The Challenge of Measuring Media Exposure: Reply to Dilliplane, Goldman, and Mutz

Political communication research has long been plagued by inaccurate self-reports of media exposure. Dilliplane, Goldman, and Mutz (2013) propose a new survey-based measure of "televised exposure to politics" that avoids some of the features that lead to self-report error and that has already been adopted by the American National Election Study. Yet the validity of the new measure has not been independently tested. An analysis reveals several weaknesses. First, construct validity of the new measure is low because it does not attempt to measure the amount of exposure to news programs, news channels, or news overall. Second, its convergent validity is poor by several different criteria. For example, the new measure shows barely any increase in news exposure as the 2008 presidential election approached. Third, the authors' criterion for predictive validity is neither necessary nor sufficient. Dilliplane, Goldman, and Mutz are right that measuring the media exposure of survey respondents in a valid and reliable way is critical for progress in political communication research. But given the inability of many respondents to report their own exposure, it is necessary to monitor the media use of survey respondents automatically.