To conduct the research, we utilized a content analysis
method that integrates data collection and analytical techniques to measure
the occurrence of identifiable elements in a text or message (Keyton,
2001). The newspapers chosen for this analysis were the New York Times,
Washington Post, Los Angeles Times and Chicago Tribune. These newspapers
were chosen because of their geographical location, and because of their
known influence on key decision makers within the government and Department
of Defense. The New York Times and Washington Post are very powerful when
it comes to influencing politicians and business leaders and both are
located in the East. These two newspapers can also be described as, "newspapers
of record," since they help set the agendas for other newspapers
in the country. The Chicago Tribune and LA Times are chosen due to their
geographical locations of Mid-West and West.
The analysis focused on newspaper articles relating to military combat
operations contained within the "local," "national,"
and "international" sections including the front page of the
newspaper. The entire article was analyzed. Stories were collected using
the Lexis-Nexis database or gathered from microfilm from the University
of Oklahoma's Bizzel Library. Each article was analyzed to determine the
tone of the story and the frame employed.
The research featured one primary independent variable. This independent
variable was the newspaper reporters’ status as either embedded
or non-embedded. This variable is the primary focus of our research; to
see if embedded journalists produce different coverage compared to non-embedded
The other independent variable was the war being examined. We chose three
recent wars/conflicts for our analysis. Each of these wars had ground
troop movement in occupation of the foreign nation. Our analysis focuses
on combat news coverage during the ground operations within a 5-day span
starting with first day of ground operations. Our selection of wars and
dates are as follows: Operation Desert Storm; Feb. 24-28; Operation Enduring
Freedom, Oct. 7-11; Operation Iraqi Freedom, Mar. 20-24.
The unit of analysis was the stories reported in the various newspapers.
Four Department of Defense employees affiliated with the University of
Oklahoma’s Joint Communication Course conducted the coding of the
stories using a code key the team created, shown in Appendix A. Coding
norms were established during a 1-hour training session, in which 8 articles
outside the 5-day span were analyzed. To achieve inter-coder reliability,
the researchers coded articles together to establish norms and then coded
20 articles in split groups to increase the chance that coding would be
done with the same judgment. The analysis was used to evaluate the articles
origin, content, tone and tenor, depiction of military members and the
framing of the article for 291 stories.
The study featured three dependent variables. The first variable was overall
tone/attitude, which was assessed using an attitude scale similar to the
one employed by Pfau, et al. (2001), to assess the overall tone of the
article toward the military. According to Pfau, et al. (2001), this scale
has been used for 20 years with an alpha reliability of better than .90.
Attitude was assessed via six bipolar adjective pairs, including: Positive
to negative; wise to foolish; valuable to worthless; favorable to unfavorable;
good to bad; acceptable to unacceptable.
The second variable was the trust of the troops covered in the news reports.
Trust was operationalized as scores on the individualized trust scale.
The scale was a 5-interval metric to determine variable ranges on the
following items: trustworthy to untrustworthy; candid to deceptive; honest
to dishonest; and sincere to insincere. This scale has been used by Wheeless
and Grotz (1977), who reported a split-half reliability of .92 for the
15 item ITS, Van Leer and Tujillo (1986) used four items from the ITS
and reported an alpha of .82. These scales were used to determine the
tone of the portrayal of military troops/unit in the newspaper articles.
The study used an attitude scale similar to the one employed by Pfau,
et al. (2001), to assess the overall tone of the article toward the military.
According to Pfau, et al. (2001), this scale has been used for 20 years
with an alpha reliability of better than .90. Attitude was assessed via
six bipolar adjective pairs, including: Positive to negative; wise to
foolish; valuable to worthless; favorable to unfavorable; good to bad;
acceptable to unacceptable.
Finally, the study employed a 5-interval scale to determine the extent
to which an article was more or less episodic or thematic. Each article
could possess varying degrees of each based on the framing and priming
approach. Each variable, episodic and thematic, was given a 5-interval
scale with 5 representing more of the variable. Articles can obtain both
episodic and thematic framing or very little. This was used to determine
if embedded journalists will produce more episodic articles than non-embeds.
In the first phase of the analysis, a 3 (embedded, non-embedded, and unknown)
x 3 (Desert Storm, Operation Enduring Freedom, and Operation Iraqi Freedom)
multivariate analysis of covariance (MANCOVA) was computed on the dependent
measures of attitude, trust, episodic and thematic.