Assessing the Electoral Impact of the 2010 Oregon Citizens’ Initiative Review

Abstract

The Oregon Citizens’ Initiative Review (CIR) distinguishes itself by linking a small deliberative body to the larger electoral process. Since 2010, CIR citizen panels have been a legislatively authorized part of Oregon general elections to promote a more informed electorate. The CIR gathers a representative cross-section of two dozen voters for 5 days of deliberation on a single ballot measure. The process culminates in the citizen panelists writing a Citizens’ Statement that the secretary of state inserts into the official Voters’ Pamphlet sent to each registered voter. This study analyzes the effect of one such Citizens’ Statement from the 2010 general election. In Study 1, an online survey experiment found that reading this Statement influenced Oregon voters’ values trade-offs, issue knowledge, and vote intentions. In Study 2, regression analysis of a cross-sectional phone survey found a parallel association between the Statement’s use and voting choices but yielded some mixed findings.

Keywords

deliberation direct democracy election reform initiatives and referenda political knowledge public opinion

Over the past three decades, the study of deliberative democracy has evolved from a principally theoretical enterprise to an empirical and practical one. Early formulations of deliberation had an abstract quality that contrasted an idealized form of speech against conventional political discourse (Cohen, 1989; Habermas, 1989; Mansbridge, 1983). This critique spurred an interest in the actual workings of either proto-deliberative public spaces (Jacobs, Cook, & Delli Carpini, 2009; Mutz, 2006; Neblo, Esterling, Kennedy, Lazer, & Sokhey, 2010) or civic innovations designed to produce high-quality deliberation (Nabatchi, Gastil, Weiksner, & Leighninger, 2012; Ryfe, 2005). Meanwhile, new discussion models, such as Deliberative Polls (Fishkin, 2009) and Citizens’ Juries (Crosby, 1995; G. Smith & Wales, 2000), made deliberative democracy into a practical approach to political reform (Grönlund, Bachtiger, & Setälä, 2014; Leighninger, 2006).

Nonetheless, a problem of scale remains for both deliberative theorists and practitioners (Gastil, 2008; Goodin & Dryzek, 2006; Lukensmeyer, Goldman, & Brigham, 2005). One can hear a reflective and articulate public voice through highly structured small-scale deliberation (Fishkin, 2009; Nabatchi et al., 2012), but can such processes effectively influence a wider public? Doubts about the potential for forums and discussions to scale up have helped spur a more systemic approach to deliberation that examines how macro-level institutions can embody or promote deliberation (Parkinson & Mansbridge, 2012), or tie back to micro-level processes, such as citizen panels or juries (Gastil, 2000). For example, one approach to bridging this gap was the British Columbia Citizens’ Assembly, which used a small deliberative body to draft an electoral reform put to a public vote (Warren & Pearse, 2008). Although the Assembly’s recommendation did not meet a supermajority requirement, it did earn 57% support and suggested the potential for making such linkages.

A new opportunity to study the connection between micro-level deliberation and macro-level institutions comes from a unique electoral reform in the state of Oregon. On June 26, 2009, that state’s governor signed House Bill 2895, which authorized Citizens’ Initiative Review (CIR) panels for the general election. In the summer of 2010, two stratified random samples of 24 Oregon citizens deliberated for 5 days on two separate ballot initiatives (Knobloch, Gastil, Reedy, & Cramer Walsh, 2013). At the end of each week, these CIR panelists produced a written Citizens’ Statement that the secretary of state prominently placed in the Voters’ Pamphlet sent to each registered voter. Official state voter guides constitute a popular platform that can potentially sway a large portion of the electorate (Bowler & Donovan, 1998), though they are used more often by more interested and knowledgeable voters (Mummolo & Peterson, 2016). A CIR Statement could have greater influence in Oregon because it is a vote-by-mail state where one’s ballot and pamphlet arrive in sync. Thus, the Oregon CIR presents a critical case in which officially sanctioned small-scale deliberation has a mechanism for influencing elections.

This study examines the sturdiness of the bridge the CIR aims to build between an intensive weeklong deliberation and the less reflective, or more heuristic (Lupia, 2015; Popkin, 1994), behavior of a large electorate. We use experimental and cross-sectional surveys to examine associations between CIR Statement use and voters’ policy-relevant knowledge, value considerations, and voting choices. To examine these questions, we provide a wider theoretical context for linking micro- and macro-level deliberation in initiative elections. We conclude by reviewing our findings’ implications for deliberative democratic theory, the CIR in particular, and other electoral reforms that aim to connect deliberative processes at different social scales.

Deliberative Designs and Direct Democracy

Deliberation scholarship fits within a broader reformist tradition in democratic theory (Chambers, 2003; Dryzek, 2010). As Robert Dahl (2000) wrote in On Democracy, “One of the imperative needs of democratic countries is to improve citizens’ capacities to engage intelligently in political life,” and “older institutions will need to be enhanced by new means for civic education, political participation, information, and deliberation” that fit modern society (pp. 187-188).

This call for new institutions has been answered by ambitious projects, such as the Australian Citizens’ Parliament (Carson, Gastil, Hartz-Karp, & Lubensky, 2013). Relatively few of these efforts, however, have exercised legal authority (Barrett, Wyman, & Coelho, 2012), with prominent exceptions being Canadian Citizens’ Assemblies (Warren & Pearse, 2008) and Deliberative Polls in China (He & Warren, 2012; Leib & He, 2006). One feature those exceptions have in common is that they focus on a narrow question preset by the same governmental body that authorized the deliberation. This takes agenda-setting power away from the deliberative body, but in that respect, it parallels a venerable form of citizen deliberation in which juries answer the narrow questions a judge puts before them (Vidmar & Hans, 2007).

The Oregon CIR also has a narrow agenda, even if it stands atop a petition process that lets the public put issues on the ballot (Altman, 2010; Bowler & Donovan, 1998). Another essential feature the Oregon CIR shares with Citizens’ Assemblies is cross-level deliberation: What happens on the small scale is designed to influence the process and outcomes in a large-scale election (Ingham & Levin, 2017; Warren & Gastil, 2015). When looking across levels, the core meaning of deliberation remains the same (i.e., a process of learning, moral reflection, and considered judgment) even as its behavioral meaning shifts away from face-to-face discussion at the micro level to impersonal information flows and decisions at the macro level (Gastil, 2008).

Initiative elections provide a particularly important context in which to study cross-level deliberation. In spite of their broad popularity (Collingwood, 2012), initiative elections have received mixed reviews for their fidelity with public preferences (Flavin, 2015; Matsusaka, 2008; Nai, 2015) and their implications for minorities, in particular (Hajnal, Gerber, & Louch, 2002; Lewis, 2011; Moore & Ravishankar, 2012). There is also reason to be concerned about the quality of information on which voters judge such laws (Broder, 2000; Gastil, Reedy, & Wells, 2007; Milic, 2015; Reedy, Wells, & Gastil, 2014; Saris & Sniderman, 2004). Voters, however, can learn new information during elections—information that can influence voting decisions (Lavine, Johnston, & Steenbergen, 2012; Rogers & Middleton, 2014; D. A. Smith & Tolbert, 2004). Moreover, the relative power of motivated reasoning depends on the political environment. As Leeper and Slothuus (2014) explain, “people adopt different reasoning strategies when motivated to obtain different end states” (p. 142).

Thus, the problem may lie not with initiative elections in general but with the availability of accessible and trustworthy information for voters (MacKenzie & Warren, 2012; Warren & Gastil, 2015). Even voters with favorable attitudes toward direct democracy recognize this problem and support reforming the process to reduce the net influence of conventional, and often misleading, campaign advertising (Baldassare, 2013; Dyck & Baldassare, 2012). The Oregon CIR aims to provide such an intervention, which voters could come to perceive as a neutral and accessible information source.

