National Policies,Agendas,and Polarization in American State Legislatures: 2011 to 2014

Abstract

Why are some states polarized and others not? This article argues that state legislators are provided with more information by lobbyists and the media about national policies, or state-level bills that are prominent in the national political discourse. Compared with state-specific issues, this additional information encourages legislators to vote along party lines to secure reelection or prepare for a run for higher office. It identifies national policies using lobbying registrations in state legislatures and Congress to show there is more party difference on roll-call votes on national policies in 25 states over 2011 to 2014. It also argues that the notoriety of national issues may encourage party leaders to put these bills on the agenda to build their party brand, or for individual legislators to raise their profiles. It finds that states with more national agendas have more polarized sessions.

Keywords

state legislatures Congress polarization lobbying interest groups

Introduction

In 2011, Ohio State Senator Shannon Jones introduced Senate Bill 5 (SB 5) to reduce the collective bargaining rights of public sector employees. She told the Dayton Daily News the bill would increase local budget flexibility and not “punish unions.”¹ But unions, led by the American Federation of State, County and Municipal Employees, immediately organized protests, which were met by counter protests by the tea party group FreedomWorks. After acrimonious debate, the bill was passed, nearly on party lines. The unions and Democrats then collected signatures for a citizen referendum, and despite millions of dollars spent in support of the bill by Citizens United and the National Federation of Independent Business, SB 5 was repealed later that year.

This article argues that bills with national implications such as SB 5 differ from the rest of the relatively anonymous state legislative agenda. State legislators are poor at estimating the opinion of their constituents (Broockman & Skovron, 2013; Uslaner & Weber, 1979), but protests, media coverage, and, especially, contact from interest group lobbyists can clarify sentiment on a bill. As the national parties resemble collections of interest groups (Bawn et al., 2012), information from these groups is like hearing from the national parties, which is consequential for state legislators interested in reelection or upward advancement. Therefore, state legislators should be more likely to vote on party lines on national issues.

Evaluating this argument is a challenge because prior literature does not offer a consensus definition of national issues, or state legislative bills that are prominent in the national political discourse. Scholars frequently discuss issues the federal government sends to the states (Karch, 2006; Lowery, Gray, & Baumgartner, 2011; McCann, Shipan, & Volden, 2015), but that is not a necessary condition for a state-level policy to be national. Operationalizing the definition necessitates capturing the perspective of a state legislator. Consider a Pennsylvania state legislator’s view of education. In 2014, President Obama made national news mentioning his “Race to the Top” policy in the State of the Union, and the American Federation of Teachers lobbied 40 state legislatures.² But primary and secondary education is often overseen by local school boards, and most of the groups that lobbied on education, such as the Education Consortium of the Upper Allegheny and the Philadelphia School Advocacy Partners, did not lobby before Congress or any other states. Therefore, is it a state-specific or national issue?

To answer this question, I focus on the information transmitted by interest groups and measure the lobbying population in each policy area. Matching lobbying registrations from several states with Congress, I calculate how many of the interest groups that lobby in a state also lobby in Congress, where they participate in the national political discourse. Issues that are lobbied by a higher share of national interest groups are deemed more national. In Pennsylvania, 49% of the interest groups that report lobbying on education also lobbied in Congress, which is well below the state’s average of 68%, making education one of the most state-specific issues. Consistent with intuition, the most national policies—abortion, guns, and gender issues—were featured in national political debate. Meanwhile, the most state-specific policies were the budget, municipal and county issues, and education. These are all policies characterized by local provision.

To establish a relationship between national policies and party polarization, I constructed a data set of more than 106,000 floor votes over the period 2011 to 2014 from the 25 states that systematically report the subject matter of their legislation. I then estimate the average party difference, which highlights votes where the parties are unified in opposition to one another, for 31 issue areas. The issues that most divide the parties—abortion, gun control, and campaign finance and elections—are national issues by my definition. More broadly, there is a positive effect of nationalization on polarization: The more nationalized an issue is, the more polarized it is.

Partisan behavior on national issues helps answer a persistent question in the study of contemporary American state legislatures: Why are some polarized and others not? Shor and McCarty (2011) found that on average, state legislatures are less polarized than Congress. However, the average obscures the fact that polarization varies wildly over states: The distance between the party medians in Rhode Island is a third of that distance of Congress, while in California, it is twice as long. Previous work primarily focuses on the electoral connection to explain this heterogeneity. Shor and McCarty (2011) and Kirkland (2014) contend that the polarization in the states reflects polarization in the electorate. Others have considered variation in electoral structure such as primaries (McGhee, Masket, Shor, Rogers, & McCarty, 2014), campaign finance (Barber, 2016; Harden & Kirkland, 2016; Masket & Miller, 2015), or gerrymandering (McCarty, Poole, & Rosenthal, 2009). These lines of research have all produced mixed findings, in part, because it is not clear that voters sufficiently monitor and respond to the behavior of their state legislators (Rogers, 2013).

More than elections, this account unpacks the roll-call record to explain variation between legislative party polarization in different states. I argue that the relevance of national issues for stakeholders and voters will encourage rank-and-file legislators to push such bills to raise their profiles, even if it runs against the wishes of party leadership. Because I found national issues to be more polarized, states that have more national agendas should be more polarized by party. I score how national each state legislative session is and calculate its party difference. I find the parties are more united and opposed to one another in sessions with more national agendas. This result holds when controlling for other explanations of cross-state polarization, including the polarization of citizen’s opinion within each state, majority party control over the agenda, and how professionalized a legislature is. These results call for more attention on how agendas are formed, and necessitates focus on interest groups that work to place their issues on the agenda and rank-and-file legislators that propose and promote these causes.

Theory

The American people are not boiling with concern about the workings of their state government. (V.O. Key, 1966)

An unmistakable fact about the American federal system is that the bulk of political attention is paid to the national government in Washington, D.C. This disparity is evident in multiple indicators. Citizens know little about their state-level representatives (Delli Carpini, Keeter, & Kennamer, 1994). Newspapers employ more reporters in D.C. than in all 50 state houses combined (Enda, Matsa & Boyles, 2014). The interest groups that operate at the national level tend to be better resourced (Miller, 2010). And as the inscription from Key (1966) indicates, this disparity is not a new phenomenon.

This lack of attention is a two-way street. State legislators have fewer information gathering resources, such as polling or staff, than members of Congress (Jones & Baumgartner, 2005). As a result, according to Uslaner and Weber (1979, p. 564) “legislators, particularly at the state level, are simply not in a good position to estimate public opinion,” which Broockman and Skovron (2013) also found in contemporary times. This lack of information is consequential because if provided information about constituent preferences, as Butler and Nickerson (2011) did for New Mexico legislators in a field experiment, legislator positions track their districts’.

A consequence of this disparity is that Washington, D.C., casts a long shadow over state capitals. So when policies with natio’nal implications are considered at the state level, the larger media and interest group presence provides much more information to state legislators on these topics. But for this information to influence legislator behavior, it needs to help legislators reach their goals. Classic legislative studies indicate that legislators are driven by reelection (Mayhew, 1974) and advancing their career, both within and beyond their current chamber (Fenno, 1973). The information that lobbyists provide—such as opinion polling, policy research, or intelligence on the likelihood of a bill passing the chamber—helps their legislative allies reach common goals (Hall & Deardorff, 2006).

