|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
FRAMING ANALYSIS OF THE NEW YORK TIMES AND LE MONDE FOLLOWING
THE ATTACKS OF SEPTEMBER 11
ALLISON IRENE AIKEN
A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF ARTS IN MASS COMMUNICATIONS
UNIVERSITY OF FLORIDA
I would like to first and foremost thank my committee chair, Dr. Kurt Kent, for his
continued support of this thesis the past two years. I would also like to thank Dr. Leonard
Tipton and Dr. Ido Oren for their support. I would additionally like to thank Dr. Julie
Dodd for her continued support over the past two years on this research and my general
TABLE OF CONTENTS
A C K N O W L E D G M E N T S ............. ...............................................................................II
LIST OF TABLES ............. .............................. ....... .. ...... ............ VI
AB STRA C T ... .................... .............................................. ................ V II
1 IN TR OD U CTION ............................................... .. ......................... ..
W hat Is the R ole of the Press? ............... ... ...................... ...............................
The Decline of International News in the American Press .........................................5
A m erican and French Press ........................................ .......................................8
D e fin itio n s ........................................................... ................ 9
Justification for Study ..................................................... ... .. ............ 10
2 REV IEW O F LITER A TU RE ........................................................................ .....12
F ra m in g ................................................................................................................. 1 2
The Elite Press ......................................................... ....... ...............18
Description of Newspapers Used for this Study ....................................................20
L e M o n d e ....................................................... ................ 2 0
The N ew York Tim es........... ...................................................... ............... 22
S u m m a ry .......................................................................................................2 4
H ypotheses ................................................. 24
3 M E T H O D .............................................................................2 8
C o n ten t A n aly sis ................................................................................................... 2 8
Quantitative Content Analysis ...................... ............... 30
Qualitative Content Analysis ................ ........ .... .........30
Study Materials ......................................................................... ........... ......... .........31
V a ria b le s ....................................................................................................... 3 3
F ra m e s ........................................................................................................... 3 3
S o u rc e s ................................................................3 7
A n a ly sis ..............................................................................3 9
R e lia b ility .............................................................................4 0
V a lid ity ..............................................................................4 1
4 F IN D IN G S ......... .........................................................................................4 3
Brief Overview ..................................................................43
Results of Application of Method ................ ........ .................................... 45
N nature of Sam ple ............. .............................................................................47
R liability A naly sis ................ ..... ... ............................................47
Descriptive Analysis and Hypotheses Results................................. ...............48
5 DISCUSSION ....................................................... ........... ................. 51
R results of A application of M ethod................................................................... ......51
Descriptive Analysis Discussion ......................................................53
Post-hoc analysis ......................................... ... .............. ...... 56
Sum m ary of H ypotheses .................. ..................................... ...............60
6 CONCLUSION..................... ..................64
S u m m ary ......................................................................................................... 6 4
C o n c lu sio n s.....................................................................................................6 7
L im stations ...................................................................................................... ....... 69
F utu re R research .............................................................................................7 1
Implications ...................... .......... ............ .................73
A C O D IN G PR O TO C O L ......................................................................... .............74
I. Project D description .................. .......................... .... .. ... .. ........ .... 74
II. S am p le ...........................................................7 5
III. Length to be coded ............................ ..... ... ...... .. ... .. ............. 75
IV C oding L og ........................................................75
V Filling in the coding sheets ....................................................... ......... 76
B D A T A ...................................... .................................................... 8 1
LIST O F R EFEREN CE S ..................................................... ................... 90
B IO G R A PH IC A L SK E T C H ...................................................................... ..................93
LIST OF TABLES
T able 3.1 V variable definitions ............................ ...................................... ................. .34
Table 4.1 N um ber of articles each day ........................................ ......................... 44
Table 4.2 Fram e usage ................ .................. ........................... .... ..... 44
T able 4.3 Source eliteness............. .......................................................... ...... .... ..... 45
Table 4.4 Source in relation to event ................................................... ..................45
Table B.1 The N ew York Tim es .................................................................... 81
Table B .2 L e M onde .............................................. .. .. .......................85
Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Master of Arts in Mass Communication
FRAMING ANALYSIS OF THE NEW YORK TIMES AND LE MONDE FOLLOWING
THE ATTACKS OF SEPTEMBER 11
Allison Irene Aiken
Chair: Kurt Kent
Major Department: Journalism and Communications
The terrorist attacks of September 11 shocked the entire world. The media frenzy
that ensued was unlike anything ever witnessed. The present study looks at the first ten
days after the attacks to discover, using framing theory, reporting differences between the
United States and France.
Using The New York Times and Le Monde, the present study content analyzed a
ten-day period of articles about the attacks. Five media frames defined in previous
research were employed as a means of categorization for the articles. In addition to
analyzing the frames used in each article, the present study also analyzed the type of
source and the eliteness of the source used in each article.
There was a significant difference in the way the two newspapers used the five
media frames and elite sources. Le Monde used the economic consequences frame more
significantly than The New York Times did, and The New York Times used the human-
interest frame and conflict frame significantly more than Le Monde used them. Le Monde
also used elite sources significantly more than The New York Times used them.
Additional analysis was done looking at the data in two time periods different
from the period under investigation. The data were looked at for the first two days under
investigation (September 12-13) and for the eight-day period following the immediacy of
the attacks (September 14-21). The two-day analysis showed no significant differences in
the newspapers' use of frames or sources. The eight-day analysis showed stronger
differences in usage of frames and sources than the analysis of the whole ten-day period
The tragic events of September 11 not only shocked Americans, they shocked the
entire world as well. Media all over the world were suddenly faced with a huge story that
needed daily coverage. The media were faced with the daunting task of quickly providing
audiences with what they wanted and needed to know. As Navasky (2002) put it in the
foreword to Journalism after September 11, "[it] would be a mistake to minimize the
difficulties the media faced covering the uniquely traumatic and unprecedented events of
September 11 and their aftermath" (xiii). He also adds that it would be a mistake not to
recognize the achievements of The New York Times and others that came out of this crisis
The pressure on the media came not only because these were traumatic and
unprecedented events, but also because they were where hundreds of thousands of people
would be getting their information of what was going on in the world and right around
them. As Navasky (2002) put it, "it's based largely on journalism that we make up our
national mind" (p. xiii). And with that being said, the pressure was on for the press to get
out there and tell the world "what to think about" (Cohen, 1963, p. 13) this international
What Is the Role of the Press?
In this study, the researcher looked at frames used in The New York Times and Le
Monde in news articles about the attacks of September 11, 2001. The study also
observed the types of sources the newspapers used when reporting on the event. These
items were observed for a ten-day period after the event occurred. They were observed
mainly to see if the newspapers remained true to their traditional roles in society during a
time of international crisis. Studies have been done in the past that have become the basis
for communication research showing a definite role of the press in society. But, what has
been the role of The New York Times and Le Monde in the United States and France
As Chomsky (1989) argues in Necessary Illusions, a problem prevalent in
democracies since the beginning of democracy persists: "Decision making power over
central areas of life resides in private hands, with large-scale effects throughout the social
order" (p. vii). He continues by arguing that in advanced democratic industrial societies,
this problem is often approached by depriving democratic political structures of
substantive content, while still technically leaving them intact. Chomsky believes
institutions like the media, which channel thought and attitudes within acceptable bounds,
take on a large part of this task. So, would The New York Times and Le Monde do this in
their reporting of the September 11 attacks? Would they deprive their audiences of the
content they wanted, or would they break from Chomsky's mold?
It is difficult to come by just one single answer to this question mainly because of
the differences between the press in these two countries. The United States had been
experiencing a serious cutback in international news coverage. Cunningham (2001)
reports that U.S. newspaper space devoted to international news had dropped from 10
percent in 1971 to a mere 6 percent in 1995. This being said, American newspapers were
now going to have to devote more space to international news than they had in the
previous three decades. Since non-American media systems had not been cutting back
their international coverage as the American systems had, the non-American systems
could have more page space and air time devoted to pieces about the effects of the attacks
felt in other countries than their own if they wanted to. Their space and time could be
devoted to emotional/analytical concepts, while the American media was going to have to
devote more time and space to the traditional who, what, when, and where. In fact,
foreign news sources that were available in America, such as the BBC, experienced
increased audiences. Those seeking to escape what James (2001) called "tunnel vision"
turned ever more to sources like BBC World News and ITN World News for Public
Television (The New York Times, 11/9/01).
In a November 2001 poll by Gallup Europe, 62.9 percent of the French surveyed
used their own national press to keep informed about the events going on in Afghanistan,
Pakistan, and the United States and the impact they were having on the rest of the world.
Only 15.9 percent of those surveyed by the same poll in France were using the press of
other countries. However, 21.4 percent of Germans surveyed were reading the press of
other countries for the same information. The French seemed to be relying on their own
national press to bring them the information they wanted to know. The French press,
perhaps, had a different role to play than that of the press of other European countries.
Le Monde was on the scene in Washington and New York as soon as the attacks
happened. They were on hand and able to report to their readers about the events and
aftermath just as The New York Times'reporters were. Both papers were there to answer
questions being raised in the minds of their readers. Thirty percent of French surveyed by
Gallup Europe in November 2001 were fearful of "imminent terrorist acts," more than in
any other EU country, while 39 percent of Americans polled by CBS and The New York
Times said they were very concerned about an attack where they live (retrieved on April
22, 2002 from http://www.americans-
world.org/digest/global_issues/terrorism_emoResp.cfm). It seemed both papers had an
important job in front of them to fulfill the needs of their audiences, who were fearful
of more attacks.
A 1992 Chang and Lee study found that only when there was a perceived impact
on American security and national interest was an international news story selected for
U.S. daily newspapers. If it takes a threat to national security for the American press to
print international news, then the attacks of September 11 were exactly the type of threat
that was going to demand newspapers print international news. Even though this was an
international event, the American people were going to be demanding information
because it happened on American soil and directly threatened American national security
in a very prominent way. The New York Times would have to send out reporters
immediately to cover the ensuing events since they knew what their audience wanted to
know. But what would Le Monde do? Because of the attacks, they would also have to
send out more reporters to different countries to cover the crisis.
Sending out many reporters to so many different locations across the globe would
cost time and money for these two newspapers. Time and money would have to come
from newsholes and budgets that the media in America had been cutting back. Parks
(2002) argues that most U.S. newspapers would have to commit more space to
international news and hire editors knowledgeable about the world to pull together
packages from wires; TV station would have to give up that crime story in the evening
news to make room for a longer foreign story; and networks would have to commit
correspondents, producers, crews, and time on their main news programs to keep up with
what the American audiences want to read and hear about.
Perhaps, now more than a year after the attacks, it is time to think about what
lesson all this has taught the American media system. In a cjr [formerly known as the
Columbia Journalism Review] article, Andrew Kohut (2002) argues that the attacks and
the ensuing responsibility placed on the media have re-emphasized the importance of the
public to the media. He says it shows that the "public's need to know trumps everything
else" (pg. 54).
The Decline of International News in the American Press
This unexpected and huge demand for information posed a large challenge for
newspapers, especially in America, because over time the amount of international news
in American newspapers had begun to decline. Newspaper studies showed that
international news coverage had dropped significantly in the past two decades
(Cunningham, 2001). Parks (2002) discusses a 2001 Newspaper Advertising Bureau
study that showed coverage of international news in newspapers in America before
September 11 was only at 2 percent, which was down from 10 percent in 1971. With the
closing of international bureaus, the newsholes in newspapers and magazines began
simply to be filled with more domestic and entertainment news. Therefore, the American
people were beginning to know less about international events and developments than
they had in previous years. Cunningham refers to Nina Burleigh, a former war
correspondent for Time, commenting that "the foreign news blackout means that the rest
of the world knows far more about America than we know about ourselves" (p. 110).
Some argue that this decline in international news had created a new isolationism
within the United States. Parks (2002) report that Celinda Lake, a Democratic pollster,
told a 1997 conference on the issue that the media only cover instability abroad "and
have made international involvement look very undesirable" (p. 56). Parks also report on
a Gallup poll that found those who described themselves as "hardly interested" in
international affairs went from 3 percent to 22 percent between the 1990 and 1998 studies
by the Chicago Council on Foreign Relations. It seems that after the Cold War ended, the
United States felt safe within its borders.
However, after the attacks, for papers to simply report on the "what" was not
enough for the public to truly understand what was happening, nor was it what the public
seemed to be saying it wanted. In a September 27-28, 2001, Newsweek poll, 15 percent of
those surveyed felt like their life would never return to normal (retrieved on January 3,
2003 from http://people-press.org/reports/print.php3?PageID=22). The public needed and
wanted to know more about the "why." Reporters and editors were faced with the
daunting challenge of filling in these gaps for their readers in a very quick manner.
Papers had to get reporters to the scenes of the events quickly and keep up with the up-to-
the-minute changes the U.S. government was making. As Zeiler and Allan (2002) put it
in the introduction to Journalism after September 11, throughuh it all, they scrambled to
provide breaking information, offset panic, and make sense of events that had devastated
most existing interpretive schema" (p. 3).
This coverage was not going to be any ordinary international news coverage for
the press. It would entail more than simply going to the White House for daily briefings
from the press secretary. It would entail hard news gathered from primary sources in a
very timely manner because audiences were looking for reasons and explanations to
questions that they had never had before. Parks (2002) argues that many American news
organizations began to play "catch-up" (p. 52).
But why was the American press playing catch-up? Parks (2002) suggests the
threat of an Islamic fundamentalist attack had been clear for years with the World Trade
Center attacks in 1993, Air Force housing in Saudi Arabia attacks in 1996, 1998 U.S.
embassy attacks in Kenya and Tanzania, and U.S.S. Cole attack in 2000. Shouldn't the
American people already have known about the "who" of these attacks? Parks suggests
the answer is "not really" because the previous attacks were only covered episodically
and had little investigative and followup reporting. Since there was little and mostly thin
reporting on the previous terrorist attacks, the "who" of the September 11 attacks was
quite unknown to an American audience.
