Fluoridation advocacy in referenda where media coverage is balanced yet biased

Background. Despite supporting scientific evidence, community water fluoridation (CWF) often fails in public referenda. To understand why, the authors quantitatively analyzed text from news media coverage of CWF referenda. Methods. The authors analyzed text from 234 articles covering 11 CWF referenda conducted in 3 US cities from 1956 through 2013. The authors used cluster analysis to identify each article's core rhetoric and classified it according to sentiment and tone. The authors used multilevel count regression models to measure the use of positive and negative words regarding CWF. Results. Media coverage more closely resembled core rhetoric used by fluoridation opponents than the rhetoric used by fluoridation proponents. Despite the scientific evidence, the media reports were balanced in tone and sentiment for and against CWF. However, in articles emphasizing children, greater negative sentiment was associated with CWF rejection. Conclusions. Media coverage depicted an artificial balance of evidence and tone in favor of and against CWF. The focus on children was associated with more negative tone in cities where voters rejected CWF.