A recent set of survey experiments by Boudreau and MacKenzie (2014) suggest the potential efficacy of such an approach. During the 2010 California general election, these investigators measured baseline attitudes toward an initiative on the legislature’s budget process, then conducted a 2 × 2 experiment. A “party cue” treatment then told voters where the Democratic and Republican parties stood on the issue, whereas a “policy information” treatment provided information about a fiscal hazard in the status quo that the initiative would alleviate. The results showed that “rather than blindly follow their party, citizens shift their opinions away from their party’s positions when policy information provides a compelling reason for doing so” (Boudreau & MacKenzie, 2014, p. 60). A more recent experiment found that a mock minipublic had modest influence on public attitudes toward Social Security (Ingham & Levin, 2017).

In theory, therefore, the Oregon CIR could serve as an effective source of policy information, generated by a small deliberative body for the benefit of a mass public. Although it does not exercise legislative authority, the Oregon CIR could have substantial influence on the electorate through the publication of its Citizens’ Statement, which first appeared in the official 2010 Oregon Voters’ Pamphlet. The focal question of this study is whether that mechanism operates as intended by giving voters information and analysis they put to actual use when deliberating on the corresponding ballot measures.

Hypothesizing the CIR’s Impact

To explain the potential impact of the CIR, we begin by describing the process itself in greater detail. After the Oregon legislature established the CIR in 2009, separate citizen panels were assembled for two statewide ballot measures in the 2010 general election. This article focuses on the Citizens’ Statement written by the first CIR panel, which studied Measure 73, an initiative that set a 25-year minimum sentence for multiple counts of certain felony sex crimes and toughened the penalties for repeat DUIs (driving under the influence). The second panel looked at a measure establishing medical marijuana dispensaries. Both CIR panels constituted stratified samples of 24 Oregon voters who had their expenses covered and were compensated at a rate equal to the state’s average wage.

An intensive field study found that these panels met high standards for deliberative quality in terms of analytic rigor, democratic discussion, and well-reasoned decision making (Knobloch et al., 2013). The 2010 CIR panel on mandatory minimum sentencing met for 5 consecutive days, using a process adapted from the Citizens’ Jury model (Crosby, 1995). The citizen panelists received extensive process training, met with advocates and policy experts, and still had considerable time for facilitated deliberation—both in smaller subgroups and as a full body—before writing their official CIR Statement for the Voters’ Pamphlet. The panelists collectively wrote the Key Findings section of their Statement, which contained initiative-relevant empirical claims that a supermajority of panelists believed to be factually accurate and important for voters to consider when casting their ballots.

On the final day, panelists divided into pro and con caucuses to write rationales for supporting or opposing the measure, but the full panel reviewed even these separate sections before settling on the final version. The Statement additionally included a brief description of the CIR process and showed the number of panelists voting for or against the measure. Afterward, the Oregon secretary of state put the CIR Statement into the Voters’ Pamphlet (see supplementary material). The Statement had a favorable location in the Pamphlet, as it appeared before the paid pro and con arguments submitted by organizations and individuals.

Even so, the Oregon CIR would fail to achieve its intended purpose if voters ignored or dismissed entirely the CIR Statement in the Voters’ Pamphlet. Such minimal electoral influence was a distinct possibility. The 2010 initiative campaigns were low-visibility affairs, with limited campaign spending both pro and con in a non-Presidential election year.¹ In that election, voters also had no prior frame of reference for the CIR. Voters generally view official guides as useful sources of information for ballot measures (Bowler, 2015; Bowler & Donovan, 1998; Canary, 2003), but the same information-seeking voters who use these guides could also prove least likely to be persuaded by the addition of the CIR to their information pool (Mummolo & Peterson, 2016; Valentino, Hutchings, & Williams, 2004).

If a CIR Statement does influence voters, it could do so in at least three ways, as summarized in Table 1. First, the recommendation of the majority of CIR panelists could serve as a powerful heuristic for initiative voting (Lupia, 2015), and such cues have been shown to have impacts even on unsophisticated voters (Boudreau, 2009; Goren, 2004). Second, the pro/con sections of the Statement could influence readers’ values trade-offs and voting choices (Lau & Redlawsk, 2006). Third, the Key Findings portion of the Statement could improve the accuracy of voters’ understanding of empirical issues relevant to the initiatives (Estlund, 2009; Luskin, Fishkin, & Jowell, 2002).

Table 1.

Hypothesized Paths of CIR Influence on Initiative Voting Choices.

Element of CIR Statement	Hypothesized effect	Empirical evidence of this influence for CIR Statement readers versus nonreaders, controlling for other influences
The final vote of the CIR panelists	Direct voting cue	Readers learn how panelists split on the measure and become more likely to vote in alignment with the majority
Values addressed in the Pro/Con sections	Shift in judgments about values trade-off	Readers change their stances on the perceived trade-offs among values that influence their voting choices
Values addressed in the Pro/Con sections	Values-voting linkage strengthened	Readers develop values trade-off judgments more predictive of their voting choices
Factual claims validated/rejected in Key Findings and Pro/Con sections	More accurate empirical beliefs	Readers develop more accurate beliefs on the initiative-relevant empirical claims that influence their voting choices
	Empirical beliefs-voting linkage strengthened	Readers’ final vote preferences become more dependent on the balance of their relevant empirical beliefs

Note. CIR = Citizens’ Initiative Review.

As to the first of these effects, when the CIR panelists do not split evenly (and they broke 21 to 3 on the sentencing issue studied herein), the balance of panel votes could serve as a powerful signal. A substantial majority of voters claim they need more accurate electoral information (Baldassare, 2013; Canary, 2003), and those lacking both political knowledge and partisan allegiances are particularly rudderless in initiative elections (Gastil, 2000). In those cases, endorsement messages can have considerable sway (Bowler & Donovan, 1998; Burnett & Kogan, 2015; Burnett & Parry, 2014; Lupia, 2015), so the implicit advice of voters’ peers could persuade those seeking a trustworthy recommendation.

Second, when the arguments in the CIR Statements invoke values, they could shift how voters judge corresponding value trade-offs in the initiatives (Yankelovich, 1991). Such judgmental shifts could alter the credence or priority that voters give to opposing values claims. On such values questions, voters’ cultural orientation (Gastil, Braman, Kahan, & Slovic, 2011; Kahan, Braman, Gastil, & Slovic, 2007) and their liberal-conservative self-identification (Zaller, 1992) provide strong guidance, but there can be some slippage between voters’ values and their choices (Lau & Redlawsk, 2006). The Statement’s Pro and Con section could expose voters to relatively clear and often value-based arguments that increase the values–votes correspondence. The CIR Statements are written by lay citizens, whom voters might view as especially credible (Gastil, 2000; Leighninger, 2006), and CIR panelists are likely to use language that parallels that of the general public, thereby making it easier for readers to synchronize their values with one or the other side of the initiative. In sum, the Statement could shift voters’ value trade-off judgments while strengthening the ties between their values and their votes.

Finally, reading the Key Findings and Pro and Con arguments in the CIR Statement could cause voters to consider new information (Cappella, Price, & Nir, 2002). Individuals’ issue-relevant empirical beliefs can become distorted through the lenses of their prior political and cultural commitments (Jerit, Barabas, & Bolsen, 2006; Kahan et al., 2007; Kuklinski, Quirk, Schwieder, & Rich, 1998) and even resist corrective messages (Nyhan & Reifler, 2010). This can occur even in low-information environments, such as with statewide ballot measures (Reedy et al., 2014; Wells, Reedy, Gastil, & Lee, 2009).

To the extent that the neutral portion of the CIR Statement provides empirical content and analysis, it could serve as an information conduit that overrides the more biased claims typical of initiative elections (Boudreau, 2009; Boudreau & MacKenzie, 2014; Broder, 2000; Ellis, 2002) and the ideological cues found in ostensibly neutral fact-checking efforts (Garrett, Nisbet, & Lynch, 2013). Knowledge gains of this sort have been observed in previous public forums with less intensive deliberative designs (Farrar et al., 2010; Grönlund, Setälä, & Herne, 2010). As with values, the net effect could include not only more accurate beliefs but also stronger links between empirical beliefs and the voting choices they buttress.