Had the American media been failing their audience? Parks (2002) reports that
Bill Wheatley, vice president of NBC News, was self-critical about the situation. "We all
have done a good job since September 11," Wheatley says, "but I and a lot of others wish
we had done more to help the public understand the intensity of feelings, the anger,
among the radical Islamic fundamentalists" (p. 53). Parks also reports that Edward Seton,
former president of the American Society of Newspaper Editors, said there is no question
that they failed their readers.
Parks (2002) argues that the media failed their audiences by not acknowledging
that, even before the attacks, the American people saw the threat of global terrorism as
the country's greatest concern. Parks (2002) mentions polls that reported in the late 1990s
that American people thought the 21st century would be even bloodier than the 20th and
that protecting the United States from foreign terrorist attacks should be a top priority.
"Simply put, most news organizations failed to cover what a substantial number of their
readers and viewers believed was vitally important the danger posed to the United
States by global terrorism" (p. 53).
American and French Press
If the American media had been failing the public in covering international news
for at least a decade, how different would the American newspaper coverage of the
attacks be from the French newspaper coverage? Would they not report on the same exact
issues since this was actually an "attack on the West?" How would these two seemingly
different countries portray, or frame, the same event to their newspaper readers?
It is without question that each world power has at least one "elite" newspaper
that stands out from the rest. According to Pool (1970), these newspapers are "usually
semiofficial, always intimate with the government ... read by public officials, journalists,
scholars, and business leaders" (p. 62). Pool categorizes Le Monde as "semiofficial." He
acknowledges that The New York Times often doesn't follow the line of the government,
yet says it still retains some of the characteristics of semiofficial newspapers. Concerning
Le Monde, Salinger (1982) notes that "[t]he role of General de Gaulle in the founding of
the newspaper has not been without effect over the years in the treatment of the news"
(pp. 82-83). The differences between The New York Times and Le Monde may not be
blatantly obvious, but Salinger makes the point:
For an American reader, Le Monde (and for that matter the rest of the French press)
differs from what one is accustomed to reading in the written media in the United
States. American newspapers make an effort, mostly successful, to separate opinion
and information. Opinion is labeled as such, often confined to special pages of the
newspaper.... The rest of the newspaper is devoted to information. ... American
newspapers stress pure information. (p. 83)
The French press is more interested in analyzing the news, while the American
press is more interested in objectively presenting the news. But articles on the attacks
were not the typical news story for either press system, so would the attacks produce
typical analytical pieces from the French and typical objective news from the Americans?
While the French press has been noted for being a "government feeler" or an "unofficial
expression of policy" (Pool 1970, pp. 65-66), would the attacks create a similar air in the
American press system? This study attempts to discern the differences in the use of
frames in The New York Times and Le Monde in the reporting after the attacks of
The following terms are used within the present study. The explanations that follow
come from a variety of sources.
Bureau "A news office away from the main newsroom of a newspaper or wire
service" (Harrigan, p. 407).
Newshole "Space left for news and editorial matter after ads have been placed on
pages" (Baskette et al., p. 435).
Frame The parameters within which an author sets or "frames" a story. For the
purposes of the present study, there are five main frames that newspaper reporters use
when writing stories: conflict, attribution of responsibility, morality, economic
consequences, and human interest (Semetko and Valkenburg, 2000).
Elite newspaper A newspaper aimed at a specific elite audience, which usually
is better educated and has a greater interest in public affairs than the readers of the
popular press (Merrill, 1968).
News story An article that is about a recent event that has been designated as
news by the editorial staff. Usually, a news story contains mostly objective facts and
Editorial ... [T]he unsigned, staff-written statement that runs on the editorial
page, stating the newspaper's official position on issues" (Harrigan, 1993, p. 411).
Column "A piece of writing ... that strongly shows the writer's opinion or
personal style" (Harrigan, 1997, p. 408).
Broadsheet "Term used to describe a full-sized newspaper page as opposed to a
tabloid" (Burkette et al (Eds.), 1997, p. 427).
Source "A person or document that provides information for a story" (Harrigan,
1993 p. 422).
Justification for Study
This study will look at the differences between the elite press of the United States
and of France with respect to their traditional roles in society. More specifically, the
study will investigate the difference between two elite newspapers in regards to the use of
media frames used when reporting soon after the attacks of September 11. The study will
look at the elite press of these two countries to content analyze the press coverage of the
September 11 attacks on the United States. The results of the study will add to the general
body of knowledge about the press system within the two countries. The comparisons and
contrasts made will help researchers understand the press systems and how they operate
in crisis situations. The coverage of the attacks has been used for the study to help put the
comparisons and contrasts on the same level, and the event was chosen because it was an
event that had international effects; therefore international newspapers would cover it.
Using the same event to do the research also helps to keep the results balanced.
The study remains balanced in the sense that the event being reported on was a
global event having global consequences. Even though the attacks were only in one
country, the entire globe was affected economically, politically, emotionally, religiously,
and in countless other ways. The events continue to have a global impact even today, one
and a half years later. In France, a new Muslim Council has been formed in hopes of
creating better relations between the country's five million Muslims and the French
government. In the United States, a new Department of Homeland Security has been
formed. These are just two examples of ways each country in the study continues to be
affected by the attacks.
REVIEW OF LITERATURE
Newspaper analysis is nothing new to the field of social science. Social scientists
have come up with varying ways of doing this analysis. In the 1950s, Erving Goffman
elaborated the concept of framing as a way to study interactions. Since then, framing has
been used in a number of studies on newspapers, and more generally media, coverage of
events in the news.
This particular study deals with news articles on the terrorist attacks of September
11. This study does not include analysis of any photos, or illustrations regarding the
event. These types of information were not analyzed because the purpose of the present
study was to analyze only text printed on the attacks.
How a reporter approaches an event and puts it into words has become commonly
referred to as how the reporter has "framed" the story. Framing has become increasingly
important in mass communication research. Goffman (1974) defines framing as a way of
organization that will govern (social) events, and it has often been linked together with
agendasetting in the theoretical world of communication studies. Tuchman (1978)
suggests that "mass media actively set the frames of reference that readers or viewers use
to interpret and discuss public events" (p. xi). In other words, instead of telling readers
what to think about as agenda setting does, frames go a step further and tell readers how,
or in what terms, to think about the issue. A frame gives a reader a specific way in which
to look at a reality taken from many different ways in which to look at that same reality.
Over the years of research, two concepts of framing can be identified: media
frames and individual frames. Kinder and Sanders (1990) suggest that media frames are
rooted in political discourse and individual frames are structures made within the mind.
Gamson and Modigliani (1987) define a media frame as "a central organizing idea or
story line that provides meaning to an unfolding strip of events ... the essence of the
issue" (p. 143).
Entman (1993) says frames essentially involve selection and salience. In other
[T]o frame is to select some aspects of a perceived reality and make them more
salient in a communicating text, in such a way as to promote a particular problem
definition, causal interpretation, moral evaluation, and/or treatment
recommendation for the item described. (p. 52)
He adds that frames diagnose, evaluate, and prescribe an idea or concept. With
respect to these functions, Entman (1993) says frames define problems by deciding who
is doing what with what causes, usually measured in terms of cultural norms. They also
diagnose the cause of these defined problems, make moral judgments about the causal
agents of these problems, and suggest remedies to these problems. Frames can be
presented in a single sentence or throughout an entire text, and they may not include all
of the aforementioned functions.
Frames also have at least four locations in the communication process, according
to Entman (1993). The communicators make framing judgments, whether conscious or
unconscious, when deciding what to say. The text has frames that are emphasized or de-
emphasized by the presence or absence of certain words, phrases, images, and sources.
The receivers' thinking and/or conclusions may or may not reflect the frames in the text
and the communicator's intention. Additionally, culture is the stock of commonly used
frames. "Framing in all four locations includes similar functions: selection and
highlighting, and use of the highlighted elements to construct an argument about
problems and their causation, evaluation, and/or solution" (p. 53).
But the question remains: How do frames work? Entman (1993) argues they
elevate some bit of information's salience. He defines salience as "making a piece of
information more noticeable, meaningful, or memorable to audiences" (p. 53). Therefore,
when the salience of an issue is increased in a certain manner, it increases the probability
that the audience will remember the issue in that certain way. Entman argues that even a
small appearance of a frame can be important if it works with the receiver's pre-existing
belief systems. Even the choice of words is important when relating to frames. Kahneman
and Tversky (1984) experimented with word usage in a survey and found that frames
determine whether most people notice and how they understand and remember a
problem, as well as how they evaluate and choose to act in regards to that problem. They
also found that the exclusion of other ideas or frames was just as significant as inclusion
of ideas or frames. The frame used gives meaning to the event or issue being reported on.
It's the giving of meaning to the event that has been turned into the five major
media frames looked at in research: conflict, human interest, economic consequences,
morality, responsibility. Semetko and Valkenburg (2000) suggest "a reliable set of
content analytic indicators is necessary for studying developments in the news over time
and similarities and differences in the ways in which politics and other topics of national
and international importance are framed in the news in different countries" (p. 94). In
other words, these five categories are important because they help researchers study
communication processes over time in a comparable fashion.
Semetko and Valkenburg (2000) also mention that other research and literature
dealing with content analysis of the nature of the news in the United States and Europe
has confirmed that the conflict and attribution of responsibility frames were most
commonly used, which also adds to the ability to compare communication across national
boundaries. The ability to compare across national boundaries also permits the present
study to use this deductive approach of coding articles into one particular frame category
using the five aforementioned frames. A brief explanation of them follows.
The conflict frame "emphasizes conflict between individuals, groups, or
institutions as a means of capturing audience interest" (p. 95). The human-interest frame
emphasizes the human side of a story in an "effort to personalize the news.., in order to
retain audience interest" (p. 96). The economic consequences frame puts the issue in
terms of the future economic consequences the event will have on groups and/or
individuals. The morality frame "puts the event ... in the context of religious tenets or
moral prescriptions" (p. 96). Finally, the attribution of responsibility frame tells the story
in terms of who or what was responsible for the issue.
This division into five categories of frames has helped communication researchers
do testable and re-testable research. Since the Shoemaker and Reese (1996) statement
that "news is a socially created product, not a reflection of an objective reality" (p. 21),
we have agreed that there is some subjectivity in the news. Norris (1995) argues that
journalists often work within news frames to simplify, prioritize, and structure the flow of
their stories. Entman (1991) argues that these routines journalist go through when
producing their product create the frames from which the public draws its opinion on
social or political movements of the day; hence, the power of the news frame. Entman
even goes as far as saying that news organizations shape their reports to get positive
reactions from their publics, and these anticipated reactions affect political elites, who are
the main "sponsors" of the news frames. The only difference comes when the news event
is something "breaking" and new. Then, the journalists frame the issues on their own.
When something is breaking and new, journalists also use their elite sources "to make
frame-confirming data more salient in the news text and to de-emphasize contradictory
data" (Entman 1991, p. 8).
It's because these frames appear to be "natural" that they are often difficult to
detect. Entman (1991) goes on to say that "frames reside in the specific properties of the
news narrative that encourage those perceiving and thinking about events to develop
particular understandings of them" (p. 7). The understanding that the public has of these
frames is implemented in their minds through the continued use of the frame. Entman
argues that through the repeating and reinforcing of words that are in reference to some
ideas and not others, "frames work to make some ideas more salient in the text, others
less so and others entirely invisible" (p. 7). This reinforcement of the frame makes the
idea easily identified, understood, and remembered in the minds of the public. But, it's
not important that everyone identify, understand, or remember the story in the same way,
it's only important that a significant majority do so, according to Entman. And media
professionals can enlarge a frame so much that it penetrates the consciousness of the
mass public, or they can shrink the frame so that the public is only minimally aware of
the issue or event. News frames are embodied in "key words, metaphors, concepts,
symbols, and visual images emphasized in a news narrative;" (p. 7) therefore, it's
important to note that the news media can play a powerful role in determining the success
of failure of certain social and political movements, such as the "War on Terrorism."
However, as Entman (1993) states, "whatever the specific use, the concept of
framing consistently offers a way to describe the power of a communicating text" (p. 51).
This power is important to remember when researching political subject matter, such as
the "War on Terrorism." Entman says this power of the text is important because frames
call attention to certain aspects of reality and also obscure certain aspects of reality.
Therefore, he argues, political elites fight over the use of certain frames with each other
and with journalists. In other words, the frame in a news text reflects the power that
dominated the fight. In the same vein, Gameson (1992) argues that a frame can have
social power when used with a widely accepted term, so that to not use that term is to risk
losing target audiences. For example, the American press has picked up on the term "War
on Terrorism." If journalists stopped using that term to describe activities of American
troops in Afghanistan, they might lose some readers because those readers are only
interested in wars and/or terrorism or because some readers might consider the use of
another term as being anti-American.
The five frames defined by Semetko and Valkenburg (2000) can be loosely
related to the idea of news values, which are taught to students of j journalism. According
to Mencher (2000), at least three-fourths of all news stories fall into one of the following
categories of news values: conflict; timeliness; impact, consequences, importance;
proximity to readers; the unusual nature of the event; and currency or "the sudden interest
people have in an ongoing situation" (p. 68). When reporters choose a subject for an
article, they usually are concerned with whether the topic has one of these values. The
five frames defined by Semetko and Valkenburg can help reporters to decide on a way in
which to present their chosen topic that will appeal to their readers. For example, if a
reporter decides the news event has the news value of impact (news value), then they may
decide to frame their article in terms of the economic consequences (frame) of the news
event. Additionally, if the reporter decides the news value of the event is the conflict it is
causing (news value), the reporter may decide to frame the article in terms of the conflict
going on (frame). Similarly, if the reporter determines that the news event has the news
value of being unusual in nature, then the reporter may decide to frame the article in
terms of human interest (frame). There is no real match for the news values of timeliness,
proximity, and currency. And there is no real match for the frames of attribution of
responsibility or morality. The present study will look to find which of the five frames are
used most often.
The Elite Press
All around the world, in every country, some aspect of the press has emerged into
what is now called the "elite press," newspapers read by the elites of their countries and
by the elites of other countries as well. Merrill (1968) suggests that these elite
newspapers express a "significant segment of international elite opinion" (p. 11). This
elite press is aimed at a specific elite audience, which usually is better educated and has a
greater interest in public affairs than do the readers of the popular press. "Through the
elite press is disseminated either the thoughtful, pluralistic, and sophisticated dialogue of
a free society, or the necessary social and political guidance of the closed society" (p. 11).
