This research addresses one aspect of the above mentioned process, i.e., the international news coverage of a mainstream American newspaper, namely The Wall Street Journal. The goal of the project is to content analyze the news coverage of six regions (i.e., Americas & the Caribbean, Asia, Africa, Middle East, Western Europe and Eastern Europe) in an effort to understand the process underlying international news coverage. Given the economical and political dynamics of these regions in relation to the United States, this study was designed to address the following questions regarding international news coverage in The Wall Street Journal: Who gets covered and how much, what gets covered, and what are the factors that determine the news coverage?
Many studies have investigated the issue of foreign news coverage, and the determining factors, in the United States' media. Stimulated by UNESCO's discussion on the issue of "imbalances" in the international news sources, topics, and distribution, studies have investigated the amount of coverage for different countries or regions in the world.4 The findings suggest an imbalance in global information attributed to global communication being controlled by a few Western-based international agencies, and media multinationals distributing relatively much less news about Third World" countries. Additionally, news about Third World countries is invariably framed in a Western ideological or cultural perspective, which in part leads to highly stereotyped accounts of only a few types of events such as coups and earthquakes.5 Research also shows that in the U.S. media, foreign news is limited to events that involve the United States, such as Vietnam in the 1960s and Central America and the Middle East in the 1970s and 1980s.6
With the dramatic change in the international political system, marked by the collapse of Communism in Central and Eastern Europe, and the demise of the Soviet Union, the economic aspect of foreign policy is increasingly emphasized.7 Therefore, it is important to investigate how "economy/business"-oriented newspapers cover international news. To that end, The Wall Street Journal, a mainstream business-oriented newspaper with national/international prominence and broad circulation, was chosen for the present analysis.
To preclude any concerns that The Wall Street Journal may deal strictly with economic news, the task of the present analysis is to classify the various types of news that are covered in The Wall Street Journal. After such a classification is accomplished and it is shown that a wide range of topics are indeed covered, it then becomes important to identify the determining factors that can explain such international news coverage in The Wall Street Journal.
Others have studied the contextual constraints that influence the content and structure of foreign news coverage in newspapers: Charles, Shore and Todd analyzed the New York Times and noted that trade is a major predictor of media attention in international news coverage;8 and Hester elaborated further by suggesting four determinants of information flow based on international relations - the hierarchy of nations (i.e., geographic size, population, economic development, and length of existence as a sovereign nation), cultural affinities (i.e., shared language, the amount of migration, intermarriage, travel between countries, and historical relationships such as mother country-colony status), economic associations (international trade, the amount of foreign aid, and business investment) between nations, and news and information conflicts.9 Other determining factors include international news agencies, foreign correspondents,10 newspaper size and ownership,11 marketplace pressure or audience factors,12 deviance of an event, relevance to the United States, potential for social change, and geographical distance.13
In sum, factors that affect the international news can be divided into two classes: 1) internal determinants such as in editorial weighting, and marketplace pressure or audience factors, major international news agencies, and newspaper size and ownership, and 2) external determinants such as geographic size, population, economic development, international trade, business investment, and the amount of foreign aid. For this study, since the analysis is based on a single business-oriented newspaper, internal factors such as newspaper size and ownership have little meaning. Instead external determinants (e.g., population size, area, gross national product, export, and import) are focused upon to identify factors that determine foreign news coverage.
Therefore, the research questions are:
Research Question 1: What is the international news coverage pattern? Specifically, RQ1-1: What is the distribution of news coverage for different regions in the world?
RQ1-2: What topics get covered for different regions in the world, is there a discernable pattern, and does it differ from one region to another?
Research Question 2: What are the factors that determine the distribution of international news coverage?
Specifically, RQ2: how well can factors such as gross national product, export, import, population size, and geographical size of a country, predict the amount of coverage?