To test these potential impacts, we present two studies. The first assesses causal influence through an online survey experiment with likely voters exposed to different stimuli. The second uses a cross-sectional phone survey of Oregonians who had already voted to estimate the independent association between reading the CIR Statement and voters’ attitudes toward the corresponding ballot measure.

Study 1

An online survey experiment was designed to assess whether the CIR Statement was even capable of changing voters’ preferences, attitudes, and beliefs. Measure 73 (hereafter called the “sentencing measure”) provided the clearest opportunity for CIR influence. This measure’s ballot title said that it “requires increased minimum sentences for certain repeated sex crimes, incarceration for repeated driving under influence.” This included raising “major felony sex crime” minimums from 70 to 100 months up to 300 months and setting a 90-day minimum class C felony sentence for a second offense of driving under influence of intoxicants (DUII).

A September phone survey pegged statewide voter support for the measure at 67% to 73%,² but the CIR panelists wrote a scathing critique and sided against it 21 to 3. The testimony against the measure carried considerable sway during the CIR’s 5-day deliberations, which revealed potential unintended consequences of the proposed law even its proponents had not considered fully (Knobloch et al., 2013).

This contrast created the opportunity for the CIR Statement to have an impact when it appeared in the October Voters’ Pamphlet. Support for the sentencing measure dropped in the final days of the election to just 57% of the vote, though it is common for initiatives to lose a degree of support as the election approaches—even without a strong opposition campaign (Bowler & Donovan, 1998). Our survey experiment tested whether the CIR could have accounted for some of the measure’s lost support.

Method

Survey sample

We collected a sample of 415 respondents from an online poll conducted by YouGov/Polimetrix from October 22 to November 1, 2010. The target population was registered Oregon voters who said they were likely to participate in the election, excluding those who had already voted or read the Voters’ Pamphlet. The RR3 response rate was 41% and approximated Oregon’s party registration and ideological profile. (Statistical power and missing data imputation are discussed at the end of this section.)

Experimental treatment

The experimental manipulation came at the front of the survey, immediately following screening questions. Respondents were assigned at random to one of four groups:

Those in a control group received no further instruction and proceeded to the survey;

Those in a modified control group were shown an innocuous letter from the secretary of state introducing the Voters’ Pamphlet;

Those in the third group saw the official Summary and Fiscal Statement on the sentencing measure—the same content that appeared in the Pamphlet; and

Those in the fourth group saw the full CIR Statement on the sentencing measure (see supplementary material).

Survey measures

After the experimental treatment, the survey posed the following question:

One of the issues in this year’s general election is statewide Initiative Measure 73, which would increase mandatory minimum sentences for certain sex crimes and DUI charges. Do you plan to vote yes or no on Measure 73, or have you not decided yet?

Those who declared themselves undecided were asked the follow-up, “If the election were being held today and you had to decide, would you probably vote yes or no on Measure 73?” Those who initially gave an answer of “Yes” or “No” were asked, “Are you fairly certain you will vote [Yes/No] on Measure 73, or is there a chance you could change your mind?” These questions yielded a dichotomous measure of Sentencing support (yes = 1, no = 0) and a 7-point Sentencing support certainty scale that ranged from −3 (certain to oppose) to +3 (certain to support).

To test whether reading the CIR Statement boosted voters’ confidence in their decision, a simple yes/no question came next: “Would you say you’ve received enough information on Measure 73 to make a well-informed vote, or not?” To probe the CIR’s utility as a voting cue, the survey also asked respondents if they could locate the CIR panelists’ position on the sentencing measure on a 5-point scale from strong support to strong opposition. (The supplementary material provides complete wording for this and other survey items.)

Subsequent batteries of randomized items measured values and empirical beliefs based on preliminary analysis of the campaign arguments advanced for and against the sentencing measure. The four values items presented arguments for and against the measure as trade-offs among conflicting goods (e.g., “Even for potentially violent crimes, mandatory minimum sentencing is unjust because it fails to consider individual circumstances”). When not analyzed individually, these items combined into a pro-sentencing values scale (α = .71), with a range of −1.5 (strongly disagree) to +1.5 (strongly agree), M = 2.84, SD = 0.69.

The survey also asked respondents if they believed each of six empirical statements (e.g., “Mandatory minimum sentencing has already raised Oregon’s incarceration rate well above the national average”). Those who responded that they were “not sure” were prompted to state whether they believed the statement was “probably true” or “probably false.” This yielded a 7-point scale from −3 (definitely false) to +3 (definitely true). The supplementary material shows how these six statements related to the content of the CIR Statement, but in each case, one or more Statement sentences could warrant an inference about these statements’ veracity. For later regression analyses, these items were combined into an index that aligned beliefs based on whether they buttressed or undermined arguments for the sentencing measure. Beliefs that supported the measure were coded as 1, those opposing as −1, and “not sure” as 0. Averaging scores on the items yielded a pro-sentencing empirical beliefs index with M = −0.16, SD = 0.29.

Power and missing data analysis

With a minimum cell size of 96 in the four experimental conditions, this study had ample statistical power to detect even small effects (Cohen, 1988). Missing data owing to nonresponse occurred in only 10 cases (2.4%) for the main dependent variable (attitude toward sentencing measure) and less often for all other measures except the values trade-off items, where between 9.6% and 16.1% of respondents declined to state a position on a given item. For analyses using those items, a linear regression model with five imputations was employed. Results were approximately equivalent to nonimputed analyses, with any changes in statistical significance noted in the text.

Results

Overall impact

Analysis began with a straightforward cross-tabulation of the experimental treatment by the measure of Sentencing support certainty. As shown in Table 2, the distribution of responses on the 7-point scale varied significantly across the four experimental conditions, χ²(18, N = 405) = 48.5, p < .001.

Table 2.

Certainty of Voting Position on Sentencing Initiative by Experimental Conditions in Study 1.

Level of sentencing support certainty	Experimental condition
Level of sentencing support certainty	Control group	Modified control group*	Read summary and fiscal statement	Read CIR Statement
Oppose, certain	11.2%	17.6%	17.9%	22.8%
Oppose, could change mind	5.6%	2.9%	9.5%	5.0%
Undecided, probably oppose	8.4%	6.9%	5.3%	20.8%
Undecided, not leaning	23.4%	19.6%	8.4%	19.8%
Undecided, probably favor	14.0%	10.8%	11.6%	14.9%
Favor, could change mind	7.5%	11.8%	8.4%	6.9%
Favor, certain	29.9%	30.4%	38.9%	9.9%
Total valid responses	107	102	95	101
Declined to answer question	4	2	1	3

Note. * The modified control group read an innocuous pdf, which was a letter from the secretary of state about the administration of the election and the Voters’ Pamphlet. CIR = Citizens’ Initiative Review.

The most striking differences were between the CIR condition and all others. Those reading the CIR statement swung against the measure: 27.8% said they would oppose it, and another 20.8% who expressed uncertainty initially leaned against it. Across the other three conditions, the comparable figures were 21.4% initially opposed and only 6.9% leaning against. A corresponding drop in strong support also occurred, with only 9.9% of CIR readers certain in their support, compared with 32.9% of all others.

Treating the 7-point Sentencing support certainty metric as a continuous variable, the CIR Statement condition yielded lower average scores (M = −0.41, SD = 1.91) than for the control (M = 0.65, SD = 2.02), secretary of state letter (M = 0.60, SD = 2.19), and summary and fiscal statement conditions (M = 0.67, SD = 2.37), F(3, 401) = 6.19, p < .001. Post hoc Tukey’s honestly significant difference (HSD) tests showed significant contrasts between each condition and the CIR treatment (max. p = .005).

Looking just at the binary measure of Sentencing support, a majority of respondents were in favor of the measure in the control condition (67.1%), the modified control (65.9%), and the summary and fiscal statement condition (64.4%). In the CIR condition, only 39.5% intended to vote for the measure—a drop of more than 25 percentage points, χ²(3, N = 332) = 18.2, p < .001.

