A Longitudinal Study of Agenda Setting
for the Issue of Environmental Pollution
Christine R. Ader
School of Journalism
Southern Illinois University at Carbondale
Presented at the Mass Communication and Society Divison
of the Association for Education in Journalism and Mass Co
Kansas City, MO, August 13, 1993
The agenda-setting hypothesis, which has drawn increasing attention
in the mass communications literature, posits a relationship between
the relative emphasis given by the media to various topics and the
degree of salience these topics have for the general public. Individuals
note the amount of and distribution of media coverage among issues
and this determines the salience of each issue for the individuals.
According to the agenda-setting hypothesis, the media do not mirror
public priorities as much as they influence them.
Most of the agenda-setting research to date follows the historical
lead of examining the relationship between the results of quantitative
content analysis of issues in the media and survey of the public's
issue agenda. Variations of this design exist usually in the type
and number of issues and the duration of development of the issue
in the public agenda. Recent studies also have examined the relationship
between the public policy agenda, the media agenda, and the public
agenda. One often ignored, yet important, component for such studies
is the real-world condition of the agenda item. The real-world condition
is the actual prominence of the specific issue in reality or the
indicators of current conditions.
_The Media Agenda and Real-World Conditions_
Lippmann (1922) noted that the media shapes people's perceptions
of things which they cannot experience directly. Those issues which
individuals have little personal contact and for which they rely
on the media as the primary, and sometimes, only source of information
are termed unobtrusive.
Numerous researchers have observed that unobtrusive issues demonstrate
a strong agenda-setting effect (Eyal, 1979; Zucker, 1978; Behr &
Iyengar, 1985). The environment is one such issue. In fact, in a
study using factor analysis to distinguish obtrusive issues from
unobtrusive issues, Eyal (1979) found that the issue of the environment
loaded strongest on the unobtrusive factor.
Zucker (1978) demonstrated that agenda setting took place for
three unobtrusive issues--pollution, drug abuse, and the energy
crisis--but did not occur with three obtrusive issues--the cost
of living, unemployment, and crime. Agenda setting did not occur
for the obtrusive issues because individuals can rely on real-world
conditions and interpersonal discussion for information, while for
unobtrusive issues, the individuals only have information from the
media to rely on.
Behr and Iyengar (1985) concluded that obtrusive issues--issues
with tangible consequences for individuals, such as energy, unemployment,
or inflation--not only affect the public by news media coverage,
but also through real-world events and conditions. They observed
that real-world conditions should be examined because they serve
two purposes: first, they can help assess the sensitivity of the
media agenda to current conditions and events. Second, they can
help distinguish between the effects of news coverage and real-world
conditions on the public agenda (p. 40). Behr and Iyengar (1985)
also pointed out that:
Moreover, since news coverage of issues is to a
significant extent determined by actual conditions,
analyses of media agenda setting that ignore real-world
conditions will arrive at severely inflated estimates of media influence
McLeod, Becker, and Byrnes (1974, p. 141) observed that if a
valid assessment of real-world conditions was possible, the relative
similarity of the audience's perceived saliencies to external reality
might serve as a control against which to evaluate the similarity
of the public agenda to the media agenda.
Many agenda-setting studies have only analyzed the media agenda
and the public agenda. By doing so, a researcher may discover a
significant relationship but this relationship might be dependent
on some exogenous variable (Erbring, Goldenberg, & Miller, 1980,
p. 19). Also they discovered that real-world conditions may affect
an individual's awareness of the media agenda. They proposed an
audience-contingent effects model of media in which the public agenda
is determined by the media agenda interacting with the audience's
In this study, real-world conditions along with the public agenda
and the media agenda were examined and this achieved two important
ends. The first important end was to determine if the media agenda
corresponds to real-world conditions. The second goal was to distinguish
between the effects of news coverage and real-world conditions on
the public agenda. The relationship between real-world and the public
agenda serves as a control against the relationship between the
public agenda and the media agenda.