Lobbyists representing interest groups that operate at both the state and national levels are likely to incorporate information from both levels of the federal system. For example, in 2016, Ohio Right to Life, an affiliate of the National Right to Life Committee, advised Ohio Governor John Kasich to veto an extreme abortion bill and sign a moderate one because it would better position the issue before the Supreme Court. The group’s president said, “It’s not just the Ohio strategy but the national strategy.”³ This national perspective essentially expands the scope of conflict on an issue (Wolfe, 2012), to include the national parties, which have aligned against each other on nearly every major issue in Congress (Jochim & Jones, 2013; Layman, Carsey, Green, Herrera, & Cooperman, 2010) and interest groups, which have increasingly joined with one of the major parties to form two ideological teams (Koger, Masket, & Noel, 2009).

The information from these groups is consequential for state legislators because many are progressively ambitious, meaning they are interested in a later run for Congress.⁴ Thomsen (2014) reports that ideological fit with the national party has increasingly become a prerequisite for state legislators looking to make that jump. Bawn et al. (2012) argue that parties are best conceived as a collection of intense policy demanders, so pleasing these groups is of particular importance.

Although not all state legislators want a national job, nearly all legislators are interested in reelection. Recent survey evidence suggests that voters selecting their members of Congress are primarily interested in representation on salient issues (Lapinski, Levendusky, Winneg, & Jamieson, 2016). The deluge of information from national media is likely to make these issues particularly salient to constituents. Therefore, legislators can be expected to be in tune with their constituents’ preferences on these issues.

The amount and style of media coverage on national issues are likely to polarize voters’ preferences along party lines. Citizens tend to hold more extreme positions on political figures subject to a large amount of attention (Kiousis & McCombs, 2004). And the coverage itself is more partisan. In comparison with the state-level outlets, national media sources compete in a crowded marketplace that encourages them to emphasize novelty and conflict, especially partisan conflict (Groeling, 2010). Furthermore, the number of local newspapers is in decline (Darr, 2016), and the digital sources that have generally replaced them tend to appeal to niche audiences and be more overtly partisan (Baum & Groeling, 2008).

Tying these strings together, when legislators face a roll-call vote, they are likely to take a position that furthers their goals. If interest groups, the national parties, and voters are all looking for the same thing, legislator positions are likely to follow. During the Obama administration, which this study examines, the national parties were aligned against each other on nearly every major issue, so when national issues were on state-level agendas, they can be expected to be partisan.

Hypothesis 1 (H1): National policies will be more polarized by party than state-specific policies.

How Are State Agendas Nationalized?

Legislators are also likely to pursue their goals when they choose which issues to put on the agenda or associate with. In 2013, Democratic Texas State Senator Wendy Davis filibustered an abortion bill in the Texas legislature, which endeared her to the abortion rights groups, increased her national notoriety, helped her raise campaign finances, and propelled her to her party’s 2014 Gubernatorial nomination. This episode shows two features of national issues. First, they can attract the support of resource-rich interest groups, which is helpful for ambitious politicians. Second, legislators like to work on issues that are salient and lack complexity, as their constituents are likely to both care about these issues and understand them (Eshbaugh-Soha, 2006; Gormley, 1986). Therefore, majority party leaders, who have the most control over which bills receive consideration for passage, may favor putting these bills up for a vote as it helps the party build a brand that helps its members get elected (Cox & McCubbins, 2005).

However, the notoriety that national policies attract can also worry majority party leaders. Majority party leadership often looks to preserve its majority above all other considerations (Aldrich & Rohde, 2000), and moderate legislators are often punished for taking partisan votes in general elections (Carson, Koger, Lebo, & Young, 2010; Nyhan, McGhee, Sides, Masket, & Greene, 2012). Therefore, leaders may actually attempt to keep salient and partisan issues off the agenda. In this line, Lindstädt and Vander Wielen (2014) find that partisan issues rarely come up for a vote as elections approach.

Backbenchers of both parties, especially in safe seats, may still want to deal with national issues, which can create a tension with party leaders. For example, in the session after Ohio’s SB 5 was rejected by referendum, Republican state representatives Kristina Rogner and Ron Magg reintroduced “Workplace Freedom” legislation that would have started the fight again. However, the bill was opposed by Republican Speaker William Batchelder and Senate President Keith Faber, who said, “The only purpose this discussion serves right now is to generate a bunch of breathless fundraising appeals from the Ohio Democratic Party.”⁵ SB 5 was a notable loss for the Republican leadership that it did not want to relive and threaten its majority status.

Regardless of party leader preferences on national issues, their agenda control is not absolute. Even in the U.S. House of Representatives, where party leadership has long had strong statutory agenda powers, Schickler and Pearson (2009) found a number of bills in the mid-20th century that made it to the floor over Speaker Sam Rayburn’s wishes. In the states, these formal agenda control powers are usually weaker. Jackman (2014) reports that about 80% of chambers allow a majority of legislators, regardless of party, to advance a bill for floor consideration over their leaders’ wishes. Some states go even further, for example, in 1988, Colorado passed the GAVEL act to “give a vote to every legislator” that guarantees each bill consideration and vote, such that 80% of introduced bills receive a roll-call vote (M. Binder, Kogan, & Kousser, 2011). These majoritarian institutions allow rank-and-file legislators to push bills to final consideration.

As a result, agendas can develop independently of leadership, and outside actors can play a part. Kingdon (1984) developed a classic framework that determined which issues make the agenda, arguing that the attention of public officials needs to be focused on a problem, a policy solution to that problem needs to be available, and policymakers have to have the motivation and the opportunity to take action on that issue. McCann et al. (2015) note that Congressional attention often makes policies more likely to be placed on state legislative agendas because it highlights problems and solutions, and allows legislators to gauge how popular they are with constituents. Determining which issues in society are worthy of attention, and proposing answers to those questions, can also be performed by actors outside the legislature. For example, model bills drafted by the conservative American Legislative Executive Council (ALEC) have been introduced, nearly verbatim, in many states (Garrett & Jansa, 2015; Hertel-Fernandez, 2014).

Taken together, an agenda has three characteristics. First, it is primarily selected by the majority party’s leadership. Second, it deserves to be considered an independent manifestation that also reflects the wishes of rank-and-file legislators and outside interest groups. Third, it is a powerful determinant of how much partisan behavior is observed in a legislative session. With those characteristics in mind, the logic of the first hypothesis can be expanded to the agenda at large. If national issues, which I expect to be more polarized by party, make up more a state’s policy agenda, that session should be more polarized by party.

Hypothesis 2 (H2): State legislative sessions with more national agendas will be more polarized by party.

The second hypothesis essentially argues that policy is driving the politics in state legislative chambers. This is an instance of a theory that has a long history in the field (Lowi, 1972; Schattschneider, 1935) and has recently been applied to the polarization of the U.S. Senate by Lee (2008). Lee shows that as economic issues have taken up more of the Senate’s agenda, the parties have become more divided. However, these changes reflect the decisions of party leaders, which is more of a case of politicians using policies to drive politics. It is a difficult, often circular, process to untangle. In the next section, I will describe how observable characteristics of state legislatures helps address some of these issues. But first, I will offer an empirical definition of national policies to test the hypothesis that they are more partisan.

