Introduction

From a distance, debates in the United States about how state judges should be chosen may seem unduly fierce. After all, the competing selection systems on which the debates focus share attributes that distinguish them from the systems used for most courts in most nations. For one thing, people in the United States who serve as judges in the judicial branch of government are chosen after working in other positions rather than joining a judicial corps at the beginning of their careers (Georgeakopoulos Reference Georgeakopoulos2000). Further, all the competing systems in the states give central roles in the selection process to people in the political sector (the chief executive, legislators, and the general public) as distinguished from the legal sector (lawyers and judges). Finally, the various systems differ less in practice than their formal rules suggest.Footnote ²

But it would be a mistake to view the debates as empty. Even within the narrow range of the systems in the United States, alternative formal rules make a difference. Systems that give the other branches of government a monopoly over selection of judges and those that give voters substantial power over selection differ in significant ways. Among the states that elect judges, the choice to list candidates' party affiliations on the ballot has considerable impact on voters' choices and on the election process as a whole. Ultimately, the selection of one system rather than another helps determine not only which people become judges but the policies that courts make. Even though the differences among selection systems in the United States are often exaggerated, scholarly and political debates over how to choose judges are worth the efforts that their participants devote to them.

In broad terms, the choices are among three types of systems that states have adopted. One is popular election of judges. The second is selection by governors, legislators, or—as in nomination by the president and confirmation by the Senate at the national level—a combination of the two. The third is the Missouri Plan, often called merit selection: a commission nominates candidates for a judgeship, the governor selects from the nominees, and the governor's choice undergoes periodic elections in which voters vote “yes” or “no” on retention of the judge in office.

On the details of selection systems, there is great variation among states.Footnote ³ Each of the three types takes multiple forms. I have already noted the distinction between partisan and nonpartisan election. To take another example, the rules for operation of Missouri Plan commissions differ a good deal. Some states use hybrids of two selection systems, and it is common for a state to employ one system for its trial courts and another for its appellate courts—most often, election for the former and the Missouri Plan for the latter.

The specifics of debates about judicial selection have shifted considerably over time, but the driving forces have remained the same motivations that shape other debates about governmental structure and procedure.Footnote ⁴ One motivation (which can be labeled “values”) is an effort to improve the performance of courts, primarily by fostering what advocates view as desirable forms and degrees of accountability for judges. The other motivation (which can be labeled “interests”) is an effort to advance the political and policy goals of participants in the debates. Although values typically dominate public discourse about alternative selection systems, on the whole interests play a more powerful part in determining the outcomes of debates over those systems.

The configurations of values and interests that prevailed in successive eras have resulted in an evolution of selection systems over time. In the mid-nineteenth century, dissatisfaction with systems of gubernatorial and legislative judicial appointments led to advocacy of popular elections of judges. That movement was successful, with new states and some existing states adopting judicial elections. As a result of that success, appointive systems were (and are) largely confined to the East Coast.

The elective systems that states adopted during that period were formally partisan. By the late nineteenth century, the prominence of parties in those systems became a matter of widespread concern. Some states switched to nonpartisan elections in the early twentieth century, and others made the switch later in the century. Because of this movement, states that use a nonpartisan ballot now substantially outnumber those with a partisan ballot.Footnote ⁵

In the early twentieth century, Progressives and others attacked the election of judges in itself. One Progressive group, the American Judicature Society (AJS), helped develop an early version of what later became the Missouri Plan (Shugerman Reference Shugerman2012, 174–76).Footnote ⁶ After a very slow start, many states adopted the Missouri Plan for at least some of their courts in the late twentieth century. Jed Shugerman (Reference Shugerman2012, 179–80, 197, 210–11, 223, 226, 233, 239) has documented the key part that business groups played in the movement to adopt the Missouri Plan, a movement that achieved considerable success. That effort was based on the perception of group leaders that elections were putting into office supreme court justices who were unfavorable to business interests.

Although a variety of proposals for change in selection systems have been considered and adopted over the past century, across that period the central question in debates over judicial selection has remained whether judges should be chosen through partisan or nonpartisan elections—what might be called “regular” judicial elections—as opposed to alternatives such as the Missouri Plan. Beginning in the 1980s, this question has attracted increased interest. The spur for that interest has been a perception that judicial election campaigns have changed. As Roy Schotland (Reference Schotland1985, 76) put it amid the early signs of change, judicial elections “have now become noisier, nastier and costlier.” In the opinion of many observers, time has confirmed Schotland's perceptions: elections have become more competitive, spending by candidates and on their behalf has grown considerably, and appeals to voters by candidates and their supporters have become hard-edged and more policy oriented.

For many legal scholars and commentators who have weighed in during the past three decades, these changes have made an undesirable but perhaps tolerable method of selecting judges considerably less tolerable. As they see it, several undesirable things have happened. Fear of defeat has eroded judicial independence by causing judges to take the prospect of electoral opposition into account when they decide cases (American Bar Association 2003, 18–19; Geyh Reference Geyh2003, 49–51; Porter Reference Porter2016, 1069–71). Changes in the content of candidates' appeals to the voters have weakened public confidence in the courts (Republican Party v. White 2002, 818). Large contributions to candidates and independent spending by interest groups with a stake in court decisions have also weakened that confidence by creating a perception that justice is being bought, and the fact that justice can be bought through elections is very undesirable in itself (Behrens and Silverman Reference Behrens and Silverman2002, 282–83; Carrington Reference Carrington2011, 1982–85; Rankin Reference Rankin2013).

