Many good books are published in political science, but rarely is one so powerful that it restructures the study of politics in multiple subfields. Stephen Nicholson's Voting the Agenda may be such a book, as it forces scholars to reconsider how we study not only voting and elections but also Congress, the presidency, state legislatures, governors, direct democracy, the media, and political campaigns.

The puzzle structuring Nicholson's research is simple—if voters are relatively uninformed about politics, why would we expect the typical voter to link candidates for office (up and down the ballot) with only the issues they raise in their campaigns, or the institutional responsibilities of their office? Can voters sort out the issues that go with each candidate office even though they know little about politics in general?

Nicholson argues that “elections are bustling affairs involving many different issues floating around with most voters receiving and processing electoral information in a haphazard manner” (p. 9). Although candidates do their best to set the agenda, they often fail in defining the big issues. Especially in low-information elections, candidates regularly must contend with issues not of their making. Controversial ballot propositions are an increasingly important way for interest groups and political parties to set the electoral agenda and to introduce common elements into voters' candidate judgments. Ballot measures can play a central role in the information environment available to voters. By helping to define the electoral agenda, ballot propositions can cast the election in favorable terms for one political party or another. The author shows us that ballot measures regularly help set the agenda in candidate contests.

Nicholson contends that agendas (and ballot propositions) affect voting behavior, and that this happens because voters do not pay much attention to politics. As relatively undiscriminating consumers of politics, voters use information available at the lowest cost. If abortion is a widely discussed issue in the media, abortion becomes an important issue. Policy issues (from ballot measure contests or candidate campaigns) that receive the most attention define the agenda. Agendas prime voters to evaluate candidates. Priming is the mechanism by which agendas shape political judgments. Spillover effects of the issues to candidate evaluations occur because priming influences political judgments indiscriminately, regardless of whether the issue is related to the candidate. Finally, partisan stereotypes do the “heavy lifting” of linking agenda issues to candidate evaluations. Voters ascribe issue positions to candidates because the Democratic or Republican Parties have reputations based on issues—voters strongly associate parties with solving certain problems or issues. Given these partisanship stereotypes, agendas may shape voting decisions along party lines.

Given the important role of partisanship in the priming effect of ballot measures, it is useful to consider whether the relationship between ballots and candidate choice is exogenous, or whether voting on ballot propositions is endogenous to individual partisanship. In Nicholson's words, “partisanship stereotypes are so powerful that voters nevertheless assign certain issue positions to candidates based on stereotypes, even when a candidate holds an issue position contrary to that of the party” (p. 16). Yet in a later chapter he says, “I believe that direct legislation triggers popular interest (its effects are exogenous) rather than reflects it (it effects are endogenous)” (p. 39). More research should be done to better understand the causal relationships among partisanship, ballot propositions, and candidate vote choice.

Not all ballot measures are expected to play an important role in candidate races, since not all have strong partisan overtones. For Nicholson, two factors stand out as accounting for whether ballot propositions achieve agenda status, and thus have a priming effect on candidate races: high levels of campaign spending and media coverage. Both of these factors are usually associated with controversial policy issues. But there may be others, including whether partisan elites have taken long-standing opposing positions on the issue.

Nicholson accurately points out that ballot propositions are rarely considered as a determinant of vote choice in candidate elections: “While research on direct legislation voting typically looks at politics and candidates as cheap sources of information, students of candidate voting seldom acknowledge ballot measures” (p. 33). Scholars lose sight of how voting decisions—low- and high-level offices, as well as issue—belong to a single family by sharing common space on the ballot, the same information environment, and election agenda. The assumption in the field is that the act of voting is largely not comparable across different types of offices, so that organizing the study of voting behavior by office is necessary. But this ignores the spillover effects of issues from ballot measures to candidate races, and how issues from one candidate race affect another.

The empirical analysis is robust, providing strong evidence that ballot measures affect the agenda for congressional candidates (Chapter 4), and prime vote choice in U.S. Senate and House races, as well as gubernatorial elections (Chapter 5), drawing on national survey data. In the latter chapter, we find that nuclear freeze ballot measures (and associated campaigns and media coverage) primed voters to choose various Democratic candidates, occurring across different offices. We also see that ballot measures on abortion and taxes (Chapter 4) and on illegal immigration and affirmative action (Chapter 6) primed voters to choose Republican candidates. Throughout the book, sophisticated empirical data analysis draws on national and state survey data, with probit coefficients converted into probabilities that allow the reader to understand the substantive magnitude of results. One query is why the statistical models do not control for standard demographic factors, such as the age, income, education, or gender of the respondent, which would strengthen the findings even more.

Nicholson has given political scientists much to ponder. However, more empirical research is needed to understand the effects of ballot propositions on candidate elections, especially high-information presidential elections not covered in this research. Future research should also try to isolate the unique priming effects that occur when issues appear on statewide ballots, compared to policy concerns in general, possibly using natural and manipulated experiments. There may also be important interactions or causal processes involving citizen ideology, partisanship, political interest, the media, campaign effects, and exposure to initiatives and referenda to consider.

Scholars like Shaun Bowler, Todd Donovan, and Daniel Smith have recently examined the indirect (or spillover) effects of ballot measures on citizen participation and attitudes, showing that individuals exposed to ballot initiatives are more likely to vote, contribute money to interest groups, have confidence in government (political efficacy), and be better informed about politics. Nicholson now shows us that initiatives and referenda can also have spillover effects on candidate elections. His research breaks the mold, opening new avenues of research for scholars of voting and elections of both candidates and issues. This beautifully written book is concise, contains exceptional content (both theoretical and empirical), and is a must-read for anyone interested in American politics and democracy.