Anonymous attribution and official news sources in the reporting of Bosnia and Somalia : a study of conflict coverage in American newspapers

This study examines source attribution in press reports of military conflict. Several issues in media sociology are addressed: (1) Whether anonymous attribution is more prevalent when American forces are committed to an active area; (2) whether certain types of news sources tend to be veiled more frequently; (3) whether differences exist across news organizations in the use of veiled sources; (4) whether reports originating from American or international locales contain more anonymous attribution; and (5) whether differences exist across Republican and Democratic administrations in the release of anonymous information.

In addition, the study takes a macroscopic look at the extent to which journalists rely on official news sources for information about military conflict. At issue is the homogenous news that can develop when a narrow band of sources provide information to several news organizations.

Reports from the Associated Press, the Los Angeles Times and the Washington Post concerning the military conflicts in Bosnia and Somalia were examined. Nearly 9,000 news paragraphs from the years 1992, 1993 and 1994 were coded for six categorical variables--Administration, Conflict, Dateline, Level of Attribution, Newspaper and Source Type. Among the central findings were significant differences across news organizations in the use of anonymous attribution, a dominance by officials as Western news sources, and consistency across news organizations in the frequency of Source Type.

The study makes a methodological contribution to the mass communication literature by using logistic regression analysis to test the effects of four predictor variables on an ordinal-level response variable. Specifically, fractions were created in which the number of paragraphs containing anonymous attribution in a given article served as the numerator and the number of paragraphs containing anonymous attribution plus the number of paragraphs identifying the source served as the denominator. The fractions were truncated, rounded down to the nearest whole number, and the resulting values formed the dependent measure. Scholars who examine source attribution in Western newspapers are encouraged to use the method in future analyses.