Content uploaded by Jesse Crosson
Author content
All content in this area was uploaded by Jesse Crosson on Aug 28, 2021
Content may be subject to copyright.
Group Bills and Polarized Policymaking:
Evidence from Proposal Locations in the California
State Legislature
Jesse M. Crosson*
Patricia A. Kirkland†
Mary A. Kroeger‡
September 10, 2020
Abstract
A growing body of literature within the study of state legislatures cites interest groups as a
major source of legislation. In spite of these findings, less is known about the actual nature
of group-written legislation, including how it differs from traditional, member-driven legisla-
tion or how agenda scarcity alters groups’ bill-writing strategies. In this study, we generate an
original dataset of bill proposal and status quo location estimates in California in order to in-
vestigate how group-sponsored bills differ from traditional member-written legislation. To do
so, we jointly scale data on bill-specific interest group position-taking, member cosponsorship
and roll call behavior, developing a large set of proposal and status quo location estimates on
the same preference scale as ideal points for both interest groups and members of the legis-
lature. Using these scores, we investigate how group-sponsored bills compared to traditional
legislation in their ideological extremity and sensitivity.
*Assistant Professor, Trinity University (jcrosson@trinity.edu)
†Assistant Professor, Princeton University (p.kirkland@princeton.edu)
‡Assistant Professor, University of North Carolina at Chapel Hill (mkroeger@email.unc.edu)
In The Semisovereign People, Schattschneider aptly summarizes a critique of American politics
that has animated hundreds of both scholarly and journalistic accounts of representation in the
U.S. According to Schattschneider, “the flaw in the pluralist heaven is that the heavenly chorus
sings with a strong upper-class accent.” In his famous critique of pluralistic theories of American
representation, Olson (1965) echoes a similar sentiment with regard to concentrated interests’ (such
as businesses’) advantages in organizating for political advocacy: “in short,” writes Olson, “the
larger the group, the less it will further its common interests.”
Taken together, these accounts depict inevitable biases in the perspectives presented to elected
officials, and in policymaking involvement more broadly. Importantly, the implication of these and
many other accounts of American politics is that policy outcomes in U.S. legislatures are distorted,
namely toward the interests of the richest and most active groups. Yet the biases of differen-
tial resources and mobilization are not the only sources of policymaking distortion in American
politics, according to these and similar accounts. That is, several institutional features that have
developed throughout American history have served to accentuate existing biases and distortions.
Some such features, like separation of powers which Baumgartner, Berry, Hojnacki, Leech, and
Kimball (2009) argue biases policymaking toward those who benefit most from the status quo,
are foundational to the American system. Others, such as term limits and low legislative capacity,
which some scholars have argued advantage organized interests at the state level (Polsby 1993;
Hertel-Fernandez 2019), have developed only in recent decades.
Regardless of the precise institutional feature said to distort policymaking and/or representa-
tion, most research on interest group influence has struggled to trace these distortions through the
policymaking process to actual policy outcomes. In a few cases, studies have focused on individual
cases or issue area and successfully investigated interest group influence on policy outcomes. Nev-
ertheless, measurement and data challenges have frequently hampered scholars of interest groups,
policymaking, and representation from assessing the influence of interest groups over policymak-
ing outcomes across different issue areas and institutional contexts.
In this paper, we address several of these longstanding measurement challenges using novel
1
data and on interest group bill sponsorship, bill proposal and status quo locations, and group pref-
erence estimates in the state of California. With these data, we examine whether one particular in-
stitution thought to empower interest groups—ballot initiatives and referenda—has enabled groups
to influence policy outcomes. In particular, we investigate 1) whether and to what extent interest
groups active in funding ballot initiatives differ in their preferences from those that are not and 2)
whether the legislative proposals sponsored by these groups differ from those offered by initiative-
inactive groups and legislators. To do so, we join our data on proposal and status quo locations with
California’s unique system of interest-group bill sponsorship, allowing us to measure differences
in group and legislator bill proposals directly.
We proceed as follows. First, we document how, in spite of widespread concerns over dis-
tortions of the policymaking process, few previous studies have been able to measure groups’
influence on policy outcomes directly. We then underscore a similar pattern with respect to bal-
lot initiatives, which some have argued empower special interests who may appeal to the under-
educated public. Based on these accounts, we then detail empirical patterns that should obtain if
interest groups do influence policy outcomes—and whether the initiative process may aid in this
process. Thereafter, we detail our measurement strategy and data sources, describing how Cali-
fornia’s institutional features and transparency statutes provide an excellent opportunity for testing
claims about interest group influence. We then present preliminary results from our investigation.
We conclude by outlining our plans for further investigation into the role of interest groups and
ballot initiatives on policy outcomes and representation, highlighting the especially important task
of accurately capturing public opinion on groups’ targeted policy changes.
Related Literature
Interest Groups in Legislative Politics
The role of interest groups in shaping policy is widely debated in popular discourse and schol-
arly works. Several topics dominate this field. The search for measures of interest group-sponsored
bill locations compared to legislator generated legislation are both motivated and informed by this
2
literature.
The relative contribution of different groups, organized or otherwise, to the overall composition
of U.S. policy is at the core of many interest group studies. Gilens and Page (2014) summarize the
different traditions of thought about who has influence over policy–Majoritarian Electoral Democ-
racy, Economic-Elite Domination, Majoritarian Pluralism, and Biased Pluralism (pg. 564). These
summarize major strains of literature that attribute policy change or stasis to the public, a particular
group of the affluent public, organized interest groups writ large, or business interest groups. Many
of these works use survey data of the various categories and assess if policy change happens in the
preferred direction of the respective group. Our paper makes gains in this measurement endeavor
by looking at the specific pieces of legislation that groups promote and push. Moreover, by plac-
ing bills on the same scale as legislators and the public, we can more carefully assess the relative
ideological and policy successes of groups versus legislators and certain groups in relation to other
groups.
Previous work on interest group influence, while motivated by the desire to identify how inter-
est group influence alters the quality of representation, has encountered a variety of empirical and
measurements challenges. When research has been able to circumvent some of these problems, it
has frequently found only weak effects of interest groups on representation. In fact, recent empiri-
cal work finds null effects in terms of the relationship between giving and policy change (Fowler,
Garro, and Spenkuch 2020). However, this is not universally the case: previous research has found
that that groups strategically give to legislators (Fouirnaies and Hall 2018; Fouirnaies 2018), and
campaign contributions help interest groups obtain access (Kalla and Broockman 2016). While not
directly related to measuring interest group influence, we contribute to the study of how group’s
ability to access the legislative agenda shapes outcomes.
In addition to assessing group influence on policymaking itself, scholars are also interested in
the role that groups play in polarizing legislative institutions. Most of this work concentrates on
the role of campaign contributions from individuals, groups, or party organizations to assess the
effect of groups on polarization (Barber 2016; Grumbach 2020). We also contribute to this line of
3
research by assessing the ideological direction of the policies preferred by interest groups, relative
to preferences of the various policymaking actors.
Finally, most related to this paper is the study of the role of interest groups in the agenda-setting
process. Hacker and Pierson (2014) call on researchers to focus on the role of interest groups in
crafting public policy. This paper takes the role of interest groups in the policymaking process
seriously by assessing the ideological demands of interest groups direclty.