To examine these questions, a total of 74,520 news articles that appeared in The Wall Street Journal for the period of April 1, 1990 through March 31, 1992, were analyzed.14 The newspaper was available on-line through TIPSTER Information Retrieval Text Research Collection.15 A computer-assisted content analysis was performed with Document Retrieval using LINguistic Knowledge (DR-LINK), which is a natural language text retrieval system that retrieves documents on the basis of "what is meant" in a query, not simply the words in the query. The use of computer-assisted content analysis is not new; the effective capabilities of the machine coding method have been demonstrated with considerable reliability when compared to human coding.16 The use of computer-assisted methods offers promising possibilities in analyzing large, on-line, electronic data bases of news articles, particularly for its speed and consistency of coding, especially when the goal is to obtain aggregated measures of content.
International News. News articles that mention at least one country name other than the United States are defined as international news.
Coverage Distribution and Size. The amount of coverage was measured by:
1. Number of articles: the number of articles mentioning a given country (or region).
2. Word-count: size of the articles, as measured by word-count, for a given country (or region).
3. Average article size: total word-count divided by the number of articles dealing with a given country (or region).
Geographic Regions. To analyze the geographic range of international news coverage, following Hess's classification, the world was divided into six regions on the basis of cultural, and political geography:17 Americas/Caribbean, Asia, Africa, Middle East, Western Europe, and Eastern Europe. Given that the study was concerned with international news and was based on American press, the United States was excluded from the analysis. In this study Africa refers to the sub-Saharan Africa, with the northern tier of the continent (Morocco, Algeria, Tunisia, Libya, and Egypt) assigned to the Middle East. Eastern Europe consists of the countries that were communist before 1989, i.e., Armenia, Azerbaijan, Bosnia, Ukraine, etc.
The Subject Categories. Twelve topics were addressed:
1. Accident/disasters: Earthquakes, famine/droughts, floods, building collapses, and airplane crashes.
2. Military affairs: References to national security, armed conflict or threat of conflict, peace movements, negotiations, settlements, army, navy, air-force, combat, war, defense, arms deals, and weapons.
3. Crime: References to nonpolitical crime, police, judicial and penal activity, court proceedings, the accused, victims, or homicide.
4. Foreign relations/diplomacy/political activity between countries: Reference to related international negotiation, foreign delegation, and negotiation.
5. Domestic politics/government: References to politics related to, or any politician or a person running for office, elections, campaigns, appointments, the performance of elected and appointed officials, leadership, internal conflict or crisis, government changes, and legislation.
6. Economic matters: References to money matters, business activities, international trade, imports, exports, trade balances, capital investment, stock issues, state investments, economic performances, industrial projects, factories, dams, ports, roads, agricultural matters, employment, and infrastructure.
7. Sports: References to any sports-related information, international sports events, athletes, and sports games.
8. Science and Technology: References to computers, scientific breakthroughs, information, technology and environmental news.
9. Culture/Education: References to cultural artifacts, festivals, cultural history; public, elementary, secondary and post secondary education, college, university, and programs affecting school systems.
10. Health: References to health problems, disease, emergency medical services, doctors, hospitals, and cures.
11. Human rights: References to human rights and political amnesty.
12. Social Issues: References to other social issues that do not belong to any of the above categories, such as feminist movement, racism, etc.
To investigate the factors that determine the international news coverage in The Wall Street Journal, a multivariate and a univariate regression analysis was conducted. A multicollinearity test was conducted on the predictor variables (below) and the highly collinear variables were combined into a single variable.18 This was sufficient to allow for the development of a multivariate regression model for the prediction of coverage based on the predictor variables. However, the problem of multicollinearity was not entirely resolved in this way because the remaining variables were still correlated with one another, though to a lower degree. Further reduction of the collinearity would call for taking further linear combinations of the remaining variables. This, however, would yield variables with no apparent interpretation. For such reasons, no further reduction of the collinearity was attempted, leading to the following four predictor variables for news coverage:
1. Population: Figures are estimates from the Bureau of the Census based on statistics from population censuses, the vital statistics registration system, or sample surveys pertaining to the recent past and to assumptions about future trends.
2. Area: Total area is the sum of all land and water areas delimited by international boundaries and/or coastlines.
3. Gross National Product (GNP): The value of all final goods and services produced within a nation in a given year, plus income earned abroad, minus income earned by foreigners from domestic production.