Measuring National Issues and Polarization

Defining National Issues

A national issue is a state legislative bill that is prominent in the national discourse. This discourse takes place between public officials, policy stakeholders, the media, and the general public. A challenge with operationalizing this definition is choosing which group of actors to observe to determine which state-level issues are national during the period in question. One could focus on the national policy agenda, and identify policies that are discussed or implemented at the federal level that are later considered at the state level, as McCann et al. (2015) did with antismoking regulations. This is also known as top–down federalism.

But top–down federalism ignores the ability of states to test policies that could later be enacted for the country as a whole, for which U.S. Supreme Court Justice Louis Brandeis labeled them the “laboratories of democracy” (Tarr, 2001). For example, in 2006, Massachusetts Governor Mitt Romney enlisted the support of U.S. Senator Ted Kennedy as he tried to pass a health care bill. Kennedy saw the bill as a possible blueprint for national legislation, and his involvement with the bill made it clear to Massachusetts legislators they were working on a national policy. Because a state-level bill can enter the national political discourse before it is introduced at the national level, top–down federalism is not a necessary condition for a state-level bill to be national.

National media sources can bring state-level bills into the national discourse, essentially giving them the ability to identify national issues at the state level. For example, the New York Times editorialized about SB 5 in 2011.⁶ S. A. Binder (2003) constructed a measure of the national agenda using unsigned Times editorials such as these. However, although national media attention may be a sufficient condition for a policy to be in the national discourse, it is not a necessary condition.⁷ The Times is not the paper of record in every state, and bills can develop national relevance for policy stakeholders and office holders outside the view of national media outlets.

The approach that I choose to operationalize this definition focuses on interest groups. I observe lobbyists representing these groups for four reasons. First, interest groups have the ability to operate at the state and/or national levels simultaneously, so they may be involved with issues in either a top–down or bottom–up federal sequence. Second, these groups interface with each of the members of the political discourse. By the nature of their lobbying, these groups are interacting with sitting legislators. Interest groups can also drive media attention, Berry (1999) found that interest groups are overrepresented in press coverage compared with the role they actually play in hearings. And the public sustains nonprofit interest groups through membership and donations (Skocpol, 2004), so they are a rough proxy of public interest. Third, the capacity of these groups to pay attention to policies is not as limited by space as the media is, so a group is likely to lobby on every issue it is interested in, whereas media can only cover so many issues per day.

Most important, interest group lobbyists are a vital information source for legislators; so, by observing this population, we are able to reasonably capture a state legislator’s view of an issue. If more of the groups that are lobbying on an issue also lobby in Congress, that issue is more likely to be in the national political discourse. Therefore, I deem policy areas that are lobbied by a higher share of national interest groups to be more national.

Independent Variable: National Policy Score

To construct this measure, I collected the universe of lobbyist registrations from three states: Indiana, Pennsylvania, and Tennessee. I also collected the lobbyist registrations from the U.S. Congress over the period 2011 to 2014. I aggregate the lobbying organizations by each subject area and calculate the share of those organizations in each state that also lobby the U.S. Congress. This allows me to create a bridge between the federal and state level, and calculate how national each issue is.

Collecting and processing the lobbying data are intensive processes, which limited me to three states. The case selection for these states had two goals. First, the states need to be “most similar” cases (Seawright & Gerring, 2008) with regard to their lobbyist reporting procedures. Although many states require lobbyists to declare their interests in words, these states require lobbyists to fill out a checklist, which allows for better identification of similar interests. This allows me to then infer that differences would be due not to reporting procedures but in the interests of the lobbyists.

Second, to generalize to the other states in my sample, these states were chosen to be “typical” cases of the country at large (Seawright & Gerring, 2008). These three states are typical along a number of key dimensions. Economically, Pennsylvania is the sixth largest state economy, but Indiana and Tennessee are more middle of the road, ranked 16 and 19, respectively, in Gross State Product in 2015. The states contain both large urban centers (such as Philadelphia and Pittsburgh) as well as influential rural sectors. Ideologically, the three states are fairly moderate, legislators in Indiana and Pennsylvania are slightly more conservative than the U.S. Congress and Tennessee is slightly more liberal (Shor & McCarty, 2011). Finally, they are not extreme geographic outliers in a way that would attract an odd set of lobbyists.

In Appendix A, I validate that the issues that are national in one of the sample states are national in the other states. This consistency allows for generalization from my sample states to the other 22 out-of-sample states that I use data from.

A number of national organizations that lobby at the state level are businesses or governmental agencies, rather than interest groups. This is a problem because my measure is meant to represent the public agenda. The public sustains nonprofit interest groups through membership and donations (Skocpol, 2004), whereas businesses only lobby on behalf of private interests. Although agencies may lobby on public issues, they are more likely lobbying for bureaucratic ends, such as funding. Unfortunately, interest groups do not identify themselves in the data. I take two steps that deal with this issue. First, I limit my sample to nonprofit groups, as this is a necessary condition of being an interest group.⁸ Second, I estimate how many nonprofits in each issue area are advocacy organizations. Project Vote Smart maintains a database of groups that engage in electioneering activities, such as endorsing candidates or rating incumbent legislators based on their roll-call votes. This allows me to differentiate between interest groups that are clearly representing an issue on the public agenda (e.g., Planned Parenthood or the National Rifle Association) against a professional organization (e.g., the National Association of Optometrists and Opticians).

In Appendix A, I report the results of a factor analysis that balances the two measures (% national and % advocacy) to maximize consistency and create a single national policy score. I report these scores in Table 1. The list generally has good face validity; the most national issues, such as abortion and gun control, were fixtures in the political debate in Washington during this period. It also reflects some realities of state politics. For example, because of largely path-dependent reasons, gambling policies, such as casino regulation or lotteries, are administered at the state level.

Table 1.

National Policy Scores: 2011-2014.

	Subject	NP
1	Abortion	0.52
2	Guns	0.26
3	Sexual orientation and gender	0.20
4	Drugs	0.19
5	Elections	0.19
6	Insurance	0.17
7	Civil rights and civil liberties	0.16
8	Labor	0.13
9	Crime	0.10
10	Health	0.05
11	Judiciary	0.05
12	Senior issues	0.04
13	State government	0.03
14	Other	0.02
15	Immigration	0.01
16	Family and children issues	0.00
17	Gambling and gaming	0.00
18	Environmental	−0.01
19	Agriculture and food	−0.02
20	Technology and communication	−0.02
21	Welfare and poverty	−0.03
22	Public services	−0.04
23	Business and consumers	−0.04
24	Energy	−0.06
25	Commerce	−0.06
26	Housing and property	−0.06
27	Social issues	−0.06
28	Transportation	−0.08
29	Budget spending and taxes	−0.09
30	Education	−0.11
31	Municipal and county	−0.13

Note. Results are standardized (i = subject) = $({\hat{X}}_{i} - {\bar{X}}_{i}) / S D_{i}$ . NP = national policy score.

