1
Chapter 1: Introduction to the Study
The top-two primary is a relatively new election system currently used
in only four U.S. states: Washington, California, and, on a limited basis in
Nebraska (Jackson, 2012; Ballotpedia, n.d.-c). This system allows all voters,
regardless of party preference, to vote for their candidate among all
candidates on a primary election ballot. The top-two vote earners in each
race advance to the general election, regardless of their party affiliation. The
result of this top-two advancement is that in some jurisdictions, two
Republicans or two Democrats may be featured on the general election
ballot. Proponents of the top-two primary system assert that it brings a
corrective balance to the tension among three competing interests in the
American electoral system: voters, candidates, and political parties (Open
Primaries, n.d.). Proponents have also asserted that the toptwo primary
system results in more moderately aligned elected officials and less party
polarization (Alvarez & Sinclair, 2015).
Although there is extensive and varied literature on elections in
general, the body of research on the top-two primary is relatively sparse. For
example, Dancey and Sheagley (2013) studied the heuristic value of political
party labels for voters. Neiheisel and Niebler (2013) researched the strategic
2
value of political party identification for candidates. There is research
suggesting that the type of primary has an impact on voter turnout as well as
on the ideologies expressed by candidates once in office (Calcagno &
Westley, 2008; Levendusky, 2010). There is also research suggesting that,
under the toptwo primary system, political parties have less control over
candidate party identification going into the primary, and thus, parties may
be incentivized to discourage some candidates from running under this
system to avoid diluting the vote (Beck &
Henrickson, 2013).
Despite this varied research, no literature was found that provided a
perspective on how political parties respond to the top-two primary, leaving
the public policy development process in a disadvantaged state. Literature is
lacking on whether or how the top-two primary has an impact on party
strategy or engagement with candidates or voters. There has been increasing
recent interest in adopting the top-two primary by members of
Congress and several states’ legislators and initiative writers. Thus, the
system may be on
the rise (Ballotpedia, n.d.-d).
3
The subsequent sections of this chapter include a review of the
background of the study, a statement of the problem and the purpose of the
study. Additionally, I will review key features of the study, including
theoretical foundation. Finally, assumptions limitations, and considerations
of the significance of this inquiry are also included in this chapter.
Background of the Study
As a highly unique form of primary election, a working understanding
of the differences of this system—as compared to other primary election
systems—is key to understanding the nuances and background of this study.
First, I present the context in which the top-two primary operates. Next, I
discuss the sampling and summary of a selection of journal articles that
informed the conception of this study. More on the context of the top-two
primary and the relevant literature is found in Chapter 2.
Context
There are several forms of primary election system in the United
States. In their study of how the form of primary election system impacts
voter turnout, Calcagno and Westley (2008) provided a concise run-down of
the types and features of primary systems in the United States. Open primary
systems are those in which all registered voters may vote for any candidate,
4
and the candidate with the most votes in each party advances to the general
election (Calcagno & Westley, 2008). Fifteen U.S. states hold open primaries
(National Conference of State Legislatures, n.d.). Closed primary systems
are those in which only voters registered as members of the party may vote
for candidates running in that party (Calcagno & Westley, 2008). Nine U.S.
states hold closed primaries (National Conference of State Legislatures,
n.d.). Twenty-one states fall somewhere between open and closed primaries,
with considerable state-level variability. Yet all advance one candidate (the
highest vote earner) per political party to the general election (National
Conference of State Legislatures, 2011).
Top-two primaries are held in just four U.S. states: Washington,
California, and for a limited number of races, in Nebraska. In top-two
primary states, all filed candidates (candidates who have officially registered
with their county auditor or clerk or local equivalent) to appear on the ballot
are listed on ballots accessible to all registered voters, and the top-two vote
earners progress to the general election, regardless of party affiliation
(National Conference of State Legislatures, n.d.). In some jurisdictions, this
results in two Republicans or two Democrats advancing to the general
election. In an interesting observation, Zhang (n.d.) noted that a top-two
5
primary “converts a traditional primary into a general election, and a
traditional general election into a runoff election” (p. 617).
Approach
This topic can be addressed from three angles. One angle of inquiry is
the impact of the top-two primary on voters in terms of both turnout and
voting behaviors. Based on literature about the heuristics used by voters to
make decisions on their ballots, it is possible that a particular top-two
primary result (i.e., two Republicans or two Democrats advancing to the
general election ballot) might disrupt the mental short-cut voters use to
choose whom to vote for (Dancey & Sheagley, 2013; Levendusky, 2010).
Additionally, there is research that suggests that the type of primary, as well
as the degree of difference between candidates, have an impact on voter
turnout (Calcagno & Westley, 2008;
Hortala-Vallve & Esteve-Volart, 2011).
Another angle of inquiry is the impact on candidates. Research about
candidates’ use of political party labels when advantageous to a campaign
(McGhee et al., 2014; Neiheisel & Niebler, 2013), coupled with the statistics
that certain jurisdictions are consistent in supporting one party’s candidates
over the other raise a concern about candidate motivation. In a top-two
6
primary environment, it is possible that a candidate might align with a party
in order to win, and not necessarily because the candidate embraces the
values, positions or ideologies of that party.
A third angle of inquiry is the impact on political parties. In the top-
two primary, political parties have lost the straightforward ability to limit
who bears the party name.
Pre-primary candidate vetting and party endorsement processes are party
responses to the top-two primary. Research suggests that political parties are
incentivized to discourage excess candidates from running (Beck &
Henrickson, 2013) and that the minority party officials elected under the top-
two primary tend to be more centrally aligned (Alvarez & Sinclair, 2012). This
may or may not please the political parties. There may be other ways political
parties are working to maintain their influence over election outcomes or at
least candidates. Of these three approaches, I focused on the political parties
at the county level in Washington State.
Resources
In reviewing the background for this inquiry, I used several studies
from the literature for the context of this study. Alvarez and Sinclair (2012)
build on the understanding that a variety of legislative institutions “constrain
7
the space of possible outcomes” of legislative activity, meaning that the
institutional context in which a legislative body operates sets boundaries and
parameters around the laws that will actually be passed (p. 544). Alvarez and
Sinclair studied the nonpartisan blanket primary (known in the present study
as the top-two primary) and reviewed for potential impacts of that electoral
institution on legislative behavior. They found that first-year legislators are
more likely to agree with other legislators in their party if elected under the
top-two primary.
Following their first study of how the institution of the top-two
primary impacts legislative behavior, Alvarez and Sinclair (2013) turned
their attention to so-called decline to state (DTS) voters, those who choose
not to declare a party preference in order to vote a primary election ballot.
Prior to the establishment of the top-two primary in California, these
unaffiliated voters were not permitted to vote in the primary. At the outset,
they caution against assuming DTS voters are a uniform group of
independents and essentially nonpartisan, and their study bears this out. Only
a third of DTS voters able to vote in the top-two primary identified as
independents, while the other two-thirds acknowledged they lean Democrat
8
or Republican, sometimes (10% for Democrats and 5% for Republicans)
strongly so.
Beck and Henrickson (2013) reviewed the legislative primaries for
2004, 2006, 2008, and 2010, with a view for the number of candidates in
each primary. The top-two primary was first used in 2008. Their results were
mixed, with the observation that the top-two primary system was affiliated
with fewer Democrats per primary, while there was no change to the number
of Republicans.
Hortala-Vallve and Esteve-Volart (2011), like many others, focused
their investigation on voter turnout. They focused on the policy differences
between candidates and the impact those differences might have on turnout.
Although their study does not directly address the top-two primary, their
conclusion that the more that voters perceive policy differences between
candidates, the more likely a voter is to vote. Conversely, when candidates
are perceived as having highly similar policies, voter turnout is decreased.
Their study’s relevance to the present one hinges on the premise held by
toptwo primary advocates that this election system incentivizes more
moderate candidates into the field, rewards them with public office, and
results in more moderate politicians.
9
McGhee et al. (2014), addressing nominating systems and legislator
ideology, reviewed whether differences in primary system in a state is
predictive of the degree of polarization of its legislators. They found that
polarization is essentially the same across all systems. The type of primary
system does not seem to follow a pattern and that primary election systems
“have little consistent effect on legislator ideology” (McGhee et al., 2014, p.
347).
Calcagno and Westley (2008) conducted their study before the
adoption of the top-two primary in California and Washington but focused
on how the primary type influences voter turnout. Calcagno and Westley
(2008) noted that the literature they drew upon suggests that “voter turnout
will be greater in the general election when the state has a more open
primary” (p. 97). The top-two primary is more open even than the open
primary of before, as it allows the top-two vote earners, regardless of party,
to advance to the general election. In this study of gubernatorial races, they
conclude that indeed a more open primary leads to greater general election
turnout.
Dancey and Sheagley (2013) opened their article with a discussion of
the importance of candidate party affiliation as a shortcut or heuristic for
10
voters. For the present study, this article is beneficial because under a top-
two primary in Washington State, candidates choose their own party
affiliation without approval or endorsement from the party. Dancey and
Sheagley determined that heuristics such as party affiliation are useful when
correct but may be destructive or detrimental if they do not convey accurate
information.
Proponents of the top-two primary state that this primary system will
result in more moderate elected officials. McGhee and Shor (2017)
concluded that the moderating effect is modest and inconsistent, and they
point out that concurrent with the policy change was a legislative
redistricting in Washington State. They determined that the toptwo has
mixed results when it comes to mitigating political polarization.
Although there has been a fair amount of research that is topically
relevant to the top-two primary, there remains a noteworthy gap in
knowledge surrounding the response of political party organizations to this
system. As political parties are an historic and current influence on all
aspects of elections, I conducted this study to address this gap.
11
Problem Statement
The problem I examined is that the top-two primary has created
uncertainty and confusion among county political parties. Currently, political
parties have no uniform method of identifying their chosen candidates nor a
way to consistently share that information with the voting public. In some
jurisdictions in Washington State, some primary races are effectively races
between several Democrats or multiple Republicans, without the opposite
party represented on the ballot (B. Galarza, personal communication,
October 19, 2019). One county party with which I am familiar attempted to
clarify its choice for voters through vetting surveys and interviews grounded
in the grassroots-adopted party platform and a rigorous endorsement process,
but prior to this study it was unknown whether this was a unique response or
whether other county party organizations have developed their own
responses. There has not been a systematic or scientific documentation of
county party response to the top-two primary, creating a vacuum of
knowledge with implications for future public policy development as well as
the research surrounding electoral systems. Well-intentioned but poorly
informed legislation does not serve the goals of free and fair elections.
12
The Washington State Republican Party, the Washington State
Democratic Party, and the state Libertarian Party were co-plaintiffs in the
case Washington State Grange v.
Washington State Republican Party, opposing the top-two primary at its
inception
(Washington State Grange v. Washington State Republican Party, 552 U.S.
442, 2008). The plaintiffs lost, with the Supreme Court upholding the top-
two system, and for political parties in Washington, the historical ability to
choose which candidates bear the party’s name was stripped away. Although
the top-two primary has been in effect for fifteen years, the scholarly
literature is silent on how or whether political parties have found a
workaround for the perceived hit to their First Amendment rights of
association. This study contributes to the body of knowledge needed to
address this problem by documenting, summarizing, and evaluating county
political party response to the top-two primary in Washington State.
Purpose of the Study
The purpose of this qualitative study was to gather the perspectives of
political party leadership in order to explore and understand Washington
State county political party response to the top-two primary election system.
Accessibility and equality in elections are of central social importance in any
13
representative democracy. Political parties and voters have the right to full
and unimpeded participation in the elections system. Candidates have the
right to choose their own party identification. By their nature, political
parties are associations protected by the First Amendment in the United
States Bill of Rights (U.S. Const., amend I, n.d.). Opponents of the top-two
primary cite infringement on this right of association, and this study
considered whether and how political parties may be responding to their
decreased ability to determine who bears the party name. The results of this
study may be used by all those involved in the public policy development
process, determining whether participants view this system as better, worse,
or neutral compared to other primary election systems, thus leading to
positive social—and societal—change.
Research Question
The central research question of this study was: What has been the
response by county-level political party leadership in Washington State to
the top-two primary?
Theoretical Foundation
Policy feedback theory conceives of public policy development in the
complex context of already existing public policy, evolving as a result of
14
interaction with multiple streams, in turn influencing future policy
innovations or decisions (Mettler & SoRelle, 2014). Public policy at Time A
may influence the meaning of citizenship (such as voter turnout and voter
behavior), the form of governance (the behavior of candidates and elected
officials), the power of groups (such as political parties), and the definition
of policy problems or political agendas (expansion or elimination of the top-
two primary into other jurisdictions), all of which interact to impact public
policy decisions or innovations at Time B (Mettler & SoRelle, 2014). By
evaluating one of these streams, that of the power of groups—namely county
political parties—this policy can be evaluated at Time A to influence public
policy development towards Time B (Mettler & SoRelle, 2014). Policy
feedback theory has its roots in earlier research, adding depth and dimension
to the widely recognized phenomenon in which “policy, once enacted,
restructures subsequent political processes” (Skocpol, 1991, p. 58).
Additional attention to policy feedback theory is given in Chapter 2.
I addressed both policy and political party response to policy; as a
result, it was important to identify a theoretical framework that could
account for both the political nature and the policy nature of the questions
being raised. I chose policy feedback theory, as articulated by Mettler and
15
SoRelle (2014), because this theory, above other considered theoretical
frameworks, acknowledges the status of political organizations within the
policy development process.
Policy feedback theory was also selected because it is in its relatively
early developmental stages. In their recommendations for future research,
Mettler and SoRelle (2014) stated, “Policy Feedback Theory is indispensable
for scholars trying to understand how policies, once developed, reshape
politics” (p. 175). In this study, I stretched the application of the theory
beyond its previous status to new concepts. Additionally, given the scarcity
of qualitative studies using policy feedback theory as a framework, I also
expanded the application methodologically.
Washington State, the focus of this study, has been characterized as
having “many elements of populism, political independence, and indeed,
overt anti-partyism” (Appleton & Grosse, 2004, p. 28). Additionally, the
framers of the Washington State Constitution, like other states joining the
union in the Progressive Era, were “strong advocates of popular sovereignty
and more direct forms of democracy” (Cornell & Meyer, 2004, p. 96). These
characteristics resonate with the tenets of policy feedback theory that
suggests that political climate influences policy, which in turn influences
16
political climate, thusly influencing policy. Washington’s current top-two
primary is an understandable outgrowth of the state’s beginnings and history,
another reason policy feedback theory was an appropriate option.
Attributing early work in this area to Schattschneider (1935), Mettler and
SoRelle
(2014) quoted him as saying, “new policies create new politics” (p. 152).
Lowi (1972) is also credited with pioneering work in this area. Mettler and
SoRelle (2014) stated that most studies have been single-policy case studies,
and the focus seems to be on either policy disbursement beneficiaries or the
dispersing agencies. As a result, the degree to which the current study fits
within the previous body of policy feedback theory’s historical application
was somewhat limited. I expanded that applicability.
Within policy feedback theory, two types of effects are identified:
resource effects and interpretive effects. Mettler and SoRelle (2014)
described resource effects as those effects that are attributable to a group
receiving some monetary or resource benefit. Resources are tangible and
quantifiable. Resource effects are not particularly relevant to the present
study, as there are no beneficiaries of concrete resources (as in the case of
welfare or public housing). Interpretive effects, however, are at play.
17
According to Mettler and SoRelle, interpretive effects “may be
fostered…through features of policy design and implementation….these
may convey messages to people about government or their relationship to
it…[which] may shape people’s subsequent participation” (2014, p. 168). In
the top-two primary, voters are not required to state a party affiliation at the
time of registration, nor are candidates chosen by a political party. The top-
two policy is less concrete than Social Security, business-sector dominance,
or even suffrage. It is about voters and candidates not having to be locked
into a party affiliation, a long-standing value held in the political culture in
Washington State.
