Conventional Wisdom? Analyzing Public Support for a State Constitutional Convention Referendum

Abstract

Although fourteen American states periodically hold automatic referendums on whether to hold a state constitutional convention, no state has approved a constitutional convention referendum since 1984. This study explores the puzzle of why voters would oppose an opportunity to broadly reform state government and the factors that underlie these attitudes. Analyses of two statewide surveys of registered voters in New York during the 2017 Constitutional Convention Referendum campaign reveal that campaign framing, elite cues, and instrumental concerns have led voters to take risk-averse positions in order to minimize potential losses that could result from a constitutional convention.

Keywords

constitutional convention referendum institutional reform voting behavior

Voters are rarely given the opportunity to significantly reform state governments, especially in the twenty-six states that do not allow direct initiatives. Institutional reform is reserved for elected officials that often have a stake in maintaining the current system and voters in recent decades have repeatedly voted against referendums to call constitutional conventions. Indeed, states have held 236 constitutional conventions since 1776, but none since 1992 (Snider 2017a). Among the fourteen states that require a referendum on whether to hold a constitutional convention every ten to twenty years, a convention has not been called since 1984 in Rhode Island (Dinan 2010), which covers thirty-one failed referendums in the remaining thirteen states. Since 2002, referendums held in these states have averaged just 36 percent support for holding a constitutional convention (see Online Supplemental Table S1). So, why have voters been so hesitant to take advantage of these opportunities for institutional reform? What factors shape their attitudes on holding a constitutional convention?

Surprisingly, few studies have assessed public attitudes toward American state constitutional convention referendums. A study of voting patterns across counties in five states during the 1960s highlighted the importance of referendum context and elite cues, but this analysis only examined aggregate patterns rather than individual attitudes (Goodman et al. 1973). More recent studies also tend to focus on elites, including state elected officials and organized interests, as the primary determinants of the decline in state conventions (e.g., Dinan 2010; Kogan 2010; Snider 2017a). Yet ultimately the decisions to hold a constitutional convention in many states ultimately lie with the public. Conventional wisdom suggests that these votes show that either majorities of the public in these elections were supportive of the status quo system or were worried about losing beneficial portions of the current institutional system, such as pension protections and collective bargaining rights. At the same time, there is ample evidence that the public remains concerned about government performance, including issues of corruption and adequate public representation. Further, many voters may not be well-served by the current system and would benefit from large-scale reform.

To move beyond the conventional wisdom about these elections, this study explores the case of New York’s (NY) 2017 Constitutional Convention Referendum (CC17). The NY State Constitution requires a public referendum on a constitutional convention to be held every twenty years. This provides New Yorkers with a once a generation opportunity to short-circuit the status quo tendencies of representative government and pursue significant institutional reform. Yet as with other states in recent decades, the past three referendums in NY (1977, 1997, and 2017) failed, with large majorities opposed to holding a convention each time. To better understand how voters formed their attitudes toward holding a constitutional convention, this study analyzes two statewide surveys of registered voters in NY in the lead up to CC17. It begins with a brief overview of CC17, followed by review of the literature on voting in referendums, and then analyzes survey on the attitudes of registered voters in NY.

NY CC17

The NY State Constitution provides the opportunity every twenty years for its voters to determine whether to hold a convention to revise and amend the constitution (NY Const. art XIX §2). A constitutional convention allows delegates elected by the people to gather, review, and change the constitution by addressing issues that the current constitution has not focused on. All convention proposals are ballot referendums which must be put before the voters, who ultimately decide whether they will support or oppose the proposed revisions. There have been nine constitutional conventions in NY State history. The 1821, 1846, 1894, and 1938 conventions, all of whose work was approved for the most part by the voters, resulted in significant policy and institutional reforms that remain important to contemporary politics and public policy, including (Galie 1996):

environmental protections, including the “forever wild” clause which prevents state lands in the Adirondack and the Catskill mountain regions from being developed;

the Education Article, which provides the right to a sound basic education;

the requirement that the state provide aid and care to the needy;

provisions encouraging the state and municipalities to provide low-income housing for their most vulnerable residents;

a bill of rights for organized labor;

the state’s equal protection clause;

constitutional protection for public employee pension benefits; and

protections against illegal searches and seizures.

Typically, NY voters choose not to hold these conventions at all, rejecting the referendum in 1957, 1977, 1997, and 2017. Previous referendum campaigns highlighted the potential risks of holding a convention, and public support withered as a result of efforts by several organized groups who were concerned that status quo provisions such as the “forever wild clause,” the bill of rights for labor, or the Education Article would be removed. One participant in the 1997 referendum campaign explained, “The constitutional convention becomes a vessel that some pour their hopes into but when it comes time to vote to hold one, public support has proven to be unreliable” (Horner 2017).