Collecting Roll-Call Votes

Twenty-six states systematically report the subject matter of their legislation, allowing for an estimation of their policy agenda. As shown in Figure 3, these states provide a suitably representative sample of all 50 states on a number of dimensions including size, regional variation, and ideology. I collected the universe of more than 106,000 roll-call votes for all sessions in these states over the period 2011 to 2014⁹ from the Open States project.¹⁰

Open States categorizes these bills into 44 subject categories. In Table A3 in the appendix, I show how the subjects from Open States are mapped to 31 lobbying subjects.¹¹ Despite the best efforts by their coders, there is still a great deal of variation in these data across states. The variation derives from the way states report the activity of their legislatures in their journals, and also from the different practices in states. For example, almost all the legislative activity in Hawaii takes place on committee votes. There are also differences in how states report committee votes, amendment votes, final votes, and passage votes. I adapt Snyder and Groseclose’s (2000) scheme for coding roll-call votes to the state level, and add a category for committee votes.

Floor votes: Passage votes, including final passage votes and third consideration votes that would advance a bill from one chamber to another (but not to committee). This also includes the final action by a legislative body (such as overriding a Governor’s veto, or a House vote to concur on the Senate’s amendments to a bill that has previously passed third consideration).

Amendments: Nonpassage votes within a chamber that change language.

Committee votes: Votes within a committee.

Procedures: Votes on rules, motions to end debate, motions to recommit, motions to override the Speaker. This is also the catch-all category if a vote does not fit into the previous categories.

Dependent Variable: Party Difference

I measure the party difference for each roll-call vote over this period. This is an adaption of the party unity measure, or Rice score (Rice, 1925), which has a long tradition in the field (Lee, 2008; Sinclair, 1977). Although party difference is not the same as spatial estimates of ideology, party unity scores tend to roughly track ideal points calculated through NOMINATE (Lapinski, 2013). Party difference is the absolute value of the difference between the two major parties’ unity measures. This measure highlights roll-call votes where the parties are opposed to one another. This has two advantages: First, it is reliable with small samples, which is necessary for this project as I divide agendas by 31 subjects. Second, it highlights issues where the parties are coherent and divergent from one another, two of the key conditions for hyperpartisanship, the most common conception of polarization (Persily, 2015).

P a r t y D i f f e r e n c e_{i} = | (\frac{(% V o t e s_{y e s, D} - % V o t e s_{n o, D}) - (% V o t e s_{y e s, R} - % V o t e s_{n o, R})}{2}) | .

To relate my results to the literature that relies on ideal point estimation, in Appendix B, I also run the analysis using the dependent variable employed by Shor and McCarty (2011).

Estimating Party Difference

The unit of analysis in Hypothesis 1 is issue areas. The dependent variable is party difference by subject area on floor votes. Due to the structure of my data (most bills are assigned more than one subject code), it is necessary to estimate the average party difference for each subject. I estimate the party difference for each subject using Equation 1.

P a r t y D i f f_{j} = \sum_{j = 1}^{j = 30} β_{j} S u b j e c t I n d i c a t o r_{j} + ω_{j},

where j is a subject. After correcting for the excluded group (municipal and county issues), I report the results in Table 2. Estimating the party difference for each subject using a regression also allows me to gauge the uncertainty of each estimate: There are many more bills on taxes than immigration, so this equation estimates the party difference on taxes with more precision.

Table 2.

Party Difference on Floor Votes Across 25 States (2011-2014).

	Subject	PD	SE
1	Abortion	0.59	(0.018)
2	Immigration	0.40	(0.027)
3	Guns	0.28	(0.010)
4	Campaign finances and elections	0.26	(0.005)
5	Sexual orientation and gender	0.21	(0.018)
6	Budget spending and taxes	0.20	(0.002)
7	Labor	0.19	(0.003)
8	Welfare and poverty	0.18	(0.011)
9	Education	0.18	(0.003)
10	Environmental	0.17	(0.004)
11	Civil rights and civil liberties	0.17	(0.008)
12	Energy	0.17	(0.006)
13	Insurance	0.16	(0.004)
14	Social issues	0.15	(0.008)
15	State government	0.15	(0.002)
16	Public services	0.14	(0.004)
17	Housing and property	0.13	(0.004)
18	Health	0.12	(0.003)
19	Municipal and county	0.12	(0.001)
20	Crime	0.11	(0.003)
21	Transportation	0.11	(0.003)
22	Technology and communication	0.11	(0.006)
23	Agriculture and food	0.11	(0.006)
24	Judiciary	0.11	(0.004)
25	Business and consumers	0.11	(0.004)
26	Gambling and gaming	0.10	(0.010)
27	Drugs	0.09	(0.006)
28	Family and children issues	0.09	(0.004)
29	Senior issues	0.08	(0.007)
30	Other	0.08	(0.001)
31	Commerce	0.08	(0.003)

Note. PD = party difference.

I use these party difference estimates as the dependent variables in Equation 2. Because these are estimates, I use a weighted least squares model in the second stage, weighing observations by the inverse of their standard error ( $1 / ω_{j}$ ) from the first stage. Lewis and Linzer (2005) note consistency concerns with this method, and suggest using ordinary least squares (OLS) with robust standard errors without any weights. Specifically, they suggest running and reporting both models, which I do in Table 3.

Table 3.

There Is More Party Difference on Roll-Call Votes Concerning National Policies: 2011-2014.

DV: Estimated party difference by subject
Model	WLS	OLS, robust
Model	(1)	(2)
National policy score	0.253*	0.493*
National policy score	(0.097)	(0.181)
Constant	0.138**	0.145**
Constant	(0.010)	(0.014)
Observations	31	31
R ²	.189	.413

Note. Standard errors in parentheses. WLS weights from standard errors listed in Table 4. DV = dependent variable; OLS = ordinary least squares; WLS = weighted least squares.

p < .05. **p < .01.

The equation is as follows:

[\begin{matrix} {\hat{β}}_{1} \\ ⋮ \\ {\hat{β}}_{31} \end{matrix}] = γ_{0} + γ_{1} N a t i o n a l S c o r e_{j} + μ_{j} .

Results: Are National Issues More Polarized?

Figure 1 shows a positive association between national policies and party difference on floor votes from 2011 to 2014. The OLS line corresponds to column 2 in Table 3. The positive and significant coefficients in both columns allow me to reject a null hypothesis of no relationship between national policies and difference between the parties. This analysis indicates that national policies are more polarized in state legislatures. One concern highlighted by Figure 1 is that abortion is an outlier for both the independent and dependent variable. If abortion is dropped from the analysis, my estimate of $β$ is .12 (p = .18) using WLS and .19 (p = .06) using OLS.

A few observations do not fit the expected pattern. For example, the “Budget, Spending, and Taxes” policies divide the parties to a fairly high degree, the party difference is 0.20; however, it is one of the most state-specific policies. So, although national policies positively associate with party difference, they are not the only policies that polarize the parties. Furthermore, immigration seems like it would be a national policy, as it deals with an issue of national citizenship; however, my policy score determines it is more of a state-specific issue. This could reflect the fact that many groups expressing interest in immigration are providing human services. Therefore, this is likely a conservative estimate of this relationship.

Measuring Agenda-Based Polarization

Hypothesis 2 predicts that states with more national agendas will be more polarized, but factors other than the agenda could lead a state to have more partisan voting patterns. These factors can be grouped into three categories: voters, party leaders, and legislators.