_The Environment and Public Opinion_
Environmental problems became a leading item on the national
agenda at the beginning of the 1970s. This concern and interest
manifested itself in April 1970 when hundreds of thousands participated
in Earth Day events all over the country (Mitchell, 1980, p. 1).
Media coverage of the environment has increased dramatically recently
(Detjen, 1990) and so has public concern for environmental issues.
In fact, the persistence and recent revival of public concern for
environmental issues was termed a "second miracle" of public opinion
by Dunlap and Scarce (1991, p. 652).
Zucker (1978) discovered a significant relationship between the
media agenda and the public agenda for the issue of pollution. Eyal
(1979) noted agenda-setting effects for unobtrusive issues and the
issue of the environment loaded strongest on the unobtrusive factor.
Funkhouser (1973a) proposed that the amount of media coverage of
an issue, including coverage for the issue of the environment, strongly
influences its visibility to the public. Based on these finding,
it is hypothesized that:
The media agenda and the public agenda for the issue ofpollution
will be related. That is, as coverage (in mean number of column
inches) and prominence of the environment increases, the
percent of responses which name environmental pollution as
the most important problem facing the country will also increase
and vice versa.
Eaton (1989) found the media coverage of three nightly networks,
five national newspapers, and three weekly national news magazines
and the real-world conditions for the issue of pollution had a low
correlation from 1968-1976. Babcock (1979) observed that _Newsweek_Us
coverage of air pollution, water pollution, soil pollution, and
total pollution was uncorrelated respectively with the real-world
conditions for air quality, water quality, soil quality, and mean
environmental quality from 1969 to 1975. Other studies found a lack
of correspondence between the real-world conditions and media coverage
for the issue of pollution (Zucker, 1978; Funkhouser, 1973a). Based
on these findings it is hypothesized that:
The relationship between the media agenda and
real-world conditions for the issue of pollution will be unrelated.
Unobtrusive issues are those which do not have tangible consequences
for individuals. Eyal (1979) found that the environment loaded strongest
on the unobtrusive factor. Zucker (1978) noted that agenda setting
did not occur with the unobtrusive issue of the environment because
the individuals only have information from the media to rely on.
Based on these findings it is hypothesized that:
The relationship between the public agenda and
real-world conditions for the issue of pollution will be unrelated.
The methods of study was content analysis of the _New York Times_
and secondary analysis of data including Gallup poll surveys,
_Environmental Quality_, and the _Characterization of Municipal Solid Waste
in the United States_. The content analysis assessed the media agenda,
while the analysis of secondary sources provided the public agenda
and real-world conditions respectively, for the issue of pollution.
_Sample_: In view of the support for print media's agenda-setting
effect over television, the elite newspaper, the _New York Times_
was examined for this study. Massing (1984) found that this newspaper
is one of the two most elite newspapers which influence television
The time period studied was 1970 to 1990, both years inclusive,
because 1970 marks the year when the environmental movement gained
widespread interest and appeal. The _Times_ was examined three months
before and after each Gallup poll because agenda setting implies
a relationship between the media agenda and the public agenda. This
three month time period was chosen because Stone (1975) concluded
that the parameters for the agenda-setting effect extended from
two to six months prior to the public agenda measure. Shoemaker,
Wanta, and Leggett (1989) found two time periods in which media
coverage of drug issues correlates with later public concern about
drugs--one to two months and four to five months. Winter (1979),
who tested several different time lags in order to identify the
optimal timing for media content to influence the public agenda,
reported the optimal timing to be about two to four months.
The _New York Times_ Indexes was used for 1970 to 1990, inclusive.
However, the _Times_ was not published from August 10, 1978 to November
5, 1978, inclusive, because of a strike. Four categories in the
_Index_ were used: air pollution, environment, water pollution,
and waste materials and disposal. For each year, a random number
was chosen between one and nine, and this number determined the
starting point of coding for each category.