4. Trade: The sum of exports and imports19, with the former being the value of all goods and nonfactor services provided to the rest of the world, including merchandise, freight, insurance, travel, and other nonfactor services, and the latter being the value of all goods and nonfactor services provided by the rest of the world, including merchandise, freight, insurance, travel, and other non-factor services.
The original set of five predictor variables was based on previous studies.20 This information was drawn from the Central Intelligence Agencys (1994) The World Factbook.
A multivariate regression model is suitable if the aim is to develop a model for the prediction of coverage based on the predictor variables. However, if the aim is to examine the predictive strength of the predictors, then such a model is misleading at best. It is well known21 that the regression slopes in a multivariate analysis are entirely meaningless if there exists any collinearity among the predictor variables. Briefly, the reason is that any collinearity between, say, two predictors can be expressed as a linear relation between the two; as a result only a particular linear combination of the corresponding slopes can be fixed in a least-squares regression analysis. For this reason the interpretation of the multivariate regression slopes as measures of predictive strength is unreliable, at best, even when all the variables have been standardized. Only the univariate regression coefficients, or, equivalently, the correlation coefficients between the various predictors and the dependent variable, are reliable measures of predictive strength. Consequently, four univariate regression analyses were performed between each of the four predictors and the amount of coverage.
From a total of 74,520 news articles from The Wall Street Journal (April 1, 1990 through March 31, 1992), 33,159 articles dealt with international news (about 44% of total news).
The international news coverage pattern of The Wall Street Journal is presented in Table 1 which shows the distribution of news coverage for different regions in terms of the number of articles and word-count. Western Europe leads the list with 37,057 articles (37%of total international news), followed by Asia (24%), Middle East (13%), Eastern Europe (12%), and Americas/Caribbean (12%). African news occupies only 3%of the total international news coverage. The results in Table 1 are statistically significant at the 95% level, as indicated by the confidence intervals.
In regards to the amount of coverageas measured by word-count, again, Western Europe has the largest volume, following the same rank order as the number of articles (Table 1). However, when average article sizes are compared, the Middle East receives more coverage than Asia and Americas/Caribbean. For the rest of the regions, the differences are statistically nonsignificant at the 95% confidence level.
Table 2 shows the regional distribution of the topics featured. For comparison, the percentage (out of total number of articles within a region) rather than the number of articles is reported. Interestingly, The Wall Street Journal displays two patterns of dealing with topics; Western Europe, Asia, and Americas/Caribbean show one pattern, and the Middle East, Eastern Europe, and Africa, show another. In the former, economic matters dominate news coverage, ranging from 66% to 70%. Drastically smaller percentages are devoted to domestic politics or government-related news, ranging from 11% to 15%, followed by foreign relations/diplomacy related topics (about 8%), and finally, military affairs (about 4%). By contrast, for the Middle East, Eastern Europe, and Africa, economic matters occupy only 37% to 39%, about half that of the former regions. This is followed by domestic politics/government related topics (24% to 29%), and foreign relations/diplomacy related topics (16% to 21%). However, as in the first pattern, military affairs rank fourth. In short, for the former regions, economic matters dominate the news coverage while for the latter regions relatively equal weight is given to economic matters and domestic politics/government issues. Such results further alleviate concerns over the apparently economic nature of The Wall Street Journal.
To investigate the factors that determine the international news coverage in The Wall Street Journal, multivariate and univariate regression analyses were conducted on the four predictor variables (population, area, GNP, and trade) and the number of articles as dependent variable. The multivariate regression model (Table 3) accounts for 78% of the variance in the news coverage, with an F-value of 175.19, significant at p<0.0001 level.
To test whether the nature of the predictive relationship differs from region to region, multiple regression analysis was conducted separately for each region (Table 3). For Western Europe, the four predictor variables account for 74% of the varianc; for Asia, 90%, and for Americas/Caribbean, 99%, indicating that these variables can predict very accurately the news coverage patterns in The Wall Street Journal.
