| Summary: | While considerable attention has been focused in recent years on questions of selection and coverage biases in media-derived protest data, little has been said about the procedures used to sample media records and scant attention has been paid to non-English-language news sources. This article outlines a Spanish-language keyword search protocol for finding protest event articles, and systematically evaluates it in terms of search validity and efficiency using a two-year dataset constructed from a full-text article search of the Mexican national news agency Noticias de Mexico. Results suggest that this protocol generates data with a high degree of search validity with respect to locating reports of conventional protest event forms, but is problematic with respect to the identification of unconventional protest forms. The significance of these findings for the broader investigation of media-derived protest data is discussed, and directions for future research are outlined.
|
|---|