Shor and McCarty (2011) argue that polarized roll-call voting is explained by the ideology of citizens, and report a positive correlation between ideological polarization in state legislatures and ideological polarization in individual responses to the National Annenberg Election Studies from 2002 to 2008. In a similar line, Kirkland (2014) finds a positive association between the variance in citizen ideology, as measured by the 2006 Cooperative Congressional Election Survey, with 2007 state legislative polarization.

These accounts both assume legislators, also known as political elites, would take their cues from the electorate. However, if members of the public adopt the position of their party’s legislators, as specified in the elite-driven model of opinion change (Lenz, 2013), the causal arrow points in the other direction. Evidence from the related case of partisan sorting, where citizens’ partisan identification changes to match ideology, suggests this may be the case. Wright and Birkhead (2014) report that in states where the elites were more sorted in the 1970s (as measured by the Project Vote Smart National Political Awareness Test), the electorate is more sorted today.

Second, party leader control over the agenda can influence the observed partisanship of that chamber. If majority party leaders use their control over the agenda to keep their party from being rolled (Cox & McCubbins, 2005), it will increase the average party difference for that chamber as the unity of the majority party is half the equation of party difference.

Another factor that needs to be considered is the competition between the two parties. The conditional party government theory (Aldrich & Battista, 2002) predicts that rank-and-file members will be more likely to delegate authority to their leaders if they are threatened electorally, regardless of formal rules. Therefore, if a party faces more competition from the other party, leaders may just be more effective at exerting pressure on legislators to adhere to the party line. Empirically, Rosenthal (1998) noted that more evenly matched party competition is linked to a higher share of party votes in Congress, which Aldrich and Battista (2002) and Hinchcliffe and Lee (2016) also find in the states.

Finally, the legislators themselves may vary. The legislatures in the American states also vary to a great degree in the sense that being a legislator is a full-time job. This variation results in some states being “professionalized” where legislators work for longer periods of times, have more staff, and are paid more. This can change the type of individual attracted to the job. Specifically, a representative whose full-time job is being a legislator is more likely to pursue higher office later in his or her career (Maestas, 2000). Because party fit, or ideological conformity to the national party line, matters for progressively ambitious legislators (Thomsen, 2014), professionalized legislatures could be expected to be more polarized not because of the policy agenda, but because they are filled with legislators of a different type.

Therefore, a comprehensive test of Hypothesis 2 needs to take these alternative explanations into account. The unit of analysis for this test is individual sessions, and the expectation will be that roll-call voting in more national sessions will be more partisan. The next section describes the model specification to test this relationship.

Dependent Variable: Session Party Difference

I measure the party difference on each floor vote for the states in my sample.¹² I then calculate the average party difference for each state chamber and session to serve as my dependent variable. Because party difference is only observed on bills that come up for a roll-call vote, I only include bills that received a passage vote (i.e., action agenda). Figure 2 shows the lack of any clear regional pattern of party difference across the 25 states in the sample over this period.

Independent Variable: Agenda National Policy Score

I take two steps to determine how national each session’s policy agenda is. First, I average the National Policy scores of the subjects assigned to each bill. Second, I average these scores for all the bills on each session’s action agenda. This is the key independent variable, summarized by the choropleth in Figure 3. I also include control variables, described below, to represent the other explanations of polarization mentioned in the previous section. The full specification, shown in column 5 in Table 4 and in Equation 3 tests the agenda-based model of polarization:

\begin{matrix} P a r t y D i f f_{t, s, c} = β_{0} + β_{1} N a t i o n a l A g e n d a_{t, s, c} + β_{2} E l e c t o r a t e P o l a r i z a t i o n_{s} \\ + β_{3} M a j o r i t y I m b a l a n c e_{t, s, c} + β_{4} P r o f e s s i o n a l i z a t i o n_{s} + \\ β_{5} M a j o r i t y S c h e d u l i n g_{s} + \sum_{j = 6}^{j = 8} β_{j} Y e a r I n d i c a t o r s_{t} + \\ β_{9} C h a m b e r I n d i c a t o r_{c} + μ_{t, s, c}, \end{matrix}

where t = year, s = state, and c = chamber. I cluster standard errors for the 50 chambers from the 25 states in the sample. Because there is wide variation in amount of votes per chamber, 23 in Oklahoma’s 2013 special session and 1,623 in its 2014 regular session, I employ a weighted least squares model, weighing by the number of votes in each session.

Table 4.

There Is More Party Difference on Roll-Call Votes in States With More National Policies on Their Agendas.

DV: Party difference
	(1)	(2)	(3)	(4)
National agenda	6.32**		5.72**	5.93**
National agenda	(1.76)		(1.87)	(2.02)
Opinion polarization		0.24**	0.20**	0.21**
Opinion polarization		(0.08)	(0.07)	(0.07)
Institutional factors
Majority party imbalance				0.05
Majority party imbalance				(0.08)
Professionalization				0.00
Professionalization				(0.01)
Major calendar control				0.05
Major calendar control				(0.03)
Year and chamber indicators
Upper chamber	−0.03	−0.03	−0.03	−0.03
Upper chamber	(0.03)	(0.03)	(0.03)	(0.02)
2012	−0.03	0.01	−0.01	−0.02
2012	(0.04)	(0.04)	(0.03)	(0.03)
2013	−0.01	−0.00	−0.00	−0.01
2013	(0.02)	(0.02)	(0.02)	(0.01)
2014	−0.06	−0.02	−0.05	−0.06
2014	(0.04)	(0.04)	(0.03)	(0.03)
Constant	0.18**	−0.39*	−0.28	−0.34*
Constant	(0.04)	(0.16)	(0.16)	(0.14)
Observations (chamber sessions)	137	137	137	137
R ²	.23	.15	.31	.35

Note. Standard errors clustered by 50 chambers in parentheses. Weighted least squares, weighted by number of floor votes in session. DV = dependent variable.

p < .05. **p < .01.

Data for Control Variables

Professionalization

Bowen and Greene (2014) provide a dynamic measure of professionalization. It builds off the Squire Index (2007) that combined session length, legislator compensation, and staff resources. I use the latest reading of each state’s professionalization (either 2011 or 2010).

Opinion polarization in the electorate

Following an approach similar to that of Shor and McCarty (2011), I measure the ideological distance between self-reported partisans using the 2012 Cooperative Congressional Election Survey.¹³

Majority imbalance

To proxy for partisan competition, I calculate the two-party balance for each year in the sample,¹⁴ which is necessary as party balance can change apart from elections (through retirements, special elections, etc.). I use data from the National Conference of State Legislatures.¹⁵ The equation is as follows: ${M a j o r i t y I m b a l a n c e}_{c} = | (S e a t s_{G O P, c} / (S e a t s_{G O P, c} + S e a t s_{D E M, c})) - . 5 |,$ where c is a legislative chamber in each state.

Results: Are States With National Agendas More Polarized?

Figure 4 shows a positive relationship between national agendas and party difference on floor votes in chambers across the 4 years in the sample period. This bivariate relationship is also positive and significant in column 1 of Table 4. Therefore, in the most direct test, states with more national agendas are more polarized. There are some outliers in Figure 4, especially for small sessions, such as Minnesota’s 2011 special session that only had 22 floor votes. There is a degree of temporal stability as states with more state-specific agendas (e.g., Alabama), as well as national agendas (e.g., California, Maine), tend to persist.

