Understanding Congressional Coalitions: A Discourse Network Analysis of Congressional Hearings for the Every Student Succeeds Act

The purpose of this study is to investigate policy coalitions of the Every Student Succeeds Act (ESSA) at U.S. congressional hearings. This study is grounded in the advocacy coalition framework, which argues that advocacy coalitions are forged by policy actors who have similar policy preferences. To identify the coalitions, according to the policy claims articulated by policy actors, discourse network analysis was performed to examine 30 testimonies in the congressional hearings on ESSA since its passage in 2015. The policy actors fall into eight categories: (1) federal administrative and executive offices, (2) state administrative and executive offices, (3) teachers unions, (4) interest groups, (5) superintendents, (6) teacher, (7) education professor, and (8) Congress members. The results of discourse network analysis suggest four coalitions based on the actors’ policy claims on (1) equity, (2) assessment and accountability, (3) states have changed/passed legislation to align the state accountability systems with ESSA goals, and (4) the U.S. Department of Education Policy Analysis Archives, Vol. 28 No. SPECIAL ISSUE 2 Education’s state plan approval was inconsistent with the ESSA statutory provisions. The findings provide timely insights into the ongoing process of ESSA implementation at the federal, state, and local levels.


Introduction
This study aims to investigate policy coalitions of the Every Student Succeeds Act (ESSA) at U.S. congressional hearings. ESSA reauthorized the Elementary and Secondary Education Act of 1965. After decades of governance of the No Child Left Behind Act of 2001, ESSA was passed with strong bipartisan support (85 Yeas and 12 Nays at the U.S. Senate;United States Senate, 2015). Since the passage of ESSA in 2015, states and local education agencies have been implementing ESSA by developing education policies and passing legislation to align the state accountability systems with ESSA. The then-ranking member of the U.S. House Committee on Education and the Workforce (now known as House Committee on Education and Labor), Robert C. "Bobby" Scott, stated that ESSA repealed the one-size-fits-all approach of No Child Left Behind with increased flexibility for States and school districts, but with flexibility comes responsibility.
States will be tasked with developing new multi-measure accountability systems, implementing more innovative assessment systems, and providing support to school districts to ensure that low performing schools improve and meet the needs of all students (Scott, 2016, p. 4).
To understand the implementation process and national politics of ESSA, it is important to understand the policy actors and their policy coalitions. This study therefore aims to examine which policy actors have forged coalitions around ESSA, and what their policy preferences were, particularly in Congress. To do so, this study applies discourse network analysis-an emerging analytical approach built on content analysis and network analysis (Leifeld, 2017)-to uncover coalitions working in areas related to ESSA. Specifically, this study seeks to answer two research questions: (1) Which policy actors have forged coalitions in ESSA congressional hearings? and (2) What specific policy claims in ESSA discourse have been driving the coalition formation in congressional hearings? To set the stage for studying ESSA policymaking in the U.S. Congress, I first placed ESSA in a historical context, followed by reviewing the literature on how the U.S. Congress has been studied in policymaking research. Next, I conduct a discourse network analysis of statements made by witnesses at all Congressional hearings on ESSA since the passage of ESSA in 2015. The findings reveal major policy actors and their policy coalitions on equity, assessment, and accountability. This article concludes by discussing how the findings help us understand the shifting power within American education politics and policy.

Placing ESSA in Context
This article focuses on ESSA implementation process at the national level. An understanding of the context that led up to the passage of ESSA provides us with a historical perspective of ESSA implementation. Education, according to the U.S. Constitution, is not a federal responsibility, but a state and local responsibility. States have had the authority to fund and govern schools (Manna, 2007). The Elementary and Secondary Education Act (ESEA) of 1965 was signed by President Lyndon Johnson as part of his administration's War on Poverty campaign. ESEA enabled the federal government to direct national education policy by providing resources and establishing programs for disadvantaged populations of students (McGuinn, 2006). In 1979, the U.S. Department of Education (USDE) was established after Congress passed the Department of Education Organization Act .
In 1983, President Ronald Reagan's National Commission on Excellence in Education, composed of 18 members from the private sector, government, and education, published a landmark report titled A Nation at Risk, which stated that American schools were failing to meet the national need for a competitive workforce. The same report, however, was framed differently by Democrats and Republicans to advocate for their proposed education policies. On the one hand, Democrats argued that the failing education demanded great federal control; on the other hand, Republicans argued that the failing education was a testament of failure in past federal programs, calling for eliminating federal control over education.
The presidential campaign of George W. Bush in the late 1990s took notice of the unpopular Republican education proposals and repositioned the Republican Party on education by embracing "an active but reformed federal role in promoting school improvement" (McGuinn, 2006, p. 201). With the 2001 reauthorization of ESEA, the No Child Left Behind Act (NCLB), the federal government gained more control over national education policy, as NCLB required states to test students in reading and mathematics, establish proficiency standards, and track and report schoolwide progress toward meeting annual performance goals toward an ultimate goal of 100% proficiency for all students by the end of the 2013-14 school year (DeBray, 2006). NCLB substantially limited states' decision-making power over testing and school implementation efforts, evoking states' growing dissatisfaction. By 2006, eight states had defied NCLB by introducing legislation in direct opposition to NCLB, and 30 states introduced or passed legislation asking Congress to modify NCLB (Duff & Wohlstetter, 2019;Shelly, 2008). In addition to states, teachers unions and interest groups had grown dissatisfied with NCLB which pushed high-stakes, test-based accountability (McDonnell & Weatherford, 2013).
In 2007, ESEA was supposed to be reauthorized. However, Congress-with a Democratic majority in both bodies of Congress during the 109 th Congress-and the Bush administration failed to make progress in reauthorization. Both the legislative and executive branches of the government agreed that NCLB was flawed, but disagreed with how to fix it. In July 2009, the Obama administration announced the $4.35 billion Race to the Top program-a competitive grant that states competed for funding to adopt standards and assessment, improve teacher and principal effectiveness, and turn around low-performing schools (U.S. Department of Education, 2009). In March 2010, the Obama administration presented Congress with a Blueprint for ESEA reauthorization, proposing creating a teacher evaluation system tied to student growth, but the proposal received lukewarm support from Congress. The stalemate on ESEA reauthorization in the 110 th , 111 th , 112 th , and 113 th Congresses led to the failure to fix the perceived shortcomings of NCLB in a timely manner.
The Obama administration capitalized on the lack of congressional leadership on ESEA reauthorization by exerting more control over shaping education policy (Wong, 2015). The inaction at Congress opened a policy window (Kingdon, 1995) for the Obama administration: the problem stream had been developing since the enactment of the problematic provisions in NCLB; the political stream had been active in the Obama administration, and USDE was more than willing to take the lead on policy change. In September 2011, criticizing Congress for inaction on ESEA reauthorization, the Obama administration announced to offer states waivers from specific provisions in NCLB, including meeting the goal of 100% proficiency by 2014. In exchange for NCLB waivers, the Obama administration-under the leadership of Secretary of Education Arne Duncan-required states to adopt Common Core State Standards and evaluate teachers based in part on student performance, among others. Although the justification of waivers was to modify the shortcomings in NCLB, the Obama administration aggressively used its executive power to grant NCLB waivers to advance the administration's policy priorities not specified in NCLB, such as implementing teacher evaluation systems (