Nature of the Study
I used a qualitative multiple case-study approach, using standard
open-ended interviews as well as available documents to determine and
define the response by county-level political party leadership in Washington
State to the top-two primary system. Given the newness of the system and
the need to define party experience from the ground up, a qualitative
approach was best suited. There are 39 counties in Washington State, each
with a statutorily defined central committee for each major political party,
Republican and Democratic, and defined leadership roles in each committee.
18
The top-two primary system is a real-life, contemporary phenomenon that
formal groups (county political central committees) have had to contend
with; as a result, a multiple case study approach was the approach best
suited. According to Creswell
(2013), a case study is a “qualitative approach in which the investigator
explores a reallife, contemporary bounded system…through detailed, in-
depth data collections” (p. 97). The current top-two primary system and the
impact on how a subset of the 39 Washington counties are reacting to this
system was certainly appropriate for a casestudy approach.
I also considered a phenomenological approach but rejected it in favor
of the multiple case-study approach for the central reason that the research
questions focus not just on the shared, lived experiences of political party
leadership but also on more concrete outgrowths of their experiences, such
as strategies, rubrics, policies, etc. developed in response to the top-two
primary. A multiple case-study approach incorporates both types of data to
benefit the larger body of scientific literature on the subject. Themes were
drawn from participants' thoughts, reactions, answers, and contributions.
19
Definitions
County political party: Roughly synonymous with county central
committee defined under Washington State statute RCW 29A.80 (Political
Parties, 1973). Under statute, elected precinct committee officers constitute
the central committee. The RCW does acknowledge that party organizations
may adopt rules governing the organization, which may include expanding
party membership to additional appointed and/or duespaying individuals
who are not elected precinct committee officers. This expanded participation
was assumed in this study. Not included in this definition are voting
members of the general public who identify with a particular political party;
formal engagement with a defined party organization was required (Political
Parties, 1973; Stevens County Republican Central Committee 2021-2022
Bylaws, 2021).
Primary election: An election held prior to a general election where
candidates are winnowed down to two candidates that will appear on the
general election ballot (McDonald, 2014).
Top-two primary election: Also known as the nonpartisan blanket
primary, an election using a common ballot, listing all candidates on the
ballot, resulting in the toptwo vote-getters advancing to the general election,
20
regardless of candidate party identification (National Conference of State
Legislatures, n.d.).
Assumptions
I made a few assumptions in this study. Some of my assumptions were
developed as a result of my own historical experience both as a voter in
Washington State, as a political party member, and as an election official.
My first assumption was that with political parties, the qualifications,
expectations, and values of a party and its preferred candidate are fairly
homogenous, party organization to party organization. I assumed that what
makes a candidate running as a Democrat a favored, viable candidate in one
county is substantially similar to that which is favored and valued in another
county. There may be some minor differences, but the similarities are
assumed to outweigh the differences.
My second assumption was that electing candidates who espouse the
values of the party with which they say they identify is a crucial motivator
for political party organizations. An example found in one county lists as the
first item in the purpose section of their bylaws “to elect Republicans to
public office” (Stevens County
21
Republican Central Committee 2021-2022 Bylaws, 2021, Article 2.1). I
assumed that this is a consistent priority across different party organizations.
My third assumption pertained to the length of time that the top-two
primary has been in place. I assumed that it has been in effect long enough to
evaluate political party response. Specifically, I assumed that party
organizations have had the chance to consider its ramifications, develop a
response, and perhaps even evaluate and adjust that response based on
several election cycles worth of information.
Scope and Delimitations
The research question focusing on political party response was broad.
Because there is no literature on this specific topic, I purposefully chose a
broadly stated research question. In this case, any response identified by
participants was considered relevant.
The populations impacted by the top-two primary include anyone
living within a jurisdiction using the top-two primary and anyone living
within a location considering this election reform. For this study, it was not
feasible to include participants who might live in a jurisdiction actively
considering the top-two primary, but that would be a worthy group for
consideration in a future study. Transferability may also apply to other
22
election reform efforts, such as ranked choice voting—also not taken up in
this study.
Limitations
Access to and rapport with participants was not an issue, as I have
established bona fides within the Republican community. Particular care was
taken to ensure proper separation of roles between researcher and elections
official, given that I am the Republican elected county auditor in my county.
Acknowledging potential bias, both explicitly with participants and within
the collection and evaluation of participant responses, required special
attention. One anticipated challenge was gaining the full participation of the
number of participants targeted. Political party leaders are predominantly
unpaid volunteers; persuading them to participate, given the demands on
their time or interest in the study, was somewhat problematic. While the
desired number of participants was ultimately received, there was no excess
of volunteer participants.
The primary limitation of this multiple case study approach is that of
the 39 Washington counties, only a handful were selected for participation in
the study. Whereas all 39 county party organizations were approached, a
limited number of volunteering respondents were selected. A target group of
23
five to eight participant organizations constituted the cases considered. A
second limitation of this study was that only Republican organizations were
approached. This, of course, leaves a well-defined area for future research,
namely other political organizations. A third limitation of this study is that
only Washington—to the exclusion of other states utilizing the top-two
primary— was considered. These limitations are limiting factors but should
not be construed as weaknesses, as the limits were placed intentionally for
reasons of feasibility, transparency, and in recognition of potential bias on
both the part of myself as the researcher and participants.
Significance
In pursuit of positive social change, this study was significant for two
distinct reasons. There are both immediate, proximate concerns that may
have very real local impacts, and then there are more distant, large-scale
possible impacts on public policy development outside of Washington. For
the immediate, this study evaluated whether county-level political parties
have developed reasoned, philosophical criteria by which they vet and
endorse candidates or whether party response is more about popularity,
connections, or other considerations. In this way, the study is significant
24
because it tells the story of what is happening in select Washington counties
at present.
Additionally, by filling a significant gap in the literature on the top-
two primary in general, the impacts of this study may reach far beyond the
borders of Washington. The current use of this system in just four states is an
example of what Supreme Court Justice Louis Brandeis once called the
states acting as laboratories of democracy of public policy development
(New State Ice Co. v. Liebmann, 1932). In the 2023 “Alternative Voting
Methods in the United States” report published by the U.S. Election
Assistance Commission (U.S. Election Assistance Commission, 2023), more
than 20 states have seen legislation introduced related to the top-two
primary, including Alabama, Alaska, Arkansas, Florida, Hawaii, Idaho,
Illinois, Maine, Maryland, Minnesota, Mississippi,
Missouri, Nevada, North Carolina, Oklahoma, Oregon, Rhode Island, Texas,
Virginia, Wisconsin, and Wyoming. Alaska used the top-four primary
variation in its 2022 election cycle (Anderson et al., 2023). It remains
essential that the laboratory of the top-two primary publish its findings so
policymakers, voters, parties, and candidates can more fairly and rationally
decide whether this system is worth replicating elsewhere.
25
Summary and Transition
The top-two primary, a relatively new innovation in elections in the
United States, presents a number of issues worthy of study. Given the
ongoing and evolutionary nature of public policy surrounding elections in
the United States, this multiple case study— applying the framework of
policy feedback theory--fills a gap in the literature, contemplates and reports
current practice, and is contributing to positive social change by evaluating
the political party response to the top-two primary. In the next chapter, the
comprehensive review of the literature will include the following: the
development and application of policy feedback theory from its inception,
the past and present political context of Washington State elections, and the
historic, legal, and political context of the top-two primary in particular.
Chapter 2 features the literature review strategy used and the results of
that review. In addition, the theoretical foundation of the current study is
covered. Chapter 3 focuses on research methods, issues of trustworthiness,
and ethical considerations for this study. Chapter 4 features results including
data analysis, and Chapter 5 centers on discussions, conclusions, and
recommendations for future study.
26
Chapter 2: Literature Review
Washington State’s top-two primary has created uncertainty and
confusion among county political parties and the voters that rely on heuristic
cues from the parties. The top-two primary leaves parties with no uniform
method of identifying their chosen candidates, nor a way to consistently
share that information with the voting public. Although the literature on
elections, in general, is extensive, and the literature on the larger category of
primary elections is substantive, the literature specific to this relatively new
and geographically limited expression of the primary election—the top-two
primary—is scarce. In this chapter, I will place the existing literature
relevant to the toptwo primary in the larger context of elections and
primaries within the United States, the implications of public policy on the
behaviors of citizens and groups, and the role of political party identification
and the parties themselves in that same context.
The following sections include a discussion of the literature search
strategy, including database review and listing of search terms. I review and
contextualize the theoretical foundations relevant to political party response
to the top-two primary in Washington State. Finally, I present a thorough
review of the applicable literature on the topic.
27
Literature Search Strategy
Library databases that I accessed included Political Science and
Business Source Complete (combined search available through Walden
University Library), Dissertations and Theses @ Walden University, EBSCO
eBooks, Public Administrations Abstracts,
Theoreau @ Walden University Library, and Google Scholar. I used the Find
@Walden tool extensively to optimize the time spent in searching and to
acquire full-text versions of the relevant journal articles. Additionally, I also
used the reference sections of particularly relevant journal articles to refine
the search and locate articles of particular benefit.
Literature review searches in the initial stages of this dissertation
(2017-2018) were attempted on the highly specific terms top-two primary
and nonpartisan blanket primary, but given the relative newness and lack of
literature on the topic, additional search terms were used, particularly to
establish context for this particular election reform. The following terms and
combinations thereof were used: primary, open primary, closed primary,
U.S. primary elections, elections, California primary, Washington Primary,
United States elections, blanket primary, and direct primary. Later literature
searches (conducted primarily in 2019-2023) demonstrated an increasing
28
interest in this particular primary election form. As a result, top-two primary
and nonpartisan blanket primary did reveal additional contemporary
literature. Nevertheless, literature specific to political party response to the
top-two primary remains lacking.
I searched for an additional area of literature specifically on the search
terms of policy feedback theory, policy feedback, and policy feedbacks
combined with terms such as civic engagement, voting, and elections. There
was a lack of results for these terms, representing a gap in the literature and
an opportunity to expand the applicability of policy feedback theory, as
explored further below.
Theoretical Foundation: Policy Feedback Theory
The theoretical foundation of the present study is policy feedback
theory. Policy feedback theory represents the maturation of early
observations of the public policy development process that has solidified
into a concrete and well-defined theory over time. As will be shown, as time
passed and scholars engaged with questions of public policy development
with this lens, increased specificity was introduced into the framework.
Oft quoted Schattschneider (1935) may be considered a progenitor of
modern policy feedback theory. He is credited with the observation, “new
29
policies create new politics” (Schattschneider, 1935, p. 288). With that
comment, he noted a cause-and-effect relationship between policy and
politics that has become a foundational concept of policy feedback theory.
Skocpol (1996) brought significant detail and specificity to the
concept introduced by Schattschneider. Skocpol’s seminal work on the
development of social policies in the United States, such as mothers’
pensions, Civil War pensions, and minimum wage and child labor laws,
fleshed out how “previously established…policies effect subsequent
politics” (1996, p. 41), which in turn becomes the context in which new
policy is developed. Such politics referred to is often expressed through
groups that form, organize, or reorganize in response to incentives or
resources to pursue. This phenomenon has been seen across multiple public
policy sectors, resulting in Skocpol’s work being cited in a broad range of
literature.
Much of the policy feedback theory literature centers around concrete
social policies where tangible or even monetary gains to citizens are in play.
As the body of literature has expanded and policy feedback premises have
been adopted, this approach “reimagines the classic unidirectional model of
30
democracy in which politics (public opinion, elections) is the cause and
policy is the effect or outcome” (Haselswerdt, 2017,
p. 4). Studies focusing on Medicaid (Clinton & Sances, 2018), the
Affordable Care Act (Chattopadhyay, 2018; Haselswerdt, 2017), and the GI
Bill (Mettler & Welch, 2004) are a few examples where the cost or benefit of
the policy has been more evident to the citizen. Other studies have reviewed
policies that are more philosophical than monetary, such as school choice
laws (Fleming, 2014), public sector collective bargaining laws
(Flavin & Hartney, 2015), or laws pertaining to the legalization of same-sex
marriage (Kreitzer et al., 2014), education accountability policies
(McDonnell, 2013), prostitution (Kotsadam & Jakobsson, 2011) or smoking
(Pacheco, 2013).
Clearly, there has been a demonstrated applicability of policy
feedback theory to a variety of policy implementations. Campbell (2012)
summarized much of the work done up until that point in her piece on the
impact of policies on mass politics, where she stated, “public policies are not
merely products of politics but also shape the political arena and the
possibilities for further policy-making” (p. 334). In a similar vein, Mettler
(2019) wrote:
31
Policy feedback research examines how policies, in addition to
producing firstorder or intended effects, may also influence politics
itself by altering citizens’ political behavior. For example, policies
may affect citizens’ political attitudes, including their support for
policies, and they may also shape citizens’ rates of political
participation, including their likelihood of taking action to advocate
for the policies they utilize (p. 32).
Thus, policy feedback theory is highly applicable to the current study.
Mettler and Soss (2004) issued a call for expanding research using
policy feedback theory and more sophisticated application of the theory,
noting “political science stops short of exploring public policies’ influence
on what citizens want, how strongly they want it, and whether they engage
the political processes that hold the power to supply or deny it to them” (p.
56) and later noting that this field of study “should be able to explain why
some policies draw citizens into public life and others induce passivity” (p.
56). It is worth noting that studying public policy development through the
policy feedback lens does require a long view of history. Pierson (1993)
wrote “political processes can best be understood if they are studied over
time” (p. 596).
32
Goss (2010) answered that call for research and studied citizen
organizations with a policy feedback lens. As a note meant to tie this
research to the present study, political parties can be understood as citizen
organizations. Goss called attention to the particular American tendencies to
form associations, a protected right under the First Amendment, and one that
political parties in Washington State have asserted is damaged by the toptwo
primary. Goss and others recognized that policies could cause the growth of
some organizations and discourage the growth of other types of
organizations. As an example, “federal state and local social policy…favors
charitable organizations” by making grant funding available, while on the
other hand, federal laws limit lobbying and the strategies that charities may
utilize for social, economic, and political change (Goss, 2010, p. 126).
I chose policy feedback theory for this study because political
behavior and attitudes are very much in play when it comes to elections.
Social scientists have studied voter turnout, mobilization of voters, civility in
campaigns, campaign spending, moderation of elected officials, and so on.
All such areas of study are meritorious, and Mettler and Soss (2004) agreed
these areas could be effectively studied using a policy feedback lens.
33
For many citizens, voting is the core of their political behavior,
although group participation—such as active political party membership—
may also be an expression of political behavior. Political attitudes such as
Washington State’s decidedly independent nature, springing from populist
beginnings, are also both an outgrowth and an input to the ongoing policy
development surrounding elections in Washington. The literature review
below sets this study within the scientific context, but within an historical
and political context as well.
Literature Review
The following sections ground this study in the cultural norms
inherent in Washington state’s election systems over time. It is important to
first examine the historic, legal, and political context of the top-two primary.
Subsequent to the review of the context is an exhaustive review of the
existing literature on this topic.
Historic, Legal, and Political Context of the Top-two Primary
Washington State’s top-two primary must be placed within a very
specific historic, legal, and political context. While by no means
comprehensive, the following attempts to give a sense of this context. From
statehood in 1889 until 1907, Washington (like most other states) had its
34
candidates nominated by political party membership, in this case through
party conventions. Washington State’s relationship with primary elections
started with a direct primary, or what would be called a closed primary
today. Political party nominees were selected through a balloting process
wherein separate ballots were printed for each party, and a voter could only
cast a ballot specific to the party to which that voter was a member or stated
allegiance (History of Washington State
Primary Systems, n.d.). This form of primary election persisted from 1907
through 1934.