By no means is this perspective universal. For one thing, as the books I consider illustrate, political scientists as a group have been less skeptical of judicial elections than legal scholars and other lawyers. For another, in the current era, business groups (despite their earlier efforts on behalf of the Missouri Plan) and other conservative interests have fared well in state supreme court elections, the level on which they focus their efforts. This success has come in part from advantages in funding, in part from the effectiveness in some elections of charges that sitting justices and other candidates were unduly liberal on social issues (especially criminal justice). In light of that success, evaluations of the perceived changes in supreme court elections and of judicial elections in general tend to differ along ideological lines—differences that extend to the US Supreme Court.Footnote ⁷

In this essay I consider three empirical questions about the changes in contests for seats on state supreme courts, focusing on partisan and nonpartisan elections. The first question is how much change actually has occurred. The second is why that change has come about. The last is what difference it has made—for voting behavior in judicial elections, for the standing of the courts, and for judicial policy. I then consider the implications of what we know about changes in judicial election contests for the debates over alternative selection systems.

To address the three questions, I use the substantial and growing body of empirical scholarship on supreme court elections. There is a long tradition of research on judicial elections, focused primarily at the supreme court level, and perceptions of major changes in judicial campaigns have led to a burgeoning of this research. Although our understanding of supreme court elections remains incomplete, the recent research has done a good deal to address the questions that I consider.Footnote ⁸

I give primary attention to four books that were published in 2012 and 2015. Each book is a high-quality, important work of scholarship. Collectively, they add enormously to our understanding of those elections. For that reason, they merit close attention from scholars who are interested in the selection of judges, from participants in the debates over selection systems, and especially from those who fall into both categories.

Overviews of the Books

These four books are all written by political scientists, each of whom analyzes aspects of state supreme court elections in systematic and primarily quantitative terms. Each book responds to perceptions of substantial changes in the character of supreme court elections, and each addresses the policy debates about judicial elections that have been heightened by evidence of those changes in character.

Melinda Gann Hall's Attacking Judges: How Campaign Advertising Influences State Supreme Court Elections (2015) analyzes television advertising in supreme court elections, with an emphasis on “attack ads.” Making use of data that have been collected on TV ads for the period from 2002 through 2008, Hall charts patterns of advertising and analyzes their impact on voters' participation in supreme court contests and on the vote shares of incumbent justices. The book's inquiry is embedded in the scholarship on political advertising and its impact as well as in a broader perspective on developments in judicial elections, a perspective that is informed by Hall's extensive past research on those elections.Footnote ⁹ Hall gives particular attention to the implications of her inquiry and her findings for the debates over the desirability of choosing judges through elections.

James L. Gibson's Electing Judges: The Surprising Effects of Campaigning on Judicial Legitimacy (2012) is also concerned with supreme court campaigns but focuses on their impact on public attitudes toward the courts. He is a long-time student of those attitudes, carrying out a series of studies on his own and with Gregory Caldeira (e.g., Gibson Reference Gibson2007; Gibson and Caldeira Reference Gibson and Caldeira2009). In a series of articles and this book he has turned his attention to state supreme courts. The book reflects Gibson's interest in the possible impact of Republican Party v. White (2002), the Supreme Court decision that loosened constraints on the expression of policy positions by candidates for judgeships. Making use of a three-wave panel surveyFootnote ¹⁰ in Kentucky and a national survey, Gibson probes the impact of attributes of supreme court campaigns and of judicial elections in themselves on public support for state supreme courts.Footnote ¹¹

Chris W. Bonneau and Damon M. Cann's Voters' Verdicts: Citizens, Campaigns, and Institutions in State Supreme Court Elections (2015) focuses on voters. Bonneau and Cann give some attention to voter participation in state supreme court contests, but their primary concern is choices between candidates. Their analyses are based on national surveys in 2010 and 2012 and on experiments to probe the impact of incumbency and partisan information on vote choice. The book's inquiries are an extension of prior studies by Bonneau and others, including Hall, that have analyzed aggregate vote patterns in supreme court contests.Footnote ¹² Past survey-based studies have focused on one or two states, so the shift to a national survey is an important step. With this combination of individual-level data and a cross-state design, the authors could analyze the determinants of participation in supreme court elections and vote choice in new and valuable ways.

Herbert M. Kritzer's Justices on the Ballot: Continuity and Change in State Supreme Court Elections (2015) is the most comprehensive of the books in its coverage of supreme court elections. Kritzer provides a detailed picture of the formal rules for selection of justices across the states. He then addresses an array of issues: the impact of using judicial elections to select judges and of supreme court campaigns and election outcomes, the extent to which judicial elections are contested and the level of competition in those contests, the scope and content of campaigns, and partisanship in voting for supreme court candidates. The book's examinations of these issues vary in the time periods they cover, but all the periods are substantial. The analyses of partisanship, like the analyses of voting behavior in Dubois (Reference Dubois1980), make use of county-level correlations between the vote in supreme court contests and gubernatorial contests. In discussing these issues, Kritzer provides a good deal of information about campaigns and elections in specific states alongside his quantitative data. In addition to Kritzer's own research, the book presents careful summaries of the broader bodies of scholarship on issues relating to judicial elections.

How Much have Judicial Election Contests Changed?

The commentaries that point with alarm to changes in judicial election campaigns have relied heavily on illustrative rather than systematic evidence.Footnote ¹³ These commentaries often cite defeats of incumbent judges whose opponents had heavily criticized them for their decisions, vicious personal attacks on some judicial candidates, and massive interest-group spending in particular contests. Heated and well-funded contests in states such as Alabama, Ohio, and Texas in past decades attracted considerable negative attention, and the same has been true of Michigan and Wisconsin in recent years.