Direct Democracy and Interest Groups
Beyond our general interest in how lobbying influences representation, we aim to investigate
how institutions—in particular, direct democracy—do or not not distort representation. Organized
interests are at the center of key debates about direct democracy, and there is evidence to suggest
that the presence of direct democracy has implications for interest groups and their ability to influ-
ence public policy. Critics of the institution often claim that direct democracy augments the power
of special interests. For example, Cain and Miller (2001) argue that the initiative undermines repre-
sentation as it allows those with sufficient resources to essentially buy a spot on the ballot, leaving
the details of the policy up to the proposer and circumventing deliberation and accountability of
representative democracy. Others take a more sanguine view of direct democracy. Lupia and Mat-
susaka (2004) theorizes that direct democracy creates opportunities for citizens to adjust policy to
reflect their preferences when legislators fall out of step with public opinion. Though he allows
that special interests may have some advantages in the process, Lupia and Matsusaka (2004) also
reiterates that ballot measures only pass with the support of the majority of voters.
The capacity of organized interests to leverage direct democracy likely varies across groups,
and these factors may shape the landscape of interest groups. As Gerber (1999) notes, groups
have different types of resources. Business and trade groups tend to be wealthier but lack the large
membership numbers of citizen groups. A resource advantage might help business groups bring
initiatives to the ballot, but ultimately passage requires majority support. Indeed, Gerber (1999)
finds that citizen groups are more successful than economic groups at modifying policy via direct
democracy while economic groups are more readily able to limit policy change by funding oppo-
4
sition to ballot measures. These dynamics also may influence the number and diversity of interest
groups. States with direct democracy institutions tend to have a larger interest group populations
and a larger share of citizen groups compared to states that do not use ballot measures (Boehmke
2002). Based on these findings, Boehmke (2002) concludes that direct democracy institutions may
actually mitigate the well-documented bias toward business in interest group representation.
Direct democracy creates a variety of opportunities for interest groups to influence policy even
without sponsoring ballot measures. For instance, groups can try to influence voter behavior. While
critics have charged that voters lack the information to determine how to vote on specific and often
detailed policy proposals, voters tend to rely on information shortcuts to vote as if they were well
informed (Lupia 1994). Interest group support and endorsements can serve as crucial heuristics.
They may also offer interest groups the chance to influence the details of ballot measures to ensure
their support (Gerber and Phillips 2004). While there is some debate about how the effectiveness
of advertising in support of ballot measures, spending in opposition is negatively correlated with
passage (c.f., Gerber 1999; Stratmann 2006).
While ballot measures directly affect policy only when they pass, direct democracy has been
likened to a “gun behind the door” that can indirectly influence public policy. Gerber (1996)
develops a formal spatial model of legislative response to the threat of direct democracy. Proposing
a ballot measure is just one of the strategies an interest group could choose in pursuit of it’s policy
goals. The initiative process is costly and success requires majority support, so groups are most
likely to opt to consider direct democracy when policy diverges from voter preferences. Facing the
possibility of policy change via ballot measure, the legislature may opt to shift policy enough to
mitigate this threat.
Theory
As previous research has established, interest group involvement in the policymaking process
brings with it the potential for representational distortions and unfair advantages for certain inter-
ests over others. However, the extent to which institutional features—versus more fundamental re-
alities about the nature of political organization—may generate these advantages is unclear. Here,
5
detail a series of empirical expectations that, if measured properly, should provide evidence either
for or against particular modes of interest group distortion of policymaking outcomes. In partic-
ular, we establish a series of expectations that we believe offer important baseline information on
whether and to what extent interest groups leverage direct democracy and other institutions to draw
policy outcomes toward their preferences.
The first distortion, of course, is in political organizing itself. That is, if the some interests or-
ganize for political influence more easily than others, the distribution of preferences represented by
organized interest groups will differ in substantial ways from public opinion. While Olson (1965)
is typically cited as providing this critique, Gilens and Page (2014) examine the implications of
this divergence empirically, as we highlight above. As their approach implies, if interest groups
are a good reflection of public opinion, interest-group-friendly policy outcomes are not necessar-
ily concerning: following the public and following interest groups is observationally equivalent.
Olsen’s classic pluralist critique provides good reason to believe that there will be divergence be-
tween groups and the public; however, in order for one to assess whether interest groups do in fact
“distort” policy outcomes, it is worth first examining how interest groups’ preferences compare to
those of the public.
To be clear, we are not able to examine this question directly in this study. That is, we do not
(yet) have measures of public opinion on the same scale as interest groups. Nevertheless, we our
measurement strategy, as we describe below, does provide ideal points for legislators, on the same
scale as interest groups. Thus, inasmuch as legislatures constitute the “people’s” branch, we can
ask how well or poorly interest groups’ preferences reflect those exhibited by legislatures.1This
leads us to the first empirical question we seek to investigate: to what extent are interest groups
representative of popularly elected officials (in the state of California, in our case)?
Of course, these distributional differences are neither necessary nor sufficient conditions for
interest group distortions. Indeed, it could be that differences in the effectiveness of groups’ activ-
ities (or those activities’ effectiveness) drive distortions. With respect to differential activity levels,
1A tactic employ by Crosson, Furnas, and Lorenz (2020) at the federal level.
6
previous research has shown that right-leaning and business-friendly interests tend to exhibit the
highest levels of advocacy activities overall. Given our interest in examining how interest groups
may co-opt direct democracy, we add to these studies the following question, which we intend
to examine in this study: to what extent are interest groups active in funding ballot initiatives
representative of popularly elected officials?
After answering these baseline questions about the representativeness of interest groups them-
selves, we are better able to contextualize why interest groups may differ in their policy proposal
activities—our next outcome of interest. In order for interest groups to distort policymaking out-
comes, one mechanism by which they may do so is through providing legislative content. As
previous research has shown (Kroeger 2016), interest group legislation in California is more likely
to pass into law than is similar legislator-sponsored legislation. However, less is known about how
the character of group proposals differ, relative to those of legislators. Given its unique system
of direct interest group sponsorship, California provides an excellent environment for examining
how group bills may differ from their legislator-sponsored counterparts. We therefore investigate
the following question: to what extent do interest groups’ proposals differ from those offered by
legislators? Moreover, given our interest in how direct democracy may influence interest groups’
legislative activities, we also ask: do interest groups active in direct democracy differ in their
proposal activities from other groups and from legislators?
Finally, while differences in group proposal locations are themselves a potentially important
source of interest group influence, any such differences do not themselves constitute distortions
of policy outcomes. Indeed, it is possible that the most unrepresentative policy proposals actually
die in earlier stages of the legislative process, never actually becoming law. In fact, the design of
separation-of-powers systems is predicated upon resisting policy change. Thus, the final question
we investigate is a straightforward extension of the previous question: insofar as differences in
interest group and legislator proposals exist, to what extent do these differences persist as legisla-
tion progresses through the legislative process? Likewise, insofar as proposals by initiative-active
groups differ from other proposals, to what extent do these differences persist as legislation pro-
7
gresses through the legislative process?
Each of these questions, we believe, capture important descriptive components of what inter-
est group distortions of policymaking might actually look like in practice. Central to examining
each of these questions, however, are several thorny methodological challenges. We detail these
challenges below and describe the methodological strategy we adopt to address these challenges.
Data and Methods
Research Design
In order to test the observable implications of interest group mobilization bias and initiative
influence, we adopt a research design that is descriptive in nature. That is, given the multi-agent,
multi-causal nature of the policymaking process, we do not intend to claim that interest groups
clearly generate unrepresentative legislative outcomes, nor that the initiative process clearly priv-
ileges one type of interest over the other. Instead, given long-running difficulties with measuring
the content and direction of policy changes (Clinton 2017), we seek to generate careful and ac-
curate measurements of policymaking in California, enabling us to establish whether the above
empirical patterns—which we view to be consistent with theories of ideological distortion by in-
terest groups—even match the policymaking trends we uncover. Inasmuch as these patterns are
consistent with expectations, we believe that both follow-up research and future iterations of this
work may build upon the conceptual and methodological advances we seek to provide here.