From 1935 through 2003, Washington State voters had a much more
open primary system dubbed a blanket primary, under which they could vote
for “any candidate for each office, regardless of political affiliation and
without a declaration of political faith or adherence on the part of the voter”
(History of Washington State Primary Systems, n.d., 1935-2003). It is not
difficult to intuit that the stability over more than two generations of voters
had an impact on the political identification of
Washingtonians as independently minded. As Spitzer (2019) stated,
“political culture can be path dependent, meaning that practices remain
stable over time” (p. 847). He stated that “[Washington’s] open government
attitude can be traced to its populist constitution, later accelerated by
35
progressivism” (Spitzer, 2019, p. 848). Incidentally, the initial blanket
primary was brought about through an initiative to the legislature process
(Spitzer, 2019), another populist and progressive governance method (Barth
et al., 2020).
The early 2000s brought legal challenges, first to California’s blanket
primary, then to Washington State’s version. In 2000, the U.S. Supreme
Court ruled in California Democratic Party v. Jones (California Democratic
Party v. Jones, 530 U.S. 567, 2000).
Between 2000 and 2003, Washington State’s top elections officials—
Secretaries of State Ralph Munro and Sam Reed—held hearings and
published a report that clearly showed that most Washington State voters
“are independent and want to continue to participate in the primary without
having to affiliate with a political party and without being restricted to the
candidates of only one party in the primary” (Washington Secretary of State,
2001, para 7). Despite this sentiment among Washington voters,
Washington’s blanket primary was deemed unconstitutional in 2003, just as
California’s had been (Washington State Grange v. Washington State
Republican Party, 552 U.S. 442, 2008).
36
Consistent with Washington State’s populist and progressive past,
voters turned to direct democracy for their remedy by passing Initiative 872,
which passed in 2004, creating the nonpartisan blanket primary, a term
synonymous with top-two primary. (History of Washington State Primary
Systems, n.d.). Though it passed, its implementation was held up by
additional court challenges that were ultimately resolved by the United
States Supreme Court in 2008. I-872 was deemed constitutional as it does
not “impose a severe burden on the political parties’ associational rights, and
that the parties’ arguments that voters will be confused can only be evaluated
once the primary is implemented” (History of Washington State Primary
Systems, n.d., March 18, 2008).
The apparent success of the top-two primary in 2008 and 2010 led the U.S.
District Court to grant summary judgment in favor of the State. In 2012, the
U.S. Ninth Circuit Court additionally upheld the top-two primary. Thus, the
relative permanence of the top-two primary seemed to have been
established.
Candidates under the top-two primary are free to identify a party
preference, such as prefer Republican Party, but they are not required to do
so. Further, they may also identify a party preference for a political party that
37
may or may not formally exist—or identify no party preference at all. An
example from one county’s published Sample
Ballot is found in Appendix A (Sample Ballot, Primary Election, August 4,
2020, Stevens County, Washington, n.d.). Note the variety of party
identifications, particularly in the governor’s race.
Pondering that chronology, Spitzer’s comments seem quite apropos:
In Washington State, there is an observable interplay between the First
Amendment’s protection for freedom of association, the late
nineteenth-century populist constitutional ban on public assistance to
private entities, and the twentieth-century progressive goal of forcing
private political parties to open their processes to the voting public.
(Spitzer, 2019, p. 825)
This succinctly describes the context in which the top-two primary sits.
In many regards, Washington and California have similar histories—at
least, both states have arrived at the top-two primary, on a fairly similar
implementation date (2008 for Washington, 2012 for California). Both states
had a blanket primary just prior to adopting the top-two primary. However,
California’s blanket primary came online in 1996 (Spitzer, 2019), whereas
Washington’s had existed since 1935. Additionally, after losing a
38
constitutional challenge to the blanket primary, California’s legislature
instituted the top-two primary as a bipartisan deal to secure a key vote on a
budget (Caen, 2015), whereas Washington’s direct democracy process
wrested the top-two from the legislature’s hands through an initiative to the
people. Although these differences are not the focus of this proposed study,
they do suggest a lens through which to view perhaps subtle and not-so-
subtle differences between these states.
Research Specific to the Top-two Primary
As stated above, the top-two primary is a relatively new electoral
innovation. The research specific to the top-two primary is bounded to just
the last several years.
Therefore, the literature in this section is presented chronologically.
Legal court decisions and scientific inquiry are different and do not
necessarily come to the same conclusions. Though the courts were not
convinced that the top-two primary resulted in voter confusion, at least one
researcher concluded that confusion is very much in evidence. Manweller
(2011) approached the research problem from the perspective of voter
confusion. Recall that a central objection to the blanket primary (the historic
version, as well as the newer nonpartisan version) was the claim that it
39
impeded the right of political parties to choose their membership and their
standard-bearers— specifically, their candidates. With this in mind,
Manweller measured how often voters mistook a candidate under the top-
two primary who declared a preference for a major political party as that
party’s official nominee. Using sample ballots which included just those
instructions that were printed on official 2010 top-two primary ballots,
Manweller found that “80%-90% of respondents expressed a belief that
candidates were affiliated or associated with a political party, despite clear
disclaimer language on the ballot informing voters no such relationship
exists” (p. 267). While these early conclusions are certainly noteworthy, it is
also worth noting that participants were limited to 183 university students
who were relatively new voters. Whether Manweller’s results can be
duplicated with more experienced voters or a larger subject group is worth
further investigation. Working in 2013, Beck and Henrickson (2013)
studied whether the shift to the top-two primary had an impact on the
number of primary candidates filing for office. Given that in the top-two
primary, any candidate can state any party preference, candidates
representing the official choice of a political party may find themselves
competing against outsiders who may not have the blessing of the party
40
organization to which they say they identify. They found that there seemed
to be a modest “incentive for the major political parties to limit the number
of primary candidates representing their party” (Beck and Henrickson, 2013,
p. 791) for the Democrats, and no statistically significant impact was found
for the Republicans. The research conducted by Beck and Henrickson is
particularly similar to the present study, insofar as it examined party
response to the top-two primary and the incentive or disincentive to advance
official nominees for the ballot.
Writing somewhat contemporaneously with Beck and Henrickson,
William Jackson (2012) wrote a legislative note specific to the California
nonpartisan blanket primary, which passed as Proposition 14 in that state in
2011. As stated above, the toptwo primary is best understood in its historic,
political, and legal context. Jackson’s (2012) work was not necessarily
research in nature but did present a number of predictions regarding the first
use of the top-two primary in 2012. His work is presented here as context,
even though it did not assert empirical conclusions. Jackson predicted that
the top-two primary would entrench incumbents, raise the costs to
administer elections as well as increase campaign spending, and would harm
minor political parties.
41
Writing in 2016, Hill and Kousser’s research demonstrated one of the
many nexuses between topics within this literature review. Specifically, they
looked at the toptwo primary and voter turnout, focusing on the unlikely
primary voter. Beginning with the assertion that primary voter turnout is a
small and arguably unrepresentative subset of the voting population, Hill and
Kousser (2016) acknowledged that proponents of the toptwo primary and
other primary election reforms do hope to change this dynamic. However,
they believed that another institutional hurdle exists for engaging unlikely
primary voters. Specifically, they hypothesized a lack of mobilization
activity on the part of candidates. Many candidates reach out (through
mailers, calls, etc.) in the primary election period only to those voters who
have routinely voted in primary elections in the past, ignoring voters who
only engage for the general election. Their research looked first at whether
the “institutional shift to a top-two primary alone leads to an immediate
boost in [primary] participation” (Hill & Kousser, 2016, p. 415) and then
secondarily whether Get Out the Vote (GOTV) efforts additionally mobilized
voters within California’s 2014 primary. They found that the institutional
change of switching to a top-two primary did not appear to have an
appreciable effect on primary voter turnout (despite proponents’ hopes), but
42
they did have results that suggested mobilization efforts of unlikely primary
voters in a top-two primary environment might be quite effective.
Patterson, Jr. (2020) also looked at the effect of the top-two primary
on voter turnout. Responding specifically to proponents' claims that the top-
two primary would increase voter turnout by injecting “competition and thus
participation” (Patterson, Jr., 2020, p. 2) into the election process, he found
the opposite. He identified a phenomenon he called voter roll-off, which
occurs when a voter votes in those races on their ballot that included partisan
competition, but then they abstain from voting in races that were singleparty
races. Another way to describe this is selective undervoting. If the voter did
not like their choices (because their party was not competing in that race),
they opted out at a rate of 4-10%.
Fisk (2020) similarly looked at undervoting relative to the top-two
primary. He focused on orphaned voters, those voters who are being asked
to vote in a race that lacks a candidate that shares the voter’s own party
affiliation. Looking at same party congressional races, Fisk (2020) found
that “orphaned voters exhibit higher levels of undervoting than non-
orphaned voters in same-party contests” (p. 302).
43
Given the uniqueness of the top-two primary to just a handful of U.S.
States, it was unusual to encounter research by researchers in other
countries, but according to
Amorós and colleagues (2016), primary election reform is of interest outside
of the United States. Their research, grounded in game theory, concluded
that like many U.S. advocates of the top-two primary assert, this form of
election resulted in more moderately aligned elected officials. In addition, it
increased the number of districts and/or states that can be considered swing,
and was a threat to the dominant political parties in so-called safe states.
In researching a very similar set of questions to those raised by
Amorós et al., regarding the impact of a top-two primary on the degree of
moderateness espoused by officials elected under that system, McGhee and
Shor (2017) wrote “the recent experiments with primary reform in
Washington and especially California have attracted international attention.
Yet, there has been little quantitative evidence of the effects” (p. 1063).
Recognizing that primary reform has been one of the most-used measures to
“bring the parties back together” (p. 1063) and combat polarization, McGhee
and Shor (2017) analyzed the top-two primary in Washington and California
in search of any effect. They found virtually no such moderating impact in
44
Washington and only a slight moderating effect on Democrats in California
(with no impact on Republicans in that
state).
Steven Sparks presented the next several studies, all looking at the
top-two primary through slightly different lenses. In his 2018 piece, Sparks
(2018) looked at the impact of campaign spending and converting those
dollars to votes in two-party and oneparty races. Focusing on state
legislative races since the adoption in Washington (2008) and California
(2012) of the top-two primary, and specifically in light of challenger versus
incumbent dynamics, he found that campaign dollars for challengers went
twice as far in converting those dollars to votes in one-party races as
compared to two-party races. Sparks framed this within the larger context of
voter information and the voter cues often offered by party labels. In one-
party races, party label is not helpful to heuristically guide the voter—
something else is needed. In state legislative and other down-ballot races,
free media coverage is scarce, so campaign expenditures, particularly in
direct communication with voters, are required. Patterson, Jr. (2020)
likewise referenced the difficulty in low information races when he stated
“while voters have the opportunity to select the ‘lesser of two evils,’ they
45
lack the ability to do so” (p. 6) if they do not have access to information
enough to make an informed choice when partisan cues are not informative.
In Sparks’s next study, he revisited the concept of candidate
moderation under the top-two primary (2019). Given that this has been a
talking point of proponents, it was not surprising to encounter the question in
a variety of studies. Sparks stated, “While prior research finds no evidence
that voters impose moderation in one-party contests, there is reason to
expect that new electoral conditions may encourage any number of changes
in candidate behavior” (2019, p. 567). One of those electoral conditions is
moderated rhetoric from the candidates. In this study, Sparks found that
candidates in single-party state legislative races in California did
demonstrate more moderated rhetoric on their campaign websites,
demonstrating “the potential for the top-two primary to soften the tone of
hyperpolarization” (Sparks, 2019, p. 581). Having elsewhere in his article
referred to state legislative races as low information races, the moderating
effect was likely only possible if campaign rhetoric information from
websites (likely the only source voters will encounter) actually reached those
voters.
46
Publishing two studies in 2020, Sparks again highlighted that
advocates of the top-two primary system “hoped that the reform would
increase electoral competition while mitigating polarization and legislative
gridlock” (Sparks, 2020b, p. 1). In the first study focusing on quality
challengers (as defined as challengers who have previously held elected
office), Sparks noted that quality challengers are more likely to emerge
against a same-party incumbent than against someone from the opposite
party. In his second study in 2020, Sparks focused on candidates'
campaign dynamics and behaviors (Sparks, 2020a). In this case, he
compared responses to various interview questions by ten state legislative
candidates in one-party general election races in Washington State, with ten
state legislative candidates in North Carolina who, lacking the top-two
primary, necessarily ran against a candidate from another party. All subjects
were 2016 candidates. He found that the risk and uncertainty present in
single-party races in Washington (an outcome of the top-two primary in a
safe district) produced very different campaign behaviors in comparison to
their North Carolina counterparts. Specifically, in North Carolina, candidates
in safe districts believed that the outcome of their general election races was
nearly guaranteed, and as a result, there was no electoral advantage to appeal
47
to a broader range of voters. By contrast, all ten of the Washington
candidates expressed uncertainty about their races during the campaign,
even if they ultimately won by 10-30 points. This uncertainty impacted
messaging, moderating rhetoric, and impacted voter outreach such as
doorbelling and direct mailings.
Henrickson and Johnson (2019) take up two questions regarding voter
turnout somewhat simultaneously, through the lens of cost of voting, in a
study with similarities to the work of other researchers already mentioned
herein (Hortala-Vallve & Esteve-
Volart, 2011). They looked at both the effect of Washington State’s decision
to become a fully vote-by-mail state and the effect of the top-two primary on
turnout. They did not appear to drill down to the granular level of
undervoting behavior, but only the binary question of whether the voter
voted at all. They did find that the top-two primary seemed to correlate with
increased voter turnout, however.
Summary and Conclusions
Under consideration in this chapter have been the historical, legal, and
political context that led to the development of the top-two primary in
Washington State. This state’s decidedly populist underpinnings can be seen
48
in this election innovation. Current scholarly literature has given the top-two
primary its due, especially recently, but has not yet provided information on
the county political party response to the top-two primary. This study filled
that gap and extended the academic knowledge surrounding the top-two
primary and offered important data to the public policy development
process. Additionally, the applicability of policy feedback theory to that
public policy development process was explored, and its appropriateness for
use with this study was confirmed. This study expanded the use of policy
feedback theory into the qualitative realm, further extending the knowledge
of the use of this theoretical approach.
In Chapter 3, the specific methodological details of this multiple case
study approach will be explored, including participant selection,
instrumentation, and the details of the five-participant pilot study, which
helped refine procedures for the final study. Final study participation, data
collection and data analysis plan will also be reviewed. Finally, the following
chapter will also address issues of trustworthiness, including credibility and
ethical considerations.
49
Chapter 3: Research Method
The purpose of this multiple case study was to gather the perspectives
of political party leadership, in order to explore and understand Washington
State county political party response to the top-two primary election system.
I filled a substantive gap in the literature, informed future public policy
development, and effected positive social change in the process. The
following section includes discussion of the specific research question,
concepts, research tradition, and methodology.
I engaged with Republican central committee leadership in five of
Washington State’s 39 counties. This population was bounded by having to
operate in a top-two primary state. For participant selection, I considered
eastern and western counties, large and small population counties, and urban
and rural counties. A directory of party leaders was made available from the
partner organization, the Washington State Republican Party. I maintained a
sensitivity to the ethical issues inherent with case study research, particularly
when achieving informed consent, respecting participant desire for
confidentiality, and acknowledging the potential impact of sharing my own
experience with the top-two primary.
50
I used recorded web conferencing interviews and participant-supplied
and publicly available documents in data collection. Recorded interactions
(in the case of web conference interviews) or primary-source documents
constituted the data. I used transcripts for manual coding methods and the
identification of predominant themes (see Creswell, 2013). Where
appropriate, document review (aka content analysis) was also included.
Research Design and Rationale
The central research question of this study was: What has been the
response by county-level political party leadership in Washington State to
the top-two primary? Expanding on that central question, I also evaluated the
foci of any response taken by party leadership. This included any stated
goals held by party leadership in their response, and the degree to which
party leadership believed their response is effective in achieving their goals.
Central concepts and phenomena of this study were reliant on the
responses from party leadership. As stated above, I focused on documenting,
summarizing, and evaluating those responses. I gave special attention to any
instances of candidate recruitment, candidate vetting, candidate
endorsement, or other voter communication.