Therefore, Merrill (1968) puts the world elite press into two main categories: the free
press of an open society, and the restricted or managed press of a closed society. This
study deals only with the former.
The elite press is referred to by different names in different parts of the world, but
researchers often use the elite press, by whatever name, for the purpose of media study.
Merrill (1968) says these papers "open minds and stimulate discussion and intelligent
reflection" (p. 16). Perhaps it is for this reason, and that these elite newspapers impact
government policy or are a reflection of that policy, that these newspapers are used often
in studies. Merrill also makes the distinction between a quality newspaper and a prestige
newspaper. The difference lies in the prestige papers being what Merrill calls "kept
organs" of the state, while a quality paper is a "courageous, news-views-oriented journal,
published in an open society" (p. 15).
The present study concerns two quality papers, The New York Times of the United
States and Le Monde of France. Schramm (1959) differentiates between these two elite
papers, calling Le Monde an "analytical" newspaper and The New York Times a "news-
oriented" newspaper (p. 5). Merrill (1968) suggests taking these differences into
consideration when comparing or studying the two, but he also adds, "regardless of the
differences among the elite newspapers, they are all serious, concerned, intelligent, and
articulate" (p. 13).
In context with the present study, it's important to note one way that Merrill
(1968) describes the elite, quality press:
These are the papers whether they be dailies or weeklies, specialized or general,
large or small that offer hope to the world. They are the reasonable journals,
freely and courageously speaking out calmly above the din of party politics and
nationalistic drum-beating. They are urging peoples to work together for the good
of all, to consider all sides of complex issues, to refrain from emotional decisions,
to cherish that which has proved good and discard that which has been detrimental,
to consider seriously the basic issues and problems that confront mankind. (p. 16)
Description of Newspapers Used for this Study
In studies of the media of countries, researchers often use the same media from
study to study. In studies of U.S. broadcast media, the three major broadcasters are
almost always used (CBS, NBC, ABC). The same is true of studies of the print media,
with The New York Times being used often. The present study is no exception. The
researcher has chosen for study one elite newspaper from the United States and one from
France The New York Times and Le Monde.
The newspaper began in 1944 after the Nazi occupation of Paris ended. The
government under Charles de Gaulle called for a newspaper that would be respected at
home and abroad. It was built on the back of the prewar newspaper Le Temps, which was
considered one of Europe's best newspapers before World War II. It only took about a
year, according to Merrill (1968), for Le Monde to gain an international reputation and
circulation for an "intelligent, well-educated, and liberal audience" (p. 191).
There had been criticism about the loss of culture in France. The debate, as
discussed in Silverman's (1999) Facing Postmodernity, peaked just after World War II.
This debate revolved around:
the decline of a notion of culture founded on the intellect, solitary reflection,
meaning and a concept of 1'esprit and [on the other hand] the emergence of an anti-
intellectualist version of culture founded on an easy hedonism and instant
gratification of the senses; the elevation of mass and popular culture forms
(television, rock music, fashion, and so on) to the same status as classical culture;
the connivance of education in this debasement, or "dumbing down", of culture so
that the pedagogic, social and national function of the school ... is jettisoned in
favour of an approach which simply indulges the individualistic whims and desires
of young people .... (pp. 98-99)
It is in response to this debate that Le Monde began and continues to operate. The
paper was founded on the ideas of intellectualism, and it continues to operate on the basis
of analysis and objectivity in the sense of basing stories on empirical evidence.
The analysis is what sets Le Monde apart from many other Western elite
newspapers. Merrill (1968) mentions that while the straight news reports may be short,
the background and interpretation "wander through columns and columns of small type"
(p. 188). Merrill goes on to add, "Le Monde ... has an uncanny ability to foresee
developments, to predict, and to offer reasons, often days or weeks before headlines burst
out with news stories" (p. 188).
However, similar to what has happened to the media in the United States,
Silverman (1999) argues, the media and state of France have given up on trying to mold
French citizens according to a national ideal of homogeneity and high culture. Instead, he
writes, they are both pushed on by demand and consumer satisfaction. To steer clear of
being pushed in the direction bending to the demands of consumer satisfaction, Le Monde
has never affixed itself to one particular party or another. As Merrill (1968) states,
"giving a political label to Le Monde is made more difficult by the great diversity of
opinions and tendencies of the paper's collaborators or contributors" (p. 192). In general,
though, Le Monde is considered liberal, left-of-center, and internationalist, often being
recognized for its coverage and background reporting of international issues. It is also a
pacifist paper, calling for rational discussion and arbitration.
Le Monde keeps this non-label label by using a mix of news and editorial-style
writing in the newspaper. According to Salinger (1982), "Journalists (in France) are
encouraged to mix information and opinion, and those who do not have a point of view
are considered dull or unreadable" (p. 84).
It is without a doubt that Le Monde is valuable to the world. Perhaps it is best
stated as Merrill (1968) says one French newsweekly put it, "Le Monde ... represents]
today one of the last lighthouses that light for France the road of intellectual courage,
sternness of spirit and of political morality" (p. 195).
In October 2001, Le Monde voted to go public, according to the economic
newspaper The Daily Deal. The initial public offering of 20 to 25 percent of the company
was planned to aid editor-director Jean-Marie Colombani with his plans to enlarge the
main paper and create stakes in provincial papers. Colombani is credited with bringing
the paper back to life after its decline in the early 1990s. He helped restore circulation to
more than 500,000, and he helped Le Monde turn a profit for the first time in four years.
The paper's physical size is somewhat similar to that of a tabloid. It is 13 by 20
inches in size, and the front pages have few pictures larger than 2.5 inches in width. The
inside is mostly text, with sparse graphics and photos. Often, the main image on the front
page is an editorial cartoon printed in full color.
The New York Times
Started in September 1851, the New York Daily Times was published for a city of
half a million. It was a broadsheet four-page newspaper and sold for one cent a copy.
Henry Raymond was the editor, and he was determined to make the paper appeal to the
highly intelligent, who might be reading Horace Greeley's Tribune. But the paper added
something more than just a moral, conservative side to stories; the paper was more of a
nei' %-paper, presenting the reader a well-balanced and heavy diet of news especially
foreign news" (Merrill 1968, p. 270).
In the late 1800s, the paper was falling into decline. It wasn't until Adolph Ochs
came from the Tennessee hills, bringing with him "intuitive business sense, faith, and
imagination," (Merrill 1968, p. 270) that the paper began to rebound. Ochs coined and
put on the front page the famous slogan, "All the News That's Fit to Print." In 1898, he
cut the price of the paper back to one cent from the three cents it had been. Within three
years, the paper was back and had a circulation of 102,000.
Now, with a circulation of more than 1 million, according to the U.S. newspaper
trade journal Editor & Publisher, The New York Times is the United States' third highest
circulation paper in the country. According to the Web site http://www.infoplease.com,
the Times is also the world's 42nd highest circulation newspaper. Merrill (1968) claims
that "one can always expect to find a copy of the Times in leading libraries and
governmental offices throughout the world" (p. 263). He continues by stating that the
newspaper is more than a national paper; it has become an international one.
When comparing the Times to other elite world papers, Merrill (1968) argues, "it
is not as careful in typography as dailies such as Pravda ... not as tediously thorough in
certain stories or as well-documented as Le Monde ... but it goes further in combining the
worthy characteristics of all these great papers than any other single daily in the world"
(p. 264). In addition, the Times' international reporting has been considered one of its
strongest areas. In the 1920s, Ochs determined to make the Times'foreign coverage the
best in the world, and his many successors have carried this dream with them.
As with Le Monde, many find it hard to categorize the Times. Merrill (1968) says
it actually defies classification. "It is a kind of composite of all newspapers, aiming to
some degree at all audiences, except, perhaps, those seeking the lurid and sensational
journalism sought by readers of such papers as New York's Daily News ... (p. 266).
One thing that can be said for sure, according to Merrill (1968), is that the paper is
thorough, and this characteristic warrants the label "The Paper of Record" that it is so
Using the concept of framing, this study will attempt to decipher what, if any,
differences separate the elite press of the United States and France. The study will look
specifically at frames used and sources used. Results are expected to show a variety of
differences and similarities that should further the understanding of the press systems of
the United States and France.
This study will look at the frames used to present the news of the September 11
terrorist attacks on the World Trade Centers and the Pentagon. The study will also look at
the use of sources in articles.
The first variable will be frames. Comparing the use of frames between the two
papers will allow for study of the differences between the two papers and show
reflections of national culture portrayed within the newspapers.
The second variable will be source. Comparing sources between the two papers
will allow for further examination into the differences between these papers, perhaps
related to the newspapers' relationship with and to the national government. Sources will
be looked at in terms of whether or not they are elite or non-elite and what their
relationship is in relation to the event.
The following hypotheses will be tested based on the literature review:
H1) The attribution of responsibility frame will be used more in The New York
As the September 11 attacks happened in the United States, and more specifically
in New York, the Times will be more likely to report on who or what was responsible for
the attacks. The readers of Le Monde will not feel the need to constantly be reminded
who or what was responsible.
H2) The human-interest frame will also be used more frequently in The New York
Because so many of the victims of the attacks were from New York, the human-
interest frame will be more prevalent in the Times. The readers of The New York Times
will want to be able to put actual names and faces with the victims, whereas the French
readers will not be as concerned with names and faces.
H3) The morality frame will be used more often in Le Monde.
Because the French media typically seems less concerned with the traditional
American who, what, when, where, and why, the French press will have more leeway
with analyzing the morality of the case. In addition to this, the Times will not have had
time to analyze a moral aspect of the attacks in the time frame being studied. They would
have to wait until all the "who, what, when, where, and why" questions had been
answered for their audience before they could move on to analyzing, and 10 days is just
not enough time for those questions to be completely answered. Additionally, according
to Salinger (1982), "reporters for Le Monde have a tendency to moralize ... (p. 84).
H4) The economic consequences frame will also be used more often in Le Monde.
Similarly to H4, the French press will have been able to devote more time to
thinking about and analyzing the economic consequences of the attacks because they did
not have to spend as much time on the human-interest side.
H5) The New York Times will have more articles devoted to the conflict frame.
Because American newspapers were being forced to spend a great deal of time
and space on getting their readers up to date on who and what was involved in the
attacks, they would have to entice their readers in to the huge number of articles printed
by using the conflict frame. The newspapers will continue with the conflict frame to keep
their readers' attention. The Times was inundated with articles on the event, so to keep
readers' attention, the newspaper will use conflict as a point of entry into each article for
H6a) Le Monde will have a higher number of articles sourced by governmental
officials, whether French, American, or international officials.
H6b) Le Monde will have a higher number or articles sourced by elites than The
New York Times.
Herman and Chomsky's (1998) third filter argues that the media rely on
government experts. Their third filter also argues that since the media claim to be
objective, they need material that can be made to seem accurate. Yet, Herman and
Chomsky argue that another reason for governmental sources being used so often is
because they keep costs down since little investigating needs to be done. Gans (1979)
adds that news sources often represent the hierarchy of society, and as Le Monde is more
clearly for the elite of France than the Times is for the elite of the United States, it will
have more "elite" sources in its articles.
Merrill (1968) calls Le Monde "the most unremorselessly intellectual of the
world's elite newspapers" (p. 187). He adds, "Le Monde likes to keep its pages,
headlines, and type small and its ideas large" (p. 187). Concentrating on world news and
commentary, Le Monde supplements the basic details of their stories with "weighty
political and economical analysis" (p. 187). To hold that weight, this political and
economic analysis needs to be sourced by governmental source.
This study used content analysis as a means of data collection. Content analysis is
a research method that uses "a set of procedures to make valid inferences from text"
(Weber 1985, p. 9). These inferences are about the message itself, the sender of the
message, or the audience of the message. The present study compares the sender of the
messages, newspapers, and the messages they sent. As Budd et al (1967) suggest,
communicators respond and handle their messages about news events in different ways.
Some may handle the reporting of a news event by reporting on it daily on the front page
of the newspaper. Others may only report on the news event once and place the article on
page three. Some may present the news on the event with color photographs or graphics,
while others may not. Some produce long messages, some short messages, and some
don't produce at all. There is a growing need to analyze these messages and to examine
the factors in the environment in which these messages were created, and content analysis
has become an effective research tool for doing so.
There are many purposes for content analysis research. Pool et al (1970) argue
content analysis provides society with a "mirror to itself and a way of observing the
external environment in which it lives" (p. xi). Weber (1985) notes reasons for using
content analysis, including disclosing international differences in communication content,
identifying the intentions and other characteristics of the communicator, detecting the
existence of propaganda, and describing trends in communication content. The present
study is concerned with the first use listed.
Weber (1985) also notes that an important use of content analysis is the
"generation of cultural indicators that point to the state of beliefs, values, ideologies, or
other culture systems" (p. 10). Additionally, the present research probes how the
concerns of one society differ from those of another.
Weber (1985) also lists several advantages to content analysis. Some are relevant
to the present study: Communication is the central form of social interaction, and content
analysis works directly with the transcripts of human communications; the best content
analysis studies use both quantitative and qualitative measures; therefore, content
analysis can use what is usually thought of as antithetical modes of analysis; compared
with other forms of research, content analysis usually yields unobtrusive measures that
neither the sender of the message nor the receiver of the message know are being
measured, therefore leaving little room for obtrusiveness in the measurements.
Additionally, Kerlinger (1964) calls content analysis "a method of observation [that]
takes the communications that people have produced and asks questions of the
communications" (p. 544). This format usually allows the researcher to work without fear
that the attention will bias the communicator.
A central purpose of content analysis is the classifying of many words into
smaller groups or categories. That is what the present study set out to do. Weber (1985)
contends that for the investigator to make valid inferences from the text, it is important
for the classification procedure used to be reliable in terms of being consistent, but there
is no one "right way" to do content analysis.
The complicated part to content analysis, according to Weber (1985), comes after
the job is done. The finished product raises many questions often found in other forms of
research: What do the results mean? Are there competing interpretations? How do we
decide whether the interpretation is in some sense correct? Unfortunately, there are no
authoritative answers to these questions.