Regarding the strength of the predictors, as previously mentioned, the regression slopes do not represent the relative importance of the predictors. For that purpose, the appropriate quantities are the univariate regression slopes, or equivalently, the correlation coefficients between the respective predictors and the amount of coverage. It is found that the best predictor is trade (r2 = 0.74), followed by GNP (r2 = 0.26), population22 (r2 = 0.21), and area (r2 = 0.11). As expected, the amount of variance explained by each univariate model is less than that of the multivariate model (R2=0.78), though trade alone does explain a considerable amount of the variance. All of these results are statistically significant at the p<0.0001 level.
Summary and Discussion
The present examination of the international news of The Wall Street Journal confirms that a large amount of news coverage is on Western Europe. Asia, ranks second in total coverage, followed by the Middle East, Eastern Europe, the Americas & the Caribbean, and Africa. This confirms earlier findings that the distribution of international news coverage by region is unbalanced in that First World (or developed) countries are reported disproportionately more.23
As for the topics, the international news in The Wall Street Journal shows two distinct patterns. For Western Europe, Asia, and the Americsa/Caribbean, the newspaper shows a narrow scope of foreign news choices, namely economic matters. To a much lesser degree the next topic relates to domestic politics/government news. Other broader and more balanced arrays of subjects such as education, culture, and health are mentioned rarely. This strong emphasis on economic matters for the these regions can be due to the regions' economic growth and economic significance in international relations with the United States. On the other hand, the heavy coverage given to the Middle East, Eastern Europe, and Africa, focuses on domestic political news/government or foreign relations.
As for the determining factors for the international news coverage, while a multiple regression analysis shows that the four predictor variables accurately predict the overall international news coverage (R2=.78), a univariate regression analysis shows that trade is the best predictor, followed by GNP. The population and area of a country have less of a contribution to predicting the coverage. The multivariate models developed for each of the six regions are found to have comparable or better success in predicting coverage in the respective regions.
For a better understanding of these findings, one would need to a) extend the analysis to other newspapers, b) extend the analysis to data including the period before and after the collapse of communism, and c) include other determining factors such as internal factors (e.g., editorial weighting, marketplace pressure, and newspaper size and ownership, etc.) that may dictate the coverage of international news in the media.
1. David L. Protess, and Maxwell McCombs, Agenda Setting: Readings on Media, Public Opinion, and Policy making (NJ: Lawrence Erlbaum Associates).
2. Everett M. Rogers, James W. Dearing, and Soonbum Chang, "Aids in the 1980s: The Agenda-Setting Process for a Public Issue," Journalism Monographs 126, (April 1991).
3. Herbert J. Gans, Deciding What's News: A Study of CBS Evening News, NBC Nightly News, Newsweek and Time (NY: Pantheon Books, 1979), 78-115; Doris A. Graber, Processing the News: How the People Tame the Information Tide, 2d ed. (NY: Longman, 1988), 252-54; Bruce Russett, Controlling the Sword: The Democratic Governance of National Security (MA: Harvard University Press, 1990); Thomas Risse-Kappen, "Public Opinion, Domestic Structure, and Foreign Policy in Liberal Democracies," World Politics, 43 (July 1991), 479-512.
4. UNESCO, Many Voices, One World (Paris: UNESCO, 1980); Teun A. van Dijik, News Analysis: Case Studies of International and National News in the Press (NJ: Lawrence Erlbaum Associates, 1988); Erwin Atwood, Suart Bullion, and Sharon Murphy, International Perspectives on the News (IL: Southern Illinois University Press, 1982); Altaf Gauhar, "Free Flow of Information: Myths and Shibboleths," in Third World Strategy: Economic and Political Cohesion in the South, ed. Altaf Gauhar (NY: Praeger, 1983), 66-90.; George Gerbner and George Marvanyi, "The Many Worlds of the World's Press," Journal of Communication 27 (Winter 1977): 52-66; D. R. Mankear, Whose Freedom? Whose Order? A Plea for a New International Information Order by the Third World (New Delhi: Clarion Books, 1985).