51
I chose a qualitative approach for this study for a couple of reasons.
First, the relative newness and limited application in just four U.S. states of
the top-two primary suggested the open-ended exploratory approach inherent
in qualitative research. Secondly, there are few qualitative studies in which
policy feedback theory was used as the theoretical framework; expanding its
applicability to qualitative research seemed appropriate. Finally, the fact that
the political party response to this primary election format is a contemporary,
currently occurring process, happening in the natural setting of
Washington State party politics made a qualitative approach most
appropriate (see Creswell, 2009).
I chose the multiple case study approach because of the distinct and
somewhat siloed nature of county political party organizations. Each of
Washington’s counties has a statutorily defined central committee (Political
Parties, 1973) for each of the two major political parties, and each of those
central committees elects representatives (one state committeeman and one
state committeewoman) to represent the county party organization at the
state party level. Yet despite the grassroots-up representation and the degree
of communal thought that might arise from working together in support of a
statewide organization, there is no statutorily formalized means for the state
52
party organization to inform or enforce consistency of operations county-by-
county, except in very specific circumstances (convention rules or the
nomination of presidential candidates being some examples). As a result, for
this study, I approached each county party organization as its own
organization, distinct from other organizations, and treated individually as its
own case study.
I considered using a phenomenological approach. I rejected this
approach because I wanted a more pragmatic and practical understanding of
the top-two primary, beyond the shared experiences of individuals existing
within the top-two primary experience in Washington State (see Creswell,
2013). Other qualitative approaches were similarly rejected for their inability
to provide the type or depth of data and thereby understanding needed to
respond to the research questions at hand.
Role of the Researcher
Frankfort-Nachmias and Nachmias (2008) drew distinctions between
the complete participant and the participant-as-observer in their reflection on
field research within the qualitative approach. The former was defined as a
member of a group under observation, wherein the “observer role is wholly
concealed; the research objectives are unknown to the observed, and the
53
research attempts to become a member of the group under observation”
(Frankfort-Nachmias & Nachmias, 2008, p. 258). The latter, by contrast (the
most oft-used approach, according to these authors) is an approach in which
the group understands that the participant-observer has a research goal that is
explicitly identified. In this present study, I used the participant-as-observer
approach for several reasons.
First, as mentioned in Chapter 1, I have well-established Republican
party bona fides. I served as an executive board officer for my county
political central committee for more than 8 years (in secretary, vicechair,
chair, and at-large roles) and won election as a Republican to county political
office in 2018 as the county auditor, the de facto elections official for the
county. As a result, it was impossible to conceal these bona fides, making the
complete observer as described by Frankfort-Nachmias and Nachmias
(2008) impossible. Incidentally, this same fact influenced participant
selection, as discussed below.
Secondly, as the data collected might speak to party strategy and a
certain trust was required in order to extract that information, full
acknowledgment of a participantobserver status was important. Finally, the
research questions had both academic and implementation value to me.
54
Acknowledging the participant dynamic in the participantobserver model
was a way to guard against bias by explicitly acknowledging such potential.
Although the community of involved partisans is small, any
familiarity between participants in the main study and myself as the
researcher was deemed to be at the acquaintance level (at the closest) or
merely by reputation (in most cases). The participants and I have traveled in
the same political circles, but no currently established relationship existed
beyond that described here. In no case were any power differentials
anticipated. The siloed nature of both county parties (which reorganize with
new leadership every 2 years) and the isolation since my election to county
office proved to be a good insulation against any real or perceived difference
in power between myself and the participants.
No specific ethical issues were apparent with the participant-as-
observer model. Transparency about my background and goals at the
recruitment stage were explicitly addressed. This eliminated possible ethical
concerns with this approach.
Methodology
This multiple case study qualitative study had a bounded and discreet
population and a phenomenon with limited scope. Replication or
55
reapplication of this methodology to a similar population should be
straightforward. For instance, I recommend that Washington county
Democratic central committees be included in a future iteration of this
research. Similarly, applying the present methodology to one or more of the
other handful of states using a top-two primary could follow very similar
selection criteria and data gathering and analysis approaches.
Participant Selection Logic
The population for this study was the leadership of the 39 county
Republican central committees in Washington State. I defined leadership
primarily as the individuals elected within each central committee as either
chair or vicechair or their designee. I further defined leadership as anyone
holding the chair or vicechair position, or state
committeeman/committeewoman position, in the current organization
(spanning the 2023-2024 biennium) or any previous organization from 2011
to present, or their designees. Such designees included subcommittee chairs
pertaining to candidate recruitment, vetting, or endorsement, or committee
chairs involved in voter communication, campaign strategy, or GOTV
efforts. The population size was therefore estimated according to the
following calculations: 39 counties, multiplied by six organizational biennia,
56
multiplied by four individuals who may consent to participate in this study
per biennium, or approximately 400 individuals (if it is assumed that some
individuals fulfill the same or similar leadership role over multiple biennia).
Sampling for this study was purposeful (see Creswell, 2013) with the
following considerations in mind. First, it was acknowledged that insofar as
Republican central committee leadership was targeted for inclusion in this
study, both criterion and convenience sampling approaches were involved.
The Washington State Republican Party, with which I had some familiarity,
was able to provide contact information for current and historic county
central committee leadership.
Secondarily, a maximum variation approach within and between potential
participants was considered. After an initial outreach to all potential
participants, I chose the study sample to allow for representation from
eastern and western counties, large and small population counties, and
urban and rural counties to the maximum degree possible.
I selected five counties for final inclusion in the study, representing
feasibility, representativeness, and saturation. Such a sample size
(representing between 10 to 20% of counties) allowed a representation of a
broad range of political party response.
57
Procedures for Recruitment
The Washington State Republican Party was both able and willing to
assist in the identification of individuals who met the criteria of currently or
historically serving as either the elected statecommitteeman or
statecommitteewoman or as the county party chairperson or vice chair. Had
this method of developing a potential pool of participants proved
insufficient, the roster of elected precinct committee officers (the same
individuals who constitute the county central committee membership,
according to state statute) is a matter of public record. Public record requests
directed to each of the 39 elections officials could have filled in most if not
all gaps left in the initial approach.
Fortunately, that was not necessary.
I identified participants by the above-described process. Participants
were initially contacted by email. Some follow up phone calls occurred, in
which I answered any questions about the study that participants had prior to
committing to participate. Recruitment included a commitment to share the
broad findings of this study with participants and their organizations.
58
Instrumentation
The anchor of instrumentation in this study was a semistructured
interview. Additionally, I also used document review, as such documents
were either made available by participant organizations, or readily available
on participant organization websites. I developed both the interview
protocols and the document analysis rubric, drawing from the literature on
qualitative interviewing and augmented by a pilot study.
Interviews
According to Rubin and Rubin (2005), good qualitative research
design includes anticipating the analysis, from the time of choosing a topic
through the development of the research question and beyond. For this study,
the analysis included identifying themes and describing both the current
response by county political parties to the top-two primary, but also eliciting
the stories of how the current response came to be, and any goals that the
participants might identify that are not yet met. As a result, the interview was
structured to capture what is, how it came to be, and possible future
developments.
Rubin and Rubin (2005), in their guidance on determining if
interviewing is an appropriate instrument for a study, listed a few relevant
59
questions that are considered below. These questions were helpful in
choosing an interview method. They also provided a guide regarding the
types of questions that might best serve the current study.
• Are you looking for nuance and subtlety? (p. 47).
• Does answering the research question require you to trace how
present situations resulted from prior events? (p. 47).
• Does puzzling out the research question necessitate layers of
discovering in which initial questions are asked to discover
alternatives that are then explored in turn? (p. 48).
A quick reaction to each of these questions follows. In this study, I looked
for the nuances, similarities, and differences in county political party
response to the top-two primary and the evolutionary process that has
occurred. I used open-ended questions and encouragement to participants to
tell stories, relay events, and offer opinions.
I used the theoretical lens of policy feedback theory, and in response
to the question above about the present situation resulting from prior events,
I used a conversational and organic direction for the interviews. I asked
follow-up and clarifying questions as they came up, and the participants
were encouraged to give opinions and express hopes for the future. Of
60
course, at all times, such follow-up was done to ensure fidelity to the
approval granted by the Walden University Institutional Review Board
(IRB).
The last question posed by Rubin and Rubin (2005), addressing layers
of discovery, influenced the emphasis on open-ended questions. The original
nature of this research required that I had the ability to ask follow-up and
clarifying questions. Doing so encouraged more extended narrative
reflections from the participants.
The tone and expectations for the interviews were addressed at the
outset of each interview. A scripted introductory paragraph reminded the
participant of the purpose of the interview and encouraged the participant to
feel free to go beyond the structured questions with additional thoughts,
reflections, or relevant stories or events. Additionally, this introductory script
reassured the participant of the confidential nature of the conversation and
the recorded interview status and ensured consent. Appendix B includes the
introductory script and interview questions.
61
Content Analysis
In addition to the interview, participants were asked to share any
documents relevant to the topic. Websites for participant organizations were
also searched for relevant documents. Documents such as organizational
bylaws, policies or procedures, platforms, candidate questionnaires, etc.
were included in this category. Coding for themes found in the documents
followed those that emerged from the interviews.
Procedures for Pilot Study
The pilot study was very simple; the intent was that the pilot study
would draw on documents already in my possession and interviews with
participants already known to me. Specifically, historic vetting committee
documents and procedures, as well as interviews with current and past
members of the Stevens County Republican Central Committee, were
approached and included in this pilot study. Five individuals who have
historically participated in that organization’s candidate vetting and
endorsement process were included in the pilot study. This is the local
organization that I had served as a member for 8 years. The purpose of the
pilot study was to serve as a trial for the interview questions, as well as a
rehearsal of interview techniques with a comfortable pool of subjects. This
62
was particularly helpful in refining my delivery and approach and
establishing content validity. Each interview lasted between 55 minutes and
1 hour 25 minutes. Due to the familiarity between myself and pilot
participants, these interviews were slightly longer than what would be
expected of a participant in the main study, mostly due to informal
conversation between acquaintances.
The pilot study almost immediately revealed a need for an update to
the introductory script and interview questions. I had assumed that each pilot
study participant would have the same depth of knowledge on the top-two
primary, its history, and its implications as I. It was determined that this was
not the case. It was evident that the introductory script required additional
context and definition, so the participants and I could operate from a shared
vocabulary. As a result, text taken from the Abstract was integrated into an
introductory statement that I read to the participant. This established a
context, shared terminology, and punctuated the start of the interview with a
degree of formality that proved beneficial when interacting with participants
in the main study.
Transcripts of the five pilot interviews were used for practice, to begin
using the coding tools in the Delve software (www.delvetool.com) selected
63
for data analysis. This practice centered on gaining facility with the Delve
tool. However, pilot interviews were not used in the data analysis in Chapter
4.
Procedures for Participation and Data Collection
Interviews were conducted using a paid version of the web conference
software
Zoom that has recording capabilities (Video Conferencing, Web
Conferencing, Webinars, Screen Sharing, n.d.), associated with a login
personally owned by me, a safeguard for participant privacy. Additionally, a
Zoom-compatible transcription app was utilized for live transcription and
backup audio recording. Participants were informed that the initial interview
would last approximately one hour. I was the sole individual involved in
both data collection and analysis.
Upon completion of the interview phase of the study, participants
were explicitly informed that their participation had ended, but they would
be given the option to stay up to date on the ongoing research efforts.
Participants were invited to contact me again with questions. No ongoing
follow-up interviews were required, but participant interest in the ongoing
project was honored.
64
Regarding the content analysis portion of the study, relevant
documents were requested from participants/participant organizations in
electronic format. Email receipt was to my university email address. If
documents were available online, I retrieved those directly.
Data Analysis Plan
By way of reminder, the research question was “What has been the
response by county-level political party leadership in Washington State to
the top-two primary?” Insofar as the question is broad by design, almost any
possible verbalized response was connected to the question, and no
discrepant cases were noted. It was originally anticipated that three main
categories of response might appear, though it was my intent to follow the
data, even if it led to fewer or more categories. This is consistent with the
advice given by Miles, et al. who wrote that they “strongly advise analysis
concurrent with data collection” (2014 p. 70).
The first anticipated response category could be described as a
response reflected in the values and beliefs of the party organization, absent
specific associated action. This was expected to take the form of participant
opining on whether they prefer the top-two primary to other election forms.
Alternately, the values and beliefs might be found reflected in organization
65
bylaws or platform documents without associated action items. The
second anticipated response category could be described as action taken that
reflects a response, where the audience for those actions is limited to party
members and candidates. This is considered an internally focused action. An
example of this would include candidate vetting interviews for the benefit
and edification of the party. The third anticipated response category
could be described as action taken that had as its intention a focus outside
the organization. Public candidate endorsement and mass communication
efforts with the voting public on behalf of the party organization would be an
example of such anticipated responses in this third category. Action taken in
the form of direct or indirect candidate monetary support would also be in
this category. The interview and document data therefore connected to
the research question by allowing for a broad range of types of responses.
These response categories represented a schema for the initial
conceptualization of the coding. However, consistent with Miles and
colleagues’ assertion, “coding is analysis” (Miles et al., 2014, p. 72) these
anticipated responses categories were just the start; the coding itself took the
research in its own direction.
66
Because of the recorded and transcribed nature of the data, it was less
necessary for me to take extensive written notes while the data collection
was underway. This benefited the quality of the interaction and encouraged
depth and free response. In keeping with the analysis presented by Miles and
colleagues (2014), more than one coding cycle occurred.
Because minimal notes were taken at the time of the interview, a
preliminary precoding effort included simply exposing myself to a playback
of each of the interviews. This took place relatively quickly after the
conclusion of each interview so that broad impressions were not lost. Some
notes were taken at this stage, but mainly as an informal test to determine if
the categories described above held true in any general sense. Following this
review, coding moved into what Miles et al. refer to as “first cycle coding”
(2014, p. 74). This included (but was not limited to) the following types of
coding: descriptive coding, in vivo coding (particularly appropriate because,
as expected, the specialized organization of the political party demonstrated
having their own phrasing and vernacular for certain phenomena), and
values coding.
A transitional phase ensued, in which I coded the codes. Following
this transition, the first round coding efforts were reorganized into patterns.
67
Stated another way, “pattern coding…a way of grouping…summaries into a
smaller number of categories, themes or constructs” (Miles et al., 2014, p.
86). This was most illuminating.
Delve software was utilized to assist with hand coding for each
interview. This tool was more than adequate due to the relatively modest
number of participants. The interface was intuitive, and the exports easily
used for manual grouping/regrouping of codes.
Issues of Trustworthiness
According to Creswell (2013), the nuances of qualitative research
require a new take on the concepts of reliability and validity. Various
scholars have attempted to adapt those quantitative concepts to the
qualitative environment. Creswell notes that Lincoln and Gupa (1985) use
the words credibility, transferability, dependability, and confirmability; those
terms are adopted and addressed here.
Credibility
Understood as the qualitative expression of internal validity
(“Credibility,” 2010; Miles et al., 2014), credibility was achieved by
persuasively and compellingly telling the story of the participants. As
expected, much of the data gathered and analyzed was descriptive and
68
evaluative, and, to a lesser degree, interpretive or theoretical, though the
results may suggest additional directions for further research. Checking my
understanding with participants by allowing the interviews to be free-
flowing and allowing for follow-up questions and reflection aided in this
manner. Opting to capture to a high degree of detail or providing thick
description, confirmed that credibility.
Transferability
Each of Washington’s 39 counties has statutorily defined Republican and
Democratic county central committees, with largely similar leadership
structures. Although this study focused only on Republican organizations, a
ready structure for transferring this research to a new sample already exists.
As is known from the legal history of the top-two primary, both Republican
and Democratic organizations initially opposed the policy in court.
Additionally, transferability was supported by the inclusion of participants
from eastern and western counties, urban and rural, large and small.