Quantitative Content Analysis
The main objective of the present study was to use quantitative data gained from a
content analysis of two newspapers to discover if stereotypical national press patterns
remain true in the reporting of the attacks of September 11. Using statistical analysis,
quantitative analysis allows researchers to reduce huge amounts of data into a
comprehensible form and make inferences from the data. Quantitative content analysis is
defined by Babbie (2001) as a "numerical representation and manipulation of
observations for the purpose of describing and explaining the phenomena that those
observations reflect" (p. G8).
The present study coded articles in a ten-day time frame from The New York
Times and Le Monde according to certain variables. Those variables were frames, source
in relation to event, and use of elite or non-elite sources. This study took these coded
observations and manipulated them so as to be able to explain certain characteristics of
what was reported in the two newspapers during the ten-day period.
Qualitative Content Analysis
The present study attempted qualitative research on the same topic. Using the
basic coding results taken from this study, the researcher created qualitative
measurements to use in further research. Based on the five frames discussed below, the
researcher attempted to create new frames for use when coding articles reporting on an
international crisis such as the attacks of September 11. These new frames could be
termed "crisis frames."
Qualitative research, according to Babbie (2001), involves a continuing
interplay between data collection and theory" (p. 359). Qualitative research helps
confirm relationships among concepts. For example, in the present study, qualitative
research could confirm, perhaps, that the French press is more opinionated than the
American press. While the number of opinion pieces can be measured with quantitative
analysis, subtleties of the type or strength of the opinions can't be measured as readily
with quantitative analysis. The coding of new frames, arising from qualitative research,
would help uncover opinions that appear repeatedly within the present study. According
to Babbie, ... the aim of data analysis is the discovery of patterns among the data ..."
(p. 365), and the present study will attempt to discover new patterns.
The New York Times and Le Monde were chosen for study because they represent
the most prestigious newspapers in the United States and France. The articles selected for
analysis begin the day after the attacks and end ten days afterward, when President Bush
declared in his address to a joint meeting of Congress on September 21, "every region
now has a decision to make. Either you are with us or you are with the terrorists. From
this day forward, any nation that continues to harbor or support terrorism will be regarded
by the United States as a hostile regime" (The New York Times, 9/21/01).
The study looked at a census of articles printed in Le Monde that dealt directly
with the attacks, but it looked only at a sample of all articles from The New York Times
about the attacks during this time frame because of the sheer amount published in the
Times regarding the attacks. As Weber (1985) suggests, sampling is used for the sake of
economy. Only articles directly relating to the attacks were analyzed, and only text
material were analyzed. No photographs, editorial cartoons, drawings, or graphs were
Both papers have online databases available through Lexis/Nexis. After an initial
search in both papers, a secondary search was done to narrow down the articles for study.
The initial search terms used were "attaque" and "11 septembre" for Le Monde and
"attack" and "terror*" for The New York Times. The with "terror" was used in hopes of
finding all derivatives of the word, i.e. terrorism, terrorist, etc. These terms produced 68
and 242 articles respectively. After a thorough scanning, or reading of the headlines, of
those initial articles retrieved, a second search was done using "attentat" (bomb attack)
and "11 septembre" for Le Monde and "attack" and "terrorism" for The New York Times.
The secondary search produced 242 and 939 articles respectively, a much more thorough
census of articles printed regarding the attacks. For Le Monde, all but two articles that
were retrieved in the initial search were in the secondary search, and all but six of The
New York Times articles that were retrieved in the initial search were in the secondary
search. After a thorough scanning of those secondary articles retrieved and a deleting of
articles not directly related to the attacks, photos, and other non-related material, the total
number of articles was 176 for Le Monde and 846 for The New York Times.
Each article was regarded as the unit of analysis with the variables in each unit
being the newspaper, how each newspaper framed the events (conflict, attribution of
responsibility, morality, human interest, or economic consequences) and the types of
sources that were used.
While they did contain text, the photo stories, maps, graphs, editorial cartoons,
and charts that may have appeared were not coded because the text was only reporting on
the photos, maps, graphs, cartoons, or charts, and they were not whole articles. The
present study only deals with text articles that were found on the database.
Four main variables were coded in this study: newspaper, frames, source
eliteness, and source in relation to the event.
As established by Semetko and Valkenburg (2000), a set of five news frames was
used for analysis in this study. (See Table 3.1.) The first five paragraphs of each article
were read to determine which frame was used in the article. Just the beginnings of stories
were read because research has shown that a high percentage of readers quit reading after
the first five paragraphs (Bush, 1966). The frames looked for were: conflict, human
interest, economic, morality, and attribution of responsibility. As stated in chapter two,
the following is a brief description of each frame:
Table 3.1 Variable Definitions
Newspaper 1) The New York
2) Le Monde
Frame used 1) Conflict
2) Attribution of
5) Human Interest
Source 1) Elite
eliteness 2) Non-elite
1) A daily newspaper
based out of New York
City, NY, USA
2) A daily newspaper
based out of Paris,
Frames that emphasize:
1) conflict between
groups to grab readers'
2) who or what is
responsible for the
3) the economic
consequences of the
4) a moral aspect of the
5) a personal angle of
the news event.
1) An person or
information for a story
who is well-known and
respected in society as
an expert in some area
2) A person or
information for a story
that may not be well-
known by readers.
Articles derived from
the databases of
Lexis/Nexis for the time
period 9/12/01 to
1)The New York Times
using the search terms
2) Le Monde using the
search terms "attentat"
emphasized* see coding
manual, Appendix A.:
1) conflict between
groups involved in
events related to the
2) who or what was
responsible for the
3) the economic after
effects of the attacks.
4) a moral aspect of the
attacks and attackers.
5) a personal story
related to the attacks.
1) Sources such as rulers
of countries, presidents
of companies, those
considered experts in a
2) Sources that are
such as people standing
around on a street after
an event has taken
Table 3.1 Continued:
Source in 1) Domestic
relation to Government
event 2) Foreign
6) Business Owner
9) Victim's relative
Person or document:
1) Of the national
government in the
papers' home country.
2) Of the national
government from a
country other than that
of the papers'.
3) From a non-
4) From the police or
5) From a person
considered by the
newspapers' readers to
be an expert in the
topic that they are
being quoted about.
6) Who owns a
7) Who witnessed the
8) Who was a victim of
9) Who is related to
one of the victims of
10) Typical "man on
1) Someone such as the
president of the country.
2) Same as #1 except
from another country's
3) Someone from a
group such as the PLO.
4) Someone from a law
5) Someone such as an
6) Someone such as the
owner of a restaurant.
7) Someone who was an
eye-witness to one of
8) Someone who was
victimized in some way
from the attacks.
9) A relative of someone
who was victimized by
10) An ordinary person
who had no real
involvement with the
attacks in any way,
shape, or form.
The conflict frame "emphasizes conflict between individuals, groups, or
institutions as a means of capturing audience interest" (p. 95). An example of the conflict
frame used in an article would be a story about how Muslim-Americans were targets of
prejudice in New York City after the attacks.
The human-interest frame emphasizes the human side of a story in an "effort to
personalize the news ... in order to retain audience interest" (p. 96). A personal story of
one particular family's ways of coping with a family member's death in the attacks would
be an example of a human interest-framed article.
The economic consequences frame puts the issue in terms of the future economic
consequences the event will have on groups and/or individuals. An example of an
economic consequence-framed article would be an article about how the attacks were
going to affect the stock market when it re-opened.
The morality frame "puts the event ... in the context of religious tenets or moral
prescriptions" (p. 96). An article about the tension between the Muslim faith and
American society would be an example of a morality-framed article.
Finally, the attribution of responsibility tells the story in terms of who or what
was responsible for the issue. An article about figuring out who or what was responsible
for the attacks would be coded into the attribution of responsibility category.
Each article was placed in one and only one of these five frames in the coding
process. In the event that more than one frame was used in the article, the first frame used
was the one coded. As noted earlier, since newspaper readership research indicates that
only a small percent of audiences read past the first five paragraphs (Bush, 1966), coding
the first frame used by the article author is an appropriate course to take.
These five categories have aided media researchers in the process of test and re-
test research. Replication studies allow the results of one study to be retested by another
researcher. This repetition helps bolster the validity and reliability of each study. The
presence of similar previous research can aid each researcher in his or her own personal
research area. Since media professionals and researchers alike have agreed that there is
no objective reality (Shoemaker and Reese 1996), the acknowledgement of these five
frames has been able to further media research. In other words, since these five frames
have been used in previous research, the present study has a good foundation on which to
The framing data were used to test hypotheses one through five. Each of the
frames works differently in the two newspapers because of cultural differences between
the countries. For example, the American press attempts to be objective in reporting
news, so the morality frame is harder to get into a news story in the United States. The
French press often discusses subjective topics; therefore, a morality frame should appear
quite often within the pages of a French newspaper.
The next variable to be analyzed in the study was sources. Journalists use sources
to gather facts. Herman and Chomsky (1998) argue that there is a symbiotic relationship
between the media and powerful sources of information. They write that government and
corporate sources are used because they're recognizable, and therefore credible, since
they're of high status and prestige.
In this study, both human and organizational sources were coded, quoted or
paraphrased. And both opinion and factual pieces of information given by sources were
According to Herman and Chomsky's (1988) third filter, all government officials
and business people are considered elite sources. The elite sources considered in this
study were: government officials international and national; international organization
officials; other elite newspapers; presidents of companies; engineers; authors and/or
people considered experts in certain areas of study; and police. These groups or
individuals were considered elite in the present study because they were either directly
involved in the aftermath of the attacks, or they are perceived by the public as extremely
knowledgeable on the topic for which they are used as a source. For example, the police
were directly involved in the clean-up of the attacks, and engineers are considered
extremely knowledgeable on the topic of building construction by the public.
Those considered not elite were sources considered to be the typical "man on the
street" sources. Examples would be people who witnessed the attacks from the streets of
New York, employees of institutions housed inside the World Trade Center who may
have lost their job as a result, or attendees at a memorial service.
After the initial elite versus non-elite coding was done for the newspapers,
another round of coding was done with more specific types of sources in order to get
more detailed results. The values being looked for were called sources in relation to
event. The ten different categories were: sources from domestic government officials,
foreign government officials, NGOs, witnesses, victims, victim's relatives, law
enforcement, experts, non-experts, and business owners. It was only after the initial
coding of sources that these ten groups were noticed as those being used as sources
throughout the coverage in The New York Times and Le Monde.
Gans (1979) defines sources as the people reporters obtain news from by
interview or observation. Gans argues that while sources theoretically come from all
walks of life, they are often the elite and powerful who shape society, and "... their
recruitment and their access to journalists reflect the hierarchies of nation and society"
Sources were coded to test hypotheses six. Only the first source quoted or
paraphrased was coded for the purposes of this study. Research has shown that readers
only read the first five paragraphs of most articles; therefore, the coding of additional
sources in each story was foregone.
The statistics of data analysis were used to ensure that the results presented could
be trusted and used in further research. After the coding was finished, z-scores were run
on the data to find out the difference of proportions between newspapers from frames
used and sources in relation to event. The z-scores indicated whether or not the sample
results may be generalized to the whole population of articles, or whether any differences
may be attributed to random chance. A Mann-Whitney U test was run on the five frames
and source in relation to event to test for significance in their use in the two newspapers.
This test was run to show whether the two newspapers used frames in a significantly
different way overall. After it had been established that the newspapers did or did not
differ in the use of frames overall and source in relation to event overall, the z-scores for
the individual frames used and individual source in relation to event used were calculated
to show the significance of the differences between the two newspapers. A .05 level of
significance will be used for statistical tests.
For content analysis studies to be considered reliable, there are many things that
have to happen. It's important to make sure the statistics presented are both reliable and
valid. Babbie (2001) states that reliable statistics will show that the researcher's
"technique, applied repeatedly, yields the same results each time" (p. 140). The data
obtained must be consistent from one coder to the next. Each coder must code the text in
the same way.
Weber (1985) suggests three types of reliability that are pertinent to content
analysis: stability, reproducibility, and accuracy. "Stability refers to the extent to which
the results of content classification are invariant over time" (p. 17). Stability can be
assessed by having the same coder code the same content more than once. But stability is
the weakest form of reliability because only one person is coding. Reproducibility, often
called intercoder reliability, refers to the extent to which the same text, coded by different
coders, produces the same results. High intercoder reliability is a minimum standard for
content analysis (as opposed to stability) because it measures the consistency of shared
understandings or meanings. Accuracy is the strongest form of reliability and refers to the
extent to which the classification of text corresponds to a standard. However, researchers
seldom use accuracy as a measure of reliability because standard codings are rarely
established for texts.
For the purposes of this study, high intercoder reliability was strived for. It was
achieved by having another graduate student at the University of Florida code all the
articles along with the researcher. In order to procure accurate results in this process, the
second coder for Le Monde (after the researcher) was bilingual in French and English,
and the second coder for The New York Times was a graduate student in the College of
Journalism and Communications. As Riffe (1998) puts it in discussing reproducibility,
"Reliability in content analysis is defined as agreement among coders about categorizing
content" (p. 104). Therefore, the results of the different coders were checked for
agreement. Riffe suggested that a random sample of the population of coding logs be
tested for the level of agreement between the two coders. However, for the present study,
a thorough analysis of each article coded was done to assess the level of reliability in the
study. A minimum level of agreement of 80 percent was set for reliability. Since the use
of these five frames in content analysis is relatively new, this relatively low value was
As Budd et al (1967) suggests, there isn't much literature on validity in content
analysis. Riffe (1998) writes that" ... the social science notion of validity relates more
rigorously to procedures for obtaining information so that appropriate inferences and
interpretations may be made" (p. 135). Riffe believes that content analysis has the
potential to have very high external validity. Problems with internal validity lie in the fact
that content analysis can only show patterns. Content analysis cannot find the cause of
these patterns, per se. Weber (1985) adds that the procedure used must generate a
variable that is valid to the extent that it "measures or represents what the investigator
intends to measure" (p. 12).
Since ... scientific validation of research is necessary before that research can
have any broader meaning or importance ... (Riffe 1998, p. 145), the present study
seeks to have validity. In the present study, the research of Semetko and Valkenburg
(2000) has set a valid precedent for the use of variables measured. This previous research
enhances the measurement of validity of the present study. Riffe also says that the
external validity of research can be increased if the content being studied is important.