Evidence of CIR as voting cue

The net impact of reading the CIR Statement was substantial, but was it the result of a straightforward voting cue? Those exposed to the Statement could read that opposition to the measure was the “POSITION TAKEN BY 21 OF 24 PANELISTS.” When asked if they recalled the “position taken by the Citizens’ Initiative Review panelists,” 43.3% recalled correctly that “a large majority OPPOSED the measure,” but almost as many (40.4%) chose the response option “Not Sure/Don’t Know.” Fourteen respondents (13.5%) thought the panelists ended up, on balance, in favor of the measure. (Seventy-two percent of those in the other three experimental conditions were not even aware of the CIR, and 76.3% of those did not venture a guess as to how it had voted.)

Among those in the CIR condition, recollection of the panelists’ strong opposition proved highly predictive of vote intention. Nearly two thirds (64.4%) of those who correctly recalled the balance of panelists’ votes sided against the measure, whereas those who could not recall the panelists’ votes were split between opposing (33.3%) and supporting (35.9%) it. (The 14 people who incorrectly recalled the panelists as supporting the measure may have been doing more than guessing wildly. Eight of those 14 people supported the measure, with half as many opposing it.)

Knowledge of the CIR panelists’ lopsided vote against the sentencing measure may have influenced voters, but reading the CIR Statement did not instill confidence in one’s choices. Only 32% of those who read the Statement felt certain of their voting choices on the sentencing measure. Comparable percentages of confident voters were 41% in the control group, 48% for those reading the secretary of state letter, and a high of 57% for those who read the official summary and fiscal statements, χ²(3, N = 405) = 12.6, p = .006.

In addition, only 43% of those who read the CIR Statement said they had “sufficient information” to cast a decisive ballot—a figure comparable with those in the control and secretary of state conditions (35% and 43%, respectively). The high water mark for this question came from respondents in the summary and fiscal statement condition: Two thirds of those exposed to that information (67%) said they had heard enough, χ²(3, N = 415) = 22.3, p < .001.

In sum, reading the conventional fiscal/summary analysis provided by the Voters’ Pamphlet left voters more confident of their choices—but largely unchanged in the balance thereof. By contrast, the CIR Statement left voters more uncertain but substantially more inclined to oppose the sentencing measure.

Impact on values trade-offs and the values-voting link

The second set of analyses examined the CIR Statement’s potential effect on issue-relevant values. Regarding direct impacts on respondents’ agreement with four values trade-off statements, those randomly assigned to the Statement condition gave responses more consistent with opposing the sentencing measure than did those in every other experimental condition. Table 3 summarizes the results with the other three conditions collapsed to show that reading the Statement caused respondents to agree more often with two antisentencing values arguments (on the injustice of mandatory sentencing and the need for rehabilitation) and less often with prosentencing values arguments (tough sentencing for repeat offenders and accepting the high cost of incarceration), with two of the four differences reaching statistical significance.³

Table 3.

Mean Responses to Values Trade-off Statements Relevant to Sentencing Initiative for CIR Statement Treatment Compared With All Other Experimental Conditions in Study 1.

	Values argument(agreement scale from −1.5 to + 1.5)	Conditions1 to 3 M (SD)	CIR conditionM (SD)	t	p
Opposing mandatory minimums	Even for potentially violent crimes, mandatory minimum sentencing is unjust because it fails to consider individual circumstances	−0.09 (0.98)	0.14 (0.93)	−2.17	.030
Opposing mandatory minimums	Oregon should focus its corrections budget more on rehabilitation, even if that means reducing the length of some prison sentences	−0.13 (0.94)	0.09 (0.98)	−1.90	.059
Favoring tougher sentencing	Repeat offenders should receive increased jail time, regardless of whether their crimes are related to mental illness or addictive substances	0.39 (0.92)	0.14 (0.87)	2.38	.017
	Though it may require spending more on prisons, it is morally wrong to give repeat sex offenders lenient sentences	0.82 (0.83)	0.67 (0.80)	1.56	.119
	Pro-Sentencing Values scale	0.056 (1.02)	−0.26 (1.03)	2.90	.004

Note. Response scale options were strongly disagree (−1.5), disagree (−0.5), agree (0.5), and strongly agree (1.5). CIR = Citizens’ Initiative Review.

Did reading the CIR Statement influence values trade-offs only for those moving against the sentencing measure? Inspection of means across voting groups showed that in the case of two values statements, those in the CIR Statement condition deviated most from their counterparts in the other conditions when they remained undecided or favored the measure. Mean support for rehabilitation over tough sentencing was 0.14 higher (on a 4-point scale) in the CIR condition than the other conditions for both undecided and measure-supporting voters, compared with just 0.04 higher for those in opposition. Likewise, willingness to impose long sentences on mentally ill or addicted offenders was lower in the CIR condition than the other conditions by 0.25 for undecideds, by 0.26 for measure supporters, and by just .05 for those opposing the measure.⁴

To provide a more direct test of whether reading the CIR Statement changed the net influence of values on voting choice, linear regression analysis was conducted using the treatment dichotomy (CIR vs. others) and the four-item standardized Sentencing values scale as predictors of the 7-point Sentencing support certainty item. The first model in Table 4 shows the significant effect of values and the Statement condition (along with measure-relevant empirical beliefs, discussed below). Model 2 adds an interaction between Statement and values, which did not reach significance. In the nonimputed data analysis, however, the interaction reached conventional significance (B = −0.51, SE = .256, p = .048).

Table 4.

Linear Regression Coefficients for Sentencing Initiative Support Certainty in Study 1.

Variable	Model 1B (SE)	Model 2B (SE)
Pro-sentencing values	1.05* (.10)	1.13* (.11)
Pro-sentencing empirical beliefs	1.95* (.32)	1.99* (.37)
CIR treatment condition (1 = yes, 0 = no)	−0.48* (.20)	−0.45 (.26)
CIR × Pro-sentencing values		−0.47(.28)
CIR × Pro-sentencing beliefs		−0.29(.89)
Constant	0.82(.11)	0.83(.11)
Adjusted R²	.44*	.45*

Note. Seven-point dependent variable Sentencing support certainty ranged from −3 (certain to oppose) to +3 (certain to support). CIR = Citizens’ Initiative Review.

p < .05.

To express this dampening effect another way, when the data were split by condition, the effect of values on Sentencing support certainty in the CIR condition was B = 0.74 (SE = .24) but larger for the combination of other conditions, B = 1.18 (SE = .18). This weakening of values’ influence on voting choice ran counter to hypotheses; the combination of this and the nonsignificant findings with imputed data suggest a null result on this issue, which is revisited in the “Conclusion” section.

Empirical beliefs and voting choice

As with values, the CIR Statement had distinct effects on the knowledge items compared with the other experimental conditions. Thus, a simple dichotomous treatment contrast was employed to present the results in Table 5. The direction of belief change on all six items pointed toward opposition to mandatory minimum sentencing, and the three significant changes acknowledged the measure’s redundancy with existing law (Item 1), its effect on the corrections budget (Item 4), and the state’s high incarceration rate (Item 5).

Table 5.

Comparison of Means Responses (and Standard Deviations) With Sentencing Initiative Knowledge Items for CIR Statement Treatment Compared With All Other Experimental Conditions in Study 1.

	True/false statement(scaled from −3 definitely false to +3 definitely true)	Experimentalconditions 1 to 3M (SD)	CIR conditionM (SD)	t-value	p
Empirical beliefs buttressing mandatory minimums	1. Currently, there are NO mandatory minimum sentences for repeat drunk driving offenders. [TRUE]	1.32(2.03)	0.75(1.88)	−2.51	.013
	2. Increasing mandatory minimum sentencing would NOT add significant costs to state and local law enforcement. [FALSE]	0.11(1.89)	−0.14(1.86)	−1.17	.244
	3. Under current law, judges can give ANY felony sex offender a sentence as brief as just 1 or 2 years. [FALSE]	2.10(1.69)	1.83(1.84)	−1.40	.164
Empirical beliefs contrary to mandatory sentencing	4. Oregon has become one of the few states that spends MORE on its corrections system than on higher education. [TRUE]	2.30(1.69)	2.72(1.56)	2.23	.027
	5. Mandatory minimum sentencing has already raised Oregon’s incarceration rate well above the national average. [TRUE]	2.06(1.70)	2.52(1.72)	2.36	.019
	6. This initiative would lead directly to funding cuts for rehabilitation services. [FALSE]	1.53(1.63)	1.70(1.80)	0.90	.370

Note. CIR = Citizens’ Initiative Review.