Yet, by 2017, ethics and corruption concerns were front and center as the referendum approached, suggesting that voters might be more open to disrupting the status quo. Several leaders of the legislature, including Assembly Speaker Sheldon Silver, Senate Majority Leader Dean Skelos, and former Senate leaders Malcom Smith, Pedro Espada Jr., and John Sampson, were convicted on various corruption charges and removed from the legislature. From 2003 through 2016, twenty-nine current or former state legislators and other elected state officials were convicted of felonies, misdemeanors, or violations (Liebman 2017). Despite this pattern of corruption, the state government did not pursue meaningful ethics reform. Evan Davis (2017), former counsel to Governor Mario Cuomo, wrote, “Without a constitutional convention, corruption in Albany will just continue to be a fact of life. With the Legislature unwilling to put in place a meaningful deterrent, it is time for the people to take charge at a constitutional convention” (D1). Given the prominent ethical issues in the state, many people may have been feeling distrustful of government and looking for change consistent with their concerns related to good governance. In addition, the state lagged behind many others in the country in economic growth, especially in upstate regions of the state, providing a potential economic justification to support a convention (Blake and Anson 2017). Indeed, early polling on CC17 showed relatively high support for the convention (see Online Supplemental Figure S1), with support fluctuating between 68 percent and 62 percent from July 2016 to May 2017 (Siena College Research Institute 2018).

While the economic and ethical context of the CC17 may have sparked the public’s support of a convention, other aspects of the political climate cut against a convention. Following a saturated 2016 presidential election and inauguration, there was little media attention to CC17. A July 2017 poll found that two-thirds of New Yorkers hadn’t heard anything about the referendum. By September, this number only dropped to about 50 percent (Siena College Research Institute 2018). When actual campaigning finally began in September, spending and messaging were almost entirely on behalf of opponents—a pattern consistent with recent constitutional convention referendums held in other states (Snider 2017a). Most prominent organized interests came out against the convention, including a strong campaign by organized labor that highlighted the risk of losing pensions and benefits that are currently protected in the state constitution. Teachers unions were particularly active, with labor rights and education guarantees potentially at stake. Opponents of the convention raised US$4,123,989, while proponents raised just US$1,074,767 (National Institute for Money in State Politics 2018). Of the over US$4 million raised by opponents, 94 percent was contributed by labor groups.

Perhaps not surprisingly, support for CC17 dropped precipitously as the campaign unfolded and the election drew near. In the summer of 2017, support for a convention had declined from over 60 percent of registered voters in May to 47–44 percent in July. By October 2017, it plummeted to just 26 percent, with 54 percent opposed. Ultimately, on Election Day, 83 percent of voters rejected holding a constitutional convention. So, what factors led voters to this overwhelming rejection of the opportunity to reform state government despite widespread concerns about ethics and corruption? The following section identifies three likely sets of factors from the literature on voting in direct democracy elections.

Public Opinion and Voting on Government Reform and Referendums

Although there is relatively scant scholarly literature on public attitudes toward constitutional conventions or similarly broad governmental reform, research on voting and public opinion on ballot initiatives and referendums highlight some key considerations for voters. Bowler and Donovan (1998), in their comprehensive exploration of voting in direct democracy, identify three sets of factors shaping public attitudes on these policy proposals: information, elite cues, and instrumental concerns. These factors likely generalize to public attitudes toward governmental reform, broadly construed, and certainly inform expectations about public support for a constitutional convention referendum.

Information

One of the most common criticisms of direct democracy is that voters simply are not well-informed enough to form coherent opinions in an environment where cognitive shortcuts, like party labels, are unclear (e.g., Converse 1964; Magleby 1984). Despite the mounting evidence that voters can, through a variety of cues, vote in ways that reflect their own self-interest and ideological orientations (e.g., McKelvey and Ordeshook 1986; Lupia 1994), this concern about a lack of information to form the foundation of attitudes toward government reform certainly raises question about exactly what actually underlies stated opinions and votes on reform proposals. One answer to this question lies in the concept of loss aversion, the idea that people tend to place a higher value on potential losses from a move away from the status quo than on potential gains (Kahneman and Tversky 1984). In terms of public attitudes toward proposed policy changes and voter behavior on ballot proposals, loss aversion implies that people will have a bias toward the status quo policy or a “no” vote in direct democracy contests (Bowler and Donovan 1998). This loss aversion is likely heightened when the public has little information about the policy proposal because the only information they have is about the status quo. Without clear and concrete information regarding the benefits of a policy, the public may view the proposal as relatively risky and prefer the status quo.

Loss aversion has been explored as a counterargument to criticisms of uninformed voters, both in candidate and direct democracy elections, but it also has implications for more general attitudes on policy reforms. Again, when people have little information about a proposal or are unclear about the benefits and goals of a proposal, they may reflexively oppose it based on the perceived risk it poses. Indeed, in exploring the paradox of why plurality rule electoral institutions tend to be so stable despite producing so many electoral “losers,” Bowler and Donovan (2007) find that risk aversion and framing of reform proposals often leads to behavior consistent with risk aversion. Thus, for a broad institutional reform proposal where the benefits are not (and may not be able to be) clearly defined, risk aversion may dominate public attitude formation, resulting in opposition. Public bias toward the status quo may also be strengthened in the context of constitutional reform due to veneration for constitutions deriving from their symbolic importance (Zink and Dawes 2016), though Brown and Pope (2018) find that state constitutions tend to be viewed less reverentially than the US Constitution.