In pursuit of this design strategy, our general approach is to examine distributions of bills’
associated proposal locations. Other studies have used similar strategies to examine strategic pro-
posal moderation by committee chairs and members of the majority, for example (Woon 2008;
Peress 2013). Previous studies have not been able to observe the proposal strategies pursued by
interest groups, however, given institutional and observational constraints. To be clear, this lack
of interest-group proposal data is not due to the fact that interest groups do not “sponsor” leg-
islation in settings outside California. To the contrary, lobbyists routinely provide members of
Congress with their preferred legislative language as they subsidize members’ policymaking ef-
8
forts (Hall and Deardorff 2006); and, at the state level, a growing literature has documented the
widespread and powerful use of interest groups’ model legislation (Hertel-Fernandez 2019; Stokes
2020). Rather, even when scholars have access to information on interest group sponsorship, they
generally lack spatial location estimates that would enable them to examine the polarizing effects
of interest groups on legislatures. Previous research in California has interest group-sponsored leg-
islation passes at higher rates than does legislation sponsored by legislators alone (Kroeger 2016).
Using new data and measurement strategies, our study interrogates some of the mechanisms by
which this phenomenon may occur.
Given this general approach, we proceed by examining proposal location patterns along three
dimensions. First, after examining how interest groups and legislators compare in terms of their
preferences, we simply compare how their proposal locations differ. We do so by examining both
the “raw” spatial locations of interest groups’ preferences, and by examining whether either type
of actor more willingly departs from their ideal policy. Second, we examine whether these patterns
differ between interest groups who are and are not active in funding initiatives/referenda. If interest
groups are in fact able to co-opt the initiatives process as leverage for obtaining their preferred
policy outcomes, then we may expect to see initiative-active groups proposal legislation that is less
“compromising” in nature than their initiative-inactive counterparts. Finally, we facet each of our
comparisons by the progress through the legislative process achieved by each bill proposal. Indeed,
it is possible that moderation (or the lack thereof) could occur in some portions of the legislative
process but not others. We therefore seek to determine whether the first two sets of comparisons
change in character for bills that end in different portions of the legislative process.
Measurement Strategy
In order to assess the relative moderation exhibited by interest-group and legislator sponsors
of legislation, we require ideological estimates for bill proposals, their associated status quo loca-
tions, and sponsors’ ideal points all on the same scale. To generate these measurements, we turn to
a methodology originally developed by Peress (2013) and adapted by Thieme (2020) and Crosson,
Furnas, and Lorenz (2019). Studies of American institutions have frequently been hampered by
9
identification challenges associated with generating proposal and status quo location estimates.
That is, as Clinton (2017) summarizes in his review of strategies for measuring the content and di-
rection of policy changes, common methodologies for generating ideal point estimates (e.g. Poole
and Rosenthal 1997; Clinton, Jackman, and Rivers 2004) fall short of producing reliable estimates
for proposal and status quo locations. More specifically, although proposal and status quo loca-
tions are technically identified via the curvature of the legislators’ preference functions, an infinite
combination of proposal and status quo locations could generate any given cutpoint.
Peress’s 2013 basic methodology addresses this problem by jointly scaling roll call and cospon-
sorship data. In particular, by adopting an absolute proximity (rather than relative proximity)
model of cosponsorship, Peress (2013) is able to recover proposal locations without respect to
their associated status quos. Then, with cutpoints identified from jointly scaled roll call data, status
quo locations are also recovered. This important innovation notwithstanding, a key drawback to
Peress’s approach is that legislation changes from the time of cosponsorship to the time of roll-
call voting. Thus, cutpoints and proposal locations lack a common item, and identification breaks
down. In response, Thieme (2019) and, later, Crosson, Furnas, and Lorenz (2019) introduce a
third type of data—interest-group positon-taking into the estimation process. Here, interest group
position-taking at the time of introduction is treated as a yea-nay “vote” for/against the proposed
legislation,2allowing for cutpoint identification. As a result, the method generates estimates not
only for bills receiving roll call votes, but also for bills which never reach the floor (in addition to
ideal points for interest groups).
A major impediment to implementing this methodology, of course, is its requirement of interest-
group position-taking data. Fortunately, in California, interest group positions have been amply
catalogued over time, both by outside organizations such as Maplight (Lorenz, Furnas, and Crosson
2020) and the legislature itself, as detailed below. Using these data, we generate the above esti-
mates using the same IRT implementation as Crosson, Furnas, and Lorenz (2020), the details of
2See Crosson, Furnas, and Lorenz (2020) and Thieme (2020) for further justification of this
approach
10
which are found in Appendix A1.
Data
These estimates enable us to examine not only groups’ revealed preferences in comparison
to those of legislators, but also how interest groups differ in the strategic location of the bills
they sponsor. However, as noted above, several datasets are necessary to assess the ideological
positions of distinct sets of actors, as well as to capture groups’ relative levels of activity within
direct democracy. More specifically, in order to jointly scale legislators and interest groups, we
need the positions that groups take and the legislators’ votes. Additionally, information on the bill-
level is also needed to assess the bill trajectory, chronological detail, and co-sponsor and author
information. Finally, we require data on groups’ resources and experience with respect to ballot
initiatives.
Information about 1) the bills that are sponsored by groups, and 2) the group positions per bill
come from the legislative analyses. These analyses have a wealth of information, both about the
details of the bill and positions taken by various groups. The Office of Research and individual
committees produce bill analyses. If the bill has an outside sponsor, it is generally indicated by
the terms SOURCE or SPONSOR at the beginning or end of the bill analysis. Other times, this
information is indicated within the text of the bill analysis. We use basic textual analysis tools to
search each bill analysis for the bill’s sponsor. These analyses also list the support and opposition
that the bill has attracted. Figure 1 shows an example of a bill analysis on which four of the groups
that support the bill are cosponsors of the bill, and no groups are listed as opposed.
Figure 1: Example of group positions on a bill.
Each bill has a legislative history attached, which provide us with a several pieces of data
necessary to implement our measurement strategy. Using these histories, we gather information
11
on the date that the bill was first considered by the legislature, as well as the listed “topic” for
the bill. However, these topics do not, in reality, neatly classify bills into a small number of
categories. While these are called topics, there are 38,191 unique topics for 57,802 bills. For
the 2009-2014 period, there are 11,024 unique topics for 14,978 bills. Thus, to group the bills
into meaningful categories, we need to further merge these topics into groups. We use keyword
searches to group these topics into the Policy Agendas Project categories for bills from 2009-
2014. We use the Policy Agendas Project codebook (https://www.comparativeagendas.net/pages/
master-codebook) to place the bills into bins by conducting keyword searches of the topics. For
example, the California topic of ‘Energy: solar energy’ for A.B. 1027 introduced in the 2009-2010
session would be coded as a bill related to ”energy”. We add the category of ‘firearms, alcohol,
tobacco, and drugs’ since these topics come up frequently in California legislation, but are not
easily categorized from referencing the Policy Agendas Project codebook.
Each bill also has a “bill history” document attached to its record, in addition to the basic in-
formation provided in the legislative history. These histories include information on the date of the
bill’s introduction, and conditional upon the bill progressing past introduction, which committee
it went to (and the corresponding date), along with other actions such as votes or final passage.
We gather information from these histories on the date of bill introduction and how far the bill
advanced in the legislative process. These data allow us not only to classify legislation accord-
ing to its ultimate fate in the legislative process, but they also enable us to match interest group
position-taking to different versions of pieces of legislation (and, consequently, to roll call votes).
Legislative cosponsors per bill also come from these bill histories.
Finally, to fully identify the positions of legislators, we need their role call vote records.
