Anonymous Attribution During two Periods of Military Conflict: Using Logistic Regression to Study Veiled Sources in American Newspapers

Abstract

In this study, nearly 9,000 news paragraphs concerning the military conflicts in Bosnia and Somalia were coded for a series of variables. The use of anonymous attribution varied significantly in Associated Press, Los Angeles Times, and Washington Post content. In addition to providing valuable information about the use of sources, the study also contributes to communication methodology by using logistic regression analysis to test the effects of four predictor variables on an ordinal-level response variable. Scholars who examine source attribution in Western newspapers are encouraged to use the method in future analyses.