Dependability
Dependability, which can be understood as whether the “process of
the study is consistent, and reasonably stable over time” (Miles et al., 2014,
p. 312), was upheld by the use of an introductory script used with each
69
participant. Additionally, inclusion of not just the semistructured interviews
but also documents supported this goal. Extensive time spent in hand coding
with the Delve tool added to dependability.
Confirmability
Miles and colleagues (2014) state, regarding confirmability, “the basic
issue here can be framed as one of relative neutrality and reasonable freedom
from unacknowledged researcher biases” (p. 311). Being reflexive, focusing
on being self-aware and transparent with participants during all phases of the
research proved to be central requirements of confirmability. Retention of
recordings and transcripts, as well as coding results further support this
study’s confirmability.
Ethical Procedures
I followed all university steps outlined in the research ethics review
process by the IRB (Harris, n.d.). This study was assigned IRB approval
number 03-04-22-024673. Participants were asked to electronically sign the
IRB approved consent form, prior to scheduling their interview.
Accessing Participants and Data
As mentioned above, a variety of approaches were contemplated to
identify possible participants. Participants were first sought through the
70
partner organization, the Washington State Republican Party (WSRP).
Formal approval for this partner organization relationship was granted by the
Chairman. Following a briefing on the proposed research, delivered to the
Chairman, the organization voluntarily shared member rosters. Partnership
with the WSRP and its membership was sufficient to produce the needed
participants for this study. Whereas utilizing public records requests was
initially considered, doing so was deemed unnecessary.
Regarding document review, when requesting data in the form of
documents, participants were encouraged to gain the permission of the
organization’s responsible party (chair) before sending documents for
evaluation. An exception to the need for this explicit permission was
assumed when relevant documents were publicly available online. In this
case, permission was deemed to be unnecessary.
Participants were briefed on the nature and purpose of the study at the
time of invitation, at the time an interview appointment was set, and as a part
of the introductory script that was used. Written informed consent was
checked verbally at the start of the interview. Participants wishing to
withdraw from the study would have been immediately allowed to do so,
though no one requested to do so.
71
While the participants were known and identifiable to me, participants
were assured recorded web conference events would remain protected from
release. It was anticipated that some of the data offered by participants
would be considered sensitive. Every effort was made to protect the privacy
of that data and the participants who offered it. Personal logins for accessing
recordings, transcripts, and coding software aided in securing that privacy,
and efforts to mask the origin of documents included in content analysis
were employed.
Summary
In this chapter, specific consideration was given to the qualitative
methodology and the multiple case study design. The beneficial contribution
made as a result of the pilot study process was reviewed. Additionally,
specifics to the final study, including participant selection, recruitment, and
participation were addressed. Finally, concepts central to trustworthiness and
ethics in qualitative research were contemplated. Data collection and
analysis form the basis of Chapter 4. An exploration of the findings and
conclusions are found in Chapter 5.
72
Chapter 4: Results
The purpose of this qualitative study was to explore the political party
response to the top-two primary election system in Washington State. The
top-two primary occurs within a system that allows states to behave as
laboratories of public policy innovation. Constitutionally protected rights—
including the right of association for groups such as political parties, and the
right of the electorate to a free, full, and unimpeded franchise— were
significant to this study. The overarching research question framing this
study was:
What has been the response by county-level political party leadership in
Washington
State to the top-two primary?
The population for this study was current or recent leadership within
the Republican county central committees in Washington State counties.
Participants included five individuals with history and knowledge of their
individual counties’ response to the top-two primary. This chapter includes
discussion of the pilot study I conducted, the research setting, and participant
demographics. This chapter also includes a description and summary of the
data collection and analysis, including first- and second-round coding
73
efforts, a review of issues pertaining to trustworthiness and a final summary
of results.
Research Setting
In some cases, several months elapsed between my initial invitations
to participants and my final study, due to personal impediments. The
majority of participants who initially agreed to participate completed the
interview, so the length of time between invitation and interview did not
negatively impact the study.
Demographics
Study participants were drawn from a roster of past and current
members of the WSRP. Membership in the WSRP is determined by its
bylaws (WSRP By-Laws, n.d.), which allow for the elected party chair, and
the elected statecommitteeman and statecommitteewoman from each of the
39 counties in Washington State. The membership in the WSRP is renewed
on a 2-year cycle of reorganization, outlined in statute (Political Parties,
1973). As a result of leadership stability and/or changes at the county party
level, the longevity of participation or involvement of study participants in
their county parties varied. Some had been involved for only a few years,
whereas others had been active in party membership and/or leadership for 30
74
years or more. Invitation to participate was sent to the membership rosters
for the previous two organization cycles.
Lovrich, Jr., et al. (2018), in their writings on political culture in
Washington State, grouped Washington voters into two main categories:
Puget Sound Policy Cluster and the Other Washington Policy Cluster. These
categories were identified in studying the election outcomes in four
statewide contests in 2010. Most voters in Washington would find those
categories intuitive. Two participants in this study were from the Puget
Sound Cluster counties, and the remaining three participants from the Other
Washington
Policy Cluster. These categories could be simplified along an urban vs. rural
spectrum. However, there is also a sentiment among Washington voters that
there are politicalcultural differences east vs. west with the Cascade
Mountain range serving as the dividing line. The demographics of
participants in this study included three from eastern counties and two from
western counties.
Registered voter counts were drawn from the Washington Secretary of
State website (November 7, 2023 General Election - Voter Turnout, n.d.).
The county voter population for the participants varied from the smallest
75
county represented at approximately 23,000 voters, to the state’s largest
county at nearly 1.4 million. Three smaller (100,000 or fewer registered
voters), one medium-sized county (360,000 registered voters), and the state’s
largest county were all represented among the
participants. Finally, participants included two men and three women.
Data Collection
I collected data for this study primarily through semistructured
interviews, which I first refined through a pilot study. Participant interviews
occurred much as expected during the initial design phases of this study.
This study included five participantinterviewees. After receiving and signing
the consent form, participants were invited to self-select a convenient time
for their interviews using the online software Calendly. This cut down on the
back-and-forth emails required to schedule each interview successfully. I
spent between 60 and 75 minutes with each participant in a Zoom
teleconferencing environment. Video was optional, at the participant's
choice, a decision made presumably due to bandwidth issues, particularly in
rural areas. In addition, each Zoom environment included an auto-
transcription tool known as Otter.ai, that captured both an audio and a
transcribed representation of the interview. The ability to easily click back to
76
the actual audio, directly from the Otter.ai transcription during coding and
analysis, saved considerable time in transcript verification. Finally, the Zoom
environment was also set to record (both audio and, where applicable, video)
as a failsafe for capturing the raw data. Each participant was asked for their
approval to record prior to starting the Zoom recording.
A peripheral secondary data source was document review. In addition
to completing their interviews, two participants shared documents related to
their county party’s response to the top-two primary. Additionally, if there
were any relevant documents (platforms, bylaws, endorsement/vetting
questionnaires, etc.) available on the county party websites with which study
participants were affiliated, those were included in the document review and
content analysis. In the interest of protecting the anonymity of participants,
source documents are not explicitly cited herein. In lieu of full citation, see
Table 4 below under Data Analysis that shows a selection of certain codes
and the frequency with which those codes were applied. Counties are
masked by letters A through E. For counties for which more than one
document was available for analysis, that county’s letter is repeated.
77
Data Analysis
Data analysis started with a thorough review of coding options; the
primary resource for the scope and variety of coding approaches was drawn
from the work of Saldaña (2016). In his text, several categories of coding
were considered. I ultimately selected six methods from three categories as
most appropriate for this study. Precise methods were selected based on the
goodness-of-fit to the research question, content of raw data, and an interest
in capturing not just relevant concepts, but also the voice of participants.
Table 1 represents the coding categories and methods that ultimately
informed the data analysis during the first-round coding.
In Chapter 3, three response categories were anticipated. Although
these categories did emerge, they were not the most relevant or interesting
concepts once initial coding efforts got underway. Neither was a set list of
codes developed in advance of analysis.
Table 1
First Cycle Coding Methods Used in Current Analysis
Grammatical Methods
Elemental Methods
Affective
Methods
Attribute Coding (5)
Concept Coding
(79) In Vivo
Coding (37)
Emotion Coding
(13)
Values Coding
(12)
78
Versus Coding
(15)
Note. Adapted from Saldaña (2016), p. 68.
Taking the advice of Saldaña (2016) as well as the recommendations
of Miles, et al., (2014), a code list and some early analysis developed
concurrently, during the first pass through the interview transcripts. At the
conclusion of the first pass through the transcripts, a coherence around the
methods included in Table 1 emerged. A second review of each transcript
further confirmed these categories and methods, and some subcategories
arose as well. For instance, I applied the elemental method of concept coding
to several discrete concepts such as campaign plan benchmark, grassroots
benchmark, and money benchmark, all of which logically grouped under a
concept code labeled endorsement criteria.
I used the Delve tool to manually highlight text within a transcript and
apply an extant or new code label to that highlighted snippet of text. Upon
export, I was able to identify how many individual times each code was
applied to a snippet of text. The range was from one snippet to nine snippets
across the five interview transcripts. The frequency of snippet coding was
informative and somewhat indicative of themes common across participants,
79
but it was not sufficient to draw out the level of meaning that I sought. This
was better achieved in the second round of coding.
First round coding is known for the proliferation of a multitude of
codes. To address this proliferation, each individual code was printed on a
separate adhesive mailing label, then each label was affixed to an individual
post-it note. In this initial attempt at organization, the broad categories of
method (grammatical, elemental, affective) were used and spatially grouped
together. Saldaña used the term “lumping” (2016, p. 79) for this kind of
activity.
From this activity, snippets and their associated codes fell into two
groups: those that lumped together logically and intuitively (which will be
discussed more below), and those that did not. Those that did not were
categorized as singularly unique occurrences, or one-offs. Although these
one-off codes and their underlying concepts have been eliminated from
inclusion in the body of this present study, they may indicate a focus or lens
for a future study.
Following the first round coding and a transitional step to stem the
code proliferation, I completed a second round of coding. In the second
80
round, I focused on process coding. At this second stage, several relevant
themes in answer to the research questions emerged.
First Round Coding
In the following section, a portion of the codes that emerged in the
first-round coding—specifically the affective method coding outcomes—
merited review and appreciation. These initial coding efforts did not tell deep
stories of meaning, belief, or strategy regarding the top-two primary.
However, these affective codes did set a tone and tenor that I used to
understand the concepts illuminated during the second-round coding efforts.
Grammatical Methods – Attribute Coding
Codes assigned in this category were predictably dry and focused on
the length of involvement in party politics (ranging from just a few years to
several decades), as well as how and why the participants got involved in
their county political party. A bit of cross-over with other coding forms was
noted, such as emotion coding or in vivo coding (feeling a connection to and
supporting a particular candidate or campaign, and wanting to make a
difference), and values coding (supporting an issue). One example of this is
that one respondent got involved as a result of running for office themselves.
81
Affective Method – Emotion Coding
The most striking aspect of using emotion coding under the affective
method category was that all but one of the emotion words, terms, or phrases
could reasonably be interpreted as negative. Thirteen unique emotion codes
were assigned, with the most frequent code summarized as “the party
doesn’t like the top-two, but we are finding a way to deal with it,” with
seven instances (obviously denoting a repeated comment by at least one
participant). This one code is interpreted as just to the negative side of
neutral.
The one positive code assigned referred to a feeling of connection with a
candidate. The remaining codes are arranged in Figure 1 along a continuum
that is increasingly negative.
Figure 1
Affective Method Emotions codes, Positive/Neutral to Increasingly Negative
82
The emotion coding information set the mood for the remainder of this
section. The participants in this study seemed to associate a fair amount of
struggle, strife, and conflict with their party responses to the top-two primary.
This was seen not just in emotion coding but also in another affective coding
method, versus coding.
Affective Method – Versus Coding
If a majority of emotion codes trended towards negative emotion, then
the following (Figure 2) representation of the tension built into and between
political party responses to the top-two primary was not a surprise. There
was an ongoing tug-of-war along multiple dimensions. According to
83
Saldaña, versus codes “identify in dichotomous or binary terms
the…phenomena,…concepts…in direct conflict with each other” (2016,
p. 137). Figure 2 is a compilation of such dichotomous ideas expressed by
participants and captured when this form of affective method coding was
employed.
Figure 2
Word Cloud of Affective Method – Versus Coding Outcomes
The final aspect of first-round coding that merited attention was the
values coding outcome. Values coding was particularly rich, possibly
because the topic of political engagement is itself values laden. People are
driven to political participation, with their values at heart.
A value of fairness emerged. This included the idea that any person
has a right to run for office. Fairness was also evident in the regret or
Affective Method
–
Values Coding
84
criticism expressed by some participants that they believe the top-two
primary leaves the other party or independents out. All this, while at the
same time they are motivated and work to have Republicans elected.
A value of duty emerged. This included the idea that Republican party
representatives (namely precinct committee officers and party leadership)
have both a duty and desire to see Republican candidates elected. This duty
included ensuring that a candidate seeking office is a good enough or real
Republican.
The related values of credibility and pragmatism also emerged.
Participants emphasized the significance of maintaining the trust of the
voting public by minimizing the visibility of intra-party fighting.
Simultaneously, participants emphasized supporting candidates that can both
win and govern effectively.
Finally, and somewhat surprisingly, some conflicting values emerged
on the concept of party identification. Among participants, the idea of party
loyalty was generally framed as positive. However, so was having fidelity to
one’s individual values if that conflicted with party loyalty. In fact, one
participant who has served as the chair of the county Republican party stated
that they really identified as an independent, rather than as a Republican.
85
These affective methods uniquely give a sense of the strains and often
negative feelings that constitute the operational milieu of the Republican
party leadership in a top-two primary environment.
Transitional Phase
In first round coding, the elemental methods of in vivo coding and
concept coding resulted in the greatest number of individual codes (37 and
79, respectively). Drawing on the guidance of Saldaña, the transitional
approach of “coding the codes” (2016, p. 229) was employed to reduce
complexity. Coding the codes assisted in drawing out further and deeper
meaning as well.
For the in vivo labels, coding of the codes resulted in a couple of
categories. The first was just a capturing of terms and their explicit or
implied definitions. In this way, this subset of in vivo codes can be
understood as establishing some of the jargon specific to these participants,
and possibly to a county Republican party vernacular in general. A portion of
the in vivo codes were analyzed with a “part of speech” viewpoint in mind.
Table 2 below represents the recoding of a portion of the in vivo codes to
derive those nuances. It is noteworthy that the terms in this table are by no
means neutral. For that reason, presenting these results in this way is
86
intended to illustrate the good/bad or positive/negative tensions, even among
members of the Republican community, even in the language they use, and
the definitions buried within the terms they employ.
A second category that emerged when recoding in vivo codes is the
category of phenomena experienced by participants. For example,
Washington State is a fully voteby-mail state. In the context of questions
about how participants felt about the top-two primary, as well as other
possible electoral innovations that participants are worried about such as
ranked choice voting, in vivo coding revealed comments such as “I miss
polling places,” and “Vote by mail changes how a candidate has to engage
with the community.” Although these comments and the codes that were
applied to them are not specifically related to the top-two primary, the
presence of these comments is an indicator of how these two issues—vote-
by-mail and the top-two primary—are closely related in participants’ minds.
