reagan Bush Clinton
______________________________________________________________________________
Mean: 3.80a 4.13a 10.07b
SD: 4.32 3.74 12.34
n: 354 686 716
______________________________________________________________________________
Note: Means in the same row that do not share subscripts differ at p < .05 by
the Scheff test.
Table 6b. One-Way Analysis of Variance for Prominence of Regulation Legislation
Theme in
Wetlands News Coverage by Presidential Administration
______________________________________________________________________________
Source df ss ms F Ratio
______________________________________________________________________________
Between Groups 2 15603.19 7801.60 109.36*
Within Groups 1753 125057.29 71.34
Total 1755 140660.49
______________________________________________________________________________
* p < .0001
Table 7a. Prominence of Agriculture/Spending Theme in Wetlands News Coverage by
Presidential Administration
______________________________________________________________________________
Reagan Bush Clinton
______________________________________________________________________________
Mean: 8.15a 9.42a 11.33b
SD: 12.76 12.07 10.51
n: 354 686 716
______________________________________________________________________________
Note: Means in the same row that do not share subscripts differ at p < .05 by
the Scheff test.
Table 7b. One-Way Analysis of Variance for Prominence of Agriculture/Spending
Theme in
Wetlands News Coverage by Presidential Administration
______________________________________________________________________________
Source df ss ms F Ratio
______________________________________________________________________________
Between Groups 2 2696.11 1348.05 10.0004*
Within Groups 1753 236305.53 134.80
Total 1755 239001.64
______________________________________________________________________________
* p < .0001
Methods