The central theme of this book deals with the relationship between knowledge, representation, and political equality as expressed in opinion surveys. Scott Althaus finds that political knowledge does matter. In addition, the way that knowledge is distributed in society has an important impact on how one should interpret the collective preferences of the public, particularly when assessing the role of the less knowledgeable.

Althaus suggests that the way knowledge is distributed in a society may cause collective preferences to disproportionately reflect the opinions of some groups more than others. While in some circumstances collective preferences may represent the will of the people, he argues that it often does not. He finds it is not a shortcoming of survey research but rather the limited degree of knowledge of some respondents and the influence of others to be better informed that affect the distribution of opinion. Althaus finds that most individuals who know little about politics are often willing to express opinions on issues and policies when asked.

The author argues that opinion surveys have never really been egalitarian. He points out that 1) descriptive representation is only one kind of representation, 2) a representative sample is no guarantee that the survey results will also be representative, and 3) equality at the individual level is no guarantee of the proportionality at the group level.

Althaus raises his concern with the quality-of-interest representation provided by opinion surveys. He focuses on two questions: how accurately the opinions of survey respondents represent the interests of a population and how heavily those opinions are weighed by decision makers who act on their behalf.

His research suggests that a sample, even when it descriptively represents a population, does not guarantee that its results will accurately reflect the population. He notes that knowledgeable respondents are more affluent, educated, male, white, and partisan than ill-informed respondents and are usually overrepresented in surveys, therefore potentially carrying a disproportionate weight in surveyed opinion measures. He also finds that the preferences of the least-knowledgeable opinion givers are often overrepresented in society's collective preferences because ill-informed people tend to give similar responses to survey questions.

Althaus explores a variety of interesting statistical analyses to develop models of opinion using surveyed and simulated opinions that may be ill-informed, fully informed, or enlightened. He presents a discussion about the consequences of using simulation methods to generate a fully informed opinion sample. The author suggests the use of demographic variables to simulate their effects on collective preferences and applied these analyses to the attitudes of abortion rights, economic policy, and reducing the federal budget deficit. He found that correcting for the low levels and uneven social distribution of political knowledge often causes collective preferences to change in significant ways. He also found that some of the differences between surveyed and fully informed collective opinion could result from biases that may exist in the information environments of respondents.

Althaus found that information effects are not merely a function of cognitive ability and that they may be smaller for issues that receive prominent media attention. He suggests that the size of the information effects goes down as levels of salience and interest in politics go up. However, he takes the position that it is inappropriate to think that survey findings represent the active voice of the people, even when information effects are minimal or nonexistent.

Much of the empirical testing found in the text relies heavily on the American National Election Studies from 1988, 1992, and 1996. While these are important and well-executed studies, they present a problem for this inquiry since they were conducted during periods of presidential elections when respondents were likely to have elevated knowledge levels. An examination of other data sets to test some of the principal theoretical questions might be pursued in future work.

Althaus suggests that opinion surveys can be descriptive or prescriptive. Descriptive surveys are used to reveal the distribution of attitudes, opinion, and beliefs in the mass public for purely informational purposes. They include interpreting election results and public reactions to political events, revealing the climate of the public's opinions, the marketing of political policies and decisions, and identifying winning issues. Prescriptive surveys seek to identify a course of action for government to follow or to define the boundaries of permissible government activity. According to the author, prescriptive surveys treat collective preferences as if they were collective decisions.

Althaus believes that since individuals and populations can never be fully enlightened, and because few individuals or populations are ever fully informed, he questions whether we should call the results of opinion surveys the “will of the people.” He provides four reasons why this is problematic. They include the facts that no formal consent was given by the population to be represented in this manner, surveys are not deliberative, surveys are reactive rather than proactive, and aggregation procedures produce results that may be inconsistent with individual preferences. Due to the limitation of polls to represent collective preferences, Althaus suggests essentially a limited role of description.

Our modern society recognizes the concept of one-man-one-vote that could be expressed through surveys giving each randomly selected respondent an equal voice. However, our founders put in place a variety of institutional mechanisms to have the opinions of the masses constrained by such institutions as Congress and the Electoral College.

Althaus notes that in attempting to develop a case for the use of informed citizens when reporting opinion polls, it may be difficult to determine what facts are relevant to a given set of opinions. He rejects the simple use of filter questions for data analysis because of the depletion effects that result from “Don't Know” responses. He suggests using a few questions to tap knowledge levels on specific issues or the use of an omnibus knowledge scale that could be included at the end of a survey. This approach would allow journalists and political actors to more appropriately use statements regarding how both ill-informed and knowledgeable citizens feel about issues.

Political actors may use opinion research to promote their efforts by defining the issues or agenda items to be explored, as well as by providing analyses that are designed to support their positions, whether it be by choosing what to report or how to report it. Althaus suggests that opinion surveys may be more appropriately used to gather opinions about the larger society's goals, or the relative importance of items seeking space on the public agenda, rather than resolving specific policy issues as the will of the people.

Collective Preferences in Democratic Politics provides a strong literature review. However, the writing style lacks accessibility for most undergraduates and early graduate students. This work would be ideal for advanced graduate students and scholars since it is rich in creative concepts and provides numerous opportunities for both hypothesis-testing experiments and philosophical debate. A major strength of the work is that it provides important insights into how knowledge is distributed in society, how it is captured through opinion research, and how this distribution of knowledge impacts the development of collective preferences in a political system. These issues should be carefully assessed by all who conduct, report, and use opinion research to assure that their findings occupy an appropriate role in public discourse.