Table 2
Recoding certain In Vivo Codes using Part of Speech/definition
In Vivo Term
Part of
Speech
Definition (implied by participant/interpreted by researcher)
Patriot
noun
Adherence to the Constitution, emphasizing individual
liberties and specific rights contained therein
Conservative
noun
Possibly the same meaning as Republican but possibly
implying more right-leaning than the average Republican
87
Christian
Conservative
noun
A category within Conservative that draws definition from
religion in addition to the Constitution
Moderates
noun
Possibly the same meaning as centrist Republican, implying
more left leaning than the average Republican
Mainstreamers
Establishment
noun
Moderate Republicans, weak on principles, particularly in
comparison to the terms Conservative or Patriot
noun or
adjective
Similar to Mainstreamers, but more pragmatic, moderate in
order to win, long in control of the party momentum
Blue
adjective
Referring to anyone or any group on the political left
Jungle Primary
Popularity
Contest
noun
Synonymous with top-two primary; derogatory connotation
noun
Negative connotation: used to draw a contrast to candidates
that are more principled, more qualified, more strategic, or
better funded but who do not win
Bias
noun
Negative connotation: used to draw a contrast to the value
of fairness, also related to the concept of credibility
Fly in the
Ointment
noun
Similar to spoiler candidate and vote splitting codes,
denotes a same-party candidate meant to disrupt, possibly
with a focus on disrupting mainstreamers or establishment
Benchmark
noun
Minimum threshold(s) necessary for formal endorsement.
Might apply to campaign finance resources, grassroots
support, ground game plan adequacy, etc.
Additional recoded in vivo codes that are best understood under the
heading of phenomena include (but are not limited to) the following:
mudslinging, our brand is toxic, narrowing the field, mainstreamers have
money, fracture in county party structure, individuals are always partisan,
politics is full of egos, eating our own, and the list goes on. The variety
presented by these phenomena in vivo codes did not paint a fully clear
picture at this transitional state. However, the relative intensity of some of
88
the codes started to coalesce around certain processes and patterns once
second round coding was concluded.
Second Round Coding
First round coding illuminated both a wide range of expressed
emotions among participants, and a fair amount of conflict (as evidenced by
the results of versus coding techniques). Utilizing the transitional method
recommended by Saldaña (2016) of coding the codes, particularly the in
vivo codes, a better sense of the emotional response to the top-two primary
among participants is identified. For the purposes of this study, the emotional
response to the top-two primary was certainly of interest, but emotional
response alone is incomplete. In order for the results of this study to truly
inform a view of the top-two phenomenon from a policy feedback theory
perspective, more was required. It was essential to consider how these
emotions and apparent conflicts influence action.
Process Coding
Process coding, also known as action coding according to Saldaña
(2016), relies on a strict grammatical rule: codes must take the form of
gerunds, words ending in -ing. Process coding was primarily applied to
89
concept codes from first round coding. Particularly relevant in vivo, values
or versus coded elements were also included in this round of coding.
The following process codes emerged: vetting, endorsing,
benchmarking, supporting, recruiting, blocking, enforcing, facilitating, and
communicating. Table 3 is meant to give the reader a quick sense of the
nuances that emerged in these process codes. Of note are where these codes
overlapped and where they diverged.
It is noteworthy that not every process code was present in each
party’s response to the top-two primary. It is also noteworthy that the
weightiness of a process code within a county varied. It is this phenomenon,
specifically, that gave rise to important themes discussed below.
After second round coding had been completed on participant
interview transcripts, attention was turned to available documents and
content analysis. By this point, a robust system of coding was well
developed. Documents were reviewed and the content was coded with the
existing concept and process codes very much in mind. As mentioned
above, county documents have been masked to protect the anonymity of the
participants. Results of the content analysis on documents are shown below
in the Results section under themes of action.
90
Table 3
Process Codes, Definitions and Related Codes from First Round Coding
Process Code
Supporting Concept (First
Round Coding)
Codes/Definitions
Select Related Codes
(quotations indicate in vivo
codes)
Vetting
The process of reviewing a
candidate's adherence to
Republican ideals or platform,
qualifications, preparation, etc.
May or may not include
assessment of electability.
Real Republican, RINO
Endorsing
A formal statement of
approval or support
Benchmarking
An assessment of the
preparation for a campaign to
be competitive and/or
successful.
Money Benchmark,
Campaign
Plan Benchmark, Grassroots
Benchmark
Supporting
May include access to party
resources such as likely-voter
databases, volunteer time
(doorbelling, calling, parade
entry), monetary support,
inclusion in publications, etc.
For counties engaged in
supporting vs. facilitating,
endorsement may be a
prerequisite to receiving that
support.
Recruiting
Identification of possible
candidates and
encouragement to run
Filing week readiness
Blocking
Entry of a spoiler candidate,
meant to disrupt the success of
another candidate or party
interest; May also include
Party officials dissuading
certain candidates from
running or asking them to
withdraw.
Fly in the Ointment, Iced Out
91
Enforcing
Adherence to formal rules or
informal expectations. This
may include a party's bylaws,
vetting or endorsement rules,
etc.
Facilitating
A lighter version of endorsing
BBQ in Park for voters to
meet candidates
Communicating
Expressing the party's pick to
the public. May include
partyfinanced campaign
materials, press releases,
letters to the editor, etc.
Study Results
This study centered around the question “What has been the response
by countylevel political party leadership in Washington State to the top-two
primary?” Creswell describes the relevance of themes within qualitative
research in this way: “Themes are broad units of information that consist of
several codes aggregated to form a common idea” (2013, p. 186). This
qualitative study is no different. Within this discussion of study results,
several themes are identified, but a distinction will be drawn between
emotional or attitudinal themes (themes of sentiment) that emerged, and
externally focused or action-oriented themes (themes of action). There are
three themes of sentiment, and three of action.
92
Themes of Sentiment
Dislike but Acquiescence
An early unanticipated theme was that while there was seemingly
universal dislike for the top-two primary, no participant gave any indication
that there was sufficient momentum or sentiment to seek a legislative change
to the current system. This acquiescence seemed in conflict with a plank
included in one participant party organization platform explicitly calling for
a state primary system that would allow for partisan selection of candidates
(note continued masking for the sake of participant anonymity). All
participants reflected that their party organization disliked the system but
was finding a way to work within or around it. Tangentially, when the
interview conversations occasionally drifted towards other emergent
electoral innovations such as ranked choice voting (RCV) or the new top-
four primary/RCV general election hybrid in Alaska (Alaska Division of
Elections, n.d.), participants expressed a suspicion and a preference to
maintain the status quo. Simply stated, the sentiment seemed to be that
participants do not like it, but sure would not like it to change.
93
Fairness, Despite the Goal to Win
A second unanticipated theme emerged when values coding was used.
One value expressed by participants, seemingly undermined by the top-two
primary, was that equal competitive access be available to candidates of both
major parties, but also third-party candidates, independents, or candidates
otherwise unaffiliated. I interpreted this to be in some conflict with the
purpose and goal of the party structure to elect Republicans; two documents,
from two different party organizations included in the content analysis
above, explicitly expressed this goal. Stated another way, this desire for an
open competitive field, underscored by a value of fairness, seemed at odds
with the goal of winning elections.
It’s Fraught
A final unexpected theme was the degree of tension, strain, and
conflict within and among Republican party organizations, at least in the
view of the participants. This was most clearly seen in the versus coding
efforts (see Figure 2, above), but also when considering certain in vivo codes
that emerged. Certainly, a clash between Republican party organizations (or
their preferred candidates) and their Democratic counterparts is so normal an
expectation as to be a simple, central, and ubiquitous feature of American
94
political life. What was surprising is the degree to which othering was
expressed towards self-identifying Republicans. Whether that was seen in
the in vivo coding (patriots, mainstreamers, and the establishment) or in the
emphasis on differentiating between RINO candidates (Republicans in name
only) from candidates who are real Republicans, the lack of shared sentiment
about how to feel and think about the top-two primary—and what to do
about it—was unexpected.
Themes of Action
Figure 3 below presents the process codes that emerged during data
analysis and situates each code on two axes. The vertical axis signifies the
valence of the behaviors associated with each process code, ranging from
positive to neutral to negative, with the distance from the neutral line
indicating the degree of positivity or negativity. The horizontal axis
represents a somewhat chronological flow, aligning with the sequence of
events that candidates may experience and ways in which parties may
respond. During the data analysis section above, it was noted that not every
process code was present in every party’s response to the top-two primary.
What emerged is that county political party response is not uniform, but
rather the degree to which different process codes are present or preeminent
95
in a particular party, said party can be described as predominantly expressing
one of three strongest roles or positions, vis a vis candidates: facilitators,
recruiters, and enforcers. These three roles are the themes of action in
addition to the previous themes of sentiment. Each of these is defined further
below.
Figure 3
Process Codes – Valence and Chronological Events
Two county organizations emerged as primarily displaying the
facilitator theme. Facilitators were most likely to express a desire for fairness
and a disapproval of the cooling effect on third party or Democratic
candidates under the top-two primary. Facilitators have the least intense
vetting procedures. Their endorsement questionnaire (if it exists at all) is
simple and straight forward, and it is up to the candidate to contact the party
chair in advance of any regular meeting, in order to be considered for
Facilitators
–
It’s All About Fairness
96
endorsement by the membership. The clearest expression of the facilitator
within this study was the comment,
In the summer, we have a barbecue after filing period. We invite
everyone that has filed…we invite everyone—we frequently have
some of the city council candidates that are more liberal, they come
and they address us and they’re honest about their viewpoints. And I
think it’s interesting, I give them a lot of credit for coming.
This participant went on to describe how they publish candidate forums
organized by other organizations on their website.
The participant from the second facilitator county identified
themselves more “of an independent or libertarian than a Republican.” This
resonates with the value of fairness and access to all candidates to the
primary ballot. This participant, a former candidate, had the following to say
about the top-two primary:
I just don’t see a lot of independents and libertarians and third-party
people on the ballot in Washington State. And I do say that some of
that has to do with the toptwo primary. The second thing I don’t like
about the top-two primary is that it kind of eliminates the political
parties from participating or from nominating their candidates.
97
These two concepts are in some conflict, but the desire for fairness
undergirds the comment.
This second facilitator participant did take a slightly stricter view of
endorsement, stating their party organization does not seek out or pursue
candidates but “would expect a candidate that wants our endorsement to
come and ask for it.” This statement expressed that everyone is welcome to
address the party and ask for endorsement, but it is up to the candidate to
step up and ask. That being said, there was no intermediary between the
candidate and the general membership, a distinction between facilitator
counties and enforcer counties. Facilitators may or may not express the
process code of enforcing, but focus their energies on endorsing,
communicating, and supporting.
Recruiters – Principles Are Important
Two participants in this study can be described as coming from
counties that represent the recruiter theme. Recruiters are party organizations
that proactively seek out candidates. Ideas expressed by counties that are
heavily recruiter in nature include looking to lower, nonpartisan office for
future candidates for partisan office, and as a result, they have an interest in
the outcome of nonpartisan races, and work in some ways to support certain
98
nonpartisan candidates. One participant from a recruiter party noted that
nonpartisan interest in party support—and party interest in nonpartisan
candidates— is significantly increased compared to 15 years ago. This
participant did not explicitly tie this shift to the top-two primary, but the
timelines align.
Recruiters look to appointed positions as important gateway positions
to identifying future candidates. They are well positioned to engage with
excess candidates in a timely manner, to encourage withdrawal from filing if
too many candidates filed. Additionally, they may intervene if excess
candidates will unfavorably split the Republican vote. Recruiters are
sensitive to the perception of outsiders and dislike it when intra-party
fracture becomes evident.
Recruiters emphasize voter education and principals. Compared to
facilitators, they seem to have a stronger candidate vetting process, and take
formal action to endorse. They focus on authenticity (being a real
Republican) and electability. Recruiter organizations are more likely to focus
on fractures in the party, identify factions, and lament when these intra-party
differences come to the attention of the public. Recruiters clearly express the
process code of recruiting, but also are much stronger in their procedures for
99
vetting, benchmarking, endorsing, and supporting candidates than their
facilitator counterparts.
Enforcers – Playing By the Book
Organizations that express the enforcer theme have the most stringent
approach to candidates. Enforcers apply the heaviest requirements and have
the highest expectations.
One participant in this study can be described as coming from an enforcer
county.
In enforcer counties, support from the party of any kind—access to
voter databases, financial, volunteer time, inclusion in communications,
etc.—is dependent on endorsement. Endorsement is only offered after a
rigorous vetting process, and is a vote of the whole membership, following
recommendation from a formal report from a candidate vetting committee.
Content analysis corroborated the verbal responses given by this participant.
In the case of the enforcer county identified in this study, candidate
vetting, endorsement and support rules are outlined in the bylaws. Every
candidate running as a Republican for partisan office is proactively
contacted by the party, immediately after candidate filing week. In order to
be considered for the vetting process, the candidate is required to submit a
resume, cover letter, and a two-part vetting questionnaire. Part 1 of the
100
questionnaire requires essay-type responses to questions about qualifications
(including education and experience), past involvement in civic or political
organizations, knowledge of the office sought including references to the
Constitution or state law, and an explicit question as to whether the candidate
will refrain from endorsing any candidate other than a Republican during the
current election.
Part 2 of the questionnaire prompts the respondent to answer yes or no
to each of the 92 separate planks on the party platform. On any plank on
which the candidate responds no or chooses not to answer, the candidate is
encouraged to explain themselves. The answer may form the basis of further
questioning at the vetting interview, described below.
If a candidate fulfills all the requirements above, they are eligible to
sit for an interview with the formally constituted vetting committee. The
vetting committee applies a scoring rubric and prepares a report of those
scores. The committee then presents those scores along with a
recommendation to the general membership for an endorsement vote.
The respondent from the enforcer county also expressed a desire to
“make sure we don’t have any bias.” They stated being perceived as “not
being fair…will ruin [our]
101
credibility.” The enforcer county wants the influence but is also mindful of
perceptions.
Enforcer counties may or may not express the process code of
recruiting. They do heavily express the process codes of enforcing, blocking,
and endorsing, depending on the candidate they are serious about supporting.
In the words of the respondent from the enforcer county, “if the party isn’t
willing to endorse someone prior to the primary, they’re really—by
default—giving up what they were elected [as a precinct committee officer]
to do.”
Closing Reflection
As quoted above, “Policy Feedback Theory is indispensable for
scholars trying to understand how policies, once developed, reshape politics”
(Mettler & SoRelle, 2014, p. 175). The policy of the top-two primary has
impacted politics, particularly at the county political party level. The results
of this study suggest that although the policy impacts on politics are clearly
in evidence, the nuanced nature of that impact differs somewhat county by
county.
These themes, particularly the themes of action—facilitator (it’s all
about fairness), recruiter (principals are important) and enforcer (playing by
the book) —are drawn from the totality of the interviews and content
102
analysis. It is important to note that individual codes and the themes of
action they point to are not silos in nature. Whether a county predominantly
expresses a facilitator versus an enforcer theme is not correlated to whether
or not they have an interest (on a candidate endorsement questionnaire, for
instance) in whether a candidate has a criminal history or past political
involvement. As can be seen in Table 4 below that participants in each
category inquire after that information. (Counties are masked in Table 4 as
A, B, C, D, E to protect the anonymity of participants). In the cases of
counties A, C and D, only one document was available for content analysis,
whereas counties B and E had an additional document included.
Table 4
Document Review/Content Analysis – Select Codes arranged by County
Theme
103
Credibility
As outlined in chapter 3, telling the story of participants with fidelity
was executed according to plan. This was achieved in part through an
emphasis on and preference for in vivo coding as one of the main coding
approaches. This choice was driven by this coding method’s ability to
generate rich and meaningful descriptions.
Evidence of Trustworthiness
104
Additionally, the free-flowing aspect of the semistructured interview that
allowed for follow up and deeper exploration of a concept was beneficial and
supported this study’s credibility.
Transferability
One emergent thread, as discussed above in the Results section, was
that even though all participant interviewees have served in leadership roles
in their county Republican party, there are elements along which participants
are closely aligned, while at the same time, there are some striking
differences. Transferability is achieved by the emergence of both consistent
themes county-to-county as well as identifying where strong local
individuality exists. This dichotomy is a ready lens through which to view
transferring this research to future research.
Dependability
Dependability was achieved by using a standard script for orienting
each participant at the start of the interview. That script was refined over the
course of the pilot study. As a result, final study participants experienced a
consistent introduction to the study. Using a consistent list of questions over
the arc of each interview also supported dependability.
