Involving the public in policy setting, or public engagement, is becoming increasingly popular in many democratic societies. Among the various ways of enacting engagement, the use of the Internet to either collect opinions or conduct public debate is seen as holding out great promise. However, the benefits of using the Internet over other communication media are uncertain because empirical research is limited. This article describes a comparison of data collected through two different media in the context of a "real-life" event, the 2003 U.K. "GM Nation?" public debate. The results suggest that demographically, the samples answering the questionnaires in the different formats differed significantly. However, there were no consistent differences between the approaches on a number of response effects.