The present study explores content that is believed to be highly important, and the results
will help future researchers and the newspaper industry in making news judgments in the
The present study looked at the differences in reporting of the attacks of
September 11 in the newspapers The New York Times and Le Monde, employing framing
theory. Content analysis was done to compare differences in sources used and frames
used in these newspapers' articles on the attacks.
After all the coding of articles was complete, the data were entered into an SPSS
computer program file. From there, difference of proportions tests were run on the data.
Each hypothesis was tested and conclusions were made regarding the content of both
Overall data were gathered from the coding sheets. The data were put into three
different spreadsheets: one for The New York Times, one for Le Monde, and one for the
two newspapers combined.
More articles were printed on September 13 in Le Monde than any other day in
the ten-day period of observation. (See Table 4.1.) The sample from The New York Times
drew twenty articles from September 13 and 16 more than any of the other eight days in
the period. Therefore September 13 had the most articles overall in the present study than
any other day in the ten-day period under observation. There were no articles coded for
Le Monde on September 12. While two articles did appear on the search of Lexis/Nexis,
those two articles were not about the attacks. A hard copy of the newspaper was used to
double check that no articles were printed on September 12.
Table 4.1 Number of articles each day
Date The New York Times Percent Le Monde Percent
12 14 8.1 0 0
13 20 11.6 30 17.1
14 19 11.0 26 14.9
15 14 8.1 15 8.6
16 20 11.6 0 0
17 16 9.3 28 16.0
18 19 11.0 16 9.1
19 16 9.3 19 10.9
20 16 9.3 17 9.7
21 18 10.5 24 13.7
Total 172 99.8* 175 100.0
*percent differs from 100 due to rounding
The frame used most often overall was the conflict frame. (See Table 4.2.) Of the
347 articles coded, 109 of them were framed in terms of conflict; 49 of those came from
Le Monde and 60 came from The New York Times. The frame used least often overall
was the attribution of responsibility frame. Only 34 of the 347 articles used this frame.
Table 4.2 Frame usage
Frame The New York
1 Attribution of 11
2 Economic 30
3 Morality 19
4 Conflict 60
5 Human Interest 52
Percent Le Monde
*percent differs from
100 due to rounding
Both newspaper used elite sources the majority of the time. (See Table 4.3) Both
newspapers also had articles with no sources at all. Le Monde had 85.1 percent of the
articles with elite sources, while The New York Times had 68.8 percent of theirs.
Table 4.3 Source Eliteness
Source The New York Times Percent Le Monde Percent
Elite 118 68.6 149 85.1
Non- 40 23.3 16 9.1
none 14 8.1 10 5.7
Total 172 100.0 175 99.9*
*percent differs from 100 due to rounding
The plurality of the sources in both newspapers were from governmental sources,
whether domestic or foreign. (See Table 4.4.) Most of The New York Times government
sources were domestic, while most of Le Monde's were foreign. Le Monde only had one
article sourced by a victim or a victim's relative.
Table 4.4 Source in relation to event
Source The New York Percent Le Monde Percent
1 Domestic 43 25.0 10 5.7
2 Foreign Government 10 5.8 52 29.7
3 NGO 18 10.5 33 18.9
4 Witness 5 2.9 2 1.1
5 Victim 8 4.7 1 0.6
6 Victim's relative 6 3.5 1 0.6
7 Law Enforcement 5 2.9 1 0.6
8 Expert 30 17.4 46 26.3
9 Non-expert 19 11.0 13 7.4
10 Business Owner 14 8.1 6 3.4
None 14 8.1 10 5.7
Total 172 99.9* 175 100.0
*percent differs from 100 due to rounding
Results of Application of Method
The method chosen for the present study was content analysis. The content
analysis was done on a sample of newspaper articles from The New York Times and a
census of articles from Le Monde for the 10-day time period of September 12 through
September 21 that related directly to the attacks of September 11, 2001.
The initial search for articles in both newspapers, while successful in some
regards, led the researcher to refine the search methods. This refinement of search terms
on the Lexis/Nexis database resulted in a wider scope of articles from both newspapers
from which to code. There were a few unusual circumstances run into even after the
refinement of search terms was done.
The New York Times sample set included 18 sports articles. They were used in the
present study because the research design included all articles that directly related to the
attacks and their effects. These sports articles' topics ranged from how the athletes were
dealing with the attacks to funds being spent to increase security at ballparks. This is
noteworthy because Le Monde does not have a sports section in the paper; therefore,
there were no comparable stories coded in Le Monde.
A similar dilemma encountered was that Le Monde does not have editorials
written by the editors of the paper, as is found in U.S. newspapers. The journalists of the
newspaper offer opinions and editorial comments throughout what The New York Times
would categorize as hard news stories. This didn't pose too much of a true dilemma,
though, because the present study did include editorials from The New York Times in the
articles coded. The only resulting problem was that some of the articles in Le Monde did
not have any sources. However, this was somewhat counterbalanced by editorials from
The New York Times with no sources being included.
Another result of the use of the 10-day period was that there were no articles for
Le Monde on September 16. The newspaper does not print on Sundays; therefore, there
were no articles in the Lexis/Nexis database for that day.
Nature of Sample
The articles coded for Le Monde came from a search on Lexis/Nexis using the
search terms "attentat" and "11 septembre". From that list, articles were discarded that
did not directly relate to the attacks of September 11; that left 176 articles from the initial
242 found. For The New York Times, systematic random sampling was done to get a
sample of 176 articles. The researcher flipped a coin to decide which article from the
Lexis/Nexis search of The New York Times that fell within the first skip interval would be
the first article to be coded. The second item from the list was selected.
Of the 176 Le Monde articles initially chosen to be coded, one had to be discarded
for not directly relating to the research study. It was about the Israel/Palestinian conflict
and not the attacks in New York or Washington. Four New York Times articles had to be
discarded as well, resulting in a total of 175 Le Monde articles and 172 The New York
Times articles being coded. These four discarded articles were tables of contents lists for
the day's edition of the newspaper. All articles were then printed out in their entirety
from the Lexis/Nexis database in order to be coded.
Of the 172 The New York Times articles coded, fourteen had no sources. Nine of
those 14 were editorials. Sports articles accounted for four of the fourteen, and one was
what the Times called a "Notebook," consisting of first-hand reports from its reporters.
All but one of the ten Le Monde articles with no sources were essays written by Le
Monde reporters. The one that was not, was a simple description of the sequence of
events that took place during the attacks.
Two second coders were used to test the reliability of the coding for the two
newspapers. Once the researcher coded all the articles for both newspapers, a second
coder was chosen to code the same set of articles. Two different people were chosen
because the amount of articles needing to be coded was large and one set of articles was
written in French, so finding a person who was able to read and understand French was
difficult to come by. Both second coders were graduate students at the University of
Florida. The second coder for Le Monde was a native French speaker who was a graduate
student in Anthropology, while the second coder for The New York Times was a graduate
student in Journalism.
The results of analysis of the second coders' coding logs showed a 82 percent
agreement of frames used between coders for The New York Times and an 83 percent
agreement of frames used between coders for Le Monde. The reliability level was
considered satisfactory. There was 87 percent agreement for eliteness of source and
source used in relation to event between coders for The New York Times, and 98 percent
agreement for eliteness of source between coders for Le Monde. The second coder for Le
Monde did not code for source in relation to event as that variable was added after the
Descriptive Analysis and Hypotheses Results
After all the data were entered into SPSS, a Mann-Whitney U test was run to see
if there was any statistical significance in the overall use of frames and source in relation
to event by each newspaper. The test for frame used resulted in a significance of .000.
The test for source in relation to even resulted in a significance of .822. And the test for
eliteness of source resulted in a significance of .015. After these tests of significance were
run, tests of the individual variables were done to see if there was significance there. Z-
scores were used to test the significance of difference for each variable.
The z-scores for frame used were computed in order to test the hypotheses stated
in Chapter 2. The z-score for the attribution of responsibility frame was -3.68. The
negative result means that H1 was not supported. There were, contrary to the hypothesis,
more articles in Le Monde framed by using attribution of responsibility than there were in
The New York Times. To be more precise, there were 23 articles (13.1 percent) framed
using attribution of responsibility in Le Monde and only 11 (6.4) in The New York Times.
The difference in the use of the human-interest frame resulted in a z-score of 3.33.
The probability of this difference occurring because of chance is .00; therefore H2 was
supported by the data; The New York Times did use the human-interest frame
significantly more than Le Monde used it. There were 52 (30.2 percent) articles in The
New York Times using the human-interest frame and only 31 (17.7 percent) in Le Monde.
The test of difference for H3 ended up with a z-score of 1.33. Hypothesis three
stated Le Monde would use the morality frame significantly more than The New York
Times would use it. With a z-score of 1.33, the probability of H3 being true based merely
on chance is .09, which is larger than .05; therefore, H3 was not be supported by this
data. Although there were only nineteen morality-framed articles in The New York Times
and 29 in Le Monde, the probability of this difference occurring by chance is high enough
to be unable to conclude the results weren't due to random chance.
The z-score for frame two, economic consequences, was 2.33. H4 stated that Le
Monde would use the economic consequences frame significantly more than The New
York Times would use it. The probability of H4 being true based only on chance is .01.
Using .05 as the cutoff point for significance, H4 was supported by the data. There were
41 (23.4 percent) articles framed in terms of economic consequences of the attacks in Le
Monde, and there were 30 (17.4 percent) articles in The New York Times.
Finally, the use of the conflict frame in the two newspapers resulted in a z-score
of 1.89. H5 stated that The New York Times would use the conflict frame more than Le
Monde. The probability of this occurring only by chance is .03, which is smaller than .05;
therefore, H5 was supported by the data. There were 49 (28 percent) conflict-framed
articles in Le Monde and 60 (34.9 percent) in The New York Times.
The difference in use of governmental officials by the two newspapers resulted in
a z-score of 3.17. The probability of this difference occurring merely due to chance is .00;
therefore H6a was supported. Le Monde had 95 articles, 54.3 percent, sourced by
governmental sources, and the Times had 71 articles, 41.3 percent, sourced by
governmental sources. The difference in use of elite and non-elite sources by the two
newspapers resulted in a z-score of 4.29, which meant the probability of this difference
being due solely to chance was .00, or very unlikely. Therefore, H6b was supported by
the data. Le Monde had 149 articles sourced by elite sources (85.1 percent), and The New
York Times had 118 articles sourced by elite sources (68.6 percent).
After all the data were gathered and analyzed, they supported some, but not all, of
the hypotheses. More research needs to be done in a more detailed study to truly
understand the implications of the data results.
Results of Application of Method
The method of content analysis was a successful method to use for this study.
Analyzing the content of Le Monde and The New York Times using framing as the basis
of study worked well in the present study.
In the present study an even more specific theory of framing was used than the
traditional theory that the media frame issues in the news. The research of Semetko and
Valkenburg (2000) was the basis for using the five frame categories employed in the
coding of articles. These five categories worked fairly well in terms of categorizing the
American newspaper articles, but they were difficult to apply in some of the French
articles because some of the French articles often included more than one of the five
frames or used a frame not within the scope of the five given to choose from. For
example, an article on the French government's launching of a security plan was coded
into the "conflict" frame, but it could have also been thought of as being framed in terms
of economic consequences of the launching of the new plan. Although all the coders were
instructed to code the first frame used within the article, sometimes it was obvious that
more than one frame was going to be used within the article; therefore, it was difficult to
decide which one to enter on the coding sheet.
Additionally, some of The New York Times articles could have been coded into
more than one of the five frames given to choose from. For example, an article on the
baseball players not playing for a week was coded as being framed in terms of
"morality," but the article also discussed aspects of the economic consequences this week
off was having on the baseball industry. It was obvious from the beginning that the article
was going to discuss both aspects, but "morality" was chosen as the frame used because
that was the aspect taken on first in the article.
The Semetko and Valkenburg (2000) research was used in an attempt to make the
coding of the articles more valid, consistent, and reliable. The pre-existing frames of
Semetko and Valkenburg were extremely useful for this research.
The Lexis/Nexis computer database was also extremely useful. The program is
accessible and free to use for students at the University of Florida; therefore, it was also a
very cost-effective method of actual data gathering. Since the whole of the articles was
available online for free, the researcher did not have to pay for copies of each article from
microfiche. The ability to narrow down searches on a day-by-day basis in the program
was also very helpful. This narrow searching ability saved a great deal of time.
The decision to only code the beginning of each article was helpful in terms of
consumption of time. If the coders had been required to read each article the whole way
through, the actual coding of 347 articles would have been extremely time consuming.
Additionally, having two different second coders for the each newspaper saved time and
money. Had one person been chosen to second code both newspapers, the cost and time
element would have been much higher since one person would have had to code 374
articles as opposed to just 172 or 175. The second coder for The New York Times coded
the articles for free, and to have one person code 347 articles would have taken more
Using a ten-day period of study was a successful way to do content analysis in
both helpful in terms of making a decision on when to stop looking for articles relating to
the event and in terms of time consumption. If the time period had been much longer,
there would have been many more articles to code, and those additional articles may have
produced different results because they were published after President Bush made his
speech about being either with the United States or against the United States.
However, there really ended up being only eight days used in the analysis of Le
Monde for two reasons. No articles from September 12 garnered hits that were used with
the search terms used in Lexis/Nexis. Two articles from the 12th did appear in the results
of the search, but they did not relate to the attacks of September 11 in the United States.
One was about a trial of nationalists in France, and the other was about the attack on the
Taliban opposition leader Ahmed Shah Massoud. The second reason only eight days
worth of articles was used in the coding of Le Monde is that the newspaper does not
publish on Sundays. Therefore, there were no Le Monde articles coded for September 16.
Descriptive Analysis Discussion
The Mann-Whitney U test run on the use of frames in each newspaper resulted in
a finding of .000. Therefore, it can be stated that each newspaper used frames
significantly differently in the time period under investigation. Since the papers did, in
fact, use the frames differently, the validity of the research is supported. From this
finding, the validity of the researcher testing the difference in use of each frame on a
frame-by-frame basis was established. In other words, the hypotheses' suggestions of
differences between the two newspapers are warranted and should be tested. The Mann-
Whitney U test run on eliteness of sources resulted in a finding of .015, while the test on
source used in relation to event garnered a result of .822. Therefore, it can be said that the
two newspapers did use elite sources significantly differently, but they did not use source
in relation to event differently.