For each category for each year, every tenth story was coded
beginning with the story that was randomly chosen. In other words,
a 10 percent sample of the four categories was coded. If the story
was not about air, water, or waste pollution, then the next consecutive
story was coded. Every tenth story was coded following this story.
Jumped parts of a story were not treated as separate items.
All environmental stories were identified in this sample. Only
news sections and editorials were analyzed. Hungerford and Lemert
(1973, p. 477) defined environmental content as:
...dealing with man's positive, negative, or unknown
influence upon, or relationship with, his environment.
This influence or relationship could be past, present or
future. This definition would include such varied topics
as wildlife preservation, sewage disposal problems,
reviews of environmental 'specials' on television,
nuclear (thermal) pollution and citizen complaints about
agricultural grass field burning.
Other researchers (Atwater, Salwen, & Anderson, 1985; Salwen,
1988) have defined environmental issues as those which are related
to humanity's unintentional disruption of the ecological system.
For purposes of this study, environmental content was defined as
dealing with humanity's influence, whether positive or negative,
on the environment. This included news items relating to humanity's
unintentional disruption of the ecological system such as disposal
of wastes, air quality, and water quality.
The coding unit was the paragraph, measured in column inches.
That is, once an environmental story was located, each paragraph
was coded into three topic categories including: 1) disposal of
wastes, 2) air quality, and 3) water quality. Atwater, Salwen,
and Anderson (1985) and Salwen (1988) noted that these categories
were the most salient environmental issues in the mass media. Definitions
are given below:
1. Disposal of Wastes: stories dealing with the storage,
transport, recycling, and/or dumping of waste
2. Air Quality: articles dealing with such problems as
smog, carbon monoxide, sulfur dioxide, and other
pollutants resulting from automobile exhaust, factory
emissions, and other stationary sources; their effects
on animal health and plant life; their cost to the
economic system; and methods of control.
3. Water Quality: including references to coastal barrier
protection, ocean dumping, offshore development,
and/or wetlands. Also, each story was coded for prominence.
Prominence was operationally defined by placement, length of story,
length of headline, if the headline was above the fold or not, and if
the article had pictures or cartoons.
_Reliability_: Reliability checks were conducted after the
original data collection was completed to determine the level of
consistency in measurement. A 10 percent sub-sample was recoded
to determine inter- and intra-coder reliabilities.
Intracoder and intercoder reliability test results
Variable Intracoder Intercoder
Category 99.7% 100%
Prominence 97.6% 96.2%
Length 98.6% 92.8%
_Public Agenda_: Gallup poll survey data was used to determine
the public agenda. Since Gallup poll data does not ask respondents
about subissues of the environment, only the general issue of environmental
pollution was analyzed.
Every Gallup poll was examined from 1970 to 1990 to determine
if environmental pollution had been mentioned. This yielded a total
of 66 MIP polls.
_Real-World Conditions_: Few efforts have been made to develop
total environmental indices because of the complexity of the problem.
One researcher has suggested that there are over 100 factors which
can be ranked and grouped into 14 categories, for which indices
of environmental quality could be constructed (Pikul, 1974).
Current conditions for the issue of pollution were operationally
defined by data from _Environmental Quality_, 1970 to 1991, inclusive,
and _Characterization of Municipal Solid Waste in the United States_.
The data from _Environmental Quality_ that was used to define the
1. Air Quality: total emissions, measured in millions
of metric tons, of carbon monoxide, sulfur oxide,
total suspended particulates (particles of smoke,
dust, etc.), and nitrogen oxides added together.
Index is based on the primary National Ambient
Air Quality Standard that protects public health.
Another index, the Pollution Standard Index, is
based on NAAQS and uses all of the above air pollution
factors in calculating its index.
2. Water Quality: oil polluting incidents reported in and
around U.S. waters, measured in millions of metric tons.
_Characterization of Municipal Solid Waste in the United States_
was used to define:
% Disposal of Wastes: gross waste generated as measured
in million of metric tons.