Perhaps the most visible of all judicial contests in the past two decades was the 2004 election to the West Virginia Supreme Court (Goldberg et al. Reference Goldberg, Samis, Bender, Weiss and Rutledge2005, 4–5). Don Blankenship, the head of a coal company, spent about $3 million on behalf of Republican Brent Benjamin, who challenged Democratic incumbent Warren McGraw. Among other things, Blankenship helped to fund television commercials that criticized McGraw for his vote with the majority in a decision that granted probation to a defendant who had been convicted of sexual assault. Benjamin won the election. Blankenship's interest in the contest apparently stemmed from a high-stakes case involving his company that was pending before the state supreme court. After the election, Blankenship's company won the case by a single vote, with Benjamin joining the majority. Ultimately, the US Supreme Court ruled that because of the massive spending by Blankenship, Benjamin should not have participated in the case (Caperton v. Massey Coal Co. 2009). Among other publicity, this episode may have inspired a John Grisham novel, The Appeal (Reference Grisham2008).

This illustrative evidence provides good reason to conclude that something is going on. But it does not tell us how much state judicial election campaigns have actually changed. To start with, although commentators regularly refer to changes in judicial elections, the evidence of change they cite comes overwhelmingly from supreme court elections.Footnote ¹⁴ Our evidence on lower-court elections is far less extensive, but the evidence we do have suggests that there have been no massive changes in campaigns at those levels. A set of studies indicates that the apparent trends at the supreme-court level have reached intermediate appellate courts only to a limited degree (Streb, Frederick, and LaFrance Reference Streb, Frederick and LaFrance2007; Streb and Frederick Reference Streb and Frederick2009; Frederick and Streb Reference Frederick and Streb2011). That appears to be true of trial court elections as well (Arbour and McKenzie Reference Arbour and McKenzie2010; Nelson Reference Nelson2011; but see American Bar Association 2003, 37–39). Like the four books examined in this essay, I focus on supreme court elections.

The first issue to consider is the level of electoral competition. One common image is that the traditional pattern in the United States was one of limited competition: elections were frequently uncontested, especially when judges ran for reelection, and incumbent judges who faced opposing candidates typically won by wide margins. In the current era, it is thought, elections are contested at higher rates and sitting judges are more vulnerable to defeat.

Historical evidence makes it clear that at least in some states in some periods, supreme court elections were hotly contested (Hall Reference Hall1984). Hall (2015, 45–54) and Kritzer (2015, 108–25) provide systematic evidence on trends over the past few decades; Hall analyzes the period from 1980 through 2010 and Kritzer the period from 1946 through 2013.

These studies show that in the years they cover, there was never a time in which incumbents seldom faced opposition. They do find that competition increased from the 1970s through the 1990s, but Kritzer shows that dramatic increases were largely confined to partisan elections in southern states. As he points out, the timing of those increases supports the conclusion that growing party competition in the South was the key driver of increased competition.

The Kritzer and Hall findings suggest that even in recent years, levels of contestation have been only moderate. In the period from 2000 to 2010, according to Hall's data, incumbent justices faced competition only 69 percent of the time—and that figure includes interim appointees who were running for the first time and thus might be perceived as vulnerable. Contestation rates were below 50 percent in seven states, and there were eight states in which no incumbents lost during that period.Footnote ¹⁵ Kritzer finds that outside the partisan election states of the South, there have not been substantial increases in the numbers of close races involving incumbents in the period since the 1980s.

The rate of defeats for incumbent justices is especially important, in part because a higher level of vulnerability might affect the behavior of sitting judges. Hall charts rates of defeat in partisan and nonpartisan elections, finding no particular trend since 1980 in nonpartisan elections but a substantial increase in defeats in partisan elections that peaked in 2000 before declining substantially in the decade that followed. Kritzer's longer-term data present a pattern that is similar in most respects, and he shows that the surge in incumbent vulnerability to defeat in partisan elections—like the overall growth in competition—was mostly concentrated in the South.Footnote ¹⁶

For the 1980–2010 period, Hall (2015, 45) found that of all incumbents who ran for new terms, opposed or unopposed, 13 percent were defeated. The rates of defeat were 21 percent for partisan elections, 8 percent for nonpartisan elections. During the same period, the rate of incumbent defeats in partisan supreme court elections was somewhat higher than that for US Senators (14 percent) and governors (16 percent).Footnote ¹⁷ The high rate of defeat for justices in partisan elections, compared with the rate for senators and governors, probably reflects their limited capacity to build name recognition and to develop other advantages of incumbency, which leaves them susceptible to unfavorable partisan balances and swings in their states.

As I have noted, scholars and other commentators who are concerned about recent developments in judicial elections give particular attention to the scale and content of campaigns. Scale can be measured in monetary terms, but data on campaign spending are incomplete. Reporting requirements for candidates are a recent phenomenon and, in most states, independent spending by parties and interest groups is not reported.

Drawing from earlier studies and data collections, Kritzer (2015, 134–53) provides a picture of trends in campaign funding since 1990, with adjustments for inflation. Contributions to supreme court candidates increased quite considerably in the 1990s and leveled off in the 2000s. In Michigan and Wisconsin, the two states for which data on spending independent of the candidates are available for more than a few years, interest-group spending was nearly nonexistent and then became substantial in some election years after 2000. Although these two states are far from typical, those trends suggest the possibility that overall spending in supreme court contests continued to grow after 2000. Perhaps the most striking pattern in spending is the distinctly higher average level of contributions in partisan elections, compared with nonpartisan elections, throughout the period from 1990 to 2010.