The remainder of Table 4 accounts for other explanations of polarization that could drive this association. Column 2 shows that opinion polarization from the 2012 Congressional Election Study is positively associated with legislative polarization, replicating the findings of the electorate-based account (Kirkland, 2014; Shor & McCarty, 2011). Column 3 shows that when both the agenda and opinion polarization are taken into consideration, the agenda is still significantly related to party difference. Column 4 demonstrates that when institutional factors are controlled for, the positive association between having a national agenda and party difference on floor votes maintains. This allows me to reject a null hypothesis of agendas having no effect on polarization.

To determine the relative scale of these effects, the bottom row of this table reports R² values that convey how much of the variation in party difference the different specifications explain. In column 2, a reduced model with only opinion polarization explains about 15% of the variation in party difference across states. The national policy agenda score in column 1 explains 23% of the variation across chambers and years. Therefore, the agenda measure explains more of the variation in party difference across the states that the polarization of the electorate does. In column 4, including all the institutional variables explains approximately 20% more of the variation in party difference than just a reduced model featuring opinion polarization.

Column 4 also shows that holding all other factors equal, considering the policy agenda mitigates any statistically significant effect of majority party control of the agenda on party voting. Because Anzia and Jackman (2013) find this control reduces the rate at which the majority is rolled, it is surprising that there would be no significant effect. This finding suggests a competing force—such as rank-and-file legislators pushing more partisan legislation when they have more access to the agenda—is at play. This result is consistent with the notion in the “Theory” section that agendas are an independent manifestation of party leaders.

Discussion

To explain why some state legislatures are polarized by party and others are not, this article advances a theory that there will be more partisan conflict on a subset of the bills on state legislative agendas. Specifically, the parties will clash on national issues, or bills that are prominent in the national political discourse. This theory is built on information disparities. Compared with state-specific issues, there is much more information about national policies provided to state legislators. This influx of information comes from national media sources, which emphasize party conflict on the national stage and focus voters’ attention on these issues, and national interest groups, which inform legislators on national party positions. This information matters for legislators interested in reelection or upward advancement and it pulls them to vote on party lines on national issues.

To test this theory, I identify national policies using lobbying registrations. I determine whether a group lobbying on a state legislative issue also lobbied before Congress, calling it a national interest group if it did. I calculate the share of national interest groups for each issue and define issues with a higher share of national interest groups as more national. The most national issues—abortion, gun control, and elections—are fixtures of national debate, whereas the most state-specific issues—education and the budget—are characterized by local provision. Using this measure as an independent variable, I find there to be more party difference on national issues on a sample of more than 106,000 roll-call votes from 25 states over this same period. Because national issues draw more attention from stakeholder groups and the public, some legislators may want to associate with these bills to build their profiles. I also find that sessions with more national agendas have more partisan voting patterns.

These results show that the policy agenda of a state’s legislature affects its polarization. This agenda-based account of polarization calls for shifting attention to the policy entrepreneurs that drive the agenda-setting process (Kingdon, 1984), as well as the media, which swarm only certain issues with attention and can influence the policy agenda (Wolfe, Jones, & Baumgartner, 2013). This account builds on a literature that primarily has looked to the electorate to explain patterns of polarization. Although my analysis replicates the finding that opinion polarization affects legislative polarization, the national agenda variable explains more variation in party difference (23%) than the electorate’s opinion polarization (15%). I also successfully test this account against alternative institutional hypotheses, such as professionalization, party competition, and majority agenda control.

This account conforms to Lowi’s (1972) dictum that “policies drive politics.” Scholars have tested this theory over time in Congress. Jochim and Jones (2013) show that although many issues are on a single liberal–conservative dimension, as predicted by the classic work on polarization by Poole and Rosenthal (1985), they arrived there via a heterogeneous process. Issues that “emphasize the traditional ideological debates around redistributive and regulatory concerns” could be simplified to a single dimension before primarily distributive policies (p. 357). Looking over a longer era, Lapinski (2013) found that domestic politics is almost always highly polarized, and the pattern changes for sovereignty policies, such as citizenship. By comparing 50 legislative chambers simultaneously, I provide a novel, cross-sectional test of this theory. Although my argument is less about the substance of policies per se, it shows that the context surrounding policies matters a great deal for roll-call voting behavior.

The way activists frame issues is a key way to change the context. In 2015, Indiana Governor Mike Pence signed the Religious Freedom Restoration Act. After the Human Rights Campaign condemned it as “anti-gay,” there was a national firestorm of attention from the press and even corporations. One reason for this onslaught of attention is the issue went from being about freedom of religion to sexual orientation and gender issues. True to my model’s predictions, Democrats and Republicans were divided on the roll-call votes on this legislation.

The results also demonstrate important linkages between national and state politics. Previous work has shown that if an issue has national attention (as observed by Congressional hearings), it is likely to be on the agenda in professionalized legislatures (McCann et al., 2015), and there will be a higher density of interest groups at the state level (Baumgartner, Gray, & Lowery, 2009). This study goes beyond these findings to show that the style of politics at the national level, specifically partisan conflict, can accompany the issue to the state level as well.

This work also presents avenues for future research to determine why some state legislatures produce more national agendas. The covariates in the model, formal powers of leaders, majoritarian institutions, and professionalization, can determine how much access backbenchers and outside groups have to the agenda. Understanding this process would provide another key step to explaining partisan polarization at the state level.

In terms of polarization more generally, this finding has mixed normative implications. Prior literature emphasizes that legislators operate in a unidimensional space and suggests preferences are fixed. In the words of Poole and Rosenthal (2000), legislators “die in their ideological boots.” An agenda-based account sees legislators as being more reactive to what is placed in front of them. So, although a legislator’s hands may be tied by his or her party’s position on a bill related to abortion, this model sees more flexibility on an education bill. This flexibility could grant legislators leeway to compromise on certain issues, which could be beneficial for all citizens (Gutmann & Thompson, 2014). My account is reminiscent of W. E. Miller and Stokes’ (1963) finding that the positions on racial issues of mid-20th-century Members of Congress were tied to geography. Even though racial issues were gridlocked, those Congresses were productive on other parts of their agendas.

However, there is a more dire prediction for states as an alternative policy venue to Congress. Former Republican National Committee Chairman Ed Gillespie told the New York Times in 2014, “People who want to see policies enacted, and see things tried, are moving their activity to the states, and away from Washington. There is a sense that you can get things done there.”¹⁶ This view implies that in the face of national gridlock, the states have the potential to be productive policy venues. But when it comes to national issues, that productivity is contingent on the partisan alignment of the states, which varies widely (Hinchcliffe & Lee, 2016). Desirable states for legislative productivity would be those where one party has a large majority, which could lead to diverging ideology of policies in red and blue states.

Polarization is not good or bad by nature, it is only a feature of certain legislatures. But understanding its origin, and the role that outside groups play in influencing legislative behavior, remains an important task for scholars interested in the generation of contemporary public policy in the American states.