Hypothesis one, which stated there would be more articles framed using
attribution of responsibility in The New York Times, was not supported by the data. A
possible reason for Le Monde having more articles framed through attribution of
responsibility could be because the reporters at Le Monde thought they knew who was
responsible because they are experts in international subjects. Since many of Le Monde's
reporters have doctorates in given subjects, they often report on a topic using themselves
as authority figures on the topic. This could have been the case in this instance. Perhaps
many reporters at Le Monde felt they knew who was responsible, while reporters at The
New York Times felt they couldn't place any attribution of responsibility on any one
person or group until later in the investigations into the attacks. A day-by-day analysis of
the use of the attribution of responsibility frame shows that The New York Times only
used that frame five times in the first five days of reporting, while Le Monde used that
frame 10 times in the same amount of time.
Finding more articles framed by human interest in The New York Times than Le
Monde (H2) was no surprise. Because the attacks took place in the same city of the
newspaper's publication, the newspaper's reporters had easy access to people directly
affected by the attacks in some manner. It was much harder for Le Monde to get reporter
access to victims or first-hand witnesses of the attacks simply because there were fewer
Le Monde reporters in the whole of the United States.
Similarly, it was no surprise that there were more morality-framed articles in Le
Monde (H3). The newspaper is known for its opinion-laded pieces. Journalists in France
are often taught to include issues of morality in their articles as a way to keep the culture
of France vigorous, not trampled upon by other cultures. In contrast to that, American
journalists are taught to be objective, not including their own opinions or issues of
morality in their articles. However, the significance of the difference in the amount of
morality-framed articles between the two newspapers was not enough to support the
hypothesis. The difference in number could be due to The New York Times' reporters
feeling somewhat more emotionally attached to their articles on the attacks since the
attacks happened in New York. Perhaps there were more morality-framed articles in The
New York Times than there would have normally been if the attacks had happened in
another city. There is no real way of knowing the answer to this speculation, though.
The reason for there being more articles framed in terms of the economic
consequences in Le Monde than in The New York Times (H4) was a simple one of time.
Both newspapers would be expected to have articles framed in these terms, but because
other stories were more pressing, The New York Times simply had to hold off on some of
those articles until later after the attacks. Therefore, the articles would not be in the 10-
day time period used in the present study. Had the study included a larger time frame for
the analysis, the amount of articles framed in terms of economic consequences may have
been more equally matched between the two newspapers.
The game of "catch up" being played by The New York Times seems to be one
explanation for why they had more articles framed by conflict than in Le Monde (H5).
Because their readers did not know as much as Le Monde's readers about those
responsible for the attacks, the Times had to spend more time describing the actual
conflict between the United States and the followers of Ossama bin Laden. Therefore, it
was only natural that there be more articles in The New York Times framed in these terms
than in Le Monde.
That Le Monde had more articles sourced by elites than The New York Times did
is not a surprise either (H6a). The elite nature ofLe Monde could be an explanation for
this finding. It is to be expected that such an elite newspaper would use a higher
percentage of elite sources than most newspapers around the world. They have to use
these elite sources to keep up their elite nature. Additionally, The New York Times had
more articles that didn't necessarily call for elite sourcing. For example, The New York
Times had more human interest-framed articles than Le Monde, and those type of articles
do not need to be sourced by engineers or presidents of companies.
Finally, the fact that Le Monde used more elite and more governmental sources
than The New York Times was expected as well (H6b). The elite nature of Le Monde
could well be one explanation for this finding. As stated in Chapter Two, Le Monde
prides itself on being a paper of analyzation by elite writers. Additionally, the nature of
the actual articles in Le Monde called for more elite sources than those in The New York
Times. For example, Le Monde had more articles framed in terms of the economic
consequences of the attacks. Those types of articles called for sourcing from people such
as financial experts and governmental sources such as officials of the Federal Reserve.
In the post-test analysis of frames used in Le Monde and The New York Times,
some qualitative items were examined. Throughout the coding process, both coders for
each newspaper took notes about what they were finding. After the coding was complete,
a thorough scanning of those notes revealed some consistent items worth discussion.
It seemed Le Monde printed many articles framed in way that could have been
called "security" and many with frames that could have been called "cultural analysis."
For example, on September 13 an article was printed discussing the tightened airport and
national security in several different countries. This article didn't really fit in to any of the
five frames given to choose from. It was put in the "economic consequences" frame, but
it really didn't seem to be written with economics as the main topic for discussion. It was
framed more in terms of making sure the readers knew that the airports in Europe were
Similarly, The New York Times had articles that were simply descriptions of the
attacks. They were just minute-by-minute descriptions of what happened on September
11. There was no discussion of who was involved or why what was happening happened.
That article was coded into the "conflict" frame, but it really needed to be coded as
"description" since there was really no framing going on at all in any of the senses for
which categories were available.
Additional analysis also was done on the results that did not relate directly to the
hypotheses. Analysis was done for each frame used in each newspaper after the
immediacy of the attacks was over. In other words, which newspapers used each frame
more during the eight-day period of September 14 through September 21? Would the
results of the hypotheses testing be the same for the data after September 13?
There were among the articles originally analyzed, 145 articles printed in Le
Monde during this time frame and 138 in The New York Times. Using the same
hypotheses as those for the ten-day period as a whole, z-scores were calculated for each
frame used in the eight-day period after September 13.
The difference in use of the attribution of responsibility frame (H1) garnered a
negative z-score as it did in the testing of original hypotheses. Therefore, it seems as
though The New York Times did not use the attribution of responsibility frame
significantly more than Le Monde after the immediacy of the attacks was over. The
economic consequences frame had a z-score of 2.35, which was similar to the z-score for
the 10-day period. Hence, it can be said that Le Monde used the economic consequences
frame more than The New York Times did even after the immediacy of the attacks was
The z-score for the use of the morality frame after September 13 was 2.59.
Therefore, the probability of this occurring not due to random chance is .00, or less than 5
percent. This was not the case for the whole 10-day period. In the data for the ten-day
period, the z-score for the use of morality frame was not large enough to be able to say
the results were due to anything other than pure chance. However, after September 13,
the use of the morality frame by Le Monde was significantly different from the use of it
by The New York Times, so the hypothesis would have been supported after the 13th.
This finding could be so because perhaps Le Monde did not feel comfortable reporting on
the attacks in morality terms until after more information was discovered concerning who
was responsible for them.
The conflict frame's use in the two newspapers resulted in a z-score of 2.05. This
means that the probability of the differences being more than a matter of chance is at .02,
which is less than 5 percent. The New York Times used the conflict frame significantly
more than Le Monde used it.
The difference in use of the human-interest frame between the two newspapers
after September 13 resulted in a z-score of 3.25, which means the probability of this
difference being due to chance is .00, or very unlikely. The New York Times used the
human-interest frame significantly more than Le Monde did during this eight-day period.
This result was similar to that of the whole ten-day period of observation.
The same tests were run for the data found during only the first two days of the
study (September 12-13) to see how the hypotheses held up during the days immediately
following the attacks. There were 34 articles printed in The New York Times during this
two-day period and 30 printed in Le Monde.
The difference in use of the attribution of responsibility frame (H1) garnered a
negative z-score, which meant that H1 was also not supported during these first two days
of reporting on the event.
The z-score for the difference in use of the human interest frame (H2) between the
two newspapers during this two-day period was .24, which meant the probability of this
occurring from pure chance was .40. Therefore, H2 was not supported for the days
immediately following the attacks, although it was supported during the whole ten-day
period and during the eight days after the immediacy of the attacks was over.
The difference in use of the morality frame in the two newspapers garnered a z-
score of .30. This meant the chances of this difference being because of anything other
than chance was .38; therefore, H3 was not supported during the first two days after the
attacks. A negative z-score resulted from the use of the economic consequences and
conflict frames in the two newspapers; therefore, H4 and H5 were not supported.
When the hypotheses were tested for the two days immediately following the
attacks (September 12-13), none of them were supported. The difference in usage of all
five frames was not statistically different between the two newspapers. This sheds light
on the idea that perhaps when a major international crisis occurs, the major media around
the world turn to a universal form of reporting.
The difference in use of governmental sources in the two newspapers garnered a
z-score of .32, which meant the probability of this being due to anything other than
chance was .37. Therefore H6a was not supported by the data gathered immediately
following the attacks. The z-score for the difference in use of elite versus non-elite
sources in the two newspapers was 1.14; therefore, the probability of the difference being
due to anything but chance was only .08, and H6b was not supported for the two-day
period immediately following the attacks.
In the end, only H1 and H3 were not supported by the data for the entire ten-day
period. These two hypotheses dealt with the use of the attribution of responsibility frame
and morality frame respectively. H1 stated that The New York Times would have more
articles framed in terms of attribution of responsibility, and H3 stated that Le Monde
would have more articles framed in morality terms than the Times. When the data were
examined for the eight-day period following the immediacy of the attacks, however, H3
was supported. Both hypotheses dealing with the use of sources in each newspaper for
the entire period of study were supported. Le Monde used both more elite sources and
more governmental sources.
Summary of Hypotheses
Number of Cases
After first search 242 939
After discarding 175 172
Eight-day study 138 145
Two-day study 34 30
Mann-Whitney U test probabilities
Frame used: .000
Source in relation to event: .822
Eliteness of source: .015
The hypotheses are summarized below.
H1) The attribution of responsibility frame will be used more in The New York
10-day : z = -3.68; X not supported
8-day : z = -; X not supported
2-day : z = -: X not supported
H2) The human-interest frame will also be used more frequently in The New
10-day : z = 3.33; probability = .00; supported
8-day: z = 3.25; probability = .00; supported
2-day : z = .24; probability = .40; X not supported
H3) The morality frame will be used more often in Le Monde.
10-day : z = 1.33; probability = .09; X not supported
8-day : z = 2.59: probability = .00; Supported
2-day : z = .3: probability .38: X not supported
H4) The economic consequences frame will also be used more often in Le
10-day : z = 2.33 : probability = .01: Supported
8-day : z = 2.35 : probability = .00 : Supported
2-day : z = : X not supported
H5) The New York Times will have more articles devoted to the conflict frame.
10-day : z = 1.89 : probability = .03: Supported
8-day : z = 2.05 : probability = .02: Supported
2-day : z = : X not supported
H6a) Le Monde will have a higher number of articles sourced by governmental
officials, whether French, American, or international officials.
10-day : z = 3.17 : probability = .00 : Supported
8-day = z = 2.15 : probability = .01: Supported
2-day : z = .32: probability = .37: X not supported
H6b) Le Monde will have a higher number or articles sourced by elites than
The New York Times.
10-day: z = 4.29: probability = .00 : Supported
8-day : z = 4.37 : probability = .00 : Supported
2-day : z = 1.14 : probability = .08 : not supported
The content analysis of framing used in The New York Times and Le Monde
garnered results expected and results not expected. Overall, the hypotheses generated
from the literature were supported. The study was successful in that it lends itself to much
future research and can help the newspaper industry to see how they have reacted to crisis
situations in the past in order to figure out how to react in the future. By looking at the
present study, the newspaper industry can see what happened after September 11, and the
industry can decide whether it would like to do the same thing again or try something
The present study was a content analysis of The New York Times and Le Monde
for the ten-day period of September 12 through September 21, 2001. The analysis used
framing theory as the basis of study. More specifically, the study looked at the five media
frames defined by Semetko and Valkenburg (2000) in the articles in the ten-day period.
The topic of the attacks of September 11 was chosen because it was a recent event
that garnered massive media attention. It was obvious that there was going to be a large
number of newspaper articles to analyze on the topic. Additionally, previous research on
framing has been done using one topic to analyze in different newspapers as a means of
comparison (e.g., Entman, and Semetko and Valkenburg). This previous research helped
create the foundations of the present study and also helped make the results more
generalizable in that there could be more of an objective comparison between the two
newspapers than if the researcher had come up with frame categories by herself.
The articles used for coding in the present study were found using the computer
database of Lexis/Nexis. The search terms "attack" and "terrorism" were used when
searching in The New York Times and "attaque" and "11 septembre" for Le Monde. The
time frame of September 12 to September 21 was used when running the search, and the
search resulted in 938 hits in The New York Times and 242 hits for Le Monde. Although
the articles coded from Le Monde were a census of articles found on the Lexis/Nexis
database, there is no way of really knowing if the set was a true and accurate census of all
articles printed in the newspaper on the attacks during that time period.
After reading the headlines found in the search of both newspapers, 846 The New
York Times articles remained, and 176 Le Monde articles remained. Examples of items
discarded from the initial articles found in the search included things such as photo essays
in The New York Times and articles in Le Monde discussing movies having the word
"attaque" in the title.
During the coding process, one article was discarded from the Le Monde set of
articles. While the article did mention the attacks of September 11, it was not about the
attacks directly. It was an article about the Israel/Palestine conflict going on in Israel.
Four articles had to be discarded from the set of articles being coded for The New York
Times. The four articles were excluded because they were not actually articles; they were
lists of headlines that were played on the front page of the Times as a way of letting
readers know what was inside.
When the researcher finished the first set of coding, the articles were given over
to second coders, who were both graduate students at the University of Florida. Once the
second coders completed their coding, reliability was found to be more than 80 percent
between second coders and the researcher. After entering the data from the coding sheets
into the computer program SPSS, difference of proportions tests were done on the
variables of frame used, eliteness of sources, and source used in relation to event.
