Scott L. Althaus

This book bridges the gap between public opinion research and political theory in a very useful way. It focuses on the knowledge citizens possess and the way this knowledge becomes manifested in opinion polls. First the author argues against any notion of truth by aggregation: "because noise is unlikely to distribute uniformly in surveys of political opinion and because noise balances rather than cancels out, many of the information-pooling characteristics attributed to collective opinion fail to materialize" (p. 54). Although the analysis concerns surveys, the general point would seem to be important for democratic theory as well.

The middle part of the book contains some pretty detailed data analyses where the author attempts to calculate "fully-informed" opinions from sample polls based on knowledge-testing questions which discriminate between well-informed and ill-informed respondents. The utility of these calculations seemed dubious to me, but political scientists may find it more interesting. This part of the book can hardly be of much interest for general readers.

The third part of the book is different again. Here the author expands his analysis toward traditional questions of democratic theory and political philosophy. He notes that surveys are frequently dismissed as trivial or meaningless in both of these fields and cogently argues against this view. The key question is representation; when can a survey be considered to hold genuine representative value and when can it be utilized to guide political decision-making? The author is sufficiently familiar with the traditional requirements of political theory to build a strong argument, so I really enjoyed this part of the book.

Although the book is a bit on the long side and the middle part is quite specialized, it still provides some great theoretical insights into public opinion surveys interpreted as political knowledge.… (plus d'informations)