5. Paul Morris Hirsch, "Occupational, Organisational and Institutional Modes in Mass Communication," in Strategies for Communication Research, eds. Paul Morris Hirsch, Peter V. Miller, and Gerald F. Kline (CA: Sage, 1977), 13-42; Wilber Schramm and Erwin Atwood, Circulation of News in the Third World: A Study of Asia (Hong Kong: Chinese University Press, 1981).
6. Jim A. Hart, "Foreign News in U.S. and English Daily Newspapers" Journalism Quarterly 43 (1966): 443-448; R. D., Casy and T. H. Copeland, "Use of Foreign News by 19 Minnesota Dailies" Journalism Quarterly 35 (1958): 87-89; R. G. Hicks and A. Gordon, "Foreign News Content in Israeli and U.S. Newspapers" Journalism Quarterly 51 (1974): 638-644; John A. Lent, "Foreign News in American Media" Journal of Communication 27 (Winter 1977): 46-51; Doris A. Graber, Processing the News: How the People Tame the Information Tide, 2nd ed. (NY: Longman, 1988).
7. Michael Cox, US Foreign Policy after the Cold War: Superpower Without a Mission? (UK: The Royal Institute of International Affairs, 1995)
8. Jeff Charles, Larry Shore and Rusty Todd, "The New York Times Coverage of Equatorial and Lower Africa," Journal of Communication 29 (spring 1979): 150-157.
9. Al Hester, "Theoretical Considerations in Predicting Volume and Direction of International Information Flow," Gazette 19 (winter 1973): 239-247.
10. Robert William Desmond, The Information Process: World News Reporting to the Twentieth Century (Iowa City: University of Iowa Press, 1978); Robert William Desmond, Windows on the World (Iowa City: University of Iowa Press, 1980).
11. Gans, Deciding What's News; Tony Nnaemeka and Jim Richstad, "Internal Controls and Foreign News Coverage: Pacific Press Systems," Communication Research 8 (spring 1981): 97-135; Gertrude Joch Robinson and Vernone M. Sparks, "International News in the Canadian and American Press: A Comparative News Flow Study," Gazette 22 (1976): 203-218.
12. Nnaemeka and Richstad, "Internal Controls," 97-135; Robinson and Sparks, "International News, " 203-218.
13. Tsan-Kuo Chang, Pamela J. Shoemaker, and Nancy Brendlinger, "Determinants of International News Coverage in the U.S. Media," Communication Research 14 (Winter 1987): 396-414; Nnaemeka and Richstad, "Internal Controls," 97-135.
14. Articles from October 1990, November 1990, and September 1991 were not included in the analysis because there were not available in the TIPSTER Information Retrieval Text Research Collection provided by the Software and Intelligent Systems Technology Office of the Advanced Research Projects Agency (ARPA/SISTO).
15. TIPSTER Linguistic Data Consortium (LDC) is an open consortium of universities, companies and government research laboratories that creates, collects and distributes speech and text databases, lexicons, and other resources for research and development purposes. The University of Pennsylvania is the LDC's host institution.
16. Roderick P. Hart, Verbal Style and the Presidency. A Computer-Based Analysis (NY: Academic Press, 1984); Roderick P. Hart, "Systematic Analysis of Political Discourse. Development of DICTION," in Political Communication Yearbook, eds. Keith R. Sanders, Linda L. Kaid, and Dan Nimmo (IL: Southern Illinois University); C. E. Hicks, E. Rush, and S. M. Strong, "Content Analysis," in Subject and Information Analysis, ed. Eleanor D. Dym (NY: Marcel Dekker, 1984); Naomi Sager, "Natural Language Analysis and Processing," in Subject and Information Analysis, ed. Eleanor D. Dym (NY: Marcel Dekker, 1985); Philip J. Stone, Dexter C. Dunphy, Marshall S. Smith, and Daniel M. Ogilvie, The General Inquirer: A Computer Approach to Content Analysis (MA: M.I.T. Press, 1969).
17. Stephen Hess, International News & Foreign Correspondents (Washington, D.C.: The Brookings Institution, 1996).
18. As a way of solving the multicollinearity, Stevens suggested combining predictors that are highly correlated. James Stevens, Applied Multivariate Statistics for the Social Sciences, 3rd ed. (NJ: Lawrence Erlbaum Associates, 1996).