Legislator roll call data come from Jeff Lewis and Seth Masket (http://amypond.sscnet.ucla.edu/
california/), whose data contain easy-to-merge data on each bill for which there was a vote.
Beyond data necessary to execute our measurement strategy, our investigation requires us to
measure groups’ involvement and capacity with respect to ballot initiatives. Much as with any
other type of advocacy activity, interest groups vary widely in terms of their involvement in direct
12
democracy. In order to capture this heterogeneity, we turn to groups’ contributions to ballot ini-
tiative campaigns. These data are compiled by the National Institute on Money in State Politics
and capture the amount of money each interest gave to committees aimed at supporting or oppos-
ing ballot initiatives. After receiving the data from NIMSP, we carefully crosswalked the data to
our interest-group position-taking data, enabling us to eventually join our measurements of group
preferences and proposal activities with the level of activity in ballot initiative funding.
Results
In order for our estimation procedure to generate a bill’s spatial location estimate, that bill must
meet several refinements, based on previous literature. First, following Peress (2013), the bill must
be cosponsored by at least three legislators. This refinement ensures that the identification of pro-
posal locations is based on a reliable amount of information. Second, in order for a bill, legislator,
or interest group to be included in the matrix, each actor and item must meet a k-core filatration of
k=5, following Crosson, Furnas, and Lorenz (2020). Operationally, this means that bills must be
subject to at least 5 evaluations by groups or legislators, who themselves took positions (or votes)
on at least 5 bills, recursively. This ensures that scores are based on position-taking activity taken
from the core of the position-taking network. Finally, in order for as-introduced cutpoints to be
identified, interest group position-taking at the time of introduction must not be unanimous (or
overly one-sided)—same as typically scaling of roll-call voting (e.g., Poole and Rosenthal 1985,
Clinton, Jackman, and Rivers 2004).
Applying these refinements to our data, we were able to generate estimates for just under 1,000
bills (n = 988) and 3,314 legislators and interest groups. While the former number constitutes a
sample of bills more than large enough to conduct our analyses, it does carry with the possibility
for sampling biases. In a separate analysis, we determined that the cutpoints from our bills does
not significantly differ from the distribution cutpoints from all bills over the same time period.
Nevertheless, as we continue to collect information on bill changes throughout the policymaking
13
Figure 2: Ideal points of interest groups versus legislators in CA.
process, we anticipate that our sample of bills will grow considerably.3
Before searching for potential differences between group- and legislator-sponsored bills, we
first compare groups’ and legislators’ ideal points. Previous research (e.g., Olson 1965, Schlozman,
Verba, and Brady 2013) has suggested that businesses and other right-leaning groups face fewer
obstacles in organization for political influence, so observing these distributions is an important
first step in understanding the role that interest groups may play in representational distortions
(c.f., Crosson, Furnas, and Lorenz 2020). These distributions are depicted in Figure 3. Contrary
to typical depictions of right-leaning interest group populations, California exhibits a distribution
of interest groups that is decidedly left-leaning (with the difference in means between groups and
legislators significant at p ≈2.1e-10). Moreover, this the distribution lies even farther to the left
for groups who are active in funding ballot initiative committees (difference in means between
initiative-active and inactive groups significant at p ≈0.008). This is notable given the fact that
we measure initiative activity via initiative committee expenditures (to which businesses and other
3One challenge for our method is that interest-group position-taking may occur after a bill
reaches committee. We drop those positions from our analysis, as it is possible that a bill has
changed in committee, compared to the time of group position-taking. In many of these cases,
however, the bill has not changed, and we are effectively discarding usable data. Thus, we are in
the process of applying plagiarism software to determine which bills changed between stages of
the legislative process, thereby retaining later position-taking on bills that did not change.
14
Figure 3: Proposal locations by sponsor type and initiative expenditure.
non-grassroots organizations may contribute), and not via initiative “organizing” or capability per
se. Thus, at least from a simple distributional perspective, any skew in the distribution lies in the
leftward direction.
This, of course, does not necessarily mean that interest groups push policy in a leftward di-
rection, even if interest group bills do pass at a higher rate in California than do legislator bills
(Kroeger 2016). That is, it is possible that conservative groups are especially active relative to
liberal ones, that liberal groups moderate their proposals more, or some combination of the two.
Thus, in our first proposal comparison, Figure 3 compares differences in the spatial locations be-
tween group-sponsored and legislator-sponsored bills. Here, we again observe interest group skew
toward the left, with interest groups proposing significantly more liberal legislation than legislators
(difference in means significant at minuscule p value). Moreover, the estimates again show that
initiative-active interest groups sponsor legislation that lies somewhat leftward of initiative-inactive
groups (difference in means significant at p ≈0.07). Thus, at least in terms of raw proposal loca-
tions, locational difference in groups’ and legislators’ ideal points appear to translate to differences
in proposals.
While potentially intriguing, differences in proposals do not necessarily translate to differences
in policy outcomes. Particularly in a separation-of-powers system like that in California, large
numbers of veto players tends to force policy outcomes toward the center of the political spectrum
(Tsebelis 2002; Tsebelis, Money, Jeannette et al. 1997). In this case, while interest groups’ propos-
15
Figure 4: Proposal locations by sponsor type and progress.
als may appeal to one chamber of the legislature, it may well not pass beyond the other chamber
or the governor’s desk. Thus, in our final set of proposal depictions, we compare proposal loca-
tions by legislators, intiative-active groups, and initiative-inactive groups across difference levels
of progress through the legislative process.
As Figures 4 and 5 demonstrate visually, differences between bills sponsored by each group of
actors disappear among those pieces of legislation that progress the farthest through the legislative
process. Among bills that failed to pass beyond committee, for example, group-sponsored bills
lean farther left than do legislator-sponsored bills, p≈0.0002 (though there were no differences in
mean proposal location between initiative-active and inactive groups). However, as bills advance
through the committee, these differences diminish. Some notable differences do persist through
the committee stage, with group bills differing slightly from legislator bills (p≈0.16) and initia-
tive active groups’ from inactive groups (p≈0.14; with initative-active groups differing somewhat
more significantly from legislators at p≈0.08). Whatever differences persist through the com-
mittee stage diminish once bills reach the floor of the each chamber. Indeed, while the difference
between mean group- and legislator-sponsored bills exhibits a lower pvalue (p≈0.06), the actual
difference between group- and legislator-sponsored bills is much smaller (mean group bill located
16
Figure 5: Proposal locations and progress by sponsor type and initiative expenditure.
at -0.52 and mean legislator bill at -0.30). Differences between average initiative-active and inac-
tive group bills were also small (p≈0.54). This pattern persists for bills that advance past both
chambers.4
Discussion: The Crucial Role of Public Opinion
Thus far, our analyses have pointed to a role for interest groups in California that, if anything,
pushes policy outcomes leftward. However, we do not believe that this is due to any systematic
feature of interest groups or organizing that biases policy in any particular direction. Instead, it
appears that skew in the distribution of groups themselves carries over to skew in the distribution
of policy proposals—a skew that appears to diminish as bills work their way through the legisla-
tive process. Nevertheless, as our results do seem to suggest, the initiative process appears to
4Similar to bills that pass through just one chamber, bills that advance beyond the legislature do
exhibit a significance difference in means by sponsor type (p≈0.03); however, this differences was
quite small in real terms. It is worth noting that the difference between legislator-sponsored bills
and bills sponsored by initiative-active groups was both statistically and substantively somewhat
significant, at p≈0.006. However, the difference is nevertheless much smaller than those from
earlier stages of the legislative process.
17
attract groups on the leftward portion of the spectrum. And for some legislative outcomes, such
as passage through committee, initiative-active groups retain their left-leaning character relative to
other groups and to legislators. Thus, it remains entirely possible that referenda empower certain
groups to extract more from the legislative process, in spite of the fact that many such groups flank
legislators to the left.