Due to the research work of the Justice at Stake coalition, comprehensive data on television advertising are available for the period since 2000.Footnote ¹⁸ Kritzer's analysis of these data shows that there was an increase in the proportion of contests with TV advertising, from 33 percent in 2002 and 46 percent in 2002 to more than 60 percent in each election year from 2004 through 2012 (Kritzer 2015,165). The peak was in 2008, with a 75 percent advertising rate. The total number of airings of ads in supreme court campaigns showed a similar but sharper trend; it more than tripled between 2000 and 2008 before dropping off somewhat. There is good reason to think that the amount of television advertising after 2000 was considerably higher than in the preceding decade, but there is no systematic information prior to 2000. A significant share of spending for TV ads has come from interest groups, a share ranging from 15 percent in 2010 to 36 percent in 2012 (Kritzer 2015, 156).

The research reports of Justice at Stake also provide our only systematic data on the content of campaign appeals. These reports classify advertising content in terms of what the reports call its tone (promoting the favored candidate, making a contrast between the two candidates, or attacking the opposing candidate) as well as its subject matter (several categories, which Hall collapses into a trichotomy of traits, values, and issues). Thus the data can be used to study both subjects of interest.

Both Kritzer (2015, 157–69) and Hall (2015, 72–92) analyze the tone of advertising, and Hall analyzes its subject matter as well. On tone they find that ads promoting the favored candidate were numerically dominant, though attack ads were hardly rare. Not surprisingly, attack ads were sponsored primarily by parties and interest groups rather than by candidates themselves. Kritzer (2015, 158–60, 167–68) shows that in the period from 2000 to 2012, the overall tone of supreme court ads was far more positive than ads in presidential, gubernatorial, and congressional campaigns.

On subject matter Hall finds that a substantial proportion of ads referred to issues.Footnote ¹⁹ That proportion was correlated with tone: 29 percent of promote ads, 56 percent of contrast ads, and 85 percent of attack ads were about issues (Hall 2015, 83). Like attack ads, issue-oriented ads came disproportionately from interest groups.

In the supreme court contests in which television advertising is used extensively, it constitutes the primary form of communication to voters. Thus the Justice at Stake data provide a good sense of what candidates and their supporters are saying to voters in those contests. But the limited availability of information on the content of other forms of communication means that we have only a partial picture of campaign appeals even in the period since 2000 and a very limited picture prior to that time. In the absence of systematic information, comparison of campaign content over time must be speculative. The image of bland and low-key appeals in past periods likely has some validity, but there were campaigns that departed from that image even in the distant past (Hall Reference Hall1984), and there continue to be some decidedly bland campaigns today. Thus the extent of change in the content of campaign messages over time remains an open question.

What we can glean about trends in supreme court campaigns supports the conventional picture of change in some respects but not in others, and the nonavailability of some relevant information leaves considerable uncertainty about change in some aspects of campaigns. But this may be a matter on which perceptions of reality are more important than reality itself in some respects. If state supreme court justices believe that they have become more susceptible to attacks for their decisions and more vulnerable to defeat, that perception in itself may affect their behavior on the bench. I will consider that possibility later in this essay.

Why have Supreme Court Contests Changed?

To the extent that state supreme court contests have changed, it is important to ascertain the sources of such change. That question receives some explicit attention in these books (Kritzer 2015, 17–21, 27–28, 239; Hall 2015, 9–11; Bonneau and Cann 2015, 7–11), and some of their findings on patterns of change—as well as findings from other scholarship—tell us something about its sources.

Some sources of change lie outside the courts. As Kritzer shows, growth in electoral competition can be traced largely to the rise of the Republican Party in the South, which enabled Republican candidates to contest elections successfully. In all likelihood, general growth in campaign spending—which may result in part from reduced legal constraints on spending—has had an effect on campaigns for all high offices. The same may be true of the apparent growth in negative political advertising.Footnote ²⁰

The most proximate and most important cause for change is the perception of high stakes in the outcomes of supreme court contests. Contributors and independent spenders devote their resources to the offices that are most relevant to their goals; those who contribute or spend substantial sums would not do so if they did not see the outcomes of races for supreme court seats as important. Something had to happen to direct attention to state supreme court elections.

Historical evidence on the involvement of interest groups and other contributors to supreme court campaigns is quite limited, so it is impossible to ascertain how much groups such as business and labor participated in those campaigns prior to the current era. But it is clear that the increased financial involvement of the business community in supreme court campaigns was the primary source of growth in the scale of those campaigns.

Business leaders always had an incentive to seek influence over the policies of state supreme courts, but that incentive strengthened. Over the twentieth century, and at an accelerating pace in the late 1950s and 1960s, state supreme courts carried out a revolution in tort law in which traditional limits on recovery for personal injuries were widely abrogated (Keeton Reference Keeton1969; Baum and Canon Reference Baum, Canon, Porter and Alan Tarr1982). Meanwhile, there was a widespread perception of a damaging “litigation explosion” in torts and other fields (e.g., Olson 1991; see also Galanter Reference Galanter1986). Faced with these negative trends, the business community and its allies fought back on several fronts beginning around the 1980s. One front was legislative, an effort to overturn unfavorable doctrines and to secure limits on tort suits in state legislatures that achieved considerable success (Rustad and Koenig Reference Rustad and Koenig2002, 65–71). Another was an effort to shape public attitudes about tort cases and thereby affect both legislative action and jury verdicts (Daniels and Martin Reference Daniels and Martin2000).