Footnotes

Appendix A

Appendix B

Acknowledgements

I thank Josh Darr, Emmerich Davies, Kyle Dropp, Guy Grossman, Dan Hopkins, Carl Klarner, John Lapinski, Matt Levendusky, Seth Masket, Seth McKee, Michele Margolis, Marc Meredith, Brendan Nyhan, Steve Rogers, Stephan Stohler, Robinson Woodward-Burns, and the editor and anonymous reviewers for helpful comments.

Author’s Note

1. Previous versions of this article were presented at the 2015 American Political Science Association Annual Meeting in San Francisco, CA, 2015 State Politics and Policy Conference in Sacramento, CA, the University of Pennsylvania and Dartmouth College.

2. Author is also affiliated to The College of New Jersey, Ewing, NJ, USA.

Declaration of Conflicting Interests

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

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes

Author Biography

Alex Garlick is an assistant professor of Political Science at The College of New Jersey. His research focuses on legislative behavior and lobbying in Congress and the American state legislatures. His website is .

References

Aldrich

J. H.

Battista

J. S. C.

(2002). Conditional party government in the states. American Journal of Political Science, 46, 164-172.

Aldrich

J. H.

Rohde

D. W.

(2000). The republican revolution and the house appropriations committee. The Journal of Politics, 62, 1-33.

Ansolabehere

Rodden

Snyder

J. M.

(2008). The strength of issues: Using multiple measures to gauge preference stability, ideological constraint, and issue voting. American Political Science Review, 102, 215-232.

Anzia

S. F.

Jackman

M. C.

(2013). Legislative organization and the second face of power: Evidence from U.S. state legislatures. The Journal of Politics, 75, 210-224.

Barber

(2016). Ideological donors, contribution limits, and the polarization of state legislatures. The Journal of Politics, 78, 1-21.

Baum

M. A.

Groeling

(2008). New media and the polarization of American political discourse. Political Communication, 25, 345-365.

Baumgartner

F. R.

Gray

Lowery

(2009). Congressional influence on state lobbying activity. Political Research Quarterly, 62, 552-587.

Bawn

Cohen

Karol

Masket

Noel

Zaller

(2012). A theory of political parties: Groups, policy demands and nominations in American politics. Perspectives on Politics, 10, 571-597.

Berry

J. M.

(1999). The new liberalism. Washington, DC: Brookings Institution Press.

10.

Binder

Kogan

Kousser

(2011). How GAVEL changed party politics in Colorado’s general assembly. In Daum

Duffy

Straayer

(Eds.), State of change: Colorado politics in the twenty-first century (pp. 153-172).Denver, CO: University Press of Colorado.

11.

Binder

S. A.

(2003). Stalemate: Causes and consequences of legislative gridlock. Washington, DC: Brookings Institution Press.

12.

Bowen

D. C.

Greene

(2014). Should we measure professionalism with an index? A note on theory and practice in state legislative professionalism research. State Politics & Policy Quarterly, 14, 277-296.

13.

Broockman

D. E.

Skovron

(2013). What politicians believe about their constituents: Asymmetric misperceptions and prospects for constituency control (Working paper). University of California, Berkeley.

14.

Butler

D. M.

Nickerson

D. W.

(2011). Can learning constituency opinion affect how legislators vote? Results from a field experiment. Quarterly Journal of Political Science, 6, 55-83.

15.

Carson

J. L.

Koger

Lebo

M. J.

Young

(2010). The electoral costs of party loyalty in congress. American Journal of Political Science, 54, 598-616.

16.

Confessore

(2014, January 11). “A national strategy funds state political monopolies.” The New York Times. Retrieved from http://www.nytimes.com/2014/01/12/us/politics/a-national-strategy-funds-state-political-monopolies.html

17.

Cox

G. W.

McCubbins

M. D.

(2005). Setting the agenda: Responsible party government in the U.S. House of Representatives. New York, NY: Cambridge University Press.

18.

Darr

J. P.

(2016). Presence to press: How campaigns earn local media. Political Communication, 33, 503-522.

19.

Delli Carpini

M. X.

Keeter

Kennamer

J. D

. (1994). Effects of the news media environment on citizen knowledge of state politics and government. Journalism & Mass Communication Quarterly, 71, 443-456.

20.

Enda

Matsa

K. E.

Boyles

J. L.

(2014). America’s shifting statehouse press. Washington, DC: Pew Research Center.

21.

Eshbaugh-Soha

(2006). The conditioning effects of policy salience and complexity on American political institutions. Policy Studies Journal, 34, 223-243.

22.

Fenno

R. F.

(1973). Congressmen in committees. Boston, MA: Little, Brown.

23.

Garrett

K. N.

Jansa

J. M.

(2015). Interest group influence in policy diffusion networks. State Politics & Policy Quarterly, 15, 387-417.

24.

Gormley

W. T.

(1986). Regulatory issue networks in a federal system. Polity, 18, 595-620.

25.

Groeling

(2010). When politicians attack: Party cohesion in the media. New York, NY: Cambridge University Press.

26.

Gutmann

Thompson

D. F.

(2014). The spirit of compromise: Why governing demands it and campaigning undermines it. Princeton, NJ: Princeton University Press.

27.

Hall

R. L.

Deardorff

A. V.

(2006). Lobbying as legislative subsidy. American Political Science Review, 100, 69-84.

28.

Harden

J. J.

Kirkland

J. H.

(2016). Do campaign donors influence polarization? Evidence from public financing in the American states. Legislative Studies Quarterly, 41, 119-152.

29.

Hershey

Magan

(2011, February 10). “GOP has Statehouse clout to change collective bargaining.” Dayton Daily News, p. A5.

30.

Hertel-Fernandez

(2014). Who passes business’s “model bills”? Policy capacity and corporate influence in U.S. state politics. Perspectives on Politics, 12, 582-602.

31.

Hinchcliffe

Lee

F. E.

(2016). Party competition and conflict in state legislatures. State Politics & Policy Quarterly, 16(2), 172-197.

32.

Jackman

M. C.

(2014). Parties, median legislators, and agenda setting: How legislative institutions matter. The Journal of Politics, 76, 259-272.

33.

Jochim

A. E.

Jones

B. D.

(2013). Issue politics in a polarized congress. Political Research Quarterly, 66, 352-369.

34.

Jones

B. D.

Baumgartner

F. R.

(2005). The politics of attention: How government prioritizes problems. Chicago, IL: University of Chicago Press.

35.

Karch

(2006). National intervention and the diffusion of policy innovations. American Politics Research, 34, 403-426.

36.

Key

V. O.

(1966). American state politics: An introduction. New York, NY: Alfred A. Knopf.

37.

Kingdon

J. W.

(1984). Agendas, alternatives, and public policies, volume 45. Boston, MA: Little, Brown.

38.

Kiousis

McCombs

(2004). Agenda-setting effects and attitude strength political figures during the 1996 presidential election. Communication Research, 31, 36-57.

39.

Kirkland

J. H.

(2014). Ideological heterogeneity and legislative polarization in the United States. Political Research Quarterly, 97, 533-546.

40.

Koger

Masket

Noel

(2009). Partisan webs: Information exchange and party networks. British Journal of Political Science, 39, 633-653.

41.