105
Confirmability
My own history in political party participation and leadership required
intentional reflection on possible bias. Where appropriate, I engaged in
limited self-disclosure and mirroring in the context of the interviews. When
this occurred, it was to encourage additional reflection from the participant,
and to reassure the participant that they were heard and understood, based on
similar or decidedly different experiences.
Summary
The data collection and results analysis portion of this qualitative
study exploring political party response to the top-two primary election
system in Washington State, has resulted in a rich data set. Through a first
round of coding using grammatical (attribute coding), elemental (concept, in
vivo and process coding) and affective (emotion, values and versus coding)
coding methods, a deep and nuanced impression of the varied responses to
this primary election system emerged. Transitional coding and second round
coding resulted in useful process codes that, when applied to participant
counties and the documents available for content analysis, illuminated three
types of counties: facilitators, recruiters, and enforcers. The following
chapter will place the findings of this study in their place within the larger
106
body of scholarly literature on the top-two primary and elections in general.
It will explore the limitations of this study, recommendations for future
research, and finally explore the social change implications of this effort.
Chapter 5: Discussion, Conclusions, and Recommendations
The purpose of this multiple case study was to explore the
Washington State county-level political party response to the top-two
primary election system. States are where many policy innovations are
tried and studied. Several state legislatures have considered implementing a
system similar to Washington’s top-two primary. An understanding of this
system’s impact on political parties gives information that may influence
future public policy development and implementation. As has been
previously stated, elections that are accessible and provide an equal and
unimpeded franchise are essential to the representative democracy of the
United States.
Interpretation of Findings
Shortly after the establishment of the top-two primary, both the
Washington State Republican Party and the Washington State Democratic
Party joined forces on a legal challenge that was ultimately heard and
decided by the United States Supreme Court. The political parties objected
107
to the burden placed by the top-two primary on their First Amendment
association rights. Objections also included perceived voter confusion
(Washington State Grange v. Washington State Republican Party, 552 U.S.
442, 2008).
The case was decided such that the top-two primary was upheld, but
not without controversy. Justice Scalia wrote in his dissent, “There is no
state interest behind this law except the Washington Legislature’s dislike for
bright-colors partisanship, and its desire to blunt the ability of political
parties with noncentrist views to endorse and advocate their own candidates”
(Washington State Grange v. Washington State Republican Party,
552 U.S. 442, (2008), sec. III, para 3). Scalia’s remarks illustrate the setting
in which the top-two primary operates, and provides a lens through which to
interpret the present study considering past research on the topic.
As discussed in Chapter 2, this study of the county political party
response to the top-two primary occurred in a very particular setting,
specifically within Washington
State’s long history of populism and progressivism, and a culture of political
independence. Perhaps that independence is one reason a variety of party
responses have emerged. In Chapter 4, themes pointing to recruiter
(principals are important), facilitator
108
(it’s all about fairness) and enforcer (playing by the book) counties were
identified. The lack of uniformity county-to-county may be, unto itself, an
outgrowth of the historic, legal, and political context from which the top-two
primary sprang into existence. In this way, this study has added another
expression of Washington’s political independence.
Early research by Manweller (2011) suggested that the top-two
primary did generate some confusion. Voters assumed that party preference
expressed by a candidate was indicative of being a party’s nominee.
Confirming Manweller’s underlying concern about voter confusion, it is
possible that counties that emerged in this study as enforcer parties have
been driven to take the serious measures they have taken, in direct response
to that possible confusion.
Beck and Hendrickson (2013) investigated whether there was
incentive to political parties to attempt to limit the number of candidates
appearing on the primary ballot. Although not discussed above in Chapter 4,
it was noted that there was variability county to county whether endorsement
action was taken pre-primary (as in the case of enforcer parties wanting
heavier influence over the outcome) versus post-primary (as in the case of
109
facilitators, where the value of fairness and access to the ballot was
strongest).
This may be an area for future study.
McGhee and Shor (2017) suggested that the top-two primary did not,
as hoped for, combat polarization between parties. In this study there are
instances of intraparty disagreement and measuring whether or not both
candidates and fellow party members are considered Republican enough.
This seems to extend the conclusion of McGhee and Shor (2017) that there
was no moderating effect on candidates and raises questions of whether the
top-two primary has an influence on intra-party moderation or polarization.
Both Sparks (2019) and Patterson, Jr. (2020) raised questions about
the challenge for voters in low-information races. Specifically, they were
concerned about those races in which media deserts or competition with up-
ballot candidates for voter attention hamper the voter’s ability to choose. The
process codes of supporting and communicating emerged in the study,
suggesting that the facilitator and recruiter parties may be seeking to reach
voters in low information races.
Advocates of the top-two primary had hoped for increased
competition, according to Sparks (2020b). In the current study, several
110
participants not only mentioned but lamented the perceived narrowing of the
competition, stating their belief that the top-two primary disincentivized or
blocked out-party, third-party or independent candidates from even filing to
be on the ballot. This suggests another potential area of investigation.
Schattschneider (1935) was quoted as saying “new policies create new
politics” (p. 288). Whether that is evidenced in this study as a new interest in
nonpartisan candidates as reported by one of the recruiter counties, or an
increase in purity tests for Republican candidates and the much bandied-
about term Republican in name only (RINO) in the enforcer counties, the
relatively new top-two primary election system in use in Washington state,
does in fact appear to have its place in a policy feedback system. When the
separately formed Republican party organizations at the county level were
confronted with the loss of associational rights—namely, being able to
control which candidate’s bear the party’s designation—they were left to
decide on their response, individually. As a result, it appears that different
counties developed different approaches. Some emphasize fairness and
facilitation. Some emphasize the long game, looking to identify and prepare
future candidates from among those willing to serve at lower office. Some
111
appear to have tightened both the pathway to approval to bear the name
Republican, and also the resources available to good Republican candidates.
The current study adds to the body of knowledge on the top-two
primary, in ways that confirm, disconfirm, and extend what has previously
been illuminated in the scholarly literature, as stated above. Given that
participants expressed a low motivation to change the top-two primary
system, it is likely to remain in effect for the long term. The varied county
political party responses found in this study suggest a number of possible
future inquiries, as will be outlined below.
Limitations of the Study
In qualitative research, trustworthiness encompasses the concepts of
credibility, transferability, dependability, and confirmability. It is appropriate
to take a retrospective look at the elements of trustworthiness of this study,
and any limitations encountered.
On the whole, trustworthiness of the process and the results of this
study remained solid. For each of the following concepts known broadly as
trustworthiness, synonyms are offered by Miles and colleagues (2014).
Credibility, also described as authenticity, was detected in the ready rapport
and rich responses from participants. Dependability, aka auditability, was
112
achieved through the use of multiple tools for coding, starting with use of the
Delve tool, and reliance on manual organization and reorganization through
multiple rounds of coding using post-it notes. Additionally, a consistent
introductory script followed by a uniform list of questions was used in each
semistructured interview. Confirmability, also described as objectivity, can
be demonstrated through the recordings and text transcripts, as well as by the
retained coding outputs from Delve.
The one area of possible limitation to trustworthiness may be detected
in the area of transferability, also described by Miles, et al. (2014) as
fittingness or external validity. This limitation is not due to the methods
employed, but by the somewhat surprising range of attitudes and priorities
amongst participants. Rather than identifying a single county political party
response to the top-two primary, I concluded that several very different
responses emerged. In some cases, the county political party response was to
act as a facilitator, emphasizing a value of openness to the ballot and fairness
for all candidates. Other counties responded by being more intentional,
recruiting candidates for higher office from those who had served in lower
nonpartisan office. Still others took a more directive enforcement role in
screening and vetting candidates, gate-keeping party resources and limiting
113
access to those resources, explicitly endorsing the candidates who are
considered real Republicans and blocking candidates they do not favor. This
limitation of transferability could be remedied by additional participants.
Additional insight may also be gleaned by adding points of information
(perhaps through a mixedmethod approach; see recommendations below) to
see if quantitative data points can provide context. An example might be to
compare and contrast the attitudinal and procedural responses of political
parties based on whether the county is a swing county, or whether the
county’s voters consistently vote in favor of one party’s candidates or the
other.
Recommendations
One of the strengths of this study was the authenticity expressed in the
context of the semistructured interviews. In no case did it seem that
participants were holding back or censoring their remarks. One contributor
to this authenticity may have been that I came to the interview with both a
deep understanding of elections (as a current elections official) and a deep
understanding of Republican party operations (as a former party leader
myself). Another contributor to the authenticity may have been that the
topic—the top-two primary and its implications for party decision making
114
vis a vis a field of candidates that they cannot nominate, eliminate or
otherwise control—is one that is deeply felt by individuals who have chosen
to be active in party politics. For most party organizations, winning elections
is a central priority. Some simple recommendations for future research
include expanding the participant pool and running another round of
interviews and content analysis with additional Washington state
participants.
Expanding on the idea of running additional participants, a second
recommendation is to expand the current study to reach deeper into the party
organizations represented by this study’s participants. This study was limited
to party leadership; a future study might target rank-and-file precinct
committee officers for participation. The semistructured interview is likely to
conceal some individual bias on the part of the participant. Reaching deeper
to corroborate or find differences within individual party organizations, to
thus create a more robust data set and stronger conclusions, would be
beneficial.
Replicating this study with Republican party leadership in California
is another possible area for future research. As stated above, whereas
California has an essentially similar primary election system as Washington
115
has now, they arrived at the idea of the extra openness of the top-two
primary through a very different pathway. Whereas
Washington’s top-two precursor—the blanket primary—was in place for
nearly 70 years,
Californians voted in a closed primary system until 1998. Replicating this
study with a California population and subject pool might illuminate
whether the historic context of primary elections might contribute to
different party responses to the top-two primary.
Another recommendation for future study is to expand participants to
include county level Democratic party leadership. This study was limited to
Republican organizations, a limit chosen based on my personal and
professional qualities, as a means to ensure objectivity and an elimination of
potential bias. Following the completion of this study, it is my belief that a
partner researcher could be identified, either without the partisan history or
current nexus with county elections as I brought to the project, or possibly
someone with similar bona fides in the Democratic party. In the case of a
study with a second researcher, additional sophistication could be brought to
the activity of coding participant responses as well as content analysis, and
inter-coder reliability could be explored. Both an expansion to Democratic
party leadership, and to a second coder could bring a degree of reliability
116
and trustworthiness to a subsequent study that would surely benefit the body
of scientific literature on this topic.
Any of the above efforts could be expanded to a longitudinal design.
Checking back in with the same participant organizations over several
election cycles might provide insight not gathered here in this study. Such an
effort may also demonstrate change or evolution over time.
Although possibly premature but worth surveilling is the new system
in Alaska which is a top-four primary followed by a ranked choice general
election. This new hybrid system blends multiple electoral innovations. If it
withstands the various efforts to repeal it (Ballotpedia, n.d.-a), this system is
worthy of study as well.
One final recommendation for future study would be to integrate a
quantitative component. The top-two primary created the environment in
which two candidates who state the same party preference may advance
from the primary election ballot to the general election. In these cases, the
general election could very well be between two Democrats or two
Republicans. Bringing a mixed methods component to the concepts in this
study could reveal a correlation between enforcer counties and heavily
Republican outcomes, or facilitator counties and the swing nature of a
117
county, as an example. The quantitative data available in the form of
historical partisan voting trends might bring a sense of why to the
differences that exist between counties.
Implications
Research worth conducting is research that is oriented not just to
expanding a dataset somewhere. Research worth conducting provides
information and insight that can be used by academics and policy makers to
contribute positively to social change. In a small way, this study on the
county political party response to the top-two primary brings added
knowledge that can influence future public policy development in a way that
is informed by the voices and stories of individuals and groups impacted by
the top-two primary, and by the data collected from those individuals. In
addition, this study expands the application of policy feedback theory into
qualitative research. Finally, the themes identified in this study provide a
lens to be used by political parties, practitioners, and policy makers alike to
evaluate the extent to which the response to the top-two primary conforms to
the intent of this electoral system.
Public policy development does not occur in a vacuum. A key
component of the representative form of government in the United States is
118
the accessibility to the governance process. Whether that is access to the
ballot as a voter—or as a candidate—or through sunshine laws that insist on
transparency in government as in the case of public records laws or open
public meeting requirements, the citizen in this form of government is
invited to be involved and informed. This study gathers the voices and
perspectives of one group of citizens—political party leadership—and gives
an insight not previously available in the scientific literature.
Regarding theoretical implications, the current study brings policy
feedback theory into the qualitative arena in a way not previously in
evidence in the literature. This study gives voice to the power of groups
component within policy feedback theory and demonstrates how that power
is wielded differently depending on local circumstances and group
dynamics. This study might be used as a model for future research, whether
that includes the use of similar semistructured interviews, content analysis, a
synthesis of varied coding methods, or simply using a multiple case study
approach for future research.
This study illuminates several themes of sentiment. The first theme of
sentiment, dislike but acquiescence, suggests that the top-two primary is here
to stay. There is no organized or energetic legislative advocacy underway to
119
eliminate this primary election system. Each participant in this study is
resigned to this form of primary election. The positive social impact of
uncovering this phenomenon might include encouraging a pivot of attention
for those involved in elections at all levels—including elections practitioners
and political parties—from disagreement and discontent with the top-two
primary election system--to voter education. In my professional work as an
elections official, I often speak with voters who assert “I am a registered
Republican,” or “I am a registered
Democrat,” when in reality, a voter does not register with a party affiliation
in a top-two primary environment. The lack of understanding on this precise
topic creates frustration and fear in voters. For example, when a voter
receives a robocall or a piece of campaign literature from one candidate or
another, they might raise questions about how the source got their
information, and how those persons know how a voter has voted in the past.
By shifting energy from tolerating the top-two primary to engaging in a
robust program of voter education, citizens may achieve a higher level of
comfort with the electoral system in general—a tremendous achievement in
the current environment. The implications of such an effort could be felt at
both the individual and community levels.
120
Another theme of sentiment, the value of fairness and access to the
ballot for all candidates, has much needed implications for this space. In a
time of intense political polarization, to demonstrate that the value of a fair
fight is present and expressed within partisan groups may serve to cut that
polarization. When interparty rhetoric is focused on demonizing one’s
opponent, a shift to more robust discussions about issues and solutions could
have far-reaching implications, from the Thanksgiving table to peace among
nations.
The final theme of sentiment, which highlights conflict, a negative
view of fellow partisans, and the need for purity tests, deserves
contemplation as well. This theme leaves a bad taste in the mouth. It weighs
on the conscience. It divides. One positive implication of the current study is
that by bringing this sentiment to light, it may engender some reflection—
and possibly a more charitable view of one another in partisan environments.
In a similar way, the action-oriented themes that place county political
parties into categories of facilitator, recruiter, or enforcer counties, reveal
some opportunities for positive social change. County political parties are
not monolithic. Even when the underlying identification or affiliation of
Republican is held in common, the values, feelings, decisions, and actions
121
may differ, party organization to party organization, county to county.
Knowing that these differences and nuances exist establishes a shared
understanding, a clearer vernacular, and an invitation to grow to counties in
any of the above-mentioned categories.
The top-two primary has an impact on voters, candidates, political
parties, and the public policy development process. These were the various
angles of inquiry first introduced in Chapter 1. A final implication of this
study is that it has started answering the larger question, “what has been the
response to Washington’s top-two primary?” from one angle—the political
parties. Participant responses, whether the meaningful and descriptive in
vivo and versus codes, or the diverse action themes that describe how county
parties have responded differently, have given much to build on current
understanding and future study.
Conclusions
One online source catalogued 491 bills introduced throughout the
United States related to election systems (of which primaries in general and
the top-two in specific are included) in just one year—2023 (Ballotpedia,
n.d.-b). Policy makers are most certainly paying attention to these issues.