Finally, business groups sought to influence the membership of state supreme courts. As discussed earlier, these groups had already achieved some success in eliminating elective systems that they saw as unfavorable to their interests. However, half the states had retained elections for supreme court justices, and at least some of those states seemed highly resistant to the elimination of elections. The alternative was to become more involved in election contests in those states. The business community increasingly did so, initially through contributions to candidates and later on through substantial independent spending as well. As Shugerman (Reference Shugerman2012, 241) put it: “After spending their capital (political and financial) on campaigns for merit reforms, businesses returned to trying to win judicial elections outright….”

Issues that concern businesses are not the only spur for interest-group involvement in supreme court elections (Tarr Reference Tarr2012, 79–81). Since the 1970s, state supreme courts have played a more active role in expanding rights under state constitutions, most visibly in criminal procedure (Shaman Reference Shaman2008). A number of supreme courts have overturned state systems for funding of public education, rulings that inevitably garner considerable attention (Paris Reference Paris2010). But it is the legal issues that pit businesses (and sometimes medical professionals) against other economic interests that have been the most powerful motivation for financial participation in supreme court campaigns.

In some contests, campaigns supported or conducted by business groups have emphasized the issues that concern the business community in their appeals to voters. For the most part, however, they do not. To the extent that these campaigns discuss issues, criminal justice is a better bet because opinion on that issue is heavily weighted in favor of proprosecution policies. As a result, business-sponsored television commercials often focus on criminal justice rather than on the issues that actually concern the sponsors.Footnote ²¹ To a considerable extent, spending is used simply to provide name recognition and a positive image for the favored candidate rather than to convince voters on the issues.

When business groups mobilize, their adversaries typically do so as well. The result is to further increase the level of campaign activity. And because news media give attention to supreme court contests largely on the basis of campaign activity,Footnote ²² their coverage also increases.

The perceived stakes in election outcomes go far toward explaining why supreme court contests have changed substantially in some states but not in others. As Kritzer (2015, 7–28) shows, the court's active role in tort law was a major spur to the growth in the scope of supreme court campaigns in Wisconsin.Footnote ²³ In Minnesota, in contrast, the more limited role of the supreme court in public policy making has meant that economic interest groups have little interest in election contests.

Even in states whose supreme court has attracted attention from the business community and other groups, some contests are of more interest than others. If a particular contest is likely to be one-sided, or if the outcome would appear to have little impact on supreme court policy, then groups have less incentive to involve themselves. To a degree, the mild decline in campaign spending and advertising in the past several years probably reflects the securing of solid probusiness majorities on supreme courts in states such as Texas that had been major battlegrounds for interest groups (Cheek and Champagne Reference Cheek and Champagne2005, 38–39).

What Effects have the Changes had?

To the extent that supreme court campaigns have become larger in scale and their content has changed, these developments could have a wide range of effects. Much of the commentary on changes in supreme court contests has pointed to what the writers see as significant negative consequences of those changes, but systematic empirical evidence on most of those possible effects has been limited and largely indirect.

Some effects relate to voters as the targets of campaigns. Perhaps the most obvious possibility is that increased spending by candidates and on their behalf has stimulated participation in supreme court contests. Since few people turn out at the polls in order to choose supreme court justices,Footnote ²⁴ participation (or, more accurately, nonparticipation) is usually measured in terms of “rolloff,” the proportion of voters who skip supreme court contests. The impact of voters' information levels on rolloff is indicated by the substantially higher levels of participation in states that provide candidates' party affiliations on the ballot. Similarly, there is strong evidence that higher spending on communication to voters reduces rolloff (Hall and Bonneau Reference Hall and Bonneau2008), evidence that Hall (2015, 148–49, 156) augments and updates in her book.

Hall looks at a second question as well: the impact of campaign advertising on rolloff. Of particular interest is the impact of attack advertising, which could increase participation by giving voters an additional basis for choice or depress it by alienating voters. Hall (2015, 147–57) finds that larger numbers of airings of attack ads are associated with reduced rolloff to a meaningful degree in states that use a nonpartisan ballot but not in those with a partisan ballot. This difference almost surely reflects voters' greater need for information when they do not know the candidates' party affiliations.

Movement toward larger-scale campaigns can affect the criteria that voters use to choose between candidates by giving them more information. One possibility is that partisan voting increases in states that do not provide the candidates' party affiliations on the ballot. Kritzer (2015, Ch. 6) probes that question by analyzing county-level correlations between the parties' share of the gubernatorial vote and the division of the vote between competing supreme court candidates. He finds that partisan voting by that measure did increase in the nonpartisan states as a whole, primarily after the 1980s (there was no increase in states with a partisan ballot), though there was considerable variation among states. For instance, in Minnesota, with its quiet contests, there was no substantial increase (Kritzer 2015, 183).

Especially interesting are voters' responses to the issue-based appeals that have become more common. The Justice at Stake data allow analysis of the relationship between the use of such appeals and the candidates' vote shares, and such analyses would be useful. It would be even better to trace how effectively issue-based appeals are communicated to voters and how those appeals affect their choices, a task that requires survey data. Bonneau and Cann (2015, 59–67) provide experimental evidence that information on supreme court candidates' issue positions can cue voters to respond to the candidates on a partisan basis in the equivalent of nonpartisan elections. Surveys of Ohio voters over the years suggest that communication of issue positions on economic issues to voters is only partially effective, even in large-scale campaigns (Baum, Klein, and Streb forthcoming). In contrast, observational evidence suggests that well-funded attacks on candidates as soft on criminals, such as the West Virginia ads on behalf of Brent Benjamin in 2004, can reach and sway voters in large numbers.Footnote ²⁵ But we do not know much about the conditions that affect the success of such campaigns.