In order for us to assess the extent to which initiatives may empower special interests, we
believe we require additional information on at least two factors. First, we need to more system-
atically assess the role of the status quo in the ideological positioning of groups’ proposals. That
is, our conclusions about their proposal patterns may differ if we observe that groups systemati-
cally target a different set of status quo policies than do legislators with similar preferences. These
data are available as part of our estimation process, and they will be incorporated systematically in
future versions of this research. Beyond incorporation of the status quo, however, a crucial compo-
nent of referendum politics—the role of public opinion—will be a major focus of future versions
of this research. In order for groups to leverage the initiative process for policy gains within the
legislative process, they must be able to credibly threaten not simply to place an initiative on the
ballot, but to see that initiative pass into law. This, of course, depends upon the preferences of the
public on the issue at hand. However, even in a progressive state such as California, public opinion
is not uniformly progressive across policy areas.
Accounting for public opinion in our analyses presents several significant obstacles. Most
obviously, the public does not take positions on every piece of legislation that legislators or interest
groups sponsor. Public opinion polls offer a path forward. Surveys can offer detailed information
about public policy preferences, but survey questions often do not map perfectly onto the specific
policy proposals on the legislative agenda. While political scientists have developed methods that
bridge respondents to generate common space ideal point estimates (Tausanovitch and Warshaw
2013; Bafumi and Herron 2010), we lack the data to generate similar scores across the time period
we cover here. Fortunately, we can access considerable data on Californians’ policy preferences
across a range of issues. Though this approach is admittedly imperfect and imprecise, survey data
18
do offer valuable insight on where the public stands on key issues.
For public opinion survey data, we turn to the Public Policy Institute of California (PPIC)
Statewide Survey. The PPIC is a nonprofit, nonpartisan think tank that conducts polling on a
variety of public policy issues. First implemented in 1998, the Statewide Survey conducts multiple
surveys each year each with a sample size of at least 1,700 Californians. Some questions are asked
routinely over time while others are asked sporadically or only once. For example, queries about
the tradeoff between taxes and services, access to abortion, and spending on education appear
regularly while questions about specific ballot initiatives might appear only once or twice. Given
the strengths and limitations of these data, we use them to summarize public opinion on several
issues in policy domains that appear frequently in our dataset of group-sponsored legislation—
education, the environment, and law and crime.
We begin by examining attitudes about spending to support education. To provide estimates
across multiple years and to include as many relevant responses as possible, we pool multiple
survey questions to generate our measure of support for each policy we examine.5All responses
are dichotomous indicators of support. While some PPIC surveys include a general question about
support for public education (K-12 and/or higher education), some questions are more specific,
asking about increasing funding for public pre-school or financial aid for college.
5Specific questions, survey dates, and sample sizes are included in the supplemental informa-
tion.
19
●
●
●
●
●
2008
2010
2012
2014
2016
40.0% 50.0% 60.0% 70.0% 80.0%
Support for
Increasing Education Funding
●
●
●
●
●
●
2008
2010
2012
2014
2016
40.0% 50.0% 60.0% 70.0% 80.0%
Support for
Tax Increases to Fund Education
Figure 6: Attitudes about education funding.
Figure 6 summarizes public opinion about education funding. The points plot the mean share of
respondents (on the horizontal axis) who express support for increasing education funding in each
of the years listed on the vertical axis, and the error bars indicate the standard error of the mean. The
top panel indicates that large majorities of Californians consistently support increasing education
funding. Support for additional funding fluctuates a bit across years, ranging from a minimum of
just over 68% (2015) to nearly 84% (2010). While respondents overwhelmingly favor increasing
education funding, there appears to be less support for taxes to cover the additional spending.
The lower panel of Figure 6 illustrates levels of support for tax increases to fund education. In
some years, only a minority of respondents express support for increasing taxes to fund education.
Opinion is noticeably much more favorable in 2012 and 2016 with levels of support reaching 71%
and 62% respectively when respondents were asked if they supported raising taxes on high-income
20
earners to fund education.6
●
●
●
●
●
●
2008
2010
2012
2014
2016
40.0% 50.0% 60.0% 70.0% 80.0%
Support for
Marijuana Legalization
Figure 7: Attitudes about marijuana legalization.
Figure 7 plots support for marijuana legalization. Here, the points indicate the share of survey
respondents who support legalizing cannabis in each year. Though Californians voted to legalize
recreational use of marijuana via Proposition 30, support for legalization did fluctuate, dropping
as low as 48% in 2012. By the time Proposition 30 appeared on the ballot in 2016, support had
reached 56%. If public opinion on marijuana has been perhaps less liberal than we might have
expected, Californians appear to have especially liberal attitudes about environmental policy. The
top panel of Figure 8 shows the share of PPIC survey respondents who favor the state formulating
its own policies, separate from the federal government, to address climate change. Large majorities
ranging from 62% of respondents in 2009 to 68% of respondents in 2016 express support for
state-level action on environmental policy. The lower panel of Figure 8 plots support for stricter
emissions standards. In any given year, more than two thirds and as many as 77% of survey
respondents favor more stringent restrictions on greenhouse gas emissions.
6In 2012, survey respondents were asked about support for raising taxes on incomes over
$250,000 to fund education, a temporary tax increase implemented with the passage of Propo-
sition 30 in 2012. Respondents answered a similar question in 2016 when policymakers were
considering options to extend this tax increase.
21
●
●
●
●
●
●
●
2008
2010
2012
2014
2016
40.0% 50.0% 60.0% 70.0% 80.0%
Support for
CA Climate Policy
●
●
●
●
●
●
●
●
2008
2010
2012
2014
2016
40.0% 50.0% 60.0% 70.0% 80.0%
Support for
Emissions Restrictions
Figure 8: Attitudes about environmental policy.
Conclusion
In this paper, we bring several datasets to bear on questions of the representativeness of interest
groups, and their preferred legislation, relative to legislators. Using bills sponsored by interest
groups in the California state legislature, along with roll call voting, and cosponsorship data, we
jointly scale legislators, groups, and legislative proposals by both interest groups and legislators.
Future iterations will incorporate mass public opinion to more clearly connect the preferences
of interest groups, politicians, and the public together to assess how representative elected and
unelected policymakers are in representing the public.
The findings suggest that, relative to legislators, groups are left-leaning in the liberal California
legislature. This finding runs counter to findings on the national level that the overall mean of the
interest group population tends ideologically to the right of politicians. Likewise, the bills spon-
sored by interest groups are, on average, more liberal than those proposed by legislators. Though,
successful proposals sponsored by interest groups compared to those introduced by legislators are
nearly identical ideologically. This similarity suggests that the deliberative process works to sort
out relatively extreme interest group proposals. Our results do suggest that the initiative process
22
could play at least some role in the legislative process by giving groups another avenue to alter
policy, as active groups’ proposals differ from those of inactive groups. Our aim in future work
is to examine whether and to what extent public opinion enables some groups to achieve more
favorable policy outcomes.
By bringing together measurements of the ideological positions of groups, bills, legislators,
and the people, this paper will further the field of representative politics and answer important
questions about how interest groups shape this process.
23
References
Bafumi, Joseph, and Michael C Herron. 2010. “Leapfrog representation and extremism: A study of
American voters and their members in Congress.” American Political Science Review pp. 519–
542.
Barber, Michael J. 2016. “Ideological donors, contribution limits, and the polarization of American
legislatures.” The Journal of Politics 78(1): 296–310.
Baumgartner, Frank R, Jeffrey M Berry, Marie Hojnacki, Beth L Leech, and David C Kimball.