Inspecting belief changes within each voting group between the CIR treatment and other experimental conditions showed a general pattern of movement toward antisentencing beliefs, but not in every case. The only significant within-voting-group mean difference between the CIR and other conditions was for undecided voters. Among those remaining undecided on the sentencing measure, those in the CIR condition were more skeptical on the first knowledge item (M = 0.40, SD = 1.45) than were those in other conditions (M = 1.37, SD = 1.78), t = 2.18, p = .033. Undecideds in the CIR condition also had higher scores for the fourth knowledge item (M = 2.45, SD = 1.47) than did those in other conditions (M = 1.69, SD = 1.65), t = 1.80, p = .076.

Returning to the linear regression results in Table 4, the composite measure of empirical beliefs favoring the sentencing initiative was included in the equation both as a main effect and interaction with reading the CIR Statement (vs. all other conditions). Model 1 yielded significant main effects for the pro-sentencing empirical beliefs index; however, in Model 2, there was no significant interaction between beliefs and reading the Statement, B = −0.29 (SE = .89).

Discussion

Study 1 indicates that exposure to the CIR Statement shook the confidence of many voters in the soundness of the sentencing measure, and it left many unsure of how to vote. This experimental survey used real Oregon voters faced with an imminent voting decision on a consequential public policy question, and it provided them with precisely the kind of information they could find in their official Voters’ Pamphlet. In light of these findings, it is plausible that reading the CIR Statement could have accounted for at least some of the drop in the initiative’s support during the 2010 Oregon statewide election.

Analyses of values and knowledge items suggest that the CIR Statement moved voters toward values trade-offs that argued against the measure. Likewise, Statement exposure moved voters toward empirical beliefs that undermined the measure. These shifts reflected inferences from the text of the Statement, rather than direct recall of exact wording, which suggests that respondents were actively reconstructing their attitudes and beliefs after reading the Statement. In addition, the Statement seemed to dampen the effect of values on voting choices, but it had no such interaction with empirical beliefs—and the analysis of imputed data suggests the findings in this area are somewhat mixed.

As for the CIR functioning as a voting cue, those who read the Statement became more likely to oppose the sentencing measure. Those who accurately recalled the balance of CIR panelists’ judgments were particularly likely to turn against the measure. Thus, this piece of information may have held special salience for some voters.

Study 2

Would corresponding results obtain for Oregonians who had already voted? To answer that question, we conducted a phone survey to collect retrospective accounts of voter behavior. The cross-sectional study of Oregon voters used regression analysis to control for conventional demographics, as well as initiative-relevant values and beliefs, to determine the independent association between initiative attitudes and self-reported reading of the CIR Statement prior to voting.

Method

Survey sample

The University of Washington Survey Research Center conducted a rolling cross-sectional phone survey September 27 to November 1, 2010. This random-digit-dial survey sample of 550 likely voters in Oregon had a low RR3 response rate (9%) but created a set of respondents that, in the aggregate, approximated both the election results on the sentencing initiative and the partisan makeup of actual ballots cast in the 2010 Oregon general election.

Survey measures

Because Study 2 emphasized actual voting decisions, the key dependent variable was a simple dichotomy between marking one’s ballot “yes” or “no,” using question wording equivalent to Study 1. Those voters who had declined to vote on the sentencing measure were dropped from the analyses.

Sixteen percent of the sample reported having read the CIR Statement on the sentencing measure, with the highest rate (19.7%) coming in the last week of the election. In statistical summaries, this variable is labeled simply Read CIR for each of the initiatives.

Respondents’ initiative-relevant values were measured using the same four items from Study 1, which formed a reliable pro-sentencing values scale (α = .73), M = 0.13, SD = 0.73. Initiative-relevant empirical beliefs were measured using the same approach as in Study 1, but we augmented Study 1’s knowledge items to create a set of 12 knowledge items (see supplementary material). These were recoded as in Study 1, such that there was an aggregate index measuring pro-sentencing empirical beliefs (M = −0.093, SD = 0.23).

In addition, Sex, Educational level, Age (in years), and Political knowledge were included as control variables in the regression analyses. The latter was assessed using a six-item scale covering state and federal politics and government (α = .55, M = 4.14, SD = 1.48).

Power and missing data analysis

Of the 550 survey respondents, 38 individuals (6.9%) declined to state their attitude on the sentencing measure and were dropped from subsequent analysis. This left an effective sample size of N = 512, which was still sufficient to detect even small effect sizes in the sample (Cohen, 1988). Eighty-three percent of the remaining cases had complete data on all variables, with 2.5% of all possible data points missing. The only variable missing in more than 5% of cases was age (26 cases). A linear regression model with five imputations was used to substitute missing values, though 14 cases could not be plausibly imputed owing to nonresponse across multiple items. Results were approximately equivalent to nonimputed analyses, with any changes in statistical significance noted in the analysis.

Results

A logistic regression was run on voting intent (Sentencing support), followed by a linear regression on the 7-point Sentencing support certainty metric. Study 1 showed that reading the CIR Statement reduced support for the sentencing measure, and Table 6 shows associations consistent with those findings. Not surprisingly, voting choices on the sentencing measure were associated most strongly with initiative-relevant values and beliefs, but after controlling for these and other variables, reading the Statement was associated with lower support for the initiative (B = −1.00 [SE = .39], p = .015).⁵ Expressed in terms of an odds ratio, those reading the Statement were roughly 3 times less likely to support the initiative compared with those who read the Voters’ Pamphlet but not the Statement (Exp[B] = .37). The hypothesized interactions, however, did not reach significance.

Table 6.

Logistic Regression Coefficients for Voting in Favor of Sentencing Initiative in Study 2.

Variable	Model 1B (SE)	Model 2B (SE)
Sex (1 = female, 0 = male)	0.93* (.30)	0.95* (.30)
Educational level	−0.20(.16)	−0.20(.16)
Age	0.27* (.10)	0.27* (.10)
Political knowledge	−0.13(.08)	−0.14(.08)
Pro-sentencing values	1.50* (.27)	1.47* (.30)
Pro-sentencing empirical beliefs	4.60* (.94)	4.87* (1.10)
Read CIR on sentencing (1 = yes, 0 = no)	−1.00* (.39)	−1.51* (1.85)
Read CIR × Pro-sentencing values	—	0.15(.666)
Read CIR × Pro-sentencing beliefs	—	−1.47(2.20)
Constant	−0.21(.79)	0.23(.80)
Nagelkerke R²	.52*	.52*

Note. Dependent variable Sentencing support was dichotomous, with 1 = voting yes, 0 = voting no. CIR = Citizens’ Initiative Review.

p < .05.

Table 7 shows a similar pattern of results for a linear regression on the 7-point measure of Sentencing support certainty: CIR Statement readers were more likely to oppose the initiative (B = −0.60, SE = .25, p = .016), with stronger associations coming from the indices of pro-sentencing values and pro-sentencing empirical beliefs.⁶

Table 7.

Linear Regression Coefficients for Sentencing Initiative Support Certainty in Study 2.