19. A multicollinearity test showed that exports and imports were highly correlated (i.e., r = .98; variance inflation factor for exports was 29.63, and imports, 28.49). As a way of resolving the collinearity between exports and imports, a new variable "trade" was constructed by combining the values of each variable.; According to Myers, if any variance inflation factor exceeds 10, there is reason for concern over multicollinearity. Raymond H. Myers, Classical and Modern Regression with Applications, 2nd ed. (MA: Duxbury Press, 1990).
20. Charles, Shore and Todd, "The New York Times," 153; Hester, "Theoretical Considerations" 241.
21. Norman R. Draper and H. Smith, Applied Regression Analysis, 2nd ed. (NY: John Wiley & Sons, 1981).
22. There are two outliers (China and India) in the distribution of population, both of which were removed from the univariate analysis. Prior to the exclusion of these two data points, the value of r2 was 0.04.
23. van Dijik, News Analysis 41; Hart, "Foreign news in U.S." 446; Hicks and Gordon, "Foreign news content," 642.
International News Coverage in the Wall Street Journal, by World Region.
Number of Articles Number of Words
Region Total SE* 95% CI** Total Mean SE* 95% CI**
Western Europe 37,057 255 36558/37556 34,062,427 919 421 94/1744
Asia 23,830 56 23719/23941 22,815,805 957 76 808/1106
Middle East 12,877 139 12605/13149 14,953,173 1,161 8 1146/1176
Eastern Europe 12,063 106 11855/12271 13,187,416 1,093 103 892/1294
Americas/Caribbean 11,662 124 11418/11906 9,998,365 857 200 6571057
Africa 2,807 3 2801/2813 3,123,797 1,113 74 967/1259
* SE = Standard Error. ** CI = Confidence Interval
International News Topics, by Region (in %)
Westen Middle Eastern Americas/
Subject Europe Asia East Europe Caribbean Africa
Accidents/Disasters 1.2% 1.5% 5.2% 2.5% 1.6% 2.4%
Military Affairs 3.8 4.0 12.8 9.4 3.7 7.8
Crime 0.4 0.3 0.8 0.6 0.5 0.6
Foreign Relations/Diplomacy 7.6 7.6 16.2 15.5 8.3 20.6
Domestic Politics/Government 11.1 13.5 24.0 28.7 15.0 26.3
Economic Matters 69.5 67.2 38.2 38.8 66.2 36.8
Sports 0.8 0.7 0.9 1.1 0.8 0.5
Science & Technology 1.7 2.0 0.6 0.9 1.1 0.6
Education 1.0 0.8 0.5 0.9 0.6 0.6
Culture 1.1 0.8 0.4 0.6 0.4 1.3
Health 0.0 0.5 0.3 0.4 0.9 0.4
Social Issues 1.0 1.1 0.3 0.6 1.0 2.0
Total Number of Articles 37,057 23,830 12,877 12,063 11,662 2,807
Multiple Regression Analysis for Number of Articles as Dependent Variable.
All Regions W. Europe Asia Middle East E. EuropeAmericas/Caribbean Africa
ß t ß t ß t ß t ß t ß t ß t
Population -.05 -1.29 .22 0.99 .02 0.19 .43 1.28 .23 0.56 -.15 -4.12 -.13 -0.08
Area .23 6.28 -.03 -0.28 .03 0.35 -.42 -1.06 .49 1.63 .17 2.77 .23 -0.79
GNP .06 1.44 .06 0.49 .12 1.47 .22 0.83 .05 0.59 -.02 -0.74 -.04 1.74
Trade .80 18.64 .63 2.66 .85 10.20 .49 1.52 .24 1.07 .91 19.18 .69 -0.27
R2 .78 .74 .90 .27 .91 .99 .50
F 175.19* 17.76* 86.53* 1.18* 41.53* 643.00* 10.49*
N 198 29 43 17 21 37 46
ß= Standardized regression coefficients; t = t-value; *p<.0001
Paik is assistant professor in the Department of Communication, University of Oklahoma.