2009. Lobbying and Policy Change: Who Wins, who Loses, and Why. University of Chicago
Press.
Boehmke, Frederick J. 2002. “The effect of direct democracy on the size and diversity of state
interest group populations.” The Journal of Politics 64(3): 827–844.
Cain, Bruce, and Kenneth Miller. 2001. The Populist Legacy: Initiatives and the Under-mining of
Representative Government. Rowman and littlefield.
Clinton, Joshua D. 2017. “Coding the Ideological Direction and Content of Policies.” Annual
Review of Political Science 20: 433–450.
Clinton, Joshua, Simon Jackman, and Douglas Rivers. 2004. “The statistical analysis of roll call
data.” American Political Science Review pp. 355–370.
Crosson, Jesse M, Alexander C Furnas, and Geoffrey C Lorenz. 2019. “Estimating Proposal and
Status Quo Locations for Legislation Using Cosponsorships, Roll-Call Votes, and Interest Group
Bill Positions.” Massachusets Institute of Technology, Cambridge, Massachusets. https://polme
th. mit. edu/sites/defau lt/files/docum ents/Jesse Cross on. pdf .
Crosson, Jesse M, Alexander C Furnas, and Geoffrey C Lorenz. 2020. “Polarized Pluralism:
Organizational Preferences and Biases in the American Pressure System.” American Political
Science Review pp. 1–21.
Fouirnaies, Alexander. 2018. “When Are Agenda Setters Valuable?” American Journal of Political
Science 62(1): 176–191.
Fouirnaies, Alexander, and Andrew B Hall. 2018. “How do interest groups seek access to commit-
24
tees?” American Journal of Political Science 62(1): 132–147.
Fowler, Anthony, Haritz Garro, and J¨
org L Spenkuch. 2020. “Quid pro quo? corporate returns to
campaign contributions.” The Journal of Politics 82(3): 000–000.
Gerber, Elisabeth R. 1996. “Legislative response to the threat of popular initiatives.” American
Journal of Political Science pp. 99–128.
Gerber, Elisabeth R. 1999. The Populist Paradox: Interest Group Influence and the Promise of
Direct Legislation. Princeton University Press.
Gerber, Elisabeth R, and Justin H Phillips. 2004. “Direct democracy and land use policy: Ex-
changing public goods for development rights.” Urban Studies 41(2): 463–479.
Gilens, Martin, and Benjamin I. Page. 2014. “Testing Theories of American Politics: Elites,
Interest Groups, and Average Citizens.” Perspectives on Politics 12(3): 564–581.
Grumbach, Jacob M. 2020. “Interest Group Activists and the Polarization of State Legislatures.”
Legislative Studies Quarterly 45(1): 5–34.
Hacker, Jacob S, and Paul Pierson. 2014. “After the” Master Theory”: Downs, Schattschneider,
and the Rebirth of Policy-Focused Analysis.” Perspectives on Politics pp. 643–662.
Hall, Richard L, and Alan V Deardorff. 2006. “Lobbying as legislative subsidy.” American Political
Science Review pp. 69–84.
Hertel-Fernandez, Alex. 2019. State Capture: How Conservative Activists, Big Businesses, and
Wealthy Donors Reshaped the American States–and the Nation. Oxford University Press.
Kalla, Joshua L, and David E Broockman. 2016. “Campaign contributions facilitate access to
congressional officials: A randomized field experiment.” American Journal of Political Science
60(3): 545–558.
Kroeger, Mary A. 2016. “The Third House: Group Bills in the California State Legislature.”.
Lorenz, Geoffrey M, Alexander C Furnas, and Jesse M Crosson. 2020. “Large-N bill positions data
from MapLight. org: What can we learn from interest groups’ publicly observable legislative
positions?” Interest Groups & Advocacy pp. 1–19.
Lupia, Arthur. 1994. “Shortcuts versus encyclopedias: Information and voting behavior in Cali-
25
fornia insurance reform elections.” American Political Science Review 88(1): 63–76.
Lupia, Arthur, and John G Matsusaka. 2004. “Direct democracy: new approaches to old questions.”
Annual Review of Political Science 7: 463–482.
Olson, Mancur. 1965. “The logic of collective action Harvard University Press.” Cambridge, MA .
Peress, Michael. 2013. “Estimating proposal and status quo locations using voting and cosponsor-
ship data.” The Journal of Politics 75(3): 613–631.
Polsby, Nelson W. 1993. “Some arguments against congressional term limitations.” Harv. JL &
Pub. Pol’y 16: 101.
Poole, Keith T, and Howard Rosenthal. 1985. “A spatial model for legislative roll call analysis.”
American Journal of Political Science pp. 357–384.
Poole, Keith T, and Howard Rosenthal. 1997. Congress: A Political-economic History of Roll Call
Voting. Oxford University Press.
Schlozman, Kay Lehman, Sidney Verba, and Henry E Brady. 2013. The unheavenly chorus: Un-
equal political voice and the broken promise of American democracy. Princeton University
Press.
Stokes, Leah Cardamore. 2020. Short circuiting policy: Interest groups and the battle over clean
energy and climate policy in the American States. Oxford University Press, USA.
Stratmann, Thomas. 2006. “Is spending more potent for or against a proposition? Evidence from
ballot measures.” American Journal of Political Science 50(3): 788–801.
Tausanovitch, Chris, and Christopher Warshaw. 2013. “Measuring constituent policy preferences
in congress, state legislatures, and cities.” The Journal of Politics 75(2): 330–342.
Thieme, Sebastian. 2019. “A Direct Test of Legislative Gatekeeping.” Available at SSRN 3440496
.
Thieme, Sebastian. 2020. “Moderation or strategy? political giving by corporations and trade
groups.” The Journal of Politics 82(3): 000–000.
Tsebelis, George. 2002. Veto Players: How Political Institutions Work. Princeton University Press.
Tsebelis, George, Jeannette Money, George Money Jeannette et al. 1997. Bicameralism. Cam-
26
bridge University Press.
Woon, Jonathan. 2008. “Bill sponsorship in Congress: the moderating effect of agenda positions
on legislative proposals.” The Journal of Politics 70(1): 201–216.
27
Supporting Information
Contents
A1 EstimationDetails................................... 1
A2 Public Opinion Survey Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
A1 Estimation Details
Below, we present the JAGS code we used the generate our bill, legislator, and interest group
scores. This code includes information on the priors we use, as well as specifics on the number and
length of chains used in our estimation. Code for generating, cleaning, and filtering the underlying
is available upon request.