The scientific literature should be engaging with these essential ideas with
122
equal focus. This study provides some small contribution to this body of
knowledge, a methodology for future study and by doing so, has engaged in
the very positive social change sought at this study’s outset. The work is on-
going.
References
Alaska Division of Elections (n.d.) Division of Elections to begin
Implementation of Top
Four Primary and Ranked Choice Voting.
https://www.elections.alaska.gov/division-of-elections-to-begin-
implementationof-top-four-primary-and-ranked-choice-voting/
Alvarez, R. M., & Sinclair, B. (2012). Electoral institutions and legislative
behavior: The effects of primary processes. Political Research
Quarterly, 65(3), 544–557.
https://doi.org/10.1177/1065912911411098
Alvarez, R. M., & Sinclair, J. A. (2013). Who are California’s “decline to
state” voters?
California Journal of Politics and Policy, 5(1), 47–66.
https://doi.org/10.5070/P2KK52
Alvarez, R. M., & Sinclair, J. A. (2015). Nonpartisan primary elections
reform:
123
Mitigating mischief. Cambridge University Press.
Amorós, P., Puy, M. S., & Martínez, R. (2016). Closed primaries versus top-
two primaries. Public Choice, 167(1–2), 21–35.
https://doi.org/10.1007/s11127-016-
0328-5
Anderson, S. E., Butler, D. M., Harbridge-Yong, L., & Marshall, R. (2023).
Top-four primaries help moderate candidates via crossover voting:
The case of the 2022
Alaska election reforms. Forum (2194-6183), 21(1), 123–136.
https://doi.org/10.1515/for-2023-2001
Appleton, A., & Grosse, A. (2004). Washington State Parties. In C. W.
Clayton, L. T.
LeLoup, & N. P. Lovrich (Eds.), Washington State Government and
Politics (pp.
27–43). Washington State University Press.
Ballotpedia (n.d.-a). Alaska repeal top-four ranked-choice voting initiative
(2024).
https://ballotpedia.org/Alaska_Repeal_Top-Four_Ranked-
Choice_Voting_Initiative_(2024)
Ballotpedia. (n.d.-b). Electoral systems legislation in the United States,
2023.
https://ballotpedia.org/Electoral_systems_legislation_in_the_United_
States,_2023
124
Ballotpedia. (n.d.-c), Primary election types by state.
https://ballotpedia.org/Primary_election_types_by_state
Ballotpedia. (n.d.-d). Top-two primary. https://ballotpedia.org/Top-
two_primary
Barth, J., Burnett, C. M., & Parry, J. (2020). Direct democracy, educative
effects, and the
(mis)measurement of ballot measure awareness. Political Behavior,
42(4), 1015–
1034. https://doi.org/10.1007/s11109-019-09529-w
Beck, J. H., & Henrickson, K. E. (2013). The effect of the top two primary
on the number of primary candidates: Candidate incentives under top-two
primary. Social
Science Quarterly, 94(3), 777–794. https://doi.org/10.1111/j.1540-
6237.2012.00876.x
Caen, C. (2015, December 29). How the top-two primary changed
California’s politics. The Atlantic.
https://www.theatlantic.com/politics/archive/2015/12/california-
toptwo-open-primary/421557/
Calcagno, P., & Westley, C. (2008). An institutional analysis of voter
turnout: The role of primary type and the expressive and instrumental
voting hypotheses.
125
Constitutional Political Economy, 19(2), 94–110.
https://doi.org/10.1007/s10602-
007-9034-0
California Democratic Party v. Jones, 530 U.S. 567. (2000). Justia Law.
https://supreme.justia.com/cases/federal/us/530/567/
Campbell, A. L. (2012). Policy makes mass politics. Annual Review of
Political Science,
15, 333–351. https://doi.org/10.1146/annurev-polisci-012610-135202
Chattopadhyay, J. (2018). Is the Affordable Care Act cultivating a cross-class
constituency? Income, partisanship, and a proposal for tracing the
contingent nature of positive policy feedback effects. Journal of
Health Politics, Policy and
Law, 43(1), 19–67. https://doi.org/10.1215/03616878-4249805
Clinton, J. D., & Sances, M. W. (2018). The politics of policy: The initial
mass political effects of Medicaid expansion in the states. American
Political Science Review,
112(1), 167–185. https://doi.org/10.1017/S0003055417000430
Cornell, C., & Meyer, S. (2004). Washington’s constitution: History and the
politics of state constitutional jurisprudence. In C. W. Clayton, L. T.
LeLoup, & N. P.
126
Lovrich (Eds.), Washington State Government and Politics (pp. 93–
112).
Washington State University Press.
Credibility. (2010). In A. Mills, G. Durepos, & E. Wiebe, Encyclopedia of
Case Study
Research. SAGE Publications, Inc.
https://doi.org/10.4135/9781412957397
Creswell, J. W. (2009). Research design: Qualitative, quantitative, and
mixed methods approaches (3rd ed). Sage Publications.
Creswell, J. W. (2013). Qualitative inquiry and research design: Choosing
among five approaches (3rd ed). SAGE Publications.
Dancey, L., & Sheagley, G. (2013). Heuristics behaving badly: Party cues
and voter knowledge. American Journal of Political Science (John
Wiley & Sons, Inc.),
57(2), 312–325. https://doi.org/10.1111/j.1540-5907.2012.00621.x
Fisk, C. A. (2020). No Republican, no vote: Undervoting and consequences
of the toptwo primary system. State Politics & Policy Quarterly,
20(3), 292–312.
https://doi.org/10.1177/1532440019893688
127
Flavin, P., & Hartney, M. T. (2015). When government subsidizes its own:
Collective bargaining laws as agents of political mobilization.
American Journal of Political
Science, 59(4), 896–911. https://doi.org/10.1111/ajps.12163
Fleming, D. J. (2014). Learning from schools: School choice, political
learning, and policy feedback: school choice, political learning, and
policy feedback. Policy
Studies Journal, 42(1), 55–78. https://doi.org/10.1111/psj.12042
Frankfort-Nachmias, C., & Nachmias, D. (2008). Research methods in the
social sciences
(7th ed). Worth Publishers.
Galarza, B. (2019, October 19). [Personal communication].
Goss, K. A. (2010). Civil society and civic engagement: Towards a multi-
level theory of policy feedbacks. Journal of Civil Society, 6(2), 119–
143.
https://doi.org/10.1080/17448689.2010.506370
Harris, K. (n.d.). Academic guides: Research ethics: research ethics review
process by IRB. https://academicguides.waldenu.edu/research-
center/research-ethics/reviewprocess
Haselswerdt, J. (2017). Expanding Medicaid, expanding the electorate: The
Affordable
128
Care Act’s short-term impact on political participation. Journal of
Health Politics,
Policy and Law, 42(4), 667–695. https://doi.org/10.1215/03616878-
3856107
Henrickson, K. E., & Johnson, E. H. (2019). Increasing voter participation
by altering the costs and stakes of voting. Social Science Quarterly,
100(3), 869–884.
https://doi.org/10.1111/ssqu.12583
Hill, S. J., & Kousser, T. (2016). Turning out unlikely voters? A field
experiment in the top-two primary. Political Behavior, 38(2), 413–
432.
https://doi.org/10.1007/s11109-015-9319-3
History of Washington State primary systems. (n.d.).
https://www.sos.wa.gov/elections/research/history-of-washington-
state-primarysystems.aspx
Hortala-Vallve, R., & Esteve-Volart, B. (2011). Voter turnout in a
multidimensional policy space. Economics of Governance, 12(1), 25–
49.
https://doi.org/10.1007/s10101-010-0089-4
129
Jackson, W. B. (2012). A blanket too short and too narrow: California’s
nonpartisan blanket primary. Stanford Law & Policy Review, 23(2),
535–548. Academic
Search Complete.
Kotsadam, A., & Jakobsson, N. (2011). Do laws affect attitudes? An
assessment of the Norwegian prostitution law using longitudinal data.
International Review of Law and Economics, 31(2), 103–115.
https://doi.org/10.1016/j.irle.2011.03.001
Kreitzer, R. J., Hamilton, A. J., & Tolbert, C. J. (2014). Does policy adoption
change opinions on minority rights? The effects of legalizing same-
sex marriage.
Political Research Quarterly, 67(4), 795–808.
https://doi.org/10.1177/1065912914540483
Levendusky, M. (2010). Clearer cues, more consistent voters: A benefit of
elite polarization. Political Behavior, 32(1), 111–131.
https://doi.org/10.1007/s11109-
009-9094-0
Lincoln, Y. S., & Guba, E. G. (1985). Naturalistic inquiry. Sage
Publications.
130
Lovrich, Jr., N. P., Pierce, J. C., & Elway, H. S. (2018). Two Washingtons?
Political Culture in the Evergreen State. In Governing the Evergreen
State: Political Life in Washington. Washington State University
Press.
Lowi, T. J. (1972). Four systems of policy, politics, and choice. Public
Administration
Review, 32(4), 298. https://doi.org/10.2307/974990
Manweller, M. (2011). The very partisan nonpartisan top-two primary:
Understanding what voters don’t understand. Election Law Journal:
Rules, Politics, and Policy,
10(3), 255–271. https://doi.org/10.1089/elj.2010.0070
McDonald, P. J. (2014). The Affects of Vote by Mail on Voter Turnout, Cost
of Elections, and Ballot Completion in the State of Washington from
1992 to 2012 (Order No.
3631561). Available from Dissertations & Theses @ Walden
University.
(1566384324). https://www.proquest.com/dissertations-theses/affects-
vote-mailon-voter-turnout-cost-elections/docview/1566384324/se-2
McDonnell, L. M. (2013). Educational accountability and policy feedback.
Educational
Policy, 27(2), 170–189. https://doi.org/10.1177/0895904812465119
131
McGhee, E., Masket, S., Shor, B., Rogers, S., & McCarty, N. (2014). A
primary cause of partisanship? Nomination systems and legislator
ideology. American Journal of
Political Science, 2, 337. https://doi.org/10.1111/ajps.12070
McGhee, E., & Shor, B. (2017). Has the top two primary elected more
moderates?
Perspectives on Politics, 15(4), 1053–1066.
https://doi.org/10.1017/S1537592717002158
Mettler, S. (2019). Making what government does apparent to citizens:
Policy feedback effects, their limitations, and how they might be
facilitated. Annals of the
American Academy of Political and Social Science, 685(1), 30–46.
https://doi.org/10.1177/0002716219860108
Mettler, S., & SoRelle, M. (2014). Policy Feedback Theory. In Theories of
the Policy
Process (3rd ed., pp. 151–181). Westview Press.
Mettler, S., & Soss, J. (2004). The consequences of public policy for
democratic citizenship: Bridging policy studies and mass politics.
Perspectives on Politics,
2(1), 55–73. https://doi.org/10.1017/S1537592704000623
132
Mettler, S., & Welch, E. (2004). Civic generation: Policy feedback effects
of the GI Bill on political involvement over the life course. British
Journal of Political Science, 34, 497-518.
https://doi.org/10.1017/S0007123404000158
Miles, M. B., Huberman, A. M., & Saldaña, J. (2014). Qualitative data
analysis: A methods sourcebook (Third edition). SAGE Publications,
Inc.
National Conference of State Legislatures. (n.d.). State primary election
types.
https://www.ncsl.org/research/elections-and-campaigns/primary-
types.aspx
National Conference of State Legislatures. (2011). The Canvass. National
Conference of
State Legislatures.
https://web.archive.org/web/20221222175437/https://www.ncsl.org/d
ocuments/le gismgt/elect/Canvass_May_2011_No_20.pdf
Neiheisel, J. R., & Niebler, S. (2013). The use of party brand labels in
congressional election campaigns. Legislative Studies Quarterly,
38(3), 377–403.
https://doi.org/10.1111/lsq.12019
133
New State Ice Co. v. Liebmann, 285 U.S. 262 (1932)
https://supreme.justia.com/cases/federal/us/285/262/
November 7, 2023 General Election—Voter Turnout. (n.d.).
https://results.vote.wa.gov/results/20231107/turnout.html
Open Primaries. (n.d.). Myths and Facts.
https://web.archive.org/web/20181116105751/https://www.openprima
ries.org/my ths_and_facts
Pacheco, J. (2013). Attitudinal policy feedback and public opinion: The
impact of smoking bans on attitudes towards smokers, secondhand
smoke, and antismoking policies. Public Opinion Quarterly, 77(3),
714–734.
https://doi.org/10.1093/poq/nft027
Patterson, S. (2020). Estimating the unintended participation penalty under
top-two primaries with a discontinuity design. Electoral Studies, 68,
102231.
https://doi.org/10.1016/j.electstud.2020.102231
Pierson, P. (1993). When effect becomes cause: Policy feedback and political
change.
World Politics, 45(4), 595–628. https://doi.org/10.2307/2950710
134
Political Parties (1973). RCW 29A.80
https://app.leg.wa.gov/rcw/default.aspx?cite=29A.80&full=true
Rubin, H., & Rubin, I. (2005). Qualitative interviewing (2nd ed.): The art of
hearing data
(2nd ed.). https://doi.org/10.4135/9781452226651
Saldaña, J. (2016). The coding manual for qualitative researchers (3E [Third
edition]).
SAGE.
Sample Ballot, Primary Election, August 4, 2020, Stevens County,
Washington. (n.d.).
https://stevenscountywa.municipalcms.com/files/documents/2020Sam
plePrimary
Ballot1329114121072120AM.pdf
Schattschneider, E. E. (1935). Politics, pressures, and the tariff. Arno Press.
Skocpol, T. (1991). Targeting Within Universalism: Politically Viable
Policies to Combat Poverty in the United States. In C. Jencks & P. E.
Peterson (Eds.), The Urban Underclass (pp. 411–436). The Brookings
Institution.
Skocpol, T. (1996). Protecting soldiers and mothers: The political origins of
social policy in the united states (1. Aufl., 4. Druck). Harvard
University Press.
135
Sparks, S. (2018). Campaign spending and the top-two primary: How
challengers earn more votes per dollar in one-party contests. Electoral
Studies, 54, 56–65.
https://doi.org/10.1016/j.electstud.2018.04.018
Sparks, S. (2019). Polarization and the top-two primary: Moderating
candidate rhetoric in one-party contests. Political Communication,
36(4), 565–585.
https://doi.org/10.1080/10584609.2019.1579772
Sparks, S. (2020a). Campaign dynamics of the top-two primary: Invoking
competition and representational pressure in one-party contests.
Representation, 56(2), 229–
251. https://doi.org/10.1080/00344893.2020.1754893
Sparks, S. (2020b). Quality challenger emergence under the top-two
primary: Comparing one-party and two-party general election
contests. Electoral Studies, 65, 102136.
https://doi.org/10.1016/j.electstud.2020.102136
Spitzer, H. (2019). Be careful what you wish for: Private political parties,
public primaries, and state constitutional restrictions. Washington Law
Review, 94(2),
136
823.
Stevens County Republican Central Committee 2021-2022 Bylaws.
(2021). U.S. Const., amend I.
U.S. Election Assistance Commission, (2023). Alternative voting methods (v
1.0).
https://www.eac.gov/sites/default/files/electionofficials/Final_Alternat
ive_Voting _Methods_in_the_United_States_508.pdf
Video Conferencing, Web Conferencing, Webinars, Screen Sharing. (n.d.).
Zoom Video.
https://zoom.us/.
Washington Secretary of State. (2001, June 6). Blanket primary hearings
report.
https://web.archive.org/web/20010606145202/http://www.secstate.wa.
gov/electio ns/bp_report.htm
Washington State Grange v. Washington State Republican Party, 552 U.S.
442. (2008).
Justia Law. https://supreme.justia.com/cases/federal/us/552/442/
WSRP By-Laws. (n.d.). Washington State Republican Party.
https://wsrp.org/by-laws/
Zhang, C. (2012). Towards a more perfect election: Improving the top-two
primary for congressional and state races. Ohio State Law Journal,
73(3), 615-650.