Six hypotheses were tested. The hypothesis stating that The New York Times
would have more articles framed in terms of attribution of responsibility (H1) and the
hypothesis stating Le Monde would have more articles framed in terms of morality (H3)
were not supported by the data. There are several possible explanations as to why the H1
was not supported; perhaps the most poignant of all is that this was an extreme crisis
situation. Perhaps under normal reporting conditions, the first hypothesis would have
been supported. However, H3 was supported when the data were looked at after
September 13. A possible reason why H3 was not supported in the ten-day period could
be that since the attacks did occur in New York, perhaps the reporters and editors felt
more comfortable framing their articles in terms of morality issues than they would under
In the analysis done for the first two days of reporting, the tests of significance of
differences resulted in different scores from that of the whole ten and eight-day periods of
analysis. The first hypothesis was not supported in any of the three sets of time. However,
H2, which stated that The New York Times would have more articles using the human-
interest frame, was not supported by the data from the first two days of analysis. This was
contrary to the other two sets of time analyzed. It can be speculated this is because during
those first two days, The New York Times did not want to frame articles in terms of
human interest because it was too busy reporting on the who, what, when, where, and
why of what had happened.
It's also interesting to note that the use of the morality frame was not significantly
different between the two newspapers during the first two days of analysis or the whole
ten-day period of study, yet it was significant for the eight days after the immediacy of
the attacks was over. Perhaps the first two days after the attacks was too soon for Le
Monde to include frames of morality in the reporting.
The differences between the two newspapers' use of the economic consequences
and conflict frames were not significant during the two-day time period either. Perhaps
these two days resulted in similar uses of these frames because the event was so huge and
indeed a crisis situation, therefore, both newspapers were more concerned with reporting
than with how they were reporting.
The study is of importance for two main reasons. First, it adds to the body of
knowledge in journalism research regarding framing. The results and conclusions drawn
from the present study can help future framing researchers do better research.
While the theory of framing has been studied for decades, the amount of literature
on media framing across national borders is limited. Furthermore, the amount of literature
on media framing across national borders during an international crisis situation is even
more limited. Therefore, the topic of the present study warrants further investigation.
Because most of the hypotheses were supported, it can be broadly stated that
previous analyses of French and American newspaper reporting styles are an accurate
representation of the reality in these countries. The literature was supported in the present
study since each hypothesis was based on at least one aspect of the literature. The results
of the present study help to support the suggestion that the literature used should and
could be used again by future researchers.
The literature on the two newspapers was somewhat limited, however. There was
little printed material comparing the two newspapers directly after the 1970s, it seemed.
But, the literature that was found seemed to give an accurate picture of what the two
newspapers are still like today.
The five media frames used to analyze the articles in the present study were useful
in most respects; however, they were also limiting in other respects. They were limiting
because there were sometimes articles that did not fit into any of the five categories from
which to choose. This result leads the researcher to suggest ideas for future research on
The second reason the present research is of importance is because it can help
communication researchers in their future analysis of the coverage and framing used in
crisis reporting. The attacks of September 11 created an international crisis in terms of
national security. The reporting that resulted needs to be analyzed. As Chomsky (1998)
argues, the media in democracies are highly important in terms of shaping what the
country as a whole thinks and thinks about. The present study had found that these two
elite newspapers seem to have reported similarly in the two days immediately following
the attacks, but they drifted back toward their traditional roles of news informer for the
Times and analyzer for Le Monde. It seems the newspapers did not follow Chomsky's
traditional mold in those first two days, but they did, perhaps, go back to channeling the
thoughts and attitudes of their readers.
After this crisis, it is important to look at what the elite press was saying and,
more importantly, how they were framing what they were saying. If researchers continue
to study this subject, then future media professionals will know what has already been
done if such a crisis should ever arise again. Then, they can make a decision as to
whether or not to do the same thing that was done the previous times around. The present
study found that the two newspapers seemed to remain true to the cultures of their
countrymen. The New York Times frames articles in terms of human interest and conflict
more than Le Monde does, but Le Monde frames its articles in terms of economic
consequences and uses significantly more elite sources than The New York Times does.
Editors at these two newspapers can now look at these results and decide whether they
want to continue this pattern of reporting.
Additionally, the results of the present study can be looked at with regard to what
the U.S. government was actually doing as a means of comparison between what the
media was reporting about the attacks and how the government was responding to the
attacks. For example, The New York Times seemingly made it clear that the United States
did not know for sure who was responsible for the attacks, yet the federal government
continued to discuss the response as if it were black and white issue.
While the present study was overall successful, several limitations were
encountered along that way. The first limitation came from the computer database. There
is no way, short of conducting a separate census of the paper copies of the newspapers,
that the researcher could be positive that the search terms entered would result in a true
census of articles printed on the topic of the attacks of September 11. The researcher
could only use the best search terms available when searching for the articles to be coded.
Had time and money not been an issue, the research could have literally looked at every
day's physical edition of The New York Times and Le Monde printed in the 10-day time
period to ensure a more accurate selection of articles to be coded. However, time and
money were limitations in the present study, so the computer database was used.
Another limitation of the present study was language. While the researcher is
conversational in the French language, she is not totally bilingual. This was a limitation
because there could have been words used that were inferred to mean one thing by the
researcher, when in fact, they were intended by the author to mean something else. The
use of a bilingual second coder helped to double check the researcher's initial coding and
showed that the researcher had been successful. Since the reliability results were similar
for both newspapers, it can be inferred that the researcher's readings of the Le Monde
articles were fairly true to the actual meanings of the articles.
Time was the final limitation of the study. If there had been more time, the
researcher would have liked to use the data gathered in this round of coding to do a new
set of content analyses on the same articles but not using the five media frames. The
researcher would have picked new frames of her own choosing to use in the coding that
were based on this initial research.
Additionally, if there had been more time, the researcher would have liked to have
done better scanning of the headlines in order to prevent articles being included in the set
to be coded that did not belong there. This would have prevented the one Le Monde
article and four The New York Times articles from having to be discarded after the fact.
While the categorization of media frames into five distinct frames was very useful
in the present study, after the coding was complete, qualitative analysis implies some new
frames that could be used in future research.
Future research could be done using a selection of media frames to do coding
analysis, but researchers could use different frames than the present study used. The
results from the present study suggest that reporting of a crisis situation includes
additional media frames to the five defined by Semetko and Valkenburg (2000). For
example, many articles were framed in terms of safety or security issues to readers. This
would be a good frame category to include in future research. Both newspapers had
articles framed in ways that would make the readers feel secure in their respective
countries. There were articles about heightened security measures at airports, ballparks,
public parks, and other places where large numbers of people pass through each day. It
seemed like the idea of making readers feel safe was an important one to both
Additionally, a number of articles were simply descriptions of the events that
happened, especially in the immediate days following the attacks. It seems that the
newspapers wanted to offer their readers a simply play-by-play of what had happened.
Since both newspapers printed this type of article, perhaps a new frame called
"descriptive" could be added to future research. Perhaps the concept of framing is too
much of an American concept, and that could be why the five frames used in the present
study weren't perfect matches for all the articles. Since the analysis was being done on
articles written about an international event, the frame definitions were a bit limited.
However, the Semetko and Valkenburg (2000) research was done on media in the
Netherlands, so the five frames defined by Semetko and Valkenburg do have some basis
for use with Western media systems.
Similarly, future research could be done on the same topic without using pre-
assigned frames. The researcher could come up with definitions of the types of frames
used in each article after reading a number of articles and coming up with broad general
frame categories used.
Future research could also replicate the present study as a way to refine the
research. In other words, this same study could be done but done a little differently. For
example, the present study included sports articles found in The New York Times even
though there are no sports articles in Le Monde. Another researcher could do the same
analysis, but they could simply code sports articles for separate analysis in their study.
Another type of refined replication of this study could be done by comparing the
articles found on the Lexis/Nexis database with a careful study of the actual printed
copies of the two newspapers. Additionally, the researcher could go back and compare
the results of the first frame used (as was the purpose of the present study) to the second
frame used in articles where there was more than one frame used. Similarly, the
researcher could compare the frames usage found in the present study to the use of
frames in the headlines of the articles analyzed since readers often only look at headlines
when reading a newspaper.
Finally, a similar study could be done by using several non-American second
coders. These second coders could code the newspapers from their home countries and
then also code newspapers from a different country. The discussions that could come
from this could help the research decide whether or not the frame definitions were too
The newspaper industries in both the United States and France can use the present
study to make decisions about their routines. The New York Times and Le Monde can now
see a generalized view of how it reported on the attacks of September 11. Both
newspapers can now make better decisions as to whether they want to frame articles on
international crisis situations the same way they did during this time period.
Additionally, society can look at this study to see how the two newspapers
reported the events. By looking at the data in the study, members of both societies can see
how the two newspapers differed in the reporting of the event, and individuals can make
a more informed decision as to from which newspaper they would like to get their news.
Overall, the present study is useful to both the newspaper industry and society as
a whole because it gives both groups an objective, general picture of how the events of
September 11 were reported. This result is important because, as has been previously
stated, the media had a great deal of power in shaping public opinion. The more the
public knows about how the media have been working, perhaps the more informed
decisions they can make.
I. Project Description
The attacks of September 11, 2001 were unprecedented and created great
challenges for the press around the world. The press systems in countries all across the
globe were facing a challenge they had never faced before.
The attacks came at a time of relative peace for the United States. The media of
the country had been scaling back coverage of international news for the past two
decades. As a result, the press in the United States had to play "catch up" for audiences.
There were huge amounts of space devoted to coverage of the attacks in almost every
paper in the country. They were used to deal with everything from background
information on the attackers to the day-to-day information on the victims.
The role the news media play in today's society is of great importance. The way
the press portray events has a large impact on the way society thinks about that event.
Traditional communications researchers have used the term agenda setting to describe
this phenomenon. This agenda-setting research has lead into the concept of framing.
While the two ideas are similar, framing looks more specifically at how the media is
reporting the news.
The way different newspapers framed the attacks of September 11 demands
research since the way society views the event and ensuing events related to the attacks
can impact society's view on things such as an impending war.
The articles to be coded will be chosen differently for The New York Times and Le
Monde. All the articles coded will come from the issues published during the period of
September 12 to September 21. This time frame was chosen because it includes the
immediate response of the two papers and follows through until President Bush's speech
stating that the world was either "with us or against us."
A census of the articles appearing in Le Monde will be coded. The census equaled
176 articles. Subsequently, systematic sampling of the census of articles found (equaling
850) will be used to code 176 articles from The New York Times. The articles for both
newspapers were obtained using the databases of Lexis/Nexis. The search terms used
were similar but not identical because of obvious inconsistent language translation
between English and French.
All stories will be analyzed with the exception of letters to the editor, photo
essays, graphics, and simple lists of victims. These types of stories will be excluded
because they do not lie in within the structure of traditional frame analysis.
III. Length to be coded
Examine only the beginning of the article. The amount to be read will vary from
article to article. Simply read until you have enough information to code for all the
variables. Typically, the information being coded for will be within the first five
paragraphs. It may take longer for some articles, though.
IV. Coding Log
Keep notes on the coding sheets of items that you feel may be of interest to the
researcher. This will help others understand how and why you are making the coding
decisions you are making. Each entry should include the date, newspaper name, headline
(with English translation), number of words, frame used, and type of source used. Jot
down any themes you see appearing. Also, jot down on the coding sheet any problems or
confusion you had in any of the coding. For example, if you think one article could be
coded in to two different frames, pick one frame and make note of the second frame you
think should also be considered. This will help the researcher flush out problems. Type
up a detailed list of notes and overall description in an organized manner as well so the
researcher will have a better idea at the ideas you may have had while coding.
V. Filling in the coding sheets
Fill out one sheet per article. Analyze all of the stories given to you.
1. Article Number: For each coding sheet, a sequential integer should be entered in
the blank to the right of the page. Start with the number "1" for the first article coded,
continue with "2", and so forth.
2. Newspaper: Write the appropriate number of the newspaper being coded on the
line at the right of the page. Write "1" for The New York Times and "2" for Le Monde.
3. Date: Enter the date of the article being coded on the line to the right of the page
in a month/day format.
4. Number of words: Enter the number of words in the article given by Lexis/Nexis
on the line at the right of the page.
5. Coder: Enter the appropriate number for who you are. Enter "1" for Allison
Aiken and "2" for the second coder.
6. Frame used: Enter the appropriate number on the line at the right for the type of
frame used in the article. The frames are:
1) Attribution of Responsibility: If the article is framed with reference to who or
what was responsible for the attacks of September 11, enter a "1" on the line at the right.
2) Economic Consequences: If the article is concerned with the economic
consequences of the attacks of September 11, enter a "2" on the line at the right.
3) Morality: If the article is put in terms of religious tenets or moral ideals, enter a
"3" on the line at the right.
4) Conflict: If the article is written in terms of conflict between individuals or
groups to get the reader's attention, enter a "4" on the line at the right.
5) Human Interest: If the article focuses on some personal aspect of the attacks,
enter a "5" on the line at the right.
7. Sources: The source is the document or person providing information for the
article. Articles often have more than one source. You will just be coding the first source
you come to. You will be coding for two different categories of sources: the sources
eliteness and the type of source in relation to the event.
Elite sources are those well-known and well-respected documents or people used in
articles. Examples would be Colin Powell, police officers working on the scene of the
attacks, foreign leaders, other well-respected news agencies (ie, The Associated Press), or
leaders of international organizations such as Koffi Anan. Non-elite sources are those
often termed "man on the street." Examples of non-elite sources would be people
attending a candlelight vigil for the victims, people who were walking on the street when
the attacks happened, individuals who have lost their jobs as a result of the attacks, or
individuals who have opinions on the attacks but are not well-known to society at large.
Enter a "1" for all elite sources found and a "2" for those non-elite sources found.
8. There 10 types of sources in relation to the event. They are: domestic
government, foreign government, NGOs, witness to the attacks, victims of the attacks,
victim's relative, law enforcement, expert, non-expert or "man on the street," and
business owners. Enter the corresponding number on the line to the right of source on the
1) Domestic government refers to sourcing by any member of any government
agency, including mayors, aides, and other governmental type institution.
2) Foreign government refers to the same, only coming from any non-U.S.
3) NGOs will include any organized group not a part of the federal government.
For example, charities or church organizations would be categorized in this group.
4) Witness, 5) victim, and 6) victim relative refer to witness of, victims of, and
relatives of victims of the attacks.
7) Law enforcement includes both fire and police forces.
8) Experts are those people whose profession it is to know an extensive amount
about some topic. Examples of experts might be terrorism or financial experts
interviewed for their opinion after the attacks.