Constitutional convention referendum campaigns are typically low information environments where the status quo may have a heightened symbolic importance. As such, loss aversion is likely an important factor shaping public attitudes, especially among those who have little information about the referendum. CC17 exemplifies this description since it was held during an off-year election with relatively little public campaigning. All else equal, NY voters with less information about the referendum should have been less likely to support holding a constitutional convention (Hypothesis 1). To assess the level of information that voters had about CC17, the surveys included a question about awareness of the election (Aware of CC17) following a brief description of the referendum:

Switching gears, the last New York State Constitutional Convention was held in 1967. Since then, New Yorkers have twice voted against having a constitutional convention—in both 1977 and 1997. In 2017, New Yorkers will again have an opportunity to vote on whether or not there should be a State Constitutional Convention. How much have you heard about that vote in 2017 on whether or not to have a constitutional convention. Would you say you have heard a great deal, some, not very much, or nothing at all?¹

However, in low information environments, the public may be more vulnerable to framing that highlights loss because this framing taps their tendencies for risk aversion (Druckman and McDermott 2008). Public bias toward the status quo, then, is contingent on the information available to people (Joslyn and Haider-Markel 2000). As noted earlier, the CC17 campaign was relatively one-sided, with labor unions emphasizing the risk of losing protections and benefits embedded in the NY Constitution. Thus, the informational content of the campaign, while still relatively low, may have primed more aware voters to be risk averse in their consideration of a constitutional convention. In light of this informational environment, a competing hypothesis would expect that voters who have heard more about the referendum would be less likely to support holding a convention (Hypothesis 1A). This competing hypothesis highlights the potentially complex ways in which the informational environment may affect attitudes since voters will vary in the extent to which they are aware of the campaign which will shape the amount and nature of the information they are exposed to on an individual level. To account for the possibility that awareness of CC17 conditions the effects of other factors, the study also analyzes relatively aware voters separately from relative unaware voters.²

Elite Cues

Yet, even in low information settings, voters can overcome informational disadvantages and vote as if they were fully informed by employing cognitive shortcuts (Lupia 1992). In particular, elite cues through endorsements and messaging can signal whether a policy proposal or candidate is in voters’ interest even if they do not have much information. If a person is familiar with the cue giver—a prominent elected official or political group (a party or organized interest)—they can accurately evaluate how the vote matches up with their interests, values, or ideology.

Often, elite cues are marked clearly on the ballot through party labels. However, when the public is asked to evaluate a policy proposal or vote on a ballot measure, party labels are not as readily accessible. Instead, the public would need to use elite endorsements or messaging to serve as a cognitive heuristic. For example, in a study of a series of competing and technical ballot proposals on insurance regulation, voters in California were able to correctly identify the proposals that protected consumer interests by using interest group endorsements as cues (Lupia 1994). So, when elite cues are readily available, the public should be able to form opinions and cast votes that are consistent with their own positions. Of course, the effects of elite cues and the loss aversion framing efforts discussed above rely on people both receiving and accepting these messages (Zaller 1992). In the absence of these cues or when these messages are not accepted, loss aversion may take over and a status quo bias likely emerges. In the CC17 campaign, most elected officials did not make strong endorsements, leaving the cues to organized interests (Snider 2017b). Since most of the endorsements against a convention came most prominently from organized labor—especially the teachers’ union—respondents from union households (Union Household) should have been less likely to support a constitutional convention (Hypothesis 2).

Instrumental Concerns

The third set of factors that may underlie attitudes toward policy proposals are instrumental concerns. These encompass a wide set of attitudes that reflect people’s private interests and suggest that the public forms opinions on policy proposals, in part, on the basis of how the proposal benefits them relative to the status quo (Bowler and Donovan 1998). Broadly put, people are likely to support a policy proposal when it aligns with their interests, such as the well-documented phenomenon of economic voting (e.g., Key 1966). Indeed, Blake and Anson (2017) found that people are more likely to vote for a state constitutional convention when the economy was underperforming. The influence of instrumental concerns is also evident in studies of public support for electoral reforms. Bowler and Donovan (2007) note that a voter’s status as an electoral winner or loser shapes their support for electoral reforms and degree of risk aversion. Political losers are more likely to support the adoption of a national referendum (Smith, Tolbert, and Keller 2010), redistricting reform (Tolbert, Smith, and Green 2009), and national presidential nominating procedures (Tolbert, Redlawsk, and Bowen 2009).

The argument that losing in the political realm can shape attitudes toward policy reforms implies that voters may view policy proposals as a vehicle to address their own interests. While the studies cited above conceptualize winning and losing solely in terms of electoral outcomes, the logic may also hold for specific issues. Voters who support specific changes to policy may also be more likely to support broader institutional reforms. This may explain why many state legislatures adopted direct democracy reforms at the turn of the twentieth century (Smith and Fridkin 2008). Although these reforms undermined the legislature’s own authority over the legislative process, populist and progressive legislators nonetheless supported these institutional changes as a tactic to pass a range of new policies that had been bottled up in the legislature. In effect, direct democracy reform served as a logrolling vehicle for multiple policies. However, this logrolling dynamic may serve opposite ends as well. Kogan (2010) notes that recent trends away from constitutional conventions may be driven by a poison pill dynamic, where a raft of institutional reforms is ultimately sunk by the inclusion of unpalatable policies. Here again, the risk of loss may be more heavily weighed than any potential gains. Nonetheless, the poison pill argument centers on how specific instrumental concerns affect attitudes toward broader reform.