Changes in judicial campaigns can have an impact om voters' collective decisions. Changes in the scope and content of campaigns may affect the outcomes of supreme court contests through their impact on both participation and choice between the competing candidates. Candidates who spend appreciably more money than their opponents gain a potential advantage in persuading voters to support them. High levels of spending, even if equally distributed between the candidates, may reduce the advantage that most incumbent justices hold in name recognition. Beyond spending in itself, campaign appeals with substantive content may affect voters' decisions in the same ways.

The research on campaigns and voting in supreme court elections has focused on contests between incumbents and challengers. This research has shown a strong relationship between candidates' relative spending levels and their vote shares, and there is some evidence that the challenger's spending has a stronger relationship with vote shares than does spending by the incumbent (Bonneau Reference Bonneau2007; Bonneau and Cann Reference Bonneau and Cann2011). It is likely, however, that some of this relationship stems from the advantage of strong candidates in attracting campaign contributions. Studies cannot take independent spending on behalf of candidates into account systematically because it is not generally reported, but such spending probably has the same kinds of effects as spending by candidates themselves.

In her book, Hall (2015, 111, 119) provides additional evidence of the impact of candidate spending on vote shares. More important, she analyzes the impact of different types of advertising by incumbents and challengers and by parties and interest groups on their behalf (Hall 2015, 100–23). Her findings are rich and complex. They indicate that even when controlling for candidates' spending, advantages in advertising volume benefit candidates; to a degree, that finding may capture the impact of independent spending. Of particular interest is the impact of attack ads against incumbent justices. Hall (2015, 113) finds evidence that the volume of attack ads does depress the incumbent's vote share, though that effect—like the analogous effect on voter participation—seems to be limited to nonpartisan contests.

A quite different set of questions about the public concerns the effects of changes in supreme court campaigns on attitudes toward courts. In particular, scholars and other commentators have expressed considerable worry about the impact of attack ads and issue-based appeals to voters on perceptions of the courts' legitimacy. Scholars have done some research on the impact of elections as an institution and attributes of election campaigns on public support for the courts, including an analysis of the impact of selection system type and attack advertising on support for courts in the Kritzer book (2015, 86–92).Footnote ²⁶

Gibson has been the primary contributor of research on this set of questions, and the largest portion of his research is reported in his book (see also Gibson et al. Reference Gibson, Gottfried, Delli Carpini and Jamieson2011). He uses his Kentucky panel survey to address several questions about the relationship between elections and public support for the state supreme court, and his national survey provides additional evidence on some of those questions.

Gibson probes the impact of three elements of supreme court campaigns that have become more common in the current era: monetary contributions from litigants and interest groups, expressions of policy positions, and use of attack ads against opponents. He finds that acceptance of contributions from interested parties has a negative effect on a court's perceived legitimacy. The findings on the other two elements are mixed, with some evidence indicating that the specific content of attack ads makes a difference and that people respond negatively only to statements that sound like promises to take particular positions in cases.

These findings can be interpreted in multiple ways. Gibson's interpretation reflects the evidence from his panel study that the experience of going through a judicial election in itself enhanced support for the Kentucky Supreme Court: “even with the effects of undesirable campaign activity, the net effect of judicial elections is positive” (Gibson 2012, 140, emphasis in original).Footnote ²⁷ However, the negative effect of some kinds of campaign appeals and the decidedly negative effect of monetary contributions from people with a stake in court decisions may strengthen the concerns of some scholars and commentators about the impact of changes in supreme court campaigns. In a broader sense, of course, the implications of the findings depend on one's judgment about the importance and desirability of support for the courts and acceptance of their legitimacy.

Other potential effects of changes in supreme court elections concern court policies. One possibility is that the growth of large-scale campaigns that frequently include negative ads and issue-based appeals has given justices a stronger incentive to avoid taking positions that might be used against them in a future campaign. Evidence from studies of sentencing by trial judges (Gordon and Huber Reference Gordon and Huber2007; Berdejó and Yuchtman Reference Berdejó and Yuchtman2013) suggests that, at least in criminal justice, judges are influenced by the prospect of facing the voters in the near future. Hall (Hall Reference Hall1987, Reference Hall1992, Reference Hall1995) found evidence indicating that some elected supreme court justices responded to perceived public support for the death penalty even before changes in the scale and content of supreme court contests had proceeded very far (see Brace and Boyea Reference Brace and Boyea2008). Kritzer's (2015, 59–84) summary and analysis of the research on the effects of electoral concerns on justices' behavior identifies additional evidence of such effects (see also Kang and Shepherd Reference Kang and Shepherd2016, 942–48).

It seems likely that these effects have strengthened as attacks on justices for their decisions have become more common and some justices have suffered well-publicized defeats after they were the recipients of such attacks. However, there is little direct evidence on that question. One study of death penalty decisions did find evidence that the well-publicized defeats in retention elections of three California justices in 1986 made justices in states with retention election more likely to uphold death sentences (Canes-Wrone, Clark, and Kelly Reference Canes‐Wrone, Clark and Kelly2014). That study is also a reminder that electoral concerns are not limited to states with partisan or nonpartisan elections (see also Huber and Gordon Reference Huber and Gordon2004; Shepherd Reference Shepherd2009b).