pkgs <- c("rjags")
invisible(sapply(pkgs, require, character.only = TRUE))
response <- m1
response_c <- m2
y <- response[,1:ncol(response)]
z <- response_c[,1:ncol(response_c)]
N <- nrow(response)
N
# set total number of items for the latent trait model
K <- ncol(y)
K
#
# ---------------------------------------------------------------------- #
# Define JAGS model statement
MODEL <- "
model{
for(i in 1:N){
for(j in 1:K){
y[i,j] ~ dbern(pi[i, j])
logit(pi[i,j]) <- beta[j]*theta[i] + alpha[j]
z[i,j] ~ dbern(qi[i,j])
logit(qi[i,j]) <- (-w[i]-q[j] - rho*pow((p[j] - theta[i]), 2))
}
}
1
## Priors
# for identification purposes
theta[286] ~ dnorm(0,1)T(,0)
theta[339] ~ dnorm(0,1)T(0,)
for(i in 1:285){
theta[i] ~ dnorm(0, 1)
}
for(i in 287:338){
theta[i] ~ dnorm(0, 1)
}
for(i in 340:N){
theta[i] ~ dnorm(0,1)
}
for(i in 1:N){
w[i] ~ dnorm(0,1)
}
for(j in 1:K){
alpha[j] ~ dnorm(0, .04) # priors the same as pscl::ideal
beta[j] ~ dnorm(0, .04)
q[j] ~ dnorm(0,1)
p[j] ~ dnorm(0,1)
}
rho ~ dnorm(0,1)T(0,)
}"
# ---------------------------------------------------------------------- #
# write the file as a temporary name to then read in
write(MODEL, file="MODEL.bug")
# ---------------------------------------------------------------------- #
# create initial values for the latent variable model
# use ML scores for priors
groups1 <- groups1[order(as.numeric(groups1$group_index)),]
inits.function <- function(chain){
return(switch(chain,
"1"=list(theta=groups1$scores, beta=results_betas$Discrimination.D1, q = rnorm(K), alpha=results_betas$Difficulty, rho = runif(0,1), p = runif(K), w = rnorm(N))
"2"=list(theta=groups1$scores, beta=results_betas$Discrimination.D1, q = rnorm(K), alpha=results_betas$Difficulty, rho = runif(0,1), p = runif(K), w = rnorm(N)),
"3"=list(theta=groups1$scores, beta=results_betas$Discrimination.D1, q = rnorm(K), alpha=results_betas$Difficulty, rho = runif(0,1), p = runif(K), w = rnorm(N))#,
)
)
}
# ---------------------------------------------------------------------- #
2
# generate variables to pass to JAGS
CHAINS <- 1
ADAPT <- 200
BURNIN <- 5000
DRAWS <- 20000
THIN <- 50
# set model file for JAGS model call
MODEL.FILE <- "MODEL.bug"
# ---------------------------------------------------------------------- #
m <- jags.model(file=MODEL.FILE, data=list("y"=y, "z"=z, "N"=N, "K"=K), n.chains=CHAINS, n.adapt=ADAPT, inits = inits.function)
save(m, file="m032020.RData")
update(m, BURNIN)
M <- coda.samples(m, DRAWS, variable.names=c("theta", "alpha", "beta","q", "p", "rho", "w"), THIN)
save.image(file="workSpace032020.RData")
# ---------------------------------------------------------------------- #
# process JAGS estimates
mat1 <- as.matrix(as.mcmc(M[[1]]))
mat2 <- as.matrix(as.mcmc(M[[2]]))
mat3 <- as.matrix(as.mcmc(M[[3]]))
posterior_estimates <- rbind(mat1, mat2, mat3)
parameter.mean <- apply(posterior_estimates, 2, mean)
parameter.sd <- apply(posterior_estimates, 2, sd)
A2 Public Opinion Survey Questions
The tables below include public opinion survey questions pooled to generate the summaries
presented in the main text. For each question, possible responses are included along with the date
of the survey(s) that included the question. All surveys were designed and fielded by Public Policy
Institute of California (PPIC) Statewide Survey.
3
Table A1: Survey questions—education funding
Question Responses Survey
On another topic, do you think that the state government should or
should not fund voluntary preschool programs for all four-year-olds
in California?
1 should 2 should not
8 [VOL] don’t know 9
[VOL] refuse
April 2016
How about increasing government funding to make community col-
lege free? (Do you favor or oppose this proposal?)
1 favor 2 oppose
8 don’t know (vol-
unteered) 9 refuse
(volunteered)
December 2016
How about increasing government funding for scholarships and
grants for students attending four-year colleges and universities?
(Do you favor or oppose this proposal?)
1 favor 2 oppose
8 don’t know (vol-
unteered) 9 refuse
(volunteered)
December 2016
In his revised budget plan Governor Brown proposes increasing state
funding for the University of California system by 4 percent in each
of the next four years in return for a two-year tuition freeze. Do you
favor or oppose this proposal?
1 favor 2 oppose 8
[VOL] don’t know 9
[VOL] refuse
May 2015
Next, please tell me if you favor or oppose increasing state spending
in the following areas.
[ROTATE QUESTIONS 20 TO 23]
Q20. How about increasing state spending on K-to-12 public educa-
tion? (Do you favor or oppose this proposal?)
1 favor 2 oppose 8
[VOL] don’t know 9
[VOL] refuse
January 2014, Jan-
uary 2007, May
2005
How about increasing state spending on higher education? (Do you
favor or oppose this proposal?)
1 favor 2 oppose 8
[VOL] don’t know 9
[VOL] refuse
January 2014
How about increasing government funding available for scholar-
ships or grants for students? (Do you favor or oppose this proposal?)
1 favor 2 oppose 8
[VOL] don’t know 9
[VOL] refuse
November 2010,
November 2009,
October 2007
4
Table A2: Survey questions—taxes to fund education
Question Responses Survey
As you may know, voters passed Proposition 30 in November 2012.
It increased taxes on earnings over $250,000 dollars until 2018 and
sales taxes by one quarter cent until 2016. Do you favor or oppose
extending for 12 years the tax increase on earnings over $250,000
dollars to fund education and healthcare?
1 favor 2 oppose 8
[VOL] don’t know 9
[VOL] refuse
March 2016, April
2016, May 2016
Proposition 55 is called the “Tax Extension to Fund Education and
Healthcare. Initiative Constitutional Amendment.” It extends by
twelve years the temporary personal income tax increases enacted
in 2012 on earnings over $250,000 dollars, with revenues allocated
to K-to-12 schools, California Community Colleges, and, in certain
years, healthcare. The fiscal impacts are increased state revenues
of $4 to $9 billion dollars annually from 2019 through 2030—de-
pending on the economy and stock market—and increased funding
for schools, community colleges, health care for low-income people,
budget reserves, and debt payments. If the election were held today,
would you vote yes or no on Proposition 55?
1 yes 2 no 8 don’t
know (volunteered) 9
refuse (volunteered)
October 2016
Next, what if the state government said it needed more money to
increase funding for California’s public higher education system?
[ROTATE QUESTIONS 32 AND 33]
Q32. Would you be willing to pay higher taxes for this purpose, or
not?
1 yes 2 no 8 don’t
know (volunteered) 9
refuse (volunteered)
December 2016,
December 2014,
November 2009
Governor Brown’s proposed tax initiative on the November ballot
includes a temporary four-year half-cent increase in the state sales
tax and a temporary five-year increase in the state personal income
tax on those earning more than $250,000 dollars annually. The ini-
tiative would raise about $5 to $7 billion dollars annually with the
new revenues going to K-to-12 public schools. Do you favor or op-
pose the proposed tax initiative?
1 favor 2 oppose 8
[VOL] don’t know 9
[VOL] refuse
January 2012
Q32. How about raising state personal income taxes to provide ad-
ditional funding for K-to-12 public education? (Do you favor or
oppose this proposal?)
1 favor 2 oppose 8
[VOL] don’t know 9
[VOL] refuse
April 2012, April
2005
Q33. How about raising the state sales tax to provide additional
funding for K-to-12 public education? (Do you favor or oppose this
proposal?)
1 favor 2 oppose 8
[VOL] don’t know 9
[VOL] refuse
April 2012
Q34. How about raising the top rate of the state income tax paid by
the wealthiest Californians to provide additional funding for K-to-12
public education? (Do you favor or oppose this proposal?)
1 favor 2 oppose 8
[VOL] don’t know 9
[VOL] refuse
April 2012
Q28. Would you be willing to pay higher taxes for K-to-12 public
education, or not?
1 yes 2 no 8 [VOL]
don’t know 9 [VOL]
refuse
May 2012
Q29. Would you be willing to pay higher taxes for higher education,
or not?
1 yes 2 no 8 [VOL]
don’t know 9 [VOL]
refuse
May 2012
—continued—
5
Survey questions—taxes to fund education (Cont’d.)