General Instrumental Concerns

Thus, voters should be more likely to support a constitutional convention when they are dissatisfied with the status quo generally or prefer changes to specific policies. These instrumental concerns may take several forms. First, general perceptions of government performance may shape levels of support for reform. Each survey asked respondents whether the state is on the right track or wrong track (NY Right Track). People who believe the state is on the right track should be less likely to support a constitutional convention (Hypothesis 3). Next, studies of support for institutional and electoral reforms highlight the role of losing politically, but in NY, political winners and losers are not always clear-cut. With historically divided control of the legislature, both parties may be viewed as winners. Similarly, though upstate residents may feel overlooked due to the political weight of NY City and its suburbs, upstate regions are fairly well represented in the state Senate. As such, the analyses account for these factors without a specific expectation of how they may affect attitudes, if at all. However, racial and ethnic minority groups are often underrepresented in the state political system and constitute a “losing” group. As such, respondents who identify with a racial or ethnic minority group (nonwhite) should be more likely than respondents who identify as white to support a convention (Hypothesis 4).

Specific Instrumental Concerns

Finally, people who favor specific policy changes should be more likely to support a constitutional convention (Hypothesis 5). Both surveys included a battery of items based on the types of issues observers expected to arise at a convention: adopting term limits for the state legislature (term limits—legislature) or statewide officials (term limits—statewide), adopting the citizen initiative (initiative), closing the legal loophole which allows wealthy individuals in NY to form multiple limited liability corporations (LLCs) and contribute millions of dollars to state candidates (close LLC loophole), shifting the NY legislature from part-time to full-time (full-time legislature), expanding economic development in the Adirondack Park (Adirondack development), limiting collective bargaining rights (limit collective bargaining), protecting reproductive rights (reproductive rights—September only), banning multiple forms of discrimination (antidiscrimination—September only), and legalizing marijuana (Marijuana legalization—October only).³ Full question wording for each of these items is presented in Online Supplemental Text S1, with topline results shown in Online Supplemental Table S2.

Analyses

To test these hypotheses, data from two statewide surveys of registered voters in NY conducted by the Siena College Research Institute were analyzed. The first survey was fielded from September 25 to October 2, 2017, using a stratified dual frame probability sample of landline and cell phone telephone numbers for a total of 789 respondents. The second survey was conducted similarly from October 25 to October 29, 2017, yielding 814 respondents. Both surveys included proprietary weights constructed using demographic data to ensure a representative sample.⁴ Analyzing surveys from both September and October allows for a more nuanced examination of the campaign environment as Election Day drew closer.

The analyses center on models predicting respondents’ intended vote choice:

Specifically, the November ballot question on a Constitutional Convention will ask, “Shall there be a convention to revise the constitution and amend the same?” Will you vote “yes” to have a Constitutional Convention or “no” to not have a Constitutional Convention?

In the September survey, 44 percent of registered voters reported intending to vote “yes” and 39 percent intended vote of “no.” Support dropped to 26 percent in the October survey, with 54 percent intending to vote against the convention. Support for the constitutional convention as a function of the three sets of factors is discussed in the literature review. In addition, the models include several relevant control variables: region of residence in NY (NY City, suburbs, and upstate), age, ideology, partisan identification (Democrat, Republican, and independent), educational attainment (college graduate), employment status (unemployed), and sex (female). All of the models are estimated using logistic regression. In order to account for missing values without biasing the sample with list-wise deletion, all analyses employ multiple imputation estimates derived from chained equations to impute missing values.⁵ The first set of analyses models all registered voters and a second set splits the sample by the degree of awareness of CC17.

Results

Table 1 reports the results of the first set of logistic regressions. In September, nearly all the hypotheses are supported to some degree. First, awareness of CC17 had a negative impact on support for a convention, suggesting a heightened risk aversion driven by negative framing (Hypothesis 1A)—an effect that is inconsistent with loss aversion due to lack of information (Hypothesis 1). Holding all other variables at their typical values,⁶ the predicted probability of supporting a convention for a voter who had heard nothing about CC17 was 32 percent, but this probability is cut in half, to 16 percent, for voters who had heard a great deal about the referendum. Next, union households were less likely to support a convention, consistent with expectations about elite cues (Hypothesis 2). The probability of a member of a union household supporting a convention was 24 percent, down from 32 percent for nonunion households. Moving down the model, both general instrumental concerns—evaluation of the state and identification with a politically “losing” group—significantly affect support for CC17 (Hypothesis 3, Hypothesis 4). Voters who believe the state is on the right track are six percentage points less likely to support a convention than those who believe the state is on the wrong track. Meanwhile, nonwhites had a 42 percent probability of supporting a convention, ten points higher than white respondents. For specific instrumental concerns, support of three of the proposed policy changes predicted support for the referendum: expanding economic development in the Adirondacks, adopting the citizen initiative, and limiting collective bargaining. Surprisingly, support for closing the LLC loophole is negatively associated with attitudes toward the convention. It may be that voters who support campaign finance reform could also perceive a constitutional convention as being captured by political interests that are unlikely to adopt major campaign finance reform.

