To fulfil their representative function and to respond to public preferences, it is essential for policy makers to know what people think. In the extant research on policy responsiveness, mass opinion surveys are used to measure public preferences on a government policy. In the literature, researchers employ a variety of measures to capture what people want: preferences on public spending (Wlezien 1995, 1996), issue salience (Hobolt, Klemmensen 2005, 2008) or ideology (Powell 2000). However, we do not know what kind of preferences these measures represent and whether the indicators measure the same concept or are related to policy outputs in the same way. In this paper, we cross the measures of public preferences used in responsiveness research and investigate whether they achieve discriminant and predictive validity: Do measures correlate with each other and are they able to predict responsiveness similarly. Our results have implications for our understanding of whether governments are responsive as well as for previous findings. The analysis is based on a sample of 14 EU countries and employs data collected by the International Social Survey Program, Eurobarometer, and the Organisation for Economic Cooperation and Development. In current responsiveness research a variety of measures is employed to indicate public preferences. This paper tests whether these are interchangeable, measure the same concept and achieve discriminant and predictive validity.