A second possibility is that campaign contributions and independent spending on behalf of judicial candidates buy influence for the sources of those funds. That possibility clearly troubled Gibson's respondents, as it does many commentators as well. Those who seek to ascertain the actual influence of contributions confront the difficulty of distinguishing that influence from the tendency for contributors to support candidates who are already favorable to the contributors' positions. Scholars who study the impact of contributions to supreme court candidates typically address this difficulty by controlling for justices' ideological positions in some way (e.g., McCall Reference McCall2003; Cann Reference Cann2007; Shepherd Reference Shepherd2009c). Most studies do find an impact for contributions, though there have been a few exceptions (e.g., Cann Reference Cann2002). In his review of this body of work, Kritzer (2015, 80–84) concludes that we still lack strong evidence about the independent effects of contributions on justices' decisions. In any case, there is no direct evidence about whether growth in the costs of supreme court campaigns has given contributors more influence over the successful candidates they support.

Some funders of supreme court campaigns, especially lawyers, do seek to ingratiate themselves with the justices they support. However, by and large, the interest groups that participate in campaigns seek primarily (or solely) to populate the bench with people who share the groups' preferences. Certainly that has been true of the business community. Of course, interest groups do not pursue changes in the scope and content of campaigns for their own sake; they would be perfectly happy to help elect candidates they favor in traditional low-key campaigns. Indeed, advantages in spending that groups can provide to candidates may have their greatest effects in such campaigns by enhancing the name recognition of favored candidates.

We know that the membership of appellate courts has considerable impact on their decisions and policies, and that certainly is true of state supreme courts. Long before the development of strong ideological separation between the Republican and Democratic parties, there were substantial relationships between justices' party affiliations and their decisional tendencies (Nagel Reference Nagel1961; see Dubois Reference Dubois1980, Chs. 5–7). For the current era, my colleagues and I have found that the election of one candidate rather than another can have a substantial impact on case outcomes in criminal and tort law (Baum, Gray, and Klein forthcoming). There is fairly clear evidence of such impact in the course of supreme court membership and policy in states such as Alabama (Ware Reference Ware1999), Ohio (Baum, Klein, and Streb 2017), and Texas (Cheek and Champagne Reference Cheek and Champagne2005).

If election outcomes clearly shape judicial policy, it is more difficult to ascertain the impact of interest-group activity on election outcomes. Undoubtedly, the growth in relatively large-scale supreme court campaigns has affected which candidates are elected, but not all well-funded interest-group campaigns are successful (Tarr Reference Tarr2012, 83), and some successful efforts may reflect conditions such as favorable party tides more than they reflect the attributes of the campaigns themselves. The success of Republican candidates in Alabama and Texas owes something to the mobilization of support from conservative groups, but it may owe even more to the increasingly Republican electorates of those states. It seems certain that changes in the scale and content of supreme court campaigns have had a meaningful impact on supreme court policy, but more precise judgments about the extent of that impact cannot be made with confidence.

Change in Supreme Court Elections and Choices Among Selection Systems

Any scholar who writes about judicial elections does so in the context of debates about the desirability of elections as a means to choose judges. Each of the books that I have considered considers the implications of its findings for those debates. I have already considered Gibson's judgment that features associated with current supreme court contests do not harm the courts' legitimacy. More broadly, Gibson (2012, 129–36) concludes that elections as an institution enhance legitimacy by giving citizens a mechanism for accountability of the courts and reminding them that this mechanism exists.

Bonneau and Cann (2015, 105–07) emphasize the value of transparency (which encompasses visibility to the public and public involvement) in the process of judicial selection. They see elections as more transparent than other systems used in the United States, though they note that those other systems could be made more transparent.

Hall (2015, 169–77) concludes that, on the whole, the features of contemporary supreme court elections that have aroused criticism do not have significant negative effects. This includes negative campaigning, the primary focus of her inquiry. In particular, she emphasizes that attack ads do not depress voter participation in supreme court contests. She points to other evidence that elections work well, reinforcing the arguments and evidence in Bonneau and Hall (Reference Bonneau and Gann2009).

Kritzer (2015, 241–64) ranges most widely in his examination of choices among selection systems, giving consideration to systems that are used in the United States but that receive little attention (such as the appointments of federal magistrate judges and Illinois associate judges by other judges) as well as systems that are common elsewhere in the world. Recognizing that changes that would leave no role for the electorate in the selection of judges are highly unlikely, he suggests some ways that the existing systems could be improved—though he also recognizes that even these more limited changes are quite unlikely to be adopted.

These authors' conclusions certainly merit attention. It is also useful to consider more broadly the implications of their findings about change for the debates over judicial elections.Footnote ²⁸ The place to start is the view of many legal scholars and other reformers that judicial elections, always undesirable, have become worse with the changes that have occurred since the 1980s. As I have emphasized, most of these critics are really talking specifically about supreme court elections.

The most important lesson of research on supreme court elections, one documented by the Hall and Kritzer books, is that the extent of change in those elections has been exaggerated by many commentators. Certainly, significant changes have occurred. Most fundamentally, there has been growth in campaign spending that in turn has fueled other changes. But those changes have not been ubiquitous; they have touched some states only to a limited degree, and in other states there is considerable variation among election years and individual supreme court contests in the scale and content of campaigns. Nor have there been dramatic increases in the competitiveness of supreme court elections and the vulnerability of incumbents to defeat.