Variable	Model 1B (SE)	Model 2B (SE)
Sex (1 = female, 0 = male)	0.65* (.19)	0.71* (.19)
Educational level	−0.12(.09)	−0.11(.09)
Age	0.22* (.01)	0.21* (.01)
Political knowledge	−0.09(.06)	−0.10(.06)
Pro-sentencing values	1.19* (.15)	1.08* (.16)
Pro-sentencing empirical beliefs	3.28* (.46)	3.56* (.50)
Read CIR on sentencing (1 = yes, 0 = no)	−0.60* (.25)	−3.29* (1.38)
Read CIR × Pro-sentencing values	—	0.92(.48)
Read CIR × Pro-sentencing beliefs	—	−1.97(1.26)
Constant	−0.17(.65)	−0.10(.65)
R ²	.61*	.62*

Note. Seven-point dependent variable Sentencing support certainty ranged from −3 (certain to oppose) to +3 (certain to support). CIR = Citizens’ Initiative Review.

p < .05.

The only noteworthy difference between Tables 6 and 7 is the interaction between reading the statement and the impact of pro-sentencing values (B = 0.92, SE = .47, p = .052). Though this coefficient did not reach conventional significance, it pointed in a direction consistent with hypotheses—and opposite from the result found in Study 1. By way of illustration, when regression analyses were split by CIR Statement use, the values-voting association was stronger for those who read the Statement (B = 1.84, SE = .56) than for those who did not (B = 1.06, SE = .15).

Conclusion

The balance of evidence across these two studies suggests that reading the 2010 Oregon CIR on Measure 73 reduced support for it. One cannot estimate from the data the precise net impact on final ballot tallies in Oregon because there exists no definitive measure (beyond self-reporting) of what percentage of Oregon voters read the CIR panel’s Citizens’ Statement. The experiment in Study 1 found that reading the Statement doubled the number of voters at least leaning against the measure, and Study 2 found an equally strong association between reading the Statement and opposing Measure 73.

Although one cannot extrapolate directly to the statewide population based on the data reported herein, the experimental effect shown in Study 1 and corresponding associations in the Study 2 regressions warrants fuller explanation. One account could view the CIR as a deliberative voting cue (Gastil, 2000). After all, voters who correctly remembered the balance of opinion on the sentencing measure’s CIR panel were more likely to agree with its negative assessment. The values and empirical belief shifts found in this study, however, suggest that voters did more than glance at the CIR Statement and follow its cue. Readers might have discussed the initiative and CIR findings with fellow voters (Reedy, Gastil, & Moy, 2016), or perhaps engaged in the kind of reading and reflection, or “deliberation within,” that can be as potent as discussion itself (Goodin, 2003). Interpreted that way, the CIR panel’s one-page analysis provided an opportunity for the electorate to stop and think about key facts and pro/con arguments before arriving at an independent judgment.

Our results also suggest the potential for a more subtle effect of the CIR Statements on values and beliefs, as well as on their combined influence on voting choices. The experimental data show that reading the Statement can influence perceptions of value trade-offs and initiative-relevant empirical beliefs. In addition, our values-consistency hypothesis predicted that Statement use would better synchronize perceptions of values trade-offs with voting choices. Study 1 provided what seems to be a null finding, with no significant result in our main analysis and the opposite of the hypothesized effect in the nonimputed version of our analysis, whereas Study 2 yielded near-significant results consistent with the prediction.

Future research will be needed to further investigate this question. As noted in the section on hypothesized effects, one might expect to find a stronger values-to-vote connection in future studies on the impacts of deliberative forums. However, should the values-dampening effect of the CIR glimpsed in Study 1 recur in later studies, prior scholarship on deliberation and heuristics suggest a potential explanation for that effect. Prior studies have found that discussion can perform a values-clarification function (Gastil, Black, & Moscovitz, 2008; Gutmann & Thompson, 1996; Pearce & Littlejohn, 1997). Nevertheless, the considerable heuristic power of one’s political and cultural values (Kahan et al., 2007; Reedy et al., 2014) suggests the importance of quieting, rather than amplifying, such influences during elections.

Reading a CIR Statement could pull voters away from the voting preferences their values might have led them to without further reflection. As Dancey and Sheagley (2013) explain, there are already many extant cases where “heuristics behave badly,” and the CIR provides an alternative to traditional voting cues (p. 315). In the case of this particular measure, even voters who generally wanted to be “tough on crime” recognized that this values commitment did not require voting for this particular tough-on-crime measure, which the CIR Statement suggested was poorly written and contained costly unintended consequences.

The Oregon CIR also raises a normative question. The small sample of the CIR (and the low response rate to the initial invitations) led Fishkin (2013) to question its representational legitimacy. Likewise, when Gastil (2000) proposed designing a citizen panel roughly twice the size of the CIR, he suggested requiring a two-thirds vote to make a recommendation to cover the 14% margin of error on such a body. Were the CIR functioning in a strictly advisory capacity or substituting for the mass public’s judgment (e.g., Leib, 2005), this would be a greater concern, but the CIR’s task is more analogous to a consensus conference or planning cell (Hendriks, 2005), which draws up a set of findings that other decision makers might use. Thus, the higher priority for a deliberative institution like the CIR is ensuring a depth of issue analysis and statement drafting than is possible in a Deliberative Poll. A resolution to this problem might be removing the vote count from the Citizens’ Statement, which would leave only issue analysis.

Finally, one could question whether the CIR will yield high-quality voter judgments, which some view as the primary purpose of deliberation (Estlund, 2009). In an important sense, this cannot be known, given the difficulty of establishing an independent political judgment in a pluralist society (Ingham, 2013). That is, different judgments cannot be assessed as better or worse than others if they are all based in public reason. However, one measurable standard would be whether the CIR functions as an effective means of increasing the accuracy of voters’ policy beliefs in initiative elections (Boudreau & MacKenzie, 2014; Milic, 2015; Reedy et al., 2016) or improving the match between voters’ interests and their choices (Lau & Redlawsk, 2006; Nai, 2015). The results herein suggest that the CIR at least partially succeeds on this front: It can improve the factual accuracy of voters’ beliefs and the coherence of their values considerations related to an issue, but it is unclear how much the CIR helps voters match interests and values with vote choices.

Laying aside such normative concerns, future research on the CIR can advance our understanding of this process by moving beyond the limits of the present study. Future experimental work could attempt to disentangle the content of the CIR Statement from the source. Varying the description of the CIR process could, for instance, contrast identical content alleged to flow from a panel of policy experts, a deliberative panel, a crowdsourced process, or a bipartisan commission. The point of such comparisons would be to clarify whether the key to the CIR’s impact is the content itself, or the electoral neutrality of the source, or the origin of the content as the fruit of a deliberative minipublic (Goodin & Dryzek, 2006; Ingham & Levin, 2017; MacKenzie & Warren, 2012).

In addition, the present research employed sample sizes that were adequate for aggregate analyses, but not intensive study of subsamples. Larger samples would facilitate analysis of how subgroups read and use the Statements—an approach that would clarify whether the CIR’s influence transcends partisanship and cultural biases (Gastil et al., 2011; Kahan et al., 2007). Over time, it will also be useful to assess the CIR’s net influence across a wide range of electoral environments, including those with substantial partisan spending on both sides of an issue.

Such questions are far from theoretical. After the initial pilot of the CIR, the Oregon legislature passed and the governor signed in 2011 House Bill 2634, which established the Citizens’ Initiative Review Commission and made the CIR a permanent institution. The CIR process has continued to run in Oregon in every even-numbered year since 2010. With CIR processes piloted in Arizona, Colorado, and Massachusetts in 2014 to 2016 and legislation introduced in Washington (House Bill 1364, 2015-2016) and Massachusetts (House Bill 368, 2017), this process could spread to other states or beyond the United States. The Oregon CIR has the potential to demonstrate the viability of institutional reforms that bridge large-scale political processes with the intensive small group issue analysis possible in well-structured and diverse citizen bodies. If successful in this regard, the CIR may inspire future designs that take this idea of bridging and micro- and macro-scale deliberation even further.

Footnotes

Authors’ Note

The authors benefited from feedback on earlier versions of this article presented at the Oregon state legislature, Arizona State University, the University of Alberta, the University of Washington, Pennsylvania State University, the University of Southern California, the University of Utah, the Institute of Rural Management–Anand (India), and conferences of the National Communication Association and the National Coalition on Dialogue and Deliberation.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The research presented in this report was supported by the National Science Foundation (NSF) Directorate for Social, Behavioral and Economic Sciences’ Political Science Program (Award No. 0961774) and the University of Washington (UW) Royalty Research Fund. Any opinions, findings, conclusions, or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of NSF or UW.