Table 1.

Support for a Constitutional Convention.

Variables	September		October
Information
Aware of 2017 Constitutional Convention Referendum [±]	−.307***	(.091)	−0.268***	(.082)
Elite cues
Union household [−]	−.405*	(.219)	−0.202	(.211)
General instrumental concerns
New York right track [−]	−.141^#	(.110)	−0.078	(.108)
Nonwhite [+]	.429*	(.235)	0.747***	(.246)
Specific instrumental concerns
Adirondack development [+]	.138*	(.060)	0.090^#	(.064)
Term limits—Legislature [+]	.015	(.090)	0.097	(.111)
Term limits—Statewide [+]	−.029	(.091)	−0.047	(.096)
Initiative [+]	.115^#	(.081)	0.308***	(.088)
Close limited liability corporations loophole	−.149*	(.063)	−0.015	(.066)
Full-time legislature [+]	.060	(.066)	−0.002	(.070)
Limit collective bargaining [+]	.106*	(.060)	0.178**	(.067)
Reproductive rights [+]	.018	(.057)	—	—
Antidiscrimination [+]	.064	(.080)	—	—
Marijuana legalization [+]	—	—	0.016	(.060)
Control variables
Region—Suburbs	.057	(.239)	−0.590*	(.242)
Region—Upstate	.343^#	(.227)	−0.329	(.234)
Age	−.007^#	(.004)	0.006	(.005)
Ideology (liberal → conservative)	−.057	(.145)	−0.143	(.144)
Republican	−.208	(.272)	−0.237	(.271)
Independent	.213	(.239)	0.150	(.244)
College graduate	−.252	(.190)	−0.163	(.192)
Unemployed	.120^#	(.347)	0.021	(.427)
Female	−.320	(.184)	−0.115	(.188)
Constant	.057	(.798)	−2.294**	(.833)
N	789		814
Pseudo R ²	.096		.119
Proportional Reduction in Error	8.21%		11.69%

Note: Logistic regression estimates are mean parameters from a multiple imputation with twenty imputations. Standard errors are given in parentheses. Directional tests indicated in square brackets.

^# p < .1.

*p < .05.

**p < .01.

***p < .001.

Three other variables also significantly affect attitudes toward CC17 in September. Compared to residents of NY City, upstate voters are more likely to support holding a convention. The results also show that older voters are less supportive of a convention. Last and consistent with the effects of instrumental concerns, respondents who were unemployed are slightly more likely to favor a convention. Interestingly, partisan identification and ideology do not seem to play a significant role in shaping public attitudes on the referendum, suggesting that traditional political factors did not play an important role in shaping these attitudes.

The results are somewhat similar in the October model, with a few key differences. Awareness of the referendum continues to be associated with opposition to the convention, with the probability of a “yes” vote declining from 19 percent for those who had heard nothing about CC17 to just under 10 percent for voters who had heard a great deal about it. However, union households were no longer significantly less likely to support a convention by the end of October. This change from September may be the result of the opposition’s cues extending beyond their members and reaching a broader audience. For general instrumental concerns, nonwhite voters continue to be significantly more likely to support the referendum, but evaluations of the direction of the state are no longer significant determinants of support. In October, there are also fewer significant specific instrumental concerns. Adirondack development and campaign finance reform are no longer significant factors, but attitudes toward collecting bargaining and a citizen initiative show significant effects. One unexpected difference from the September model is the effect of region. In September, upstate voters were most likely to support a convention—possibly reflecting their perception as being part of a politically losing group. Yet, in October, NY City voters are most likely to support the referendum and Suburban voters are most opposed to it.

The results of the first set of analyses tend to support the expectations from the literature. While overall voters seem to be cautious about supporting a convention, they are more likely to do so if they are part of a minority group or prefer policy changes such as Adirondack development, the citizen initiative, and limiting collective bargaining. Yet perhaps the most consistent factor is the role of elite cues and risk averse framing that is evident in the negative effects of awareness of the referendum, union households, and the question of limiting collective bargaining rights. These significant coefficients suggest that many voters are viewing the referendum through the framework of labor rights as most prominently presented by unions in the campaign. This, in turn, suggests that effects of some of these factors related to labor rights are likely conditional upon awareness of the referendum and the accompanying campaign.

Awareness of the Referendum as a Conditioning Factor

To examine whether the factors underlying attitudes toward the convention are, indeed, shaped by awareness of the referendum, the sample was split between those who are less aware (who have heard nothing or not much about the referendum) and those who are more aware (who have heard some or a great deal). This allows for an assessment of whether the impacts of other variables are conditional on respondents’ information levels. The effects of labor-related factors should be less important in shaping the attitudes of voters who are less aware of the referendum.

Tables 2 and 3 show the results from the split models. Each model specification is consistent with the models from Table 1 but does not include the information variable since the samples are split along this variable. However, the results still show a significant direct impact of information on support for a convention. In September, the difference between less aware voters (about 71 percent of the sample) and aware voters (about 28 percent of the sample) is stark. The predicted probability of support for CC17 typical voters who were less aware of the referendum was 34 percent, but this probability drops to just 6 percent for more aware voters. In addition, nearly all the factors that are significant for less aware voters are not significant for aware voters. For those who are less aware of the referendum, attitudes are shaped by racial/ethnic group identification, region, and attitudes toward three specific issues also affect their support for a convention—Adirondack Development, the citizen initiative, and legislative professionalization. Each of these factors boosted the probability of supporting a convention by nine to twelve percentage points. As expected, union membership and attitudes were not significant factors for voters who had not heard much about CC17.

