Actual citizenries bear little resemblance to the democratic ideal. Some variation by individual, by issue, and by circumstances notwithstanding, not many people know or think much about politics (Converse 1964, Luskin 1987, Delli Carpini and Keeter 1996, Kinder 1998, Price 1999). As a rule, therefore, the public opinion revealed by everyday polling is neither particularly informed nor particularly thoughtful. But suppose people did know and think more about politics. What would public opinion be like then? Deliberative Polling addresses this question by drawing and interviewing a random sample, providing its members with briefing documents laying out the arguments for and against policy alternatives, getting them to discuss the issues in small groups, giving them opportunities to question competing experts and policy-makers, and then gauging their opinions again (Fishkin xxxx; Luskin, Fishkin, and Jowell 2002). Typically, a separate random sample, answering the same questions at the end of the process, provides a control group. 1 The enterprise, therefore, is not only a form of public consultation providing a view of more considered public opinion but also and a form of social-scientific quasi-experimentation shedding light on deliberation’s effects on information, attitudes, and behavior. 2

Until very recently, the deliberation in Deliberative Polling has been entirely face-to-face. The small group discussions and questioning of policy experts and policy makers have taken place at a common site, typically over a weekend. Advances in information technology, however, now make it possible to implement Deliberative Polling online, provided a representative sample of the whole public, not just the online population, can be obtained. Here we describe the first online Deliberative Poll, conducted in parallel with a traditional face-to-face Deliberative Poll, both about US foreign policy on the eve of our war with Iraq. The online sample, from Knowledge Networks, is a true random sample (subject only