The changes in supreme court elections that have occurred certainly are relevant to assessment of elections as a means to choose judges. For those who think that judicial elections were already functioning in undesirable ways, the changes of the past few decades may indeed make a bad situation worse. But judgments about the value of judicial elections should not give undue weight to those changes. And on the basis of the evidence that we have now, it is at least premature to generalize from trends in supreme court elections to the lower courts.Footnote ²⁹

Ultimately, for those who seek the most desirable system for selection of judges, judicial elections must be compared with the alternatives. Leaving aside other possible systems, the existing alternatives in the states are primarily the Missouri Plan and gubernatorial and legislative appointment. Of course, the Missouri Plan includes retention elections. Bonneau and Cann (2015, Ch. 5) and Kritzer (2015, Ch. 7) both analyze voting in retention elections, Bonneau and Cann with their survey data and Kritzer through analysis of county-level correlations between gubernatorial and supreme court elections. Each finds evidence of partisan voting. Bonneau and Cann show that both Republicans and Democrats were more likely to vote to retain justices who shared the voter's party affiliation. However, Democrats also had a stronger propensity to vote to keep justices in office than did Republicans: all else being equal, a Democratic voter was more likely to vote to retain a Republican justice than was a Republican voter. Kritzer's analysis of county-level correlations between supreme court and gubernatorial voting shows a tendency for those correlations to strengthen over time in absolute terms: voting in retention elections has become more partisan. The correlations also suggest that relative to Republicans, Democrats have become more likely to vote for retention over time.

Those findings reflect the same forces that have brought about changes in “regular” supreme court elections, including campaigns that charge incumbents with excessive liberalism on criminal justice and other issues. Retention elections remain much safer for incumbents than partisan or nonpartisan elections, but justices have reason to feel more vulnerable than they did in earlier periods.

The nomination and appointment stages of the Missouri Plan and the selection process in gubernatorial and legislative appointment systems have received relatively little attention from scholars, so we have only limited information on those processes. The little that we do know is consistent with what we would expect: appointment processes sometimes work in ways that would generally be considered undesirable, and considerations other than the merit of prospective judges have considerable weight in the selection of judges (e.g., Watson and Downing Reference Watson and Downing1969; Fitzpatrick Reference Fitzpatrick2009; McLeod Reference McLeod2012).

Moreover, there is evidence of changes in appointment processes similar to changes in judicial elections, including growth in partisanship and greater scrutiny of incumbent judges who come up for reappointment (Cooper Reference Cooper2006; Pérez-Peña Reference Pérez‐Peña2010). Those changes may be responsible for the findings of one study indicating that justices who face reappointment decisions are more favorable to state governments as litigants (Shepherd Reference Shepherd2009a). The evolution of judicial appointment at the federal level over the past few decades parallels what has happened in state judicial elections in important respects, in that both presidents and senators are more attentive to the policy positions of prospective and actual nominees. However, that change has been more substantial and more pervasive at the federal level (Binder and Maltzman Reference Binder and Maltzman2009; Steigerwalt Reference Steigerwalt2010).

Even a body of research as extensive and impressive as this set of four books is unlikely to have much impact on debates over alternative systems for selection of judges. The positions of proponents and opponents of judicial elections are deeply rooted in their values, their interests, or a combination of the two. Those who oppose judicial elections on principle will not be swayed by evidence that supreme court elections have changed less than they had thought. Business leaders who think they have gained more desirable judicial policy through their participation in supreme court campaigns would have no reason to give up a favorable playing field unless research demonstrated that their efforts actually had no impact. And participants in public policy debates, like other people, process new information in ways that accord with their biases.Footnote ³⁰

Moreover, there are reasons to be cautious about making policy recommendations on the basis of what we have learned about the functioning of judicial elections and how those elections have changed. Most important, there is still a good deal that we do not know. Even at the supreme-court level, for instance, information on the content of campaigns in forms other than television advertising is thin. We have a quite limited sense of how voters respond to campaigns, including the more issue-oriented and negative campaigns that have occurred in some states. Because changes in selection systems usually are not limited to supreme courts, it is significant that there is so small a body of knowledge on elections to other courts. And because judicial elections should be compared with their alternatives, the scattered information that we have on nomination and appointment in other systems makes it difficult to undertake meaningful comparisons of alternative rules for judicial selection.

The issue on which the existing research provides the strongest basis for judgment is the choice between partisan and nonpartisan elections. Kritzer (2015, Ch. 6) presents evidence that changes in supreme court elections have reduced the differences between the two forms of election by increasing the level of partisan voting in nonpartisan elections, and the findings in Bonneau and Cann (2015, Ch. 3) are consistent with that conclusion. But there are still differences between the systems, differences that stem largely from the presence or absence of candidates' party affiliations on the ballot. The absence of that information allows more room for other considerations to influence voters' choices, especially information that can be gleaned from candidates' names such as gender, ethnicity, and actual or perceived name recognition.Footnote ³¹ As Hall (2015, 168) points out, nonpartisan elections also provide a greater potential for campaign-based information such as attack advertising to influence voters.

From the perspective of most political scientists, it is better that voters make choices on the basis of an attribute that correlates with the kinds of positions a candidate would take on the bench—party affiliation—than on the basis of attributes inferred from candidates' names that have little to do with their prospective performance as judges. That is especially true for supreme courts, which make important policy choices and for which we have substantial evidence of differences in policy positions between Republicans and Democrats. But those who think that partisan identifications and ideological positions should be irrelevant to the selection of judges will not be convinced on that point. And if interest-group leaders perceive that their chances of influencing voters' judgments are enhanced by nonpartisan elections, they would see no point in adopting a less advantageous system.

Regardless of whether research on judicial elections has an impact on the debates over selection of judges, its value for our understanding of an important institution is clear. Indeed, the largest share of what we know about elections for offices that are not among the most visible ones comes from the research on state supreme courts. Certainly, continued research on candidates, campaigns, and voters in judicial elections is merited.

As scholars undertake research in the future, the books that I have reviewed provide excellent models to follow. Each book addresses important questions with care and insight. The research that they report adds considerably to the body of research on elections to state supreme courts. Thus these four books are important sources for those who are interested in the selection of judges and in campaigns and elections more generally.