Supplemental Material

Supplementary material is available for this article online.

Notes

Author Biographies

John Gastil is a professor of Communication Arts & Sciences and Political Science at the Pennsylvania State University, where he is a senior scholar at the McCourtney Institute for Democracy. His most recent books are Democracy in Small Groups, 2nd ed. (2014) and The Australian Citizens’ Parliament and the Future of Deliberative Democracy (2013).

Katherine R. Knobloch is an assistant professor in the Department of Communication Studies at Colorado State University, where she helps direct the Center for Public Deliberation.

Justin Reedy is an assistant professor in the Department of Communication at the University of Oklahoma, where he also serves as a research associate at the Center for Risk & Crisis Management.

Mark Henkels is a professor in the Department of Political Science at Western Oregon University, where he helped establish a program in Public Policy and Administration.

Katherine Cramer is a professor in the Department of Political Science at the University of Wisconsin-Madison, where she directs the Morgridge Center for Public Service. Her most recent book is The Politics of Resentment: Rural Consciousness in Wisconsin and the Rise of Scott Walker (2016).

References

Altman

(2010). Direct democracy worldwide. New York, NY: Cambridge University Press.

Baldassare

(2013). Reforming California’s initiative process. Public Policy Institute of California. Retrieved from http://www.ppic.org/content/pubs/atissue/AI_1013MBAI.pdf

Barrett

Wyman

Coelho

V. S. P.

(2012). Assessing the policy impacts of deliberative civic engagement in the health policy processes of Brazil and Canada. In Nabatchi

Gastil

Weiksner

Leighninger

(Eds.), Democracy in motion: Evaluating the practice and impact of deliberative civic engagement (pp. 181-203). New York, NY: Oxford University Press.

Boudreau

(2009). Closing the gap: When do cues eliminate differences between sophisticated and unsophisticated citizens? The Journal of Politics, 71, 964-976.

Boudreau

MacKenzie

S. A.

(2014). Informing the electorate? How party cues and policy information affect public opinion about initiatives. American Journal of Political Science, 58, 48-62.

Bowler

(2015). Information availability and information use in ballot proposition contests: Are voters over-burdened? Electoral Studies, 38, 183-191.

Bowler

Donovan

(1998). Demanding choices: Opinion, voting, and direct democracy. Ann Arbor: University of Michigan Press.

Broder

D. S.

(2000). Democracy derailed: Initiative campaigns and the power of money. New York, NY: Harcourt.

Burnett

C. M.

Kogan

(2015). When does ballot language influence voter choices? evidence from a survey experiment. Political Communication, 32, 109-126.

10.

Burnett

C. M.

Parry

J. A.

(2014). Gubernatorial endorsements and ballot measure approval. State Politics & Policy Quarterly, 14, 178-195.

11.

Canary

(2003). Know before you go: A case for publicity funded voters’ guides. Ohio State Law Journal, 64, 81-94.

12.

Cappella

J. N.

Price

Nir

(2002). Argument repertoire as a reliable and valid measure of opinion quality: Electronic dialogue during campaign 2000. Political Communication, 19, 73-93.

13.

Carson

Gastil

Hartz-Karp

Lubensky

(Eds.). (2013). The Australian citizens’ parliament and the future of deliberative democracy. University Park: Pennsylvania State University Press.

14.

Chambers

(2003). Deliberative democratic theory. Annual Review of Political Science, 6, 307-326.

15.

Cohen

Jacob

. 1988. Statistical power analysis for the behavioral sciences, 2nd ed. New York: Routledge.

16.

Cohen

(1989). Deliberation and democratic legitimacy. In Hamlin

A. P.

Pettit

P. H.

(Eds.), The good polity: Normative analysis of the state (pp. 17-34). Oxford, UK: Basil Blackwell.

17.

Collingwood

(2012). Levels of education and support for direct democracy. American Politics Research, 40, 571-602.

18.

Crosby

(1995). Citizen Juries: One solution for difficult environmental questions. In Renn

Webler

Wiedemann

(Eds.), Fairness and competence in citizen participation: Evaluating models for environmental discourse (pp. 157-174). Boston, MA: Kluwer Academic.

19.

Dahl

R. A.

(2000). On democracy. New Haven, CT: Yale University Press.

20.

Dancey

Sheagley

(2013). Heuristics behaving badly: Party cues and voter knowledge. American Journal of Political Science, 57, 312-325.

21.

Dryzek

J. S.

(2010). Foundations and frontiers of deliberative governance. New York, NY: Oxford University Press.

22.

Dyck

J. J.

Baldassare

(2012). The limits of citizen support for direct democracy: California journal of politics and policy. California Journal of Politics and Policy, 4, 1-20.

23.

Ellis

R. J.

(2002). Democratic delusions: The initiative process in America. Lawrence: University Press of Kansas.

24.

Estlund

D. M.

(2009). Democratic authority: A philosophical framework. Princeton, NJ: Princeton University Press.

25.

Farrar

Fishkin

J. S.

Green

D. P.

List

Luskin

R. C.

Levy Paluck

(2010). Disaggregating deliberation’s effects: An experiment within a deliberative poll. British Journal of Political Science, 40, 333-347.

26.

Fishkin

J. S.

(2009). When the people speak: Deliberative democracy and public consultation. New York, NY: Oxford University Press.

27.

Fishkin

J. S.

(2013). Deliberation by the people themselves: Entry points for the public voice. Election Law Journal: Rules, Politics, & Policy, 12, 490-507.

28.

Flavin

(2015). Direct democracy and political equality in the American states. Social Science Quarterly, 96, 119-132.

29.

Garrett

R. K.

Nisbet

E. C.

Lynch

E. K.

(2013). Undermining the corrective effects of media-based political fact checking? the role of contextual cues and naïve theory. Journal of Communication, 63, 617-637.

30.

Gastil

(2000). By popular demand: Revitalizing representative democracy through deliberative elections. Berkeley: University of California Press.

31.

Gastil

(2008). Political communication and deliberation. Thousand Oaks, CA: Sage.

32.

Gastil

Black

Moscovitz

(2008). Ideology, attitude change, and deliberation in small face-to-face groups. Political Communication, 25, 23-46.

33.

Gastil

Braman

Kahan

Slovic

(2011). The cultural orientation of mass political opinion. Political Science & Politics, 44, 711-714.

34.

Gastil

Reedy

Wells

(2007). When good voters make bad policies: Assessing and improving the deliberative quality of initiative elections. University of Colorado Law Review, 78, 1435-1488.

35.

Goodin

R. E.

(2003). Reflective democracy. Oxford, UK: Oxford University Press.

36.

Goodin

R. E.

Dryzek

J. S.

(2006). Deliberative impacts: The macro-political uptake of mini-publics. Politics and Society, 34, 219-244.

37.

Goren

(2004). Political sophistication and policy reasoning: A reconsideration. American Journal of Political Science, 48, 462-478.

38.

Grönlund

Bachtiger

Setälä

(Eds.). (2014). Deliberative mini-publics: Involving citizens in the democratic process. Colchester, UK: ECPR Press.

39.

Grönlund

Setälä

Herne

(2010). Deliberation and civic virtue: Lessons from a citizen deliberation experiment. European Political Science Review, 2, 95-117.

40.

Gutmann

Thompson

D. F.

(1996). Democracy and disagreement. Cambridge, MA: Harvard University Press.

41.

Habermas

(1989). The structural transformation of the public sphere: An inquiry into a category of bourgeois society. Cambridge: MIT Press.

42.

Hajnal

Z. L.

Gerber

E. R.

Louch

. (2002). Minorities and direct legislation: Evidence from California ballot proposition elections. Journal of Politics, 64, 154-177.

43.

