News media tracking of tobacco control: a review of sampling methodologies

Abstract

News media tracking can be used to understand the content of news coverage and quantify its influence on public opinion and the social environment. We hypothesize that a parsimonious sampling of public health news coverage can gather statistically equivalent information to capture a census of coverage. We tested two general approaches to sampling. First, we randomly selected articles within a timeframe. Second, we randomly selected dates during 2000. None of these strategies showed a substantial deviation from the 2000 census of articles. We conclude that sampling coverage can produce estimates statistically equivalent to a census. Researchers should utilize coverage sampling.