Table 2.

Support for a Constitutional Convention by Awareness of the Referendum, September.

Variables	Unaware Voters		Aware Voters
Elite cues
Union household [−]	−.087	(.256)	−1.873**	(0.652)
General instrumental concerns
New York right track [−]	−.066	(.123)	−0.635*	(0.293)
Nonwhite [+]	.499*	(.264)	−0.122	(0.669)
Specific instrumental concerns
Adirondack development [+]	.105^#	(.067)	0.286*	(0.147)
Term limits—Legislature [+]	.003	(.098)	0.104	(0.219)
Term limits—Statewide [+]	−.024	(.097)	−0.141	(0.223)
Initiative [+]	.144^#	(.091)	−0.091	(0.179)
Close limited liability corporations loophole	−.118	(.073)	−0.155	(0.133)
Full-time legislature [+]	.098^#	(.074)	−0.062	(0.148)
Limit collective bargaining [+]	.027	(.067)	0.446***	(0.137)
Reproductive rights [+]	.049	(.066)	−0.046	(0.126)
Antidiscrimination [+]	.051	(.087)	0.126	(0.169)
Control variables
Region—Suburbs	.074	(.259)	−0.750	(0.694)
Region—Upstate	.435*	(.259)	−0.390	(0.637)
Age	−.007	(.004)	−0.006	(0.010)
Ideology (liberal → conservative)	−.050	(.164)	0.220	(0.333)
Republican	−.314	(.305)	0.085	(0.667)
Independent	.292	(.272)	−0.070	(0.640)
College graduate	−.149	(.215)	−0.612	(0.460)
Unemployed	.045	(.403)	0.441	(0.596)
Female	−.214	(.211)	−1.445***	(0.402)
Constant	−.598	(.871)	2.366	(1.611)
N	562		227
Pseudo R ²	.058		.303
Proportional Reduction in Error	8.40%		17.9%

Note: Logistic regression estimates are mean parameters from a multiple imputation with twenty imputations. Standard errors are given in parentheses. Directional tests indicated in square brackets.

^# p < .1.

*p < .05.

**p < .01.

***p < .001.

Table 3.

Support for a Constitutional Convention by Awareness of the Referendum, October.

Variables	Unaware Voters		Aware Voters
Elite cues
Union household [−]	0.318	(0.303)	−0.706*	(0.308)
General instrumental concerns
New York right track [−]	−0.075	(0.152)	−0.091	(0.159)
Nonwhite [+]	0.525*	(0.307)	1.121**	(0.401)
Specific instrumental concerns
Adirondack development [+]	0.129^#	(0.091)	−0.006	(0.100)
Term limits—Legislature [+]	0.050	(0.154)	0.159	(0.170)
Term limits—Statewide [+]	−0.041	(0.123)	−0.043	(0.160)
Initiative [+]	0.281*	(0.127)	0.336**	(0.132)
Close limited liability corporations loophole	−0.067	(0.086)	0.031	(0.108)
Full-time legislature [+]	0.072	(0.099)	−0.108	(0.103)
Limit collective bargaining [+]	0.000	(0.091)	0.409***	(0.097)
Marijuana legalization [+]	0.049	(0.079)	−0.034	(0.092)
Control variables
Region—Suburbs	−0.752*	(0.330)	−0.527	(0.392)
Region—Upstate	−0.667*	(0.327)	−0.065	(0.371)
Age	0.000	(0.006)	0.012	(0.008)
Ideology (liberal → conservative)	−0.030	(0.193)	−0.558*	(0.231)
Republican	−0.096	(0.378)	−0.072	(0.406)
Independent	0.395	(0.341)	−0.010	(0.370)
College graduate	−0.105	(0.277)	−0.513^#	(0.271)
Unemployed	−0.678	(0.568)	1.133*	(0.604)
Female	0.138	(0.263)	−0.659*	(0.301)
Constant	−2.161*	(1.059)	−2.042	(1.285)
N	342		472
Pseudo R ²	.091		.176
Proportional Reduction in Error	20.06%		3.96%

Note: Logistic regression estimates are mean parameters from a multiple imputation with twenty imputations. Standard errors are given in parentheses. Directional tests indicated in square brackets.

^# p < .1.

*p < .05.

**p < .01.

***p < .001.

Yet, for those voters who had heard more about the referendum and likely have heard the loss aversion messages of organized labor, a nearly entirely different set of factors are at play. Belonging to a union household dropped the predicted probability of support from 6 percent to less than 1 percent. Meanwhile, those who strongly opposed limits on collective bargaining had just a 4 percent predicted probability of support, but the probability of a “yes” vote increased to 20 percent for those who strongly supported limits. The “right track/wrong track” evaluations of the state also showed a significant divide. The probability of support for those who saw the state as on the “right track” was just 3 percent, but more pessimistic voters had predicated probability of 11 percent. Although the group identities that seem to shape attitudes of less aware voters in September are not at play for more aware voters, one group does show a significant effect. Women show less support among the aware voters, perhaps owing to their overrepresentation in education, with more than three-quarters of public school teachers in NY State being women (US Department of Education 2013).