The above categories were chosen because they are the most visible
of the environmental problems. In addition to the above categories,
an index for overall environmental quality, measured in millions
of metric tons, was calculated for each year by summing the individual
subindices above. Although air quality and water quality originally
are measured in other units, they were converted to millions of
_Tests for Hypotheses_: For investigating hypothesis 1, two
Pearson r correlations were calculated: a correlation of the media
agenda three months before the poll with the poll percentage and
a correlation of the media agenda three months after the poll with
the poll percentage. This allowed for pre- and post-comparisons.
The media agenda was operationalized first by average length and
then by average prominence score.
To investigate hypothesis 2, the real-world condition pollution
indices for each year was correlated with the mean number of column
inches devoted per year to environmental pollution. The total pollution
index and the three pollution subissues of water, air, and waste
pollution each were correlated separately.
For investigating hypothesis 3, the percentage of Gallup poll
responses per year which named environmental problems/pollution
as the most important problem in America was correlated with the
air pollution index, water pollution indicator, waste pollution
indicator, and the pollution index.
_Distribution of Total Environmental Articles by Year_: A total
of 1,954 environmental stories were coded for this study, a 10 percent
sample of the total number of environmental stories listed in the
_Times Index_ under the categories of air pollution, environment,
waste and disposal, and water pollution. The year 1970 had the largest
number of articles with 161 stories and 1978 had the least with
53. Table 2 and Figure 1 illustrate the number of coded stories
for the years 1970 to 1990, inclusive.
Year versus Frequency, Average Length, and Average Prominence Score
Year Frequency Length Prominence
1970 161 10.8 1.91
1971 134 12.3 1.92
1972 119 11.5 2.02
1973 121 11.2 1.97
1974 92 9.1 1.95
1975 102 10.7 1.76
1976 110 9.2 1.88
1977 102 12.5 2.13
1978 53 9.0 1.69
1979 78 10.8 1.69
1980 71 11.0 1.90
1981 88 12.2 2.37
1982 71 13.6 2.28
1983 93 15.3 2.63
1984 69 15.0 2.36
1985 62 16.6 2.55
1986 64 15.5 2.42
1987 75 15.6 2.57
1988 92 16.9 2.86
1989 104 18.0 3.18
1990 93 18.8 3.13
_Distribution of Articles by Average Length and Prominence Score_:
Figure 2 illustrates average length of the environmental stories
versus the years 1970 to 1990, inclusive. There is no discernable
trend from 1970 to 1977 because of fluctuations and the average
length increases steadily from 1978 until 1990.
Hence, in general while the frequency of stories was decreasing
over the years, the average length and average prominence score
Figure 3 illustrates average prominence score of the articles
versus the years 1970 to 1990, inclusive. The year 1989 had the
greatest average prominence score with 3.134 and 1978 had the lowest
_Distribution of Articles by Environmental Category_: By
environmental category, there were a total of 573 air pollution articles,
766 water pollution articles, and 615 waste and disposal articles.
Distribution of articles by environmental category
Pollution Total Average Average
Category Number Length Prominence
of Articles Score
Air 573 12.401 2.079
Water 766 12.476 2.184
Waste 615 14.022 2.474
Total 1954 12.966 2.245
_Pollution Real-World Conditions by Year_: Figure 4 illustrates
year versus air pollution measured in million metric tons. The amount
of sulfur oxides, nitrogen oxides, carbon monoxide, and total suspended
particulates for each year was summed to operationalize the air
pollution index. The graph shows that since 1970, the amount of
air pollution has steadily decreased.
Figure 5 illustrates the amount of oil, measured in million metric
tons, spilled each year. The figure illustrates that there has
been many fluctuations throughout the years and this is due to the
fact that oil spills are usually caused by accidents.
Figure 6 illustrates the net amount of solid waste, measured
in million metric tons, which was disposed of each year. Since 1970
there has been an overall increase in the amount of disposed net
solid waste. Figure 7 show the total pollution index versus year.