9) The non-expert would be just any everyday person interviewed for their thoughts
or opinions on the topic.
10) Business owners are those sources whom own businesses and have been
interviewed for their thoughts and opinions on the attacks regarding their business.
9. Headline: Write the headline on the lines provided. Write as much of the
headline as is needed in order to get a good idea of what the article is about. ForLe
Monde headlines, write a rough English translation below the French version.
Coverage of the September 11 attacks in
The New York Times and Le Monde
vl. Article Number ID
v2. Newspaper NP
1) The New York Times
2) Le Monde
v4. Number of Words WORDS
v5. Coder CODER
1) Allison Aiken
v6. Frame used FRAME
1) Attribution of Responsibility
2) Economic Consequences
5) Human Interest
v7. Sources SOURCE
v8. Source in relation to event EVTSOU
1) Domestic Government
2) Foreign Government
6) Victim reality
7) Law Enforcement
10) Business Owner
Table B.1 The New York Times
Item Date Frame Eliteness Source
1 12 3 2 9
2 12 5 2 5
3 12 4 1 1
4 12 3 2 9
5 12 4 1 1
6 12 4 1 8
7 12 4 1 1
8 12 4 1 1
9 12 1 1 1
10 12 1 1 2
11 12 5 1 3
12 12 4 1 2
13 12 2 1 8
14 12 2 1 8
15 13 5 1 4
16 13 5 2 10
17 13 5 2 4
18 13 3 2 5
19 13 1 1 8
20 13 4 1 1
21 13 4 1 1
22 13 4 1 1
23 13 2 1 10
24 13 5 2 6
25 13 2 1 1
26 13 5 1 7
27 13 5 2 3
28 13 4 0 0
29 13 3 1 3
30 13 5 2 6
31 13 5 1 3
32 13 2 1 8
33 13 3 2 9
34 13 4 1 2
35 14 2 1 3
Table B.1 Continued:
Item Date Frame Eliteness Source
36 14 4 1 7
38 14 2 1 3
39 14 4 1 1
40 14 3 1 3
41 14 3 2 9
42 14 4 1 7
43 14 3 2 6
44 14 4 1 4
45 14 4 1 1
46 14 2 1 1
47 14 3 1 5
48 14 5 1 1
49 14 3 1 10
50 14 4 1 1
51 14 4 0 0
52 14 4 1 10
53 14 2 1 8
54 14 2 2 8
55 15 2 1 8
56 15 4 1 10
57 15 4 1 1
58 15 5 2 1
59 15 4 1 1
60 15 5 2 1
61 15 3 0 0
62 15 1 1 2
63 15 2 1 2
64 15 5 1 8
65 15 4 2 5
66 15 4 1 8
67 15 2 1 8
68 15 5 1 10
69 16 5 1 8
70 16 5 2 6
71 16 5 2 6
73 16 1 1 1
74 16 5 1 1
75 16 4 1 1
76 16 4 1 1
77 16 4 1 2
78 16 4 2 9
79 16 4 1 8
Table B.1 Continued:
Item Date Frame Eliteness Source
80 16 5 0 0
81 16 4 0 0
82 16 2 1 8
83 16 5 1 8
84 16 4 1 8
85 16 2 1 8
86 16 5 1 1
87 16 3 1 1
88 16 4 1 3
89 16 2 1 8
90 17 1 1 8
91 17 1 2 9
92 17 4 1 10
93 17 4 1 1
94 17 5 2 5
95 17 5 2 4
96 17 2 1 5
97 17 1 1 2
98 17 5 2 6
99 17 3 2 3
100 17 4 1 2
101 17 2 1 8
102 17 4 1 1
103 17 5 1 8
104 17 5 1 10
105 17 4 0 0
106 18 4 0 0
107 18 4 1 1
108 18 4 1 8
109 18 2 2 5
110 18 2 2 8
111 18 2 2 5
112 18 5 1 1
113 18 4 1 1
114 18 5 2 9
115 18 4 0 0
116 18 4 1 1
117 18 5 0 0
118 18 5 2 9
119 18 2 1 10
120 18 1 2 9
121 18 5 0 0
Table B.1 Continued:
Item Date Frame
122 18 2
123 18 1
124 18 5
125 19 2
126 19 4
127 19 2
128 19 4
129 19 2
130 19 5
131 19 4
132 19 5
133 19 5
134 19 5
135 19 5
136 19 4
137 19 4
138 19 5
139 19 2
140 19 5
141 20 4
142 20 3
143 20 5
144 20 5
145 20 5
146 20 3
147 20 3
148 20 5
149 20 5
150 20 4
151 20 4
152 20 3
154 20 5
155 20 4
156 20 5
157 20 5
158 21 4
159 21 2
160 21 4
161 21 4
163 21 1
164 21 4
165 21 5
Table B.1 Continued:
Item Date Frame
166 21 5
167 21 4
168 21 5
169 21 4
170 21 2
171 21 4
172 21 3
173 21 4
174 21 3
175 21 5
176 21 2
Table B.2 LeMonde
Item Date Frame
1 13 4
2 13 4
3 13 4
4 13 1
5 13 4
6 13 4
7 13 3
8 13 3
9 13 3
10 13 3
11 13 5
12 13 5
13 13 5
14 13 5
15 13 1
16 13 1
17 13 5
18 13 2
19 13 4
20 13 4
21 13 4
22 13 4
23 13 5
24 13 5
25 13 5
26 13 2
27 13 2
28 13 2
Table B.2 Continued:
Item Date Frame Eliteness Source
29 13 4 1 2
30 13 3 0 0
31 14 4 1 8
32 14 1 1 8
33 14 1 1 2
34 14 5 1 8
35 14 2 1 8
36 14 2 1 8
37 14 2 1 8
38 14 2 1 3
39 14 2 1 2
40 14 3 2 9
41 14 4 1 2
42 14 3 1 3
43 14 5 1 8
44 14 2 1 8
45 14 4 1 10
46 14 4 1 8
47 14 1 1 8
48 14 5 1 8
49 14 5 1 8
50 14 1 1 2
51 14 4 1 1
52 14 3 0 0
53 14 4 1 3
54 14 3 0 0
55 14 4 1 2
56 14 1 1 3
57 15 1 1 2
58 15 3 2 9
59 15 5 1 3
60 15 2 2 9
61 15 4 1 3
62 15 1 1 2
63 15 4 1 2
64 15 3 1 1
65 15 3 1 8
66 15 3 2 9
67 15 4 1 8
68 15 2 1 9
69 15 2 1 8
70 15 3 1 10
Table B.2 Continued:
Item Date Frame
71 15 2
72 17 5
73 17 2
74 17 5
75 17 2
76 17 2
77 17 2
78 17 2
79 17 2
80 17 3
81 17 3
82 17 3
83 17 4
84 17 1
85 17 1
86 17 4
87 17 4
88 17 4
89 17 4
90 17 5
91 17 4
92 17 5
93 17 4
94 17 4
95 17 4
96 17 4
97 17 5
98 17 1
99 17 2
100 18 2
101 18 3
102 18 4
103 18 1
104 18 1
105 18 4
106 18 3
107 18 4
108 18 3
109 18 4
110 18 4
111 18 2
112 18 2
Table B.2 Continued:
Item Date Frame
113 18 3
114 18 5
115 18 5
116 19 2
117 19 2
118 19 3
119 19 5
120 19 3
121 19 4
122 19 5
123 19 1
124 19 2
125 19 2
126 19 3
127 19 2
128 19 2
129 19 2
130 19 4
131 19 3
132 19 4
133 19 1
134 19 1
135 20 2
136 20 1
137 20 5
138 20 2
139 20 4
140 20 1
141 20 5
142 20 2
143 20 2
144 20 4
145 20 4
146 20 4
147 20 4
148 20 4
149 20 2
150 20 5
151 20 5
152 21 5
153 21 1
154 21 3
Table B.2 Continued:
Item Date Frame Eliteness Source
155 21 4 1 3
156 21 4 1 2
157 21 1 2 9
158 21 1 1 2
159 21 2 1 2
160 21 2 1 2
161 21 2 1 8
162 21 4 1 1
163 21 5 1 1
164 21 4 1 2
165 21 3 1 2
166 21 5 1 3
167 21 5 1 8
168 21 2 1 2
169 21 4 1 2
170 21 3 1 8
171 21 5 1 2
172 21 2 1 8
173 21 3 1 3
174 21 2 1 3
175 21 5 1 3
LIST OF REFERENCES
Akhavan-Majid, R. and J. Ramaprasad. (2000). Framing Beijing: Dominant Ideological
Influences on the American Press Coverage of the Fourth UN Conference on
Women and the NGO Forum. Gazette, 1, 45-59.
Americans Open to Dissenting Views on the War on Terrorism. (2001, October 14).
Retrieved August 28, 2002 from http://people-press.org/reports/print.php3?PageID=22.
Babbie, E. (2001). The Practice of Social Research. 9th Ed. Belmont, CA: Wadsworth
Baskette, F., J. Sissors, and B. Brooks. (Eds). (1997). The Art ofEditing. 6th Ed. Boston:
Allyn and Bacon.
Budd, R., R. Thorp, and L. Donohew. (1967). Content Analysis of Communications. New
York: The Macmillan Company.
Bush, C. (Ed.). (1966). "When and Why Does the Reader Stop Reading?" pp. 76-78 of
News Research for Better Newspapers. Vol. 1. New York, NY: American
Newspaper Publishers Association Foundation.
Chang, T. and J. Lee. (1992). Factors Affecting Gatekeepers' Selection of Foreign News:
A National Survey of Newspaper Editors. Journalism Quarterly, 3, 554-561.
Chomsky, N. (1989). Necessary Illusions. South End Press: Boston.
Cohen, B. (1963). The Press and Foreign Policy. Princeton, NJ: Princeton University
Cunningham, B. (2001, November/December). Global Blinders. Columbia Journalism
Review. 4, 110-111.
Entman, R. (1991). Framing U.S. Coverage of International News: Contrasts in
Narratives of the KAL and Iran Air Incidents. Journal of Communication, 4 p. 6-
Entman, R. (1993). Framing: Toward Clarification of a Fractured Paradigm. Journal of
Communication, 4, 51-58.
Flash Eurobarometer 114: Crise Internationale. Released by EOS Gallup December 2001.
Gamson, W. (1992). Talking Politics. New York: Cambridge University Press.
Gamson, W. and A. Modigliani. (1987). The Changing Culture of Affirmative Action. In
R.G. Braungart and M.M. Braungart (Eds), Research in Political Sociology. 33,
137-177. Greenwich, CT: JAI Press.
Gans, H. (1979). Deciding What's News: A Study of CBS Evening News, NBC Nightly
News, and Time. New York: Pantheon Books.
Goffman, E. (1974). Frame Analysis: An Essay on the Organization ofExperience. New
York: Harper & Row.
Harrigan, J. (1993). The Editorial Eye. New York: St. Martin's Press.
Herman, E. and Chomsky, N. (1988). Manufacturing Consent: The Political Economy of
the Mass Media. New York; Pantheon Books.
James, C. (2001, November 9). British Take Blunter Approach to Reporting War. New
York Times. Retrieved on January 3, 2003, from http://www.nytimes.com.
Kahneman, D. and A. Tversky. (1984). Choice, Values, and Frames. American
Psychologist, 39, 341-350.
Kerlinger, F. (1964). Foundations ofBehavioral Research. New York: Rinehart and
Kohut, A. (2002, January/February). A Worried Public Tunes In: The Press .\/Ne.\ in a
DarkMoment. cjr, 5, 54-55.
Kinder,D. and L.M. Sanders. (1990). Mimicking Political Debate With Survey Questions:
The Case of White Opinion on Affirmative Action for Blacks. Social Cognition, 8,
Mencher, M. (2000). News Reporting and Writing. 8th Ed. Boston, Mass.: McGraw-Hill.
Merrill, J. (1968). The Elite Press: Great Newspapers of the World. New York: Pitman
Navasky, V. (2002). "Foreword" in Journalism After September 11. Zelizer, B. and S.
Allan (Eds). (xiii-xviii). New York: Routledge.
Noris, P. (1995). The Restless Search: Network News Framing of the Post-Cold War
World. Political Communication, 12, 357-370.
Parks, M. (2002, January/February) Foreign News: What's next? cjr. 5, 52-57.
Pool, I. (1970). The Presitge Press: A Comparative Study of Political Symbols.
Cambridge: The M.I.T. Press.
Riffe, D. (1998). Analyzing Media Messages: Using Quantitative Content Analysis in
Research. Mahwah, N.J.:Lawrence Erlbaum Associates, Inc.
Salinger, P. (1982). Le Monde: De Gaulle's Only Legitimate Heir. In Rice, M. and J.A.
Cooney (Eds) Reporting U.S. -European Relations: Four Nations, Four
Newspapers. 82-113. Elmsford, NY: Pergamon Press.
Scheufele, D.(1999). Framing as a Theory of Media Effects. Journal of Communication,
Schramm, W. (1959). One Day in the World's Press: Fourteen Great Newspapers on a
Day of Crisis. Stanford, Calif: Stanford University Press.
Semetko, H. and Valkenburg, P. (2000). Framing European Politics: A Content Analysis
of Press and Television News. Journal of Communicaiton, ISSUE, 93-109.
Shoemaker P. and S. Resse. (1996). Mediating the message: Theories of Influences of
Mass Media Content. (2nd ed.) New York: Longman.
Silverman, M. (1999). Facing Postmodernity: Contemporary French Thought on Culture
and Society. London: Routledge.
Singer, J. (2001). The Metro Wide Web: Changes in Newspapers' Gatekeeping Role
Online. Journalism & Mass Communication Quarterly, 1, 66.
Tillier, A. (2001, October 22). Le Monde Sets New Course. The Daily Deal. Retrieved
November 23, 2002 from http://www.thedeal.com.
Tuchman, G. (1978). Making News: A Study in the Construction of Reality. New York:
Weber, R.P. (1985). Basic Content Analysis. Beverly Hills: Sage Publications.
Zeiler, B. and Allan, S. (2002). Introduction: When Trauma Shapes the News. In
Journalism After September 11. Zelizer, B. and S. Allan (Eds). 1-23. New York: