Frame Flow between Government and the News Media and its Effects on the Public: Framing of North Korea

Abstract

Public opinion is likely to be susceptible to the way a government and the news media frame foreign countries, because unlike domestic issues, foreign news is typically beyond a person's direct experience. How does the American public respond to foreign news when its government and the news media promote competing frames and change their prominence according to the relations between the U.S. and that foreign country? The present study shows this frame building and frame effects by using a public opinion poll and content analysis of U.S. policy statements and media coverage. North Korea was chosen because its visibility to the American public has increased since President George W. Bush designated it as one of the countries in the “axis of evil”. The results show that during a four-month period, the U.S. government and the newspaper produced three competing frames, and that the magnitude of the frames shifted as U.S.–North Korean relationships shifted. These shifts in turn made the American public choose economic sanctions over military solutions toward the country.

When it comes to U.S. foreign policy, three actors emerge: the U.S. government, the U.S. news media, and the U.S. public. However, the three actors do not equally contribute to foreign policy because it is rare for the public to exert a direct influence on either the U.S. government or the U.S. news media (Entman, 2004). Instead, the U.S. government and the U.S. news media are able to influence the American public's perceptions and attitudes toward foreign countries.

An example is President George W. Bush's State of Union address in 2002. He branded North Korea as part of the “axis of evil” along with Iran and pre-war Iraq, indicating that North Korea was a major threat to the security of U.S. citizens. After his speech, North Korea received intensive attention from both the U.S. government and the U.S. news media. The U.S. government and the news media framed North Korea in using multiple terms which directly affected public perception of such things as North Korea as a U.S. security threat, or the U.S. government's ability to productively negotiate with the country.

When they consume these policy statements and news coverage concerning North Korea, the American people receive competing frames generated by the U.S. government and the news media regarding the country. This is a “competitive situation” where people pay attention to competing frames, whether in equal or unequal amounts (Chong &], 2007). The level of attention given to the competing frames about North Korea is likely to be high because a foreign country is a complex world that is out of direct reach (Lippmann, 1922; McCombs, Danielian, & Wanta, 1995), making people rely increasingly on a government's official statements and news coverage to understand foreign issues. The implications of frame competition have started to draw scholarly attention (Chong & Druckman, 2007; Entman, 2003, 2008), but few studies have addressed the question of how the public responds to competing frames having varying magnitudes, which are targeted toward foreign countries, and specifically North Korea.

To fill this gap, the present study examines how the U.S. government and the news media frame discourse on North Korea and what specific frames are transferred from the U.S. government and the U.S. news media to the American public.

FRAMING EFFECTS AND FRAME COMPETITION

The present study is based on framing studies that directly address framing patterns of both presidents and elites, and the news media. The main focus of the study is how the U.S. government and the U.S. media frame discourse on North Korea. Gamson (1992) and Gamson and Modigliani (1989) provide a guiding definition of a frame, defining it as a core organizing idea for making sense of real-time events or issues. Specifically, Gamson and Modigliani (1987) define a frame as “a central organizing idea or story line that provides meaning to an unfolding strip of events, weaving a connection among them” (p. 143). Other scholars present similar definitions, arguing that a frame makes sense of reality for individuals and journalists (Entman, 1993; Gitlin, 1980; Scheufele, 2004; Tuchman, 1978). This definition suits the present study's purpose of conceptualizing and measuring frames built by the U.S. government and the U.S. news media. For instance, President George W. Bush's framing patterns of North Korea is likely to represent his major idea of how to organize policy statements on North Korea. The New York Times’ framing patterns of the country capture the newspaper's core idea that guides coverage of the country.

However, governmental frames are different from media frames in terms of scope and level. First, governmental frames aim to facilitate desirable policies and constrain undesirable policies (Levin, 2005). Accordingly, they reflect a government's policy preferences and try to make the public accept the preferences (Rottinghaus, 2008). By contrast, media frames base themselves on newsroom routines as to how to cover and interpret social reality (Tuchman, 1978). That is why media frames capture the characteristics of news (Entman, 1991). Second, governmental frames have advantages over media frames in crafting foreign policy because the president and high-ranking officials control information flow from governmental departments to the news media by deciding which policy statements can be released (Entman, 2003). In this sense, governmental frames are likely to guide media frames, although the latter can affect the former.

Framing scholars explain framing effects in terms of three social cognitive concepts: availability, accessibility, and applicability (Chong & Druckman, 2007; Scheufele, 2004; Scheufele & Tewksbury, 2007). Availability means that information needs to be saved in memory for retrieval for attitude formations, and accessibility is the ready availability and activation of that information in the memory. Applicability refers to making connection between framed objects and a frame. For instance, when government's policy statements focus on an applicable connection between the axis-of-evil designation and North Korea, there are overlapping meanings and implications between the two terms. Chong and Druckman (2007) propose that such a connection is most likely to occur when interpretations conveyed by a frame and a frame's persuasiveness are increasingly present. Therefore, when people store the meanings and implications of the axis of evil in memory and those pieces of information are easily accessible, people are able to make sense of the connection.

Several experimental studies found evidence of framing effects. For instance, a conflict-framed news story made people show negative feelings toward university administration (Price, Tewksbury, & Powers, 1997). Media framing of the Ku Klux Klan controversy significantly influenced tolerance for the group, with the perceived importance of different values mediating the effects (Nelson, Clawson, & Oxley, 1997). News stories focusing on either an issue or a strategy influenced individuals’ perceptions toward civic duty, and the effects were moderated by education (Valentino, Beckmann, & Buhr, 2001). When an issue such as the introduction of the euro was constructed in terms of the human interest frame, individuals’ capacity to recall specific bits of information decreased (Valkenburg, Semetko, & de Vreese, 1999). When stem cell funding was framed in terms of either a benefit frame or an ethics frame, individuals who read benefit-framed stories were more likely to support the funding than those who read ethics-framed stories (Shen, 2004).

However, previous social cognitive explanations and subsequent findings did not consider the possibility that in the real world, people receive multiple frames, which have equal or unequal magnitude. This idea leads to frame competition. Chong and Druckman (2007) link this competitive nature to framing effects, and suggest that people consume messages that are likely to contain equal portions of multiple strong frames or weak frames or unequal portions of multiple strong frames or weak frames. The strength and prevalence of the frame in messages determine to a substantial degree the effects of framing on public opinion. Druckman (2001) also points out that multiple competing frames influence the effects of each individual frame on public opinion. Saris and Sniderman (2004) found that people demonstrated higher support for government spending on the poor when consuming two different frames invoking positive and negative responses than those consuming a single frame.

FRAMING AND FOREIGN POLICY

Although these prior works serve as a useful guideline, their applications to U.S. foreign policy areas are limited because they concentrate on domestic issues such as stem-cell research or racial discrimination, and do not explain how the framing of foreign countries affects public opinion. Furthermore, in foreign affairs, there are multiple competing frames, as there are in domestic social concerns, such as immigration, the financial crisis, or crime. These social concerns can be learned by people's personal experiences or interpersonal connections. However, understanding the nature of foreign affairs is fundamentally different from that of domestic social concerns. Foreign issues are far away from people's everyday lives, therefore the American people cannot utilize their personal experiences to understand the implications of the U.S. government's policy toward North Korea, unless they themselves are policy makers.

Accordingly, an alternative theoretical work directly addressing the implications of a frame for foreign policy is needed, particularly Entman's (2003, 2004) cascading network activation model. The basic idea behind this model is that a frame constructed by government at a high level activates relevant thoughts and feelings in elites and the news media at lower levels, and finally the public at the bottom. Although a frame can be pumped from the public to government, it is frequently observed in foreign policy areas that a top-down frame flow is more likely than such a bottom-up flow. However, the government keeps an eye on the news media's coverage patterns because of their influence on public opinion. In this light, the news media have the legitimacy to influence government's foreign policy, as shown in prior research (Tallman & McKerns, 2000).

According to Entman's cascading network activation model, government and the news media have, to a greater degree than the public, the power to initiate frames and distribute them to the public by constructing an effective discourse about foreign countries. Such a discourse refers to the interactive process of developing meanings and communicating these to social actors (O'Sullivan, Hartley, Saunders, Montgomery, & Fiske, 1994). Discourse facilitated by government and the news media captures a president's announcements in press conferences, radio addresses, and media coverage. This is part of elite discourse: it is initiated and promoted by elites such as the president, the president's press secretary, or major newspapers (Zaller, 1992). As Kinder and Sanders (1990) point out, an elite discourse involves a frame, and Simon and Xenos (2000) suggest that a frame provides connections between concepts in constructing political issues. Making connections between issues in a specific context is a crucial strategy for policy makers and the news media. Specifically, government and the news media compete for a frame in foreign policy areas because a frame defines what is accepted as reality and its interpretations. Entman (2003) points out this frame competition by using the “frame contestation” between President George W. Bush and journalists on Afghan issues as an example. President Bush focused on military attacks on the Taliban government of Afghanistan as his remedy frame. In response to Bush's frame, two journalists, Seymour Hersh and Thomas Friedman reported the link of the Saudi government to terrorism and emphasized that the Saudi government could be a more serious danger than Afghanistan. In response to this frame competition, other U.S. major newspapers published follow-up news stories on the issue. This case exemplifies that a frame has counter frames and that this situation yields a continuum from dominance of a single frame to contestation among different frames to standoff between competing frames (Entman, 2003, 2008).

Therefore, it would be reasonable to argue that the U.S. government and the U.S. media are likely to create frames reflecting their organizing principles when constructing discourse on North Korea. The present study defines this stage as frame building and poses the following research questions, as no prior studies have provided clear guidelines as to the issues:

• RQ1: What frames did the U.S. government develop to construct discourse on North Korea since President George W. Bush's axis-of-evil speech?

• RQ2: What frames did The New York Times develop to construct discourse on North Korea since President George W. Bush's axis-of-evil speech?

Given this frame building, another intriguing theoretical argument can be made regarding a frame flow between government and the news media. In his cascading network activation model, Entman (2003) predicts that a frame at a governmental level activates relevant framing devices on the news media side, and that the news media also spread their specific frames to a governmental level. One study (Wanta & Kalyango, 2007) found evidence for this prediction: when the U.S. president frequently applies a terrorism frame to foreign countries, The New York Times portrays the countries consistent with the terrorism frame. This implies a significant level of correspondence between governmental frames and media frames. Accordingly, the following hypothesis is predicted.

• H1: Since President George W. Bush's axis-of-evil speech, the U.S. government frames were significantly correlated with The New York Times frames when they constructed discourse on North Korea.

Given that competing frames moderate framing effects (Druckman, 2001), an important theoretical consideration is that the American public is likely to receive multiple frames in unequal magnitudes over time from the U.S. government and the news media regarding any given country. Therefore, the first step for the success of the U.S. government's specific framing patterns of North Korea requires their dominance in the news media coverage as well as in policy statements. The question then becomes; “How could such frame dominance influence public opinion?” Entman (2003) proposes that cultural resonance and magnitude determine the effectiveness of a frame. When framing North Korea, the U.S. government used symbols, metaphors, and images that are culturally accepted among the American people. Furthermore, those symbols and other framing devices appeared frequently and repeatedly over time in President Bush's press conferences and addresses, and news coverage, which refers to the magnitude of a frame. With this in place, exposure to those framed messages does not lead to the acceptance of the U.S. government's frames because the exposure only is not sufficient for framing effects (Scheufeule & Tewksbury, 2007). North Korean issues are not directly experienced by the American public, but obtained by indirect channels, such as government policy statements and news coverage. In other words, North Korean issues are unobtrusive topical categories, which require substantial attention over time for assessing their implications. In terms of a historical context, the American public has invested a significant level of attention to North Korean news. According to a public opinion survey conducted by Harris Interactive in June 2002, 93 percent of 400 American respondents paid a substantial amount of attention to news about the relations between the U.S. and foreign countries, including North Korea.

When the American public is attentive to specific frames embedded in messages, as framing-effect studies suggest (Scheufele, 2004; Scheufele & Tewksbury, 2007), the attention to the frames likely activates relevant thoughts and consideration among the American people. They are able to connect those pieces of information and characteristics of foreign countries when deciding their positions toward the countries. This eventually leads the American public to make decisions that are consistent with frames of the U.S. government and the news media.

In the area of foreign affairs, frames promoted by government and the news media change according to current relationships between the U.S. and target foreign countries. For instance, frames that the U.S. government and The New York Times construct to describe North Korea are likely to depend on changes in U.S-North Korean relations over time. When a North Korean military threat was imminent, the U.S. government would emphasize the severity of the threat and suggest its elimination. However, when the North Korean government moved forward on talks, the U.S. government would focus on negotiation with the country.

In this regard, the more the U.S. government emphasizes the military threats of North Korea among competing frames with the bilateral relationship tilting toward confrontations, the more the American people pay attention to that frame because North Korean issues are not interpretable through their personal experiences. Then, they retrieve their prior affective evaluation of the country and the specific relationship between the two countries. This elaboration predicts attitudes toward foreign countries (Peffley & Hurwitz, 1992). As a result of the U.S. government's focus on North Korea's military threats, the American public is likely to prefer military solutions to the country.

Conversely, the more the U.S. government emphasizes the non-military aspects of North Korea, the more the American people are likely to favor non-military solutions, such as economic sanctions. The same prediction is made about The New York Times:

• H2: If the U.S. government shifts framing patterns of North Korea from a military-threat frame to a non-military frame over time, and the American public pays attention to the non-military frame, the public is likely to favor non-military solutions.

• H3: If The New York Times shifts framing patterns of North Korea from a military-threat frame to a non-military frame over time, and the American public pays attention to the non-military frame, the public is likely to favor non-military solutions.

METHOD

Content Analysis

The U.S. government's policy statements and The New York Times’ news stories were examined during a four-month period from January 29, 2002, the day when President George W. Bush delivered the axis-of-evil speech, to May 31, 2002, the day before a public opinion poll survey. The four-month period was determined with regard to the following: first, on January 29, 2002, Bush's axis-of-evil speech maximized contrasting images of South Korea and North Korea to images of the American public. The two Koreas serve as a unique case because unlike Iran and pre-war Iraq, the two Koreas have been divided for more than 50 years, with South Korea remaining as an ally and North Korea as an enemy of the U.S.. Second, the most suitable public opinion poll for answering the research questions and testing the hypotheses was the one conducted on June 1, 2002.

During the time frame, President George W. Bush's policy statements concerning North Korea originated from the website of the White House (2006) and The Weekly Compilation of Presidential Documents in Public Papers of the Presidents (2006), an annual compilation of presidential statements. The compilation is a reliable source for collecting presidents’ policy statements (Wanta & Kalyango, 2007; Wood & Peake, 1998). Daily press briefing transcripts from the website of the Department of State were also used. The U.S. government policy statements on North Korea include President Bush's press conferences, speeches, press briefings by the Department of State, and any other statements produced by the White House. A general search term, “North Korea”, was used to locate a large pool of policy statements. During the same time frame, news stories of The New York Times concerning North Korea were collected from the website of LexisNexis Academic. The New York Times was chosen because of its leading role in foreign news (Barber & Weir, 2002; Carpenter, 2007; Pfau et al., 2004). The present study used the same search term, “North Korea”, as the one used for policy statements to identify stories on North Korea, including North Korea's politics, its international policies, or relationships with South Korea.

On the basis of prior research guidelines (Riffe, Lacy, & Fico, 2005), the context unit was the entire contents of the U.S. government's policy statements and The New York Times’ news stories concerning North Korea. The coding unit was a policy statement and a story, and the unit of analysis was the number of paragraphs containing frames. There were two considerations for adopting paragraphs as the unit of analysis. First, the pretest shows that the term, “North Korea”, tended to appear in small portions of a policy statement. In contrast, The New York Times’ stories dedicated virtually the entire content to North Korea because the country was the main focus of the stories. A direct comparison between policy statements and news stories required relatively large data points. In this regard, a paragraph would yield more data points than a statement or a story. Second, a paragraph has been accepted as the smallest unit of meaning because the meaning of a sentence is identified in the context of a paragraph (Jasperson, Shah, Watts, Faber, & Fan, 1998).

Independent Variables

On the basis of prior research (Gamson, 1992; Gamson & Modigliani, 1987, 1989), a frame of policy statements and stories was defined as a core organizing idea of constructing meanings and allowing for interpretations on North Korea. The frame was likely to appear frequently and repeatedly over time, which was defined as the magnitude of the frame, as shown in prior research (Entman, 2003).

As there were no guiding frames for this study due to limited research on President Bush's axis-of-evil speech on North Korea and subsequent media coverage, the present study used the constant comparative technique (Glaser & Strauss, 1967) in order to determine frame types. The constant comparative technique started from coding each case so that it captured as many coding categories as possible until no new categories emerged. The coding categories on North Korea which emerged from this technique included: (a) terrorism, nuclear threats, and evil country; (b) North Korean leaders’ suppressions of their people, North Korea's abductions of foreigners, and hardships of North Korean defectors; and (c) North Korea's readiness to engage in dialogue with the U.S., and its compliance with international treaties and regulations. The first, second, and third set of categories was labeled as respectively the “military threat” frame, the “human rights” frame, and the “dialogue partner” frame. The human-rights frame and the dialogue-partner frame were the non-military frame as opposed to the military-threat frame.

The military-threat frame portrays North Korea as a terrorism-sponsoring country that pursued or exported missile technology and weapons of mass destruction. One of the relevant policy statements was “The President was crystal clear in his State of the Union address about the growing danger posed by North Korea, Iran, and Iraq and states like these. These are brutally repressive regimes that are determined to acquire weapons of mass destruction” (Condoleezza Rice, Assistant to the President for National Security Affairs, April 8, 2002). One of the news stories mentioned that North Korea “still exports missiles and other arms, though in far smaller quantities than a decade ago; it may still be quietly working on its long-range missile program; it might even have a ‘basement bomb’ project to develop nuclear arms”. In another example of the use of the military-threat frame, North Korea “remains heavily armed and threatening, and largely as a result of its excessive military spending, its citizens are extremely poor.”

The human-rights frame focuses on the human rights abuse of the North Korean leadership such as suppressions of their people, North Korea's abductions of foreigners, and hardships of North Korean defectors. On February 20, 2002, President Bush mentioned, “I'm troubled by a regime that tolerates starvation. I worry about a regime that is closed and not transparent. I'm deeply concerned about the people of North Korea.” The New York Times article reported, “About 200,000 North Koreans are confined in a prison camp system where conditions are so harsh that an estimated 400,000 have died over the last 30 years.” Other articles reported on a German human rights activist who was expelled from North Korea for exposing the country's human rights violations to the world.

The dialogue-partner frame describes North Korea as a country with which the United States could resolve pending issues through dialogue because the country has complied with international regulations and a self-imposed missile moratorium, unlike pre-war Iraq and Iran. The U.S. government's dialogue-partner frame indicated that his administration was ready to talk with North Korea and would not attack the country: “We’re prepared to talk with the North about steps that would lead to a better future, a future that is more hopeful and less threatening” (President Bush, February 20, 2002). Another example was “We continue to await a response from North Korea to our longstanding proposal to meet with them” (Ari Fleisher, press secretary for President Bush, April 3, 2002). The dialogue-partner frame in The New York Times articles contained such descriptions as “North Korea has shown its ability to observe international agreements” and “North Korean leader Kim Jong-il's approach is the more appropriate and promising one.”

The authors completed the coding, and two trained coders coded a randomly selected 10 percent of the entire data to test intercoder reliability. Any coding disagreement was resolved by discussion, which yielded that Krippendorff's alpha (Krippendorff, 2004) was 0.83 for the military-threat frame, 0.80 for the human-rights frame, and 0.82 for the dialogue-partner frame.

Dependent Variable

A poll contained indicators of public opinion toward North Korea. The poll was obtained from “iPOLL” dataset at the Roper Center for Public Opinion Research, University of Connecticut. The poll was conducted by Harris Interactive from June 1, 2002 to June 30, 2002 and was sponsored by the Chicago Council on Foreign Relations. Through a national probability sampling, a total of 400 men and women 18 years of age and older participated in personal in-home interviews. Because the 400 people participated in the face-to-face interviews, the response rates were 100%. To assess the level of attention of American people to foreign news, the poll asked: “When you follow the news these days, how interested are you in reports about the relations of the U.S. with other countries?” The answer categories were: 0 = “No answers”; 1 = “Very interested”; 2 = “Somewhat interested”; 3 = “Hardly interested at all”; 4 = “Don't read newspapers”; 8 = “Not sure”; and 9 = “Decline to answer” (poll number: Q300A4). The category of “9” was coded as a missing value, which was also applied to the following two questions.

To test framing effects, the present study established such a temporal order that the poll was conducted after a series of the U.S. government's policy statements and The New York Times’ stories. Therefore, the temporal order allowed for a possible causal link between frames and public opinion. Regarding non-military solutions, the poll question asked: “On another issue, do you favor or oppose the use of economic sanctions against North Korea?” The answer categories were: 0 = “No answers”; 1 = “Yes, favor sanctions”; 2 = “No, oppose sanctions”; 8 = “Not sure”; and 9 = “Decline to answer” (poll number: Q675A). Regarding military solutions, the other poll question asked: “There has been some discussion about the circumstances that might justify using U.S. troops in other parts of the world. I’d like to ask your opinion about some situations. First, would you favor or oppose the use of U.S. troops if North Korea invaded South Korea?: The answer categories were: 0 = “No answers”, 1 = “Favor”; 2 = “Oppose”; 8 = “Not sure”; and 9 = “Decline to answer” (poll number: Q535A01).

Therefore, if the American public's attitudes toward North Korea were influenced by frames used by the U.S. government and the newspaper, their attitudes were likely to reflect such changes in the magnitude of the non-military frame and the military-threat frame. The dependent variable was the American public's specific attitudes toward North Korea, which were measured as the proportions of respondents who favored non-military solution and opposed military solution. If the proportions were significantly high, they were likely to be consistent with predictions made in the second and third hypotheses.

RESULTS

In order to construct discourse on North Korea during the four-month period ranging from January 29, 2002 to May 31, 2002, the U.S. government used 171 paragraphs (4.2 paragraphs per every three days) for 43 policy statements, and The New York Times used 535 paragraphs (13.1 paragraphs per every three days) for 85 stories.

The first research question asked what frames the U.S. government developed to construct discourse on North Korea since President George W. Bush's axis-of-evil speech. Table 1 shows that during the four months, the U.S. government developed three competing frames to construct discourse on North Korea: military threat, human rights, and dialogue partner. The three competing frames were present in discourse on North Korea in a disproportionate manner over time. From January 29 to February 28 and to March, the U.S. government framed policy statements concerning North Korea in terms of military threat, and North Korea as dialogue partner remained a competing frame during the same periods. However, the dialogue-partner frame was dominant in April. The visibility of the human-rights frame drastically decreased until it resurged as a dominant frame in May. The second research question asked what frames The New York Times developed to construct discourse on North Korea since President George W. Bush's axis-of-evil speech. The New York Times showed framing patterns dissimilar and similar to Bush's. During the first month, the newspaper used military threat and dialogue partner as two competing frames, but in March, human rights surged as a major frame while the previous two competing frames’ visibility noticeably decreased. However, as in the U.S. government's framing, dialogue-partner and human-rights frames dominated the newspaper's coverage in April and in May, respectively.

Table 1

U.S. government and The New York Times framing during 4 months of 2002

	January 29 to February 28	March	April	May	Total
U.S. government
Military threat	30 (49.2%)	32 (71.1%)	5 (15.6%)	0 (0%)	67 (39.2%)
Human rights	11 (18%)	0 (0%)	2 (6.3%)	25 (75.8%)	38 (22.2%)
Dialogue partner	20 (32.8%)	13 (28.9%)	25 (78.1%)	8 (24.2%)	66 (38.6%)
Total	61 (100%)	45 (100%)	32 (100%)	33 (100%)	171 (100%)
New York Times
Military threat	106 (45.3%)	41 (32%)	2 (3.3%)	28 (24.8%)	177 (33.1%)
Human rights	24 (10.3%)	68 (53.1%)	0 (0%)	68 (60.2%)	160 (29.9%)
Dialogue partner	104 (44.4%)	19 (14.8%)	58 (96.7%)	17 (15%)	198 (37.0%)
Total	234 (100%)	128 (99.9%)	60 (100%)	113 (100%)	535 (100%)

The first hypothesis posited that since President George W. Bush's axis-of-evil speech, the U.S. government frames were significantly correlated with The New York Times frames when they constructed discourse on North Korea. To test the hypothesis, this study used Spearman's rho correlations because rankings of frames were likely to capture frame competition. For each month, rankings of the three frames used by the U.S. government were compared with those of The New York Times, yielding Table 2.

Table 2

Correlations between the U.S. government and The New York Times frames

	January 29 to February 28 NYT frames	March NYT frames	April NYT frames	May NYT frames
January 29 to February 28
Governmental frames	1.0** (<0.001)	–0.50 (0.67)	0.50 (0.67)	–0.50 (0.67)
March
Governmental frames	1.0** (<0.001)	–0.50 (0.67)	0.50 (0.67)	–0.50 (0.67)
April
Governmental frames	0.50 (0.67)	–1.0** (<0.001)	1.0** (<0.001)	–1.0** (0.001)
May
Governmental frames	–1.0** (<0.001)	0.50 (0.67)	–0.50 (0.67)	0.50 (0.67)

Note: **Correlation is significant at the 0.01 level (two-tailed) and p-values are in the parentheses.

There were perfect correlation coefficients because the sample size was only three, given that there were three frames. Nonetheless, they were significant, supporting the first hypothesis. These findings prompt the need for consideration of both Table 1 and Table 2. First, when the government dropped the human-rights frame to describe North Korea in March, The New York Times stopped using the same frame in the following month, although it was dominant in the newspaper in March. This indicates the influence of the U.S. government on the newspaper's framing patterns of North Korea. Conversely, the newspaper continued to reduce usage of the military-threat frame the first month and in April, the proportion of the military-threat frame was down to only 3.3 percent of the entire frames. The U.S. government started to reduce the proportion of the military-threat frame in April, and in May, it dropped the frame from discourse on North Korea. This shows that the newspaper influenced how the government framed North Korea. Accordingly, these findings reveal that there were reciprocal influences between the U.S. government and the newspaper.