The total pollution index was calculated by summing air pollution,
oil spills, and net solid waste for each year. This index is in
millions of metric tons. Overall, the amount of pollution has steadily
declined although there were a few years, such as 1976, 1977, and
1978, when the amount increased.
_Hypothesis 1_: This study tested the agenda-setting hypothesis
by correlating the media agenda three months before each poll with
the public poll percentage and also correlated the media agenda
three months after each poll with the poll percentage. The media
agenda was operationalized by average length and average prominence
score. The agenda-setting hypothesis was supported because the correlation
of the pre-poll media agenda with the poll percentage was statistically
significant and stronger than the correlation of the post-poll media
agenda with the poll percentage. Table 4 provides the correlations.
Correlations between poll percentages and pre- and post-poll media
Time Period Avg. Length Avg. Length Avg. Prom. Avg. Prom.
Pre-Poll Post-Poll Pre-Poll Post-Poll
70-90 0.224* 0.165 0.240* 0.0456
* correlation significant at 0.1 level
_Hypothesis 2_: Hypothesis 2 postulated that there would be
no correspondence between real-world conditions and the media agenda
for the issue of pollution. Pearson r correlations were run for
each subissue between the media agenda, operationalized by the subissue's
average article length or average prominence score, and the subissue's
real-world indicators for each year. Additionally, a correlation
was run between the overall pollution index and the average length
and average prominence score of all environmental stories for each
year. Table 5 reports the correlations.
Correlation between real-world indicators of pollution with their
respective media agendas
Pollution Category Length Prominence
Total Pollution Index -0.811* -0.796*
Air -0.702* -0.650*
Water -0.569* -0.532**
Waste 0.675* 0.815*
*correlation significant at 0.01 alpha level
**correlation significant at 0.02 alpha level
Overall, all the real-world indicators had a statistically significant
correlation with the media agenda, operationalized by either average
length or average prominence score. All of the correlations were
statistically significant and all were negative except the correlations
for the waste and disposal pollution. The total pollution index,
air pollution, and water pollution have all declined on the average
through the 21 years while the average prominence score and average
length of environmental articles have increased. Therefore, the
correlations are negative. This indicates that despite a reduction
in pollution, the media coverage has increased. The fact that waste
pollution has the only positive correlations is due to the fact
that waste pollution has steadily increased and overall, so has
the average length and average prominence scores.
_Hypothesis 3_: Hypothesis 3 postulated that there would be
no relationship between real-world conditions and the public agenda
for the issue of pollution. Table 6 illustrates the Pearson r correlations
between the Gallup poll respondents naming environmental problems
as the most important problems in America with real-world indicators
Percentage of Gallup respondents naming environmental problems as
MIP correlated with real-world indicators of pollution
Pollution Category Correlation
Air Index 0.344
Total Pollution Index 0.354
Overall, real-world indicators of pollution are statistically
not significantly correlated with the Gallup poll percentage naming
environmental problems as the MIP.
From 1970 to 1990, the agenda-setting hypothesis was supported
for the issue of environmental pollution. Hence, the amount of media
attention devoted to pollution determined the degree of public salience
for this issue. This is consistent with previous findings by researchers
that the environment is an unobtrusive issue, an issue which does
not have tangible consequences for individuals. Therefore, the public
has little personal contact with pollution and so they rely on the
media for information. This was reinforced by this studyUs finding
that real-world conditions and the public agenda for pollution were
The media agenda and real-world conditions for pollution did
have a statistically significant relationship. However, the correlations
for the total pollution index, air pollution, and water pollution
were negative. In other words, despite the overall reduction in
pollution, media coverage on pollution has increased. Probably this
increase is partially due to greater amounts of government legislature
aimed at reducing the amount of pollution. In addition, the number
of special interest groups supporting environmental protection has
However, a positive correlation was found between the media agenda
and real-world conditions for the waste and disposal category. This
result is not unexpected since the news mediaUs renewed attention
in the environment has been paralleled by an increase in the public's
interest in recycling its waste. This is due to the fact that landfills
are becoming scarce and recycling is becoming a crucial and controversial
issue. Waste pollution is directly impacting the public.