Question Responses Survey
Q29. Would you be willing to pay higher taxes for higher education,
or not?
1 yes 2 no 8 [VOL]
don’t know 9 [VOL]
refuse
May 2012
Proposition 30 is called the “Temporary Taxes to Fund Education.
Guaranteed Local Public Safety Funding. Initiative Constitutional
Amendment.” It increases taxes on earnings over $250,000 dollars
for seven years and sales taxes by a quarter cent for four years, to
fund schools. It guarantees public safety realignment funding. Fis-
cal Impact is increased state tax revenues through 2018-19, averag-
ing about $6 billion dollars annually over the next few years, rev-
enues available for funding state budget, and in 2012-13, planned
spending reductions, primarily to education programs, would not
occur. If the election were held today, would you vote yes or no
on Proposition 30?
1 yes 2 no 8 [VOL]
don’t know 9 [VOL]
refuse
September 2012,
October 2012
Q22.Proposition 38 is called the “Tax for Education and Early Child-
hood Programs. Initiative Statute.” It Increases taxes on earnings
using a sliding scale, for twelve years. Revenues go to K-to-12
schools and early childhood programs, and for four years to repay-
ing state debt. Fiscal Impact is increased state tax revenues for 12
years—roughly $10 billion dollars annually in initial years, tending
to grow over time. Funds used for schools, child care, and preschool,
as well as providing savings on state debt payments. If the election
were held today, would you vote yes or no on Proposition 38?
1 yes 2 no 8 [VOL]
don’t know 9 [VOL]
refuse”
September 2012,
October 2012
What if the state said it needed more money just to maintain current
funding for K-to-12 public education. Would you be willing to pay
higher taxes for this purpose?
1 yes 2 no 8 [VOL]
don’t know 9 [VOL]
refuse
January 2010,
April 2010, May
2010, April 2009,
January 2008,
April 2008, Jan-
uary 2004, March
2004, June 2003
What if the state said it needed more money just to maintain current
funding for public colleges and universities. Would you be willing
to pay higher taxes for this purpose?
1 yes 2 no 8 [VOL]
don’t know 9 [VOL]
refuse
January 2010,
May 2010,
November 2010,
January 2008,
November 2008
How about raising the top rate of the state income tax paid by the
wealthiest Californians? (Do you favor or oppose this proposal to
raise revenue for state school funding?)
1 favor 2 oppose 8
[VOL] don’t know 9
[VOL] refuse
April 2008,
September 2007,
April 2006
Table A3: Survey questions—CA climate policy
Question Responses Survey
Do you favor or oppose the California state government making its
own policies, separate from the federal government, to address the
issue of global warming?
1 favor 2 oppose
8 don’t know (vol-
unteered) 9 refuse
(volunteered)
July 2016, July
2015, July 2014,
July 2013, July
2012, July 2009,
July 2008, July
2005
6
Table A4: Survey questions—emissions restrictions
Question Responses Survey
Next, to address global warming, do you favor or oppose the state
law that requires California to reduce its greenhouse gas emissions
back to 1990 levels by the year 2020?
1 favor 2 oppose
8 don’t know (vol-
unteered) 9 refuse
(volunteered)
July 2016, July
2015, July 2014,
July 2013, July
2012, July 2010,
July 2009, July
2008, July 2007,
July 2006
To address global warming, the state legislature is currently consid-
ering legislation that would require California to reduce its green-
house gas emissions to 40 percent below 1990 levels by the year
2030. Overall, do you favor or oppose this proposal?
1 favor 2 oppose
8 don’t know (vol-
unteered) 9 refuse
(volunteered)
July 2016
Q40. How about setting stricter emission limits on power plants
in order to address climate change? (Do you favor or oppose this
proposal?)
1 favor 2 oppose
8 don’t know (vol-
unteered) 9 refuse
(volunteered)
July 2016, July
2015, July 2014,
July 2013
To address global warming, the state legislature is currently consid-
ering legislation that would require California to reduce its green-
house gas emissions to 80 percent below 1990 levels by the year
2050. Overall, do you favor or oppose this proposal?
1 favor 2 oppose 8
[VOL] don’t know 9
[VOL] refuse
July 2015
Do you think the government should or should not regulate the re-
lease of greenhouse gases from sources like power plants, cars, and
factories in an effort to reduce global warming?
1 should 2 should not
8 [VOL] don’t know 9
[VOL] refuse
March 2013
Q9. Would you be willing to see tougher air pollution standards on
new passenger vehicles, such as cars, trucks, and SUVs (sport-utility
vehicles), or not?
1 yes 2 no 8 [VOL]
don’t know 9 [VOL]
refuse
July 2012, July
2010, July 2009,
July 2008, July
2006, July 2005
Q9a. Would you be willing to see tougher air pollution standards on
agriculture and farm activities, or not?
1 yes 2 no 8 [VOL]
don’t know 9 [VOL]
refuse
July 2012, July
2010, July 2009,
July 2008, July
2007, July 2006,
July 2005
Q9b. Would you be willing to see tougher air pollution standards on
commercial and industrial activities, or not?
1 yes 2 no 8 [VOL]
don’t know 9 [VOL]
refuse
July 2012, July
2010, July 2009,
July 2008
Q10. Would you be willing to see tougher air pollution standards on
diesel engine vehicles, such as trucks and buses, or not?
1 yes 2 no 8 [VOL]
don’t know 9 [VOL]
refuse
July 2012, July
2010, July 2009,
July 2008
7
Table A5: Survey questions—marijuana legalization
Question Responses Survey
In general, do you think the use of marijuana should be legal, or not? 1 yes, legal 2 no, not
legal 8 [VOL] don’t
know 9 [VOL] refuse
May 2016, March
2015, May 2015,
March 2014,
September 2013,
March 2012,
December 2010
Proposition 64 is called the “Marijuana Legalization. Initiative
Statute.” It legalizes marijuana under state law, for use by adults 21
or older and imposes state taxes on sales and cultivation. It also pro-
vides for industry licensing and establishes standards for marijuana
products and allows local regulation and taxation. The fiscal impacts
are additional tax revenues ranging from high hundreds of millions
of dollars to over $1 billion dollars annually, mostly dedicated to
specific purposes and reduced criminal justice costs of tens of mil-
lions of dollars annually. If the election were held today, would you
vote yes or no on Proposition 64?
1 yes 2 no 8 don’t
know (volunteered) 9
refuse (volunteered)
October 2016
A November ballot initiative is titled, “changes California law to
legalize marijuana and allow it to be regulated and taxed.” In general,
do you think the use of marijuana should be made legal, or not?
1 yes, legal 2 no, il-
legal 8 [VOL] don’t
know 9 [VOL] refuse
May 2010,
September 2010
Regardless of what you think about the personal non-medical uses
of marijuana, do you think adults should be allowed to legally use
marijuana for medical purposes if their doctors prescribe it or do you
think that marijuana should be illegal even for medical purposes?
1 should be allowed
for medical purposes 2
should be illegal even
for medical purposes 8
[VOL] don’t know 9
[VOL] refuse
May 2010,
September 2005
Proposition 19 is called the “Legalizes Marijuana Under California
but Not Federal Law. Permits Local Governments to Regulate and
Tax Commercial Production, Distribution, and Sale of Marijuana.
Initiative Statute.” It allows people 21 years old or older to possess,
cultivate, or transport marijuana for personal use. Depending on
federal, state, and local government actions, fiscal impact is poten-
tial increased tax and fee revenues in the hundreds of millions of
dollars annually and potential correctional savings of several tens of
millions of dollars annually. If the election were held today, would
you vote yes or no on Proposition 19?
1 yes 2 no 8 [VOL]
don’t know 9 [VOL]
refuse
September 2010,
October 2010
8