The second hypothesis expected that if the U.S. government shifted framing patterns of North Korea from a military-threat frame to a non-military frame over time, and the American public paid attention to the non-military frame, the public was likely to favor non-military solutions. The third hypothesis predicted that if The New York Times shifted framing patterns of North Korea from the military-threat frame to non-military frame over time, and the American public paid attention to the non-military frame, the public was likely to favor non-military solutions. The Harris poll data found that 62 percent of 400 American respondents showed a high level of attention to news about U.S. relations with foreign countries and 31 percent of them paid a substantial level of attention to it. Instead of choosing the “no answer” category, 0.8 percent of the total respondents chose the “decline to answer” category, which was coded as missing values. Therefore, the American people were likely to follow news about U.S.–North Korean relations.

As Table 3 shows, 58.8 percent of the survey respondents favored economic sanctions, and only 32.5 percent of the respondents favored military solutions. Because the respondents answered the two questions in the same survey, however, these decisions obscured the hypothesized framing effects. Table 4 breaks down responses given by the respondents on the two questions in terms of nine attitudinal patterns. The missing values were eliminated because they remained less than 1 percent of the total responses per each question. Accordingly, the total respondents for analysis were 394.

The nine attitudes were significantly related to the two types of solutions, indicating preference for economic sanctions over military solutions [χ² (1, df= 4) = 45.277, p = 0.001]. This means that the preference for economic sanctions and opposition of military solutions (attitude 1) recorded the highest proportion (31.5 percent), followed by preference for both non-military and military solutions (23.4 percent) and opposition of the two solutions (15.5 percent). The preference for military solutions and opposition to economic sanctions (attitude 5) was only 7.1 percent of the total responses. Before the poll, the U.S. government shifted framing patterns from the military threat of North Korea to dialogue partner and human rights two months before the poll was conducted. The New York Times changed framing patterns in the same way as the U.S. government.

Table 3

American public's attitudes toward North Korea

	Favor	Oppose	Not sure	Decline to answer	Total
Economic sanctions	235 (58.8%)	99 (24.8%)	62 (15.5%)	4 (1.0%)	400 (100.1%*)
Military solutions	130 (32.5%)	218 (54.5%)	50 (12.5%)	2 (0.5%)	400 (100%)

Note: Asterisk reflects a rounding error.

Table 4

Specification of the American public's attitudes toward North Korea

Attitudes	Economic sanctions	Military solutions	Frequency (%)
Attitude 1	Favor	Oppose	124 (31.5)
Attitude 2	Favor	Favor	92 (23.4)
Attitude 3	Oppose	Oppose	61 (15.5)
Attitude 4	Not sure	Oppose	29 (7.4)
Attitude 5	Oppose	Favor	28 (7.1)
Attitude 6	Not sure	Not sure	23 (5.8)
Attitude 7	Favor	Not sure	18 (4.6)
Attitude 8	Not sure	Favor	10 (2.5)
Attitude 9	Oppose	Not sure	9 (2.3)
Total			394 (100.1*)

Note: χ² (1, df = 4) = 45.277, p = 0.001. Asterisk reflects a rounding error.

In this respect, the reason why the preference of non-military solutions to North Korea was highest was that the American public's attitudes toward the country followed changes in frames used by the government and the newspaper. The survey question (poll number: Q535A01) assumed a hypothetical situation such as North Korea invading South Korea. However, it is hardly a feasible option to attack North Korea when the country is a dialogue partner or when the country is criticized mainly for violating human rights in terms of international standards. A realistic option would be non-military solutions such as international warnings or economic sanctions. Therefore, the second and third hypotheses were supported.

DISCUSSION

The present study assesses the extent to which frame competition exists in U.S. government policy statements and a major newspaper's coverage regarding North Korea. It also examines how frame building by the two counterparts influences the American public's attitudes toward the country. The U.S.–North Korean relationship provides interesting contexts regarding frame building and framing effects. Following President Bush's axis-of-evil speech, both the U.S. government and The New York Times actively developed specific frames that were applied to North Korea. Three frames—military threat, human rights, and dialogue partner—competed in the government's and the newspaper's discourse on North Korea, with the magnitude of each frame changing according to political and security conditions of a specific period. Information on North Korean issues is not obtained through people's direct experiences. Accordingly, the American people rely on competing frames for interpretation, and reflecting changes in those competing frames, the American public preferred non-military solutions to North Korea.

These findings make three meaningful contributions to existing framing studies. First, the present study provides empirical evidence for Entman's cascading network activation model, and in particular, his proposition of two-way frame flows between government and the news media. Entman (2004) argues that hegemony and indexing models fall short of explaining the media-government relationship in the post-Cold War era. Hegemony and indexing approaches suggest that the media generally mirror or index the dialogue of elites, and thus lack independent and critical coverage of foreign policy issues. In comparison, the cascading network activation model emphasizes the increased independent influence of the news media on foreign policy issues (Entman, 2004). That is, the White House no longer holds hegemonic control over foreign policy discourse, and the news media have enhanced their power to challenge the White House's framing of foreign affairs issues, although Entman assumes that government frames have significant influence on news media frames.