Warren

M. E.

(2012). Authoritarian deliberation: The deliberative turn in Chinese political development. Perspectives on Politics, 9, 269-289.

44.

Hendriks

C. M.

(2005). Consensus conferences and planning cells: Lay citizen deliberations. In Gastil

Levine

(Eds.), The deliberative democracy handbook: Strategies for effective civic engagement in the twenty-first century (pp. 80-110). San Francisco, CA: Jossey-Bass.

45.

Ingham

(2013). Disagreement and epistemic arguments for democracy. Politics, Philosophy & Economics, 12, 136-155.

46.

Ingham

Levin

(2017). Effects of deliberative minipublics on public opinion: Experimental evidence from a survey on Social Security reform. International Journal of Public Opinion Research. Advance online publication. Retrieved from https://academic-oup-com-s.web.bisu.edu.cn/ijpor

47.

Jacobs

L. R.

Cook

F. L.

Delli Carpini

M. X.

(2009). Talking together: Public deliberation and political participation in America. Chicago, IL: The University of Chicago Press.

48.

Jerit

Barabas

Bolsen

(2006). Citizens, knowledge, and the information environment. American Journal of Political Science, 50, 266-282.

49.

Kahan

D. M.

Braman

Gastil

Slovic

(2007). Culture and identity-protective cognition: Explaining the white-male effect in risk perception. Journal of Empirical Legal Studies, 4, 465-505.

50.

Knobloch

K. R.

Gastil

Reedy

Cramer Walsh

(2013). Did they deliberate? Applying an evaluative model of democratic deliberation to the Oregon citizens’ initiative review. Journal of Applied Communication Research, 41, 105-125.

51.

Kuklinski

J. H.

Quirk

P. J.

Schwieder

D. W.

Rich

R. F.

(1998). “Just the facts, ma’am”: Political facts and public opinion. Annals of the American Academy of Political and Social Science, 560, 143-154.

52.

Lau

R. R.

Redlawsk

D. P.

(2006). How voters decide: Information processing during election campaigns. New York, NY: Cambridge University Press.

53.

Lavine

H. G.

Johnston

C. D.

Steenbergen

M. R.

(2012). The ambivalent partisan: How critical loyalty promotes democracy. New York, NY: Oxford University Press.

54.

Leeper

T. J.

Slothuus

(2014). Political parties, motivated reasoning, and public opinion formation. Political Psychology, 35, 129-156. doi:10.1111/pops.12164

55.

Leib

E. J.

(2005). Deliberative democracy in America: A proposal for a popular branch of government. University Park: Pennsylvania State University Press.

56.

Leib

E. J.

(2006). The search for deliberative democracy in China. New York, NY: Palgrave Macmillan.

57.

Leighninger

(2006). The next form of democracy: How expert rule is giving way to shared governance—And why politics will never be the same. Nashville, TN: Vanderbilt University Press.

58.

Lewis

D. C.

(2011). Direct democracy and minority rights: Same-sex marriage bans in the U.S. states. Social Science Quarterly, 92, 364-383.

59.

Lukensmeyer

C. J.

Goldman

Brigham

(2005). A town meeting for the twenty-first century. In Gastil

Levine

(Eds.), The deliberative democracy handbook (pp. 154-163). San Francisco, CA: Jossey-Bass.

60.

Lupia

(2015). Uninformed: Why people seem to know so little about politics and what we can do about it. New York, NY: Oxford University Press.

61.

Luskin

R. C.

Fishkin

J. S.

Jowell

(2002). Considered opinions: Deliberative polling in Britain. British Journal of Political Science, 32, 455-487.

62.

MacKenzie

M. K.

Warren

M. E.

(2012). Two trust-based uses of minipublics in democratic systems. In Parkinson

Mansbridge

J. J.

(Eds.), Deliberative systems: Deliberative democracy at the large scale (pp. 95-124). New York, NY: Cambridge University Press.

63.

Mansbridge

J. J.

(1983). Beyond adversary democracy. Chicago, IL: The University of Chicago Press.

64.

Matsusaka

J. G.

(2008). For the many or the few: The initiative, public policy, and American democracy. Chicago, IL: The University of Chicago Press.

65.

Milic

(2015). “For they knew what they did”—What Swiss voters did (not) know about the mass immigration initiative. Swiss Political Science Review, 21, 48-62.

66.

Moore

R. T.

Ravishankar

(2012). Who loses in direct democracy? Social Science Research, 41, 646-656.

67.

Mummolo

Peterson

(2016). How content preferences limit the reach of voting aids. American Politics Research, 45, 159-185.

68.

Mutz

D. C.

(2006). Hearing the other side: Deliberative versus participatory democracy. New York, NY: Cambridge University Press.

69.

Nabatchi

Gastil

Weiksner

Leighninger

(Eds.). (2012). Democracy in motion: Evaluating the practice and impact of deliberative civic engagement. Oxford, UK: Oxford University Press.

70.

Nai

(2015). The maze and the mirror: Voting correctly in direct democracy. Social Science Quarterly, 96, 465-486.

71.

Neblo

M. A.

Esterling

A. E.

Kennedy

K. M.

Lazer

R. P.

Sokhey

D. M. J.

(2010). Who wants to deliberate—And why? American Political Science Review, 104, 566-583.

72.

Nyhan

Reifler

(2010). When corrections fail: The persistence of political misperceptions. Political Behavior, 32, 303-330.

73.

Parkinson

Mansbridge

(Eds.). (2012). Deliberative systems: Deliberative democracy at the large scale. New York, NY: Cambridge University Press.

74.

Pearce

W. B.

Littlejohn

S. W.

(1997). Moral conflict: When social worlds collide. Thousand Oaks, CA: Sage.

75.

Popkin

(1994). The reasoning voter: Communication and persuasion in Presidential campaigns. Chicago, IL: The University of Chicago.

76.

Reedy

Gastil

Moy

(2016). From the secret ballot to the public vote: Examining voters’ experience of political discussion in vote-by-mail elections. Political Communication, 33, 39-58.

77.

Reedy

Wells

Gastil

(2014). How voters become misinformed: An investigation of the emergence and consequences of false factual beliefs. Social Science Quarterly, 95, 1399-1418.

78.

Rogers

Middleton

(2014). Are ballot initiative outcomes influenced by the campaigns of independent groups? A Precinct-randomized field experiment showing that they are. Political Behavior, 37, 567-593.

79.

Ryfe

D. M.

(2005). Does deliberative democracy work? Annual Review of Political Science, 8, 49-71.

80.

Saris

W. E.

Sniderman

P. M.

(2004). Studies in public opinion: Attitudes, nonattitudes, measurement error, and change. Princeton, NJ: Princeton University Press.

81.

Smith

D. A.

Tolbert

(2004). Educated by initiative: The effects of direct democracy on citizens and political organizations in the American states. Ann Arbor: University of Michigan Press.

82.

Smith

Wales

(2000). Citizens’ juries and deliberative democracy. Political Studies, 48, 51-65.

83.

Valentino

N. A.

Hutchings

V. L.

Williams

(2004). The impact of political advertising on knowledge, Internet information seeking, and candidate preference. Journal of Communication, 54, 337-354.

84.

Vidmar

Hans

V. P.

(2007). American juries: The verdict. Amherst, NY: Prometheus Books.

85.

Warren

M. E.

Gastil

(2015). Can deliberative minipublics address the cognitive challenges of democratic citizenship? The Journal of Politics, 77, 562-574.

86.

Warren

M. E.

Pearse

(Eds.). (2008). Designing deliberative democracy: The British Columbia Citizens’ Assembly. New York, NY: Cambridge University Press.

87.

Wells

Reedy

Gastil

Lee

(2009). Information distortion and voting choices: The origins and effects of factual beliefs in initiative elections. Political Psychology, 30, 953-969.

88.

Yankelovich

(1991). Coming to public judgment: Making democracy work in a complex world. Syracuse, NY: Syracuse University Press.

89.

Zaller

J. R.

(1992). The nature and origins of mass opinion. New York, NY: Cambridge University Press.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.20 MB