Fictitious Issues Revisited: Political Interest, Knowledge and the Generation of Nonattitudes Patrick Sturgis Patten Smith University of Southampton Ipsos-MORI, London It has long been suspected that, when asked to provide opinions on matters of public policy, significant numbers of those surveyed do so with only the vaguest understanding of the issues in question. In this article, we present the results of a study which demonstrates that a significant minority of the British public are, in fact, willing to provide evaluations of non-existent policy issues. In contrast to previous American research, which has found such responses to be most prevalent among the less educated, we find that the tendency to provide ‘pseudo-opinions’ is positively correlated with self-reported interest in politics. This effect is itself moderated by the context in which the political interest item is administered; when this question precedes the fictitious issue item, its effect is greater than when this order is reversed. Political knowledge, on the other hand, is associated with a lower probability of providing pseudo-opinions, though this effect is weaker than that observed for political interest. Our results support the view that responses to fictitious issue items are not generated at random, via some ‘mental coin flip’. Instead, respondents actively seek out what they consider to be the likely meaning of the question and then respond in their own terms, through the filter of partisan loyalties and current political discourses. Seminal accounts of democratic governance uniformly incorporate some notion of public opinion as the key link between the populace and those elected to govern (Habermas, 1989; Mill, 1937; de Tocqueville, 1835). And, despite revisionist critiques (Blumer, 1948; Gins- berg, 1986; Herbst, 1993), the opinion poll has steadily attained hegemonic status as the tool for measuring the ‘will of the people’ in modern democratic polities (Althaus, 2003; Fishkin, 1995; Page and Shapiro, 1992). Since the early days of systematic opinion research, however, pollsters and academic researchers alike have expressed concerns about the quality of the estimates they produce (Gallup, 1947; Hyman and Sheatsley, 1947). Echoing earlier debates over extensions of the franchise,pioneers of the new science of opinion polling questioned whether the average person-in-the-street could be considered ‘competent’ to provide meaningful views on matters of public controversy (Lippmann, 1922; 1925). Lacking any proper incentive to keep abreast of public policy debates,citizen responses to opinion polls were argued by some to represent little more than an ‘echo chamber’ of elite opinion; weak, labile and easily influenced by vested interest groups (Key,1961). These early speculations were soon lent substance by a plethora of empirical investigations. The general public, it was repeatedly confirmed, have only a very dim awareness of matters politic (Berelson et al., 1954; Hyman and Sheatsley, 1947), a state of affairs which appears to generalise cross-nationally (Almond and Verba, 1963; Baker et al., 1996), and to have remained largely unchanged since systematic measurements began (Delli Carpini and Keeter, 1996). In addition to being broadly unfamiliar with institutions, politicians and policies,the mass public were also found to be highly sensitive to the context and order in doi: 10.1111/j.1467-9248.2008.00773.x POLITICAL STUDIES: 2010 VOL 58, 66–84 © 2009 The Authors. Journal compilation © 2009 Political Studies Association