Lapinski

J. S.

(2013). The substance of representation: Congress, American political development, and lawmaking. Princeton, NJ: Princeton University Press.

42.

Lapinski

J. S.

Levendusky

Winneg

Jamieson

K. H.

(2016). What do citizens want from their member of congress? Political Research Quarterly, 69, 535-545.

43.

Layman

G. C.

Carsey

T. M.

Green

J. C.

Herrera

Cooperman

(2010). Activists and conflict extension in American party politics. American Political Science Review, 104, 324-346.

44.

Lee

F. E.

(2008). Agreeing to disagree: Agenda content and Senate partisanship, 1981-2004. Legislative Studies Quarterly, 33, 199-222.

45.

Lenz

G. S.

(2013). Follow the leader? How voters respond to politicians’ policies and performance. Chicago, IL: University of Chicago Press.

46.

Lewis

J. B.

Linzer

D. A.

(2005). Estimating regression models in which the dependent variable is based on estimates. Political Analysis, 13, 345-364.

47.

Lindstädt

Vander Wielen

R. J.

(2014). Dynamic elite partisanship: Party loyalty and agenda setting in the U.S. house. British Journal of Political Science, 44, 741-772.

48.

Lowery

Gray

Baumgartner

F. R.

(2011). Policy attention in state and nation: Is anyone listening to the laboratories of democracy? Publius: The Journal of Federalism, 41(2), 286-310.

49.

Lowi

T. J.

(1972). Four systems of policy, politics, and choice. Public Administration Review, 32, 298-310.

50.

Maestas

(2000). Professional legislatures and ambitious politicians: Policy responsiveness of state institutions. Legislative Studies Quarterly, 25, 663-690.

51.

Masket

S. E.

Miller

M. G.

(2015). Does public election funding create more extreme legislators? Evidence from Arizona and Maine. State Politics & Policy Quarterly, 15, 24-40.

52.

Mayhew

D. R.

(1974). Congress: The electoral connection. New Haven, CT: Yale University Press.

53.

McCann

P. J. C.

Shipan

C. R.

Volden

(2015). Top–down federalism: State policy responses to national government discussions. Publius: The Journal of Federalism, 45, 1-31.

54.

McCarty

Poole

K. T.

Rosenthal

(2009). Does gerrymandering cause polarization? American Journal of Political Science, 53, 666-680.

55.

McGhee

Masket

Shor

Rogers

McCarty

(2014). A primary cause of partisanship? Nomination systems and legislator ideology. American Journal of Political Science, 58(2), 337-351.

56.

McLaughlin

J. P.

Wolfgang

Leckrone

J. W.

Gollob

Bossie

Jennings

Atherton

M. J.

(2010). The Pennsylvania policy database project: A model for comparative analysis. State Politics & Policy Quarterly, 10(3), 320-336.

57.

Miller

L. L.

(2010). The perils of federalism: Race, poverty, and the politics of crime control. New York, NY: Oxford University Press.

58.

Miller

W. E.

Stokes

D. E.

(1963). Constituency influence in congress. American Political Science Review, 57, 45-56.

59.

Norman

G. R.

Streiner

D. L.

(2003). PDQ statistics, volume 1. Shelton, CT: People’s Medical Publishing House.

60.

Nyhan

McGhee

Sides

Masket

Greene

(2012). One vote out of step? The effects of salient roll call votes in the 2010 election. American Politics Research, 40, 844-879.

61.

Persily

(2015). Introduction in Solutions to political polarization in America. In Persily

(ed.) (pp. 3-15). New York, NY: Cambridge University Press.

62.

Poole

K. T.

Rosenthal

(1985). A spatial model for legislative roll call analysis. American Journal of Political Science, 29, 357-384.

63.

Poole

K. T.

Rosenthal

(2000). Congress: A political-economic history of roll call voting. New York, NY: Oxford University Press.

64.

Rice

S. A.

(1925). The behavior of legislative groups: A method of measurement. Political Science Quarterly, 40(1), 60-72.

65.

Rogers

(2013). Accountability in state legislatures: How parties perform in office and state legislative elections (Working paper). Vanderbilt University, Nashville, TN.

66.

Rosenthal

(1998). The decline of representative democracy. Washington, DC: CQ Press.

67.

Schattschneider

E. E.

(1935). Politics, pressures and the tariff: A study of free private enterprise in pressure politics, as shown in the 1929-1930 revision of the tariff. Upper Saddle River, NJ: Prentice Hall.

68.

Schickler

Pearson

(2009). Agenda control, majority party power, and the house committee on rules, 1937-52. Legislative Studies Quarterly, 34, 455-491.

69.

Seawright

Gerring

(2008). Case selection techniques in case study research a menu of qualitative and quantitative options. Political Research Quarterly, 61, 294-308.

70.

Shor

McCarty

(2011). The ideological mapping of American legislatures. American Political Science Review, 105, 530-551.

71.

Shor

McCarty

(2015). Aggregate state legislator Shor-McCarty ideology data, June 2015 update. doi:10.7910/DVN/K7ELHW

72.

Sinclair

B. D.

(1977). Determinants of aggregate party cohesion in the U.S. House of Representatives, 1901-1956. Legislative Studies Quarterly, 2, 155-175.

73.

Skocpol

(2004). Voice and inequality: The transformation of American civic democracy. Perspectives on Politics, 2, 3-20.

74.

Smyth

J. C.

(2013, May 1). “Ohio leaders greet right-to-work bills with caution.” Akron Beacon Journal. Retrieved from http://www.ohio.com/news/local/ohio-leaders-greet-right-to-work-bills-with-caution-1.394205

75.

Snyder

J. M.

Groseclose

(2000). Estimating party influence in congressional roll-call voting. American Journal of Political Science, 44, 193-211.

76.

Squire

(2007). Measuring state legislative professionalism: The squire index revisited. State Politics & Policy Quarterly, 7, 211-227.

77.

Tarr

G. A.

(2001). Laboratories of democracy? Brandeis, federalism, and scientific management. Publius: The Journal of Federalism, 31, 37-46.

78.

Thomsen

D. M.

(2014). Ideological moderates won’t run: How party fit matters for partisan polarization in congress. The Journal of Politics, 76, 786-797.

79.

“Union-bashing, now in Ohio. ” (2011, April 4). The New York Times, p. A22.

80.

Uslaner

E. M.

Weber

R. E.

(1979). U.S. state legislators’ opinions and perceptions of constituency attitudes. Legislative Studies Quarterly, 4, 563-585.

81.

Wilson

(2016, December 13). “Kasich signs 20-week abortion ban, vetoes ‘heartbeat’ bill.” The Hill. Retrieved from http://thehill.com/policy/healthcare/310238-kasich-signs-20-week-abortion-ban-vetoes-heartbeat-bill.

82.

Wolfe

(2012). Putting on the brakes or pressing on the gas? Media attention and the speed of policymaking. Policy Studies Journal, 40, 109-126.

83.

Wolfe

Jones

B. D.

Baumgartner

F. R.

(2013). A failure to communicate: Agenda setting in media and policy studies. Political Communication, 30, 175-192.

84.

Wright

G. C.

Birkhead

(2014). The macro sort of the state electorates. Political Research Quarterly, 67, 426-439.