Overall, by including real-world conditions, the sensitivity
of the media agenda and public agenda to current conditions was
assessed. As predicted, this study found that real-world conditions
do not influence the media or public agendas directly.
Atwater, Tony, Salwen, Michael B., & Anderson, Ronald B. (1985).
Media Agenda-Setting with Environmental Issues. Journalism Quarterly,
Babcock, William A. (1979). Environmental Pollution Coverage in
Newsweek from 1969 through 1975. Unpublished doctoral dissertation,
Southern Illinois University, Carbondale, Ill.
Behr, Roy L., & Iyengar, Shanto (1985). Television News, Real-World
Cues, and Changes in the Public Agenda. Public Opinion Quarterly,
Detjen, Jim (1990, Summer). The Traditionalist's Tools (And a Fistful
of New Ones). Gannett Center Journal: Covering the Environment,
4, 73- 83.
Dunlap, Riley E. & Scarce, Rik (1991). The Polls-Poll Trends:
Environmental Problems and Protection. Public Opinion Quarterly, 55,
Eaton Jr., Howard (1989). Agenda-Setting with Bi-Weekly Data on Content
of Three National Media. Journalism Quarterly, 66(4), 942-948, 959.
Erbring, Luis, Goldenberg, Edie N. & Miller, Arthur H. (1980). Front-Page
News and Real-World Cues: New Look at Agenda-Setting by the Media.
American Journal of Political Science, 24, 16-42.
Eyal, Chaim H. (1979). Time-Frame in Agenda-Setting Research: A Study
of the Conceptual and Methodological Factors Affecting the Time
Frame Context of the Agenda-Setting Process. Unpublished doctoral
Syracuse University, Syracuse, N.Y.
Funkhouser, G. Ray (1973a). The Issues of the Sixties: An Exploratory Study
in the Dynamics of Public Opinion. Public Opinion Quarterly, 37, 62-75.
Lippmann, Walter (1941). Public Opinion. New York: Harcourt Brace.
Massing, Michael (1984, January/February). The Network Newscasts: Still
Hot off the Presses. Channels of Communications, 47-52.
McLeod, Jack M., Becker, Lee B., & Byrnes, James E. (1974). Another
Look at the Agenda Setting Function of the Press. Communication
Research, 1, 131-166.
Mitchell, Robert C. (1980). Public Opinion on Environmental Issues:
Results of a National Public Opinion Survey. Washington, D. C.:
Resources for the Future.
Pikul, Robert (1974). Development of Environmental Indices. In John
W. Pratt (Ed.), Statistical and Mathematical Aspects of Pollution
Problems. New York: Marcel Dekker, Inc.
Salwen, Michael B. (1988). Effect of Accumulation of Coverage on
Issue Salience in Agenda Setting. Journalism Quarterly, 65, 100-106.
Shoemaker, Pamela J., Wanta, Wayne, & Leggett, Dawn (1989). Drug Coverage
and Public Opinion, 1972-1986. In Pamela J. Showmaker (Ed.), Communication
Campaigns about Drugs: Government, Media, and the Public. Hillsdale,
New Jersey: Lawrence Erlbaum Associates, Inc.
Stone, Gerald (1975). Cumulative Effects of the Media. Paper presented
at the Conference on Agenda-Setting, Syracuse University, Syracuse,
Winter, James P. (1979). An Agenda-Setting Time Frame for the Civil
Rights Issue, 1954-1976. Unpublished paper, Communications Research
Center, Syracuse University.
Zucker, Harold G. (1978). The Variable Nature of New Media Influence.
In B. D. Rubin (Ed.), Communication Yearbook No. 2 (pp. 225-245).
New Brunswick, N.J.: Transaction.
spondents naming environmental problems as MIP correlated with real-world
indicators of pollution
Pollution Category Correlation
Air Index 0.344