This is one of the most interesting books I have read on the mass media, public opinion, and policymaking. Capital punishment is an important and compelling issue in its own right, which makes the first part of the book a great read, devoid of technical detail and filled with stunning descriptions of specific cases. Moreover, the rise and staying power of the idea of innocence—that innocent people sit on death row and may be executed—is clear. There continues to be a stream of news stories and commentaries about convicted murderers making plausible appeals for DNA tests that may set them free. While the overturning of murder convictions based in new evidence or faulty defenses preceded the use of DNA testing, this testing became important since it could confirm guilt or prove innocence.

The elevation of the idea of innocence came out of the “innocence movement” or “innocence projects” of advocacy groups and legal projects. The most important of these originated at universities and had credibility because they arguably had no interest other than the search for truth. The same could be said for journalists, whose bias (my emphasis, not the book's) was reporting on issues and conflicts that were good stories in order to attract audiences. The press reported on the university truth seekers and did important investigations of its own, sometimes with leads from lawyers unable to do the legwork themselves. Ironically, current newsroom cuts mean that journalists are less able to pursue such cases.Footnote ¹ One result of the innocence movement, helped by cases of police malfeasance and by Illinois Governor George Ryan's declaring a moratorium on executions and commuting all death sentences, was a decline in public support for capital punishment and a corresponding decline in the numbers of death sentences nationally. Should this trend continue—and the data show that the numbers of executions are very small to begin with—a stronger case than ever can be made for ending capital punishment in the United States.

This study is relevant to policy issues broadly. Frank Baumgartner, Suzanna De Boef, and Amber Boydstun persuasively demonstrate the power of ideas. They describe how the death penalty issue came to be increasingly described or framed along a particular dimension—the innocence frame—to the exclusion of alternate dimensions (p. 4), and how this in turn led to changes in public opinion and, through public opinion and directly, to changes in government policy. The book is a superb social science work, from conceptualization to measurement and analytics, as it tracks and explains trends over the last half century.

The chapters systematically and near exhaustively examine death sentences and executions over time, by states and nationally. They show qualitatively and quantitatively how the capital punishment debate and its newspaper coverage (in the New York Times and confirmed in regional newspapers) shifted in terms of the use of many different arguments or frames (65 in all) and how these could be reduced to a smaller set of frames emphasizing morality, constitutionality, humanization of the defendant, vengeance (“eye for an eye”), public support, and innocence. Using what they call “evolutionary factor analysis,” the authors show how trends in frames cluster, or “cascade” together, and how the overall innocence frame is distinct from all other news frames in the past because of the number of frames that trend together to define it (its “resonance”), the high frequency of the innocence frame (“salience”), and how long it has lasted (“persistence”). With this reframing came a change in the “tone” of news coverage that increasingly challenged the death penalty. A multivariate time series analysis (using quarterly data and controlling for changes in the murder rate) shows that over the long run, since the late 1980s, what followed the change in tone was an erosion of public support for capital punishment (though the balance of public opinion continued to support it). What affect did this have on government policy? Both public opinion and the tone of news coverage had independent effects on the number of death sentences yearly, and public opinion itself was affected substantially by news coverage.

What is most interesting at the end is the causal chain at work and the mechanism by which the “discovery of innocence” affected death sentences. What began with the innocence movement led to the emergence and dominance of the innocence frame, which in turn altered the tone of news coverage, which in turn affected both public opinion and policy directly. How did the latter direct effect occur? According to the authors, the possibility of innocence evoked values of fairness and problems in the criminal justice system; this made judges and juries less likely to sentence convicted murderers to death, and this also might have affected decisions by prosecutors to seek capital punishment. It is also possible that news coverage was interpreted as a proxy for current or anticipated public opinion, and so the influence at work was perceived public opinion. In the end, these influences led to the current state of debate on the issue in which the innocence frame dominates in a way that may have lasting power. It has withstood reframing that could have occurred with the terrorist threat or the shooting rampages in recent years at schools and churches, which have evoked calls for citizens to be allowed to carry guns to protect themselves, not just calls for stronger gun control measures! It is interesting, too, that the Obama administration has not voiced very strong positions on either gun control or capital punishment, and so what happens on these issues will depend on state governments and the courts.

I am hard-pressed in critiquing this well-crafted book, having been left with going, so to speak, for the capillaries, not the jugular. First, while I think that both the opinion and policy effects of changes in media framing and tone hold up overall, the measurement of overall tone was more subjective than my own preference for coding the statements of sources speaking or cited in new stories.Footnote ² But any measurement error should be diminished for measuring change, for which errors are likely to cancel out. Second, focusing on newspaper coverage probably also leads to valid measures of changes in news framing and tone, but the public gets less news from newspapers, compared to local and national television. While different news outlets cover much of the same news, there is less on television than in the papers, but the authors did not acknowledge television. Was any attempt made to get rough measures for television news coverage—using the Vanderbilt Television News Archive or other sources?

One interesting aspect of the book is its devotion to the aggregate national level, even though the policy decisions were largely state level. That the idea of innocence transcends individual states and its effect on policy everywhere is very plausible. Also, the measures of newspaper coverage arguably pick up media effects at the state level. But the use of national public opinion data and yearly death sentences nationally raises ecological inference questions. While it is likely that state-level opinion trends parallel national ones (as “parallel publics”)Footnote ³ this is an empirical question. The book does not consider this issue nor cite the relevant state-level research literature. There has been important research on public opinion and state policy-makingFootnote ⁴ including research on capital punishment.Footnote ⁵ This leaves open ground for others to use recent innovations for estimating state public opinion to study this further.Footnote ⁶ The states are centrally important to the future of capital punishment. Even with the current power of the innocence frame, some states may tightly retain capital punishment, though making sure that those they execute are in fact guilty. We would want to know whether they do so with the support of their publics at large or particular subgroups (the book does not compare trends in support among different demographic or partisan subgroups, though they may move in parallel as well). State-level analysis can provide insight into whether “the ‘discovery’ of innocence” in fact signifies the beginning of the end of capital punishment in the United States (p. 230).