Information and Democracy: Public Policy in the News is an ambitious title for an ambitious project. Stuart Soroka and Christopher Wlezien’s book seeks to answer two fundamental questions at the heart of democratic functioning: Do the media inform the public about policy change? And, if so, how does the public respond? These questions are especially relevant given current debates over news quality, and in particular the potential harms of misinformation. But even without deliberately setting out to misinform, media outlets can chip away at democratic functioning by simply failing to give people the basic information they need to evaluate policies (and, by extension, politicians).

The book serves as a sequel of sorts to Soroka and Wlezien’s 2009 book Degrees of Democracy, which drew on both spending and public opinion data to show that elected officials respond to shifts in public opinion, and that public opinion in turn responds to changes to the federal budget. Specifically, they argue, the public responds thermostatically; supporting more spending in a particular policy area after a decrease and less spending after an increase. However, Degrees of Democracy was largely silent on how the public learns about these changes in spending. As the authors point out, the average American’s lived experience does not provide much insight into levels of, for example, defense spending. Instead, Soroka and Wlezian argue in Information and Democracy, people learn about policy change from the media.

The presumption that the media provide a (somewhat) accurate signal about policy change may seem obvious—indeed, it is precisely why many worry when local news outlets close and partisan outlets gain traction. However, systematic empirical tests of this proposition are surprisingly rare. While a great deal of research, especially in the last decade, has highlighted the ways in which media coverage can both unintentionally and deliberately mislead, fewer studies have systematically tested whether (and when) the media succeeds in accurately conveying the dynamics of public policy. One reason for this lack of research is that high-quality content analysis is challenging, particularly when trying to capture a phenomenon as nuanced as policy change. The first half of the book systematically tackles this problem by employing a dictionary-based coding system to count mentions of increases and/or decreases in government spending in these areas. The authors carefully evaluate this system against human coding (by both students and Mechanical Turker workers) as well as machine learning in Chapters 3 and 4, both of which could easily be assigned as primers on different content analytic approaches. Their analyses suggest that while the “media signal” that they measure may not perfectly capture every nuance of coverage, it is a reliable indicator of how the news discuss current and future spending.

In total, they examine nearly forty years’ worth of media coverage of five policy areas (defense, health care, the environment, education, and welfare) and three media types (newspapers, television, and Facebook). Ultimately, they find consistent evidence that when the government spends (or, more accurately, plans to spend) more or less in a particular policy area, the “media signal” reflects this change. Even apart from the analyses of public opinion that follow, this finding is an important contribution. The ability of the media to send accurate signals about public policy is at the foundation of democratic performance, and Information and Democracy offers empirical evidence across multiple issue areas and media outlets over a protracted period. The authors have also made accessible summaries of their findings and methodology (http://mediaaccuracy.net/), a helpful resource for researchers looking to build on these results and/or replicate their approach for other issue areas.

The second half of the book compares the relationship between the media signal and aggregate public opinion (based on the General Social Survey and American National Election Study) toward levels of spending. They find that, in general, the public does indeed respond thermostatically to media coverage (adjusting their preference downward when spending increases, and upward when spending decreases), though this relationship is much weaker in the domains of environment and education. While the authors suggest that this disjuncture may occur because these two issues are less salient, they also mention that the dynamics of these issues may not be captured well by “spending.” This problem speaks to a larger limitation of the book: its operationalization of policy coverage. While spending is inarguably important, it reflects only one facet of public policy, and is not an equally relevant consideration for all policy areas. For example, while the issue of immigration is highly salient for many Americans, it makes little sense to think of changes to immigration policy in terms of spending. A spending increase in immigration could imply either more money flowing to US Immigration and Customs Enforcement or more money being provided for resettlement programs. A major question raised by this book, then, is whether the media also provide accurate information about policies that are not easily understood through the shortcut of “spending.” It is possible that there is something inherently different about these policy areas (which also include abortion, gun control, and election reform) that inhibit the media’s ability to provide accurate signals about policy change. Perhaps these empirical questions are a subject for another sequel.

One of the strengths of this book is its willingness to explore and learn from variation, both among media outlets and policies. Rather than elide the areas in which their theoretical expectations are not clearly met (e.g., weaker public responsiveness in some policy areas), they use them as opportunities to explore potential other factors affecting the media’s ability to cover policy and the public’s ability to respond. For example, case studies of media coverage of the Affordable Care Act and the Deficit Reform Act offer potential explanations for why the media sometimes fail to accurately signal changes in spending. In Chapter 8, they also consider how the increasing dominance of social media may change patterns of news coverage. Drawing on data from Facebook posts, they find some initial evidence that while the content that news outlets post on social media accurately reflects policy change, the content from these outlets that is shared may be less informative.

Finally, the authors are straightforward in acknowledging the difficulty of establishing definitive causal relationships between policy change, media coverage, and public opinion. While their analyses offer compelling evidence that public opinion reacts to that media coverage of policy, Chapter 7 also explores how and why the causal arrow might be reversed. They suggest that media coverage sometimes also reflects public demand for more or less spending, thereby providing another link between opinion and policy.

There are, of course, many possible ways to operationalize policy coverage and public responsiveness, and future research should test the extent to which the patterns uncovered in this book are robust to these alternative approaches. But even if they are not, this book provides a much-welcome proof of concept that the media are capable of serving as an effective conduit of information. Despite the many obvious problems with the media, Information and Democracy shows that they are indeed able to fulfill an important role in democratic functioning by signaling changes in government policy, and the public in turn is responsive to this signal. As the authors conclude, “Even as media often fail, they also sometimes succeed” (p. 167).