In the present study, as the government reduced the magnitude of the human-rights frame, the newspaper dropped the same frame in the following month. Conversely, as the newspaper significantly decreased the usage of the military threat frame, the magnitude of the frame completely disappeared in government policy statements in a later month. In this light, in foreign policy areas, the frame flows are not one direction from government to the news media, but instead, the two actors influence each other. This reciprocal relationship is convincing because government and the news media are major actors who circulate powerful frames and meanings. It will be interesting to study conditions under which the news media initiate frame building.

Second, the present study contributes our understanding of frame competition, which Chong and Druckman (2007) and Entman (2003) suggest is a key consideration for studying framing effects. As Chong and Druckman (2007) point out, most framing effects studies have focused on a single frame, excluding implications of frame competition. For example, in examining the effects of framing stem-cell funding, Shen (2004) exposed people to only one frame such as a benefit frame or an ethics frame instead of allowing people to receive the two frames at the same time. Unlike most social concerns, which people are able to learn by observing their community and having discussions with their neighbors, interpretation of foreign affairs is a demanding task for the general American public because foreign issues are beyond their personal experiences and their interpersonal network does not help much.

Given this situation, competing frames provide the public with cues and guidelines for a better understanding of foreign countries. Thus, people pay attention to competing frames in either unequal or equal amounts. The present study is the first of its kind to analyze specific frames on North Korea from the perspective of frame competition, providing a foundation for future research on similar topics. Since Bush's axis-of-evil speech, the American people pay attention to the frames of military threat, human rights, and dialogue partner in government policy statements and news coverage. As U.S.–North Korean relations evolved into a specific situation, the stage of frame competition changed into frame dominance, such that the dialogue-partner frame was most prominent two months before the public opinion survey, followed by the dominance of the human-rights frame during the month before the survey. Interestingly, frame dominance over time is more apparent in the U.S. government than in The New York Times. In terms of its magnitude, the U.S. government preferred the military threat in March, dialogue partnering in April, and human rights before the survey. The three frames were relatively competitive in newspaper coverage during the same period. The American people who closely follow those frames are likely to store the frames in their memory. While people respond to survey questions asking them to make choices between non-military solutions and military solutions, the frames are likely to activate thoughts related to non-military solutions in the people's memory. As Simon and Xenos (2000) point out, the frames allow people to connect between the thoughts and related ideas. Here, the thoughts and ideas were of North Korean issues, and in this case, non-military solutions were preferred over military solutions.

Finally, the present study examines framing effects in real settings as opposed to experimental settings. It shows a realistic picture of how political actors in foreign policy areas exerted influence on the American people's perceptions toward North Korea. Accordingly, the findings improve the present study's ecological validity.

It should be noted, however, that the present study cannot argue for explicit causality because there was no manipulation of frame competition and exposures to frames. Prior researchers (Saris & Sniderman, 2004) conducted experiments to determine the effects of dual frames on attitudes toward government spending on the poor. This indicates that a specific time lag should be considered when establishing causality. Another limitation is that the research period of four months is rather short. Future studies can examine a longer time period to examine the effects of competing frames on foreign policy issues. Therefore, the duration of a time lag for the strength of framing effects can be identified. It will also be useful to analyze those issues before and after Bush's axis-of-evil speech. Attention to specific news media, media content consumption habits (Bennett, Flickinger, Baker, Rhine, & Bennett, 1996; Drew & Weaver, 1990; Pan, Ostman, Moy, & Reynolds, 1994), and individuals’ expertise (Peffley & Hurwitz, 1992) predict the amount of knowledge of current issues and changes in attitudes. Follow-up studies need to examine how these variables mediate framing effects on the public's attitudes and behaviors toward foreign countries or foreign leaders.

In sum, the U.S. government and a major newspaper actively engaged in two-way frame flows regarding North Korea. Over time, the U.S. government's framing patterns of the country guided a major newspaper to frame North Korea. The major newspaper also influenced how the U.S. government portrayed the country. Two-way frame flows-occur when multiple frames compete in the foreign policy arena. People receive these competing frames over time. Accordingly, people who closely follow policy statements and news coverage are likely to acquire these mediated opinions, and develop thoughts and meanings consistent with the frames.