As seen in Table 3, the differences between less aware and more aware voters generally persist later in the campaign season. However, by late October, there are more relatively aware voters (about 58 percent of the sample) than less aware voters (about 42 percent of the sample). Again, a typical less aware voter was more likely to support a convention than a typical more aware voter, with predicted probabilities of support at 18 percent and 6 percent, respectively. As in September, less aware voters seem to base their support for the convention on racial/ethnic identification, region, and attitudes toward Adirondack development and the citizen initiative. Yet by October, the regional effects have reversed course. NY City residents were now most likely to support the referendum, with predicted probability of nearly 32 percent.

For more aware voters, a wider range of factors shapes their vote choice. As expected, given the nature of the campaign, union households and those who oppose limiting collective bargaining are significantly less likely to support a convention. Belonging to a union household cuts the predicted probability of support in half, from 7 percent to 3½ percent. In addition, several demographic factors come into play. Among more aware voters, conservatives, college graduates, and women are significantly less likely to support the referendum. Last, more aware voters who are unemployed are more likely to support a convention, with a probability of support at nearly 19 percent.

On the whole, there is considerable evidence that many of the factors noted in the existing literature are relevant only among more aware voters. Thus, while a lack of information itself shields less aware voters from risk averse messaging, it also conditions the effects of other factors, including the roles of elite cues, education, and some specific instrumental concerns.

Discussion

These analyses of NY’s CC17 provide important insights into how the public forms opinions toward broad governmental reforms generally and constitutional referendums specifically. Overall, the results are consistent with tendencies toward loss and risk aversion, especially among voters who were more aware of the referendum. This result is not particularly surprising in this case, with the one-sided campaign focused on potential losses and the unclear goals of a constitutional convention. Those who had heard about the referendum—voters who were most likely to receive and accept the loss aversion framing—were less likely to support holding a convention. Similarly, some elite cues also appeared to be influential among respondents that were aware of the referendum; respondents from union households were significantly less likely to support holding a convention than others. Related to this, the analyses showed the importance of attitudes toward collective bargaining in relation to support for a convention. The analyses also showed support for the importance of instrumental concerns. This was most consistent for racial/ethnic minorities, who may see themselves as on the “losing” side of the current political system. More specific instrumental concerns seemed to vary by how much attention voters paid to the CC17 campaign. For less aware voters, attitudes toward Adirondack development and ballot initiatives were important predictor of support. For more aware voters, collective bargaining policy was the most consistent specific concern.

Not surprisingly, these results suggest that garnering support for broad governmental reform is an uphill battle. Even though the public may be unsatisfied with government and relatively supportive of specific reforms, the combination of a lack of information, loss aversion, and cues from entrenched elites all push the public against the opportunity for significant reform. However, for advocates of institutional reform, the results do highlight some ways forward. Highlighting salient institutional reforms as part of a constitutional convention is critical in overcoming loss aversion framing and tendencies. Also, underscoring how voters are “losing” under the current system may also be able to overcome tendencies toward loss aversion. Of course, these strategies depend on mounting a robust campaign in favor of a convention. Recent history demonstrates that public officials and organized interests have not pursued these campaigns with much energy or resources. As such, constitutional convention proponents need to convince both voters and elites to overcome their loss aversion.

Supplemental Material

Supplemental Material, Conventional_Wisdom_SLGR_Table_and_figure_supplement - Conventional Wisdom? Analyzing Public Support for a State Constitutional Convention Referendum

Supplemental Material, Conventional_Wisdom_SLGR_Table_and_figure_supplement for Conventional Wisdom? Analyzing Public Support for a State Constitutional Convention Referendum by Daniel C. Lewis, Jack D. Collens and Leonard Cutler in State and Local Government Review

Supplemental Material

Supplemental Material, Conventional_Wisdom_SLGR_Text_1_Supplement - Conventional Wisdom? Analyzing Public Support for a State Constitutional Convention Referendum

Supplemental Material, Conventional_Wisdom_SLGR_Text_1_Supplement for Conventional Wisdom? Analyzing Public Support for a State Constitutional Convention Referendum by Daniel C. Lewis, Jack D. Collens and Leonard Cutler in State and Local Government Review

Footnotes

Acknowledgment

We thank Don Levy and the Siena College Research Institute for their generous assistance in developing and fielding the survey items.

Declaration of Conflicting Interests

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

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was generously supported by the Howard J. Samuels State and City Policy Center at the Austin W. Marxe School of Public and International Affairs at CUNY Baruch.

ORCID iD

Daniel C. Lewis

Supplemental Material

Supplemental material for this article is available online.

Notes

References

Barth

Jay

Burnett

Craig M.

Parry

Janine

. 2019. Direct democracy, educative effects, and the (mis)measurement of ballot measure awareness. Political Behavior. doi: 10.1007/s11109-019-09529-w.

Blake

William D.

Anson

Ian G.

