Paul Burstein has published widely on US public opinion over the last thirty-five or so years. This volume brings together the wealth of experience and expertise gathered during this time to produce a truly impressive research monograph. His task was to examine systematically the effects of “public” opinion on Congress' policy decisions, and thus address the very difficult but fundamental democratic question in US government and politics: do US voters get what they want in terms of policies in a wide range of policy domains, including trade, the environment, Medicaid, energy policy, foreign policy, and so on? Or is the Congress much more strongly influenced by the advocacy of “special” interest groups and individuals who have the wherewithal – campaign contributions, expertise, staff resources, time, and so forth – to pressure Congress members most effectively?

The research effort reported in this fairly slim book is remarkable, underpinned by the author's skilful framing of relevant research questions within critiques of the existing literature, arduous data gathering, and the deployment of an impressive and often innovative methodology. Rather than conceptualizing “public opinion” as what polling organizations, lobby groups, or his own judgement define it as – which conventionally amounts to the most salient policy domains – Burstein's unit of analysis is a Congressional policy proposal. Following Gamson and Baumgarter et al., the author constructs a sample of legislative proposals that cover a wide range of policy domains and exhibit varying degrees of salience to voters and advocacy groups, and on which the Congress acted in some way. The scale and depth of these data are monumental. Beginning with an initial dataset of the 5,977 public bills introduced in the 101st Congress (1989–90) – which allows the author to trace both the origins and the fate of each policy proposal forward through time and discover whether it was enacted or not – he selected a stratified random sample of fifty House bills (excluding appropriations bills but including omnibus bills) topped up with ten other bills reported out of committee and enacted. Burstein then arduously traced the legislative history of each of these bills (backwards and forwards) to discover similar bills, the result of which was to yield 417 bills that were manifestations of the same sixty policy proposals considered by the House over almost three decades, and just twenty of which were enacted.

Similar to Alan D. Monroe's results (“Public Opinion and Public Policy, 1980–1993,” Public Opinion Quarterly, 62, 1 (1998), 6–28), Burstein finds that for the thirty-six proposals for which specific or general “public-opinion” data were available, voters' preferences and policies enacted by the House were consistent 50 per cent of the time. However, when voters opposed a House proposal, it died 86 per cent of the time whereas when they supported a measure's enactment the Congress responded favourably only 41 per cent of the time. In short, “public” opinion did not show much impact. For, once all sixty proposals were entered into the mix – including the twenty-four proposals on which no opinion data was available – Burstein's results show that Congress responded favourably to voter preferences only 31 per cent of the time. In other words, sampling bias in previous studies has led scholars to overestimate the impact of voters' preferences, which leads to the conclusion Congress has not been as responsive to “the public” as most people suppose.

What about the indirect effects of “public” opinion generated by the advocacy behaviour of various interest groups? While most voters register little interest in most issues considered in Congress, advocacy groups with a direct interest in a particular policy proposal help supply important informational needs for members of Congress, who (following Mayhew's assumption in Congress: The Electoral Connection (1974)) are primarily interested in reelection: in order to win reelection, they need detailed information that will help them enact policies that are popular within their districts – about the importance of problems they are asked to address, about the likely impact of policies proposed to address the problems, and about the likely impact of proposed policies on their reelection prospects.

Perhaps counterintuitively – and potentially controversially – Burstein finds (along with Baumgarter et al., Lobbying and Policy Change: Who Wins, Who Loses and Why (2009)) that advocacy has very little impact on Congressional policy: most policy proposals are associated with very little advocacy, at least that which is publicly visible. Using data from 116 newspapers derived from standardized keyword searches of Proquest, LexisNexis, and other databases, he identifies just 1,614 instances of visible public advocacy generated by interest groups, either in support of or opposition to the sixty Congressional policy proposals – which, although this measurement technique inevitably underestimates the extent of de facto advocacy, does not amount to a great deal of political activity. Moreover, 80 per cent of the proposals generated just 5 per cent of the total activity (thirty-one generated no visible activity whatsoever), leaving just three proposals generating just over half the activity identified. Focussing on the impact of publicly visible advocacy over the entire period that these policy proposals remained on the House's agenda, the research showed no impact at all – plausibly, Burstein suggests, because there was so little of it. Indeed, as the author's analysis of the composition of committee hearings witnesses, less than half represented private advocacy groups; many were Congress members. As for the lack of impact of those who do advocate, Burstein suggests that the relatively poor quality of testimony is a major cause: very few witnesses invoke likely electoral consequences to Congress members of supporting or proposing a proposal, or make much use of systematic research evidence to support their case. Effectively, then, on most issues, the “public” has no opinion and there is little advocacy. Ergo, if we assume that Congress will not adopt new policies unless pressured to do so – and fairly strongly – in the absence of strong public opinion and advocacy, we should expect a “status quo bias” (Baumgarter et al 2009) to kick in. Only when public opinion exists – also, probably, when there is publicly visible advocacy – does policy change result.

This book reports painstaking data collection and manipulation, impressive research effort, and a new and innovative methodology that addresses important questions in United States government and politics. Notwithstanding the innovative research methodology and the fact that Burstein's findings challenge a great deal of accepted wisdom, however, the book's findings are not that new – either for citizens or for scholars, many of whom see interest groups as powerful and influential players in US politics and society. Moreover, for many scholars – and citizens – who think about the role and influence of interest groups, questions will remain about the structural power exercised by many interests, notably many business groups, within a capitalist society. Take, for example, the deregulation of the financial services industries ratified by the 1982 Garn-St. Germain Act. Here, too, a researcher might very well employ a similar methodology to Burstein's and reach similar results, but then – without identifying important changes during the late 1970s that occurred in the US's prevailing public philosophy and the new embrace of free-market ideas – be left with the difficulty of explaining why the regulatory state built during the New Deal period was largely dismantled. Take also the sudden change in Congressional policy towards same-sex marriages/relationships, which occurred over recent years. Could that change be identified and measured through this methodology? Notwithstanding these important limitations, Burstein's book will be valuable to scholars and upper-level students of public opinion and political analysis.