. 2017. Economic voting and support for constitutional conventions. Paper read at American Political Science Association Meetings, August 31–September 3, San Francisco, CA.

Bowler

Shaun

Donovan

Todd

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

Bowler

Shaun

Donovan

Todd

. 2007. Reasoning about institutional change: Winners, losers and support for electoral reforms. British Journal of Political Science 37:455–76.

Brown

Adam R.

Pope

Jeremy C.

. 2018. Measuring and manipulating constitutional evaluations in the states: Legitimacy versus veneration. American Politics Research. doi: 10.1177/1532673X18776626.

Converse

Philip E.

1964. The nature of belief systems in mass publics. In Ideology and Discontent, edited by Apter

D. E.

, 206–61. New York: Free Press of Glencoe.

Davis

Evan A

. 2017. A path to ethics reform. Albany Times Union, February 5, D1.

Dinan

2010. The political dynamics of mandatory state constitutional convention referendums: Lessons from the 2000s regarding obstacles and pathways to their passage. Montana Law Review 71:395–32.

Druckman

James N.

McDermott

Rose

. 2008. Emotion and the framing of risky choice. Political Behavior 30:297–321.

10.

Galie

Peter J.

1996. Ordered liberty: A constitutional history of New York. New York: Fordham University Press.

11.

Goodman

Jay S.

Arseneau

Robert

Cornwell

Elmer E.

Jr R. Swanson

Wayne

. 1973. Public responses to state constitutional revision. American Journal of Political Science 17:571–96.

12.

Horner

Blair

. 2017. The fight over the constitutional convention heats up. WAMC. July 24. Accessed September 17, 2018. http://wamc.org/post/blair-horner-fight-over-constitutional-convention-heats.

13.

Joslyn

Mark R.

Haider-Markel

Donald P.

2000. Guns in the ballot box: Information, groups, and opinion in ballot initiative campaigns. American Politics Quarterly 28:355–78.

14.

Kahneman

Daniel

Tversky

Amos

. 1984. Choices, values, and frames. American Psychologist 39:341–50.

15.

Key

V. O.

1966. The responsible electorate. Boston, MA: Belknap Press of Harvard University Press.

16.

Kogan

Vladimir

. 2010. The irony of comprehensive state constitutional reform. Rutgers Law Journal 41:881–905.

17.

Liebman

Bennett

. 2017. “What are you going to do about it?” Ethics and corruption issues in the New York State Constitution. New York: Albany Law School Government Law Center.

18.

Lupia

Arthur

. 1992. Busy voters, agenda control, and the power of information. American Political Science Review 86:390–403.

19.

Lupia

Arthur

. 1994. Shortcuts versus encyclopedias: Information and voting behavior in California insurance reform elections. The American Political Science Review 88:63–76.

20.

Magleby

David B.

1984. Direct legislation: Voting on ballot propositions in the United States. Baltimore, MD: Johns Hopkins University Press.

21.

McKelvey

Richard D.

Ordeshook

Peter C.

. 1986. Information, electoral equilibria, and the democratic ideal. The Journal of Politics 48:909–37.

22.

National Institute for Money in State Politics. 2018. Election overview. FollowTheMoney.Org. Accessed March 25, 2018. https://www.followthemoney.org/tools/election-overview/?s=NY&y=2017.

23.

Siena College Research Institute. 2018. New York State public opinion. Accessed August 8, 2018. https://www.siena.edu/scri.

24.

Smith

Daniel A.

Fridkin

Dustin

. 2008. Delegating direct democracy: Interparty legislative competition and the adoption of the initiative in the American States. American Political Science Review 102:333–50.

25.

Smith

Daniel A.

Tolbert

Caroline J.

Keller

Amanda M.

. 2010. Electoral and structural losers and support for a national referendum in the U.S. Electoral Studies 29:509–20.

26.

Snider

J. H.

2017a. Does the world really belong to the living? The decline of the constitutional convention in New York and other US States, 1776–2015. American Political Thought 6:256–93.

27.

Snider

J. H.

2017b. New York State legislators on the November 7, 2017 referendum. The New York State Constitutional Convention Clearinghouse. Accessed August 8, 2018. http://newyork.concon.info/?page_id=1984.

28.

Tolbert

Caroline J.

Redlawsk

David P.

Bowen

Daniel C.

. 2009. Reforming presidential nominations: Rotating state primaries or a national primary? PS: Political Science & Politics 42:71–79.

29.

Tolbert

Caroline J.

Smith

Daniel A.

Green

John C.

. 2009. Strategic voting and legislative redistricting reform: District and statewide representational winners and losers. Political Research Quarterly 62:92–109.

30.

US Department of Education. 2013. Public school teacher data file. National Center for Education Statistics, Schools and Staffing Survey. Accessed September 17, 2018. https://nces.ed.gov/surveys/sass/tables/sass1112_2013314_t1s_002.asp.

31.

Zaller

John R.

1992. The nature and origins of mass opinion. Cambridge, UK: Cambridge University Press.

32.

Zink

James R.

Dawes

Christopher T.

. 2016. The dead hand of the past? Toward an understanding of constitutional veneration. Political Behavior 38:535–60.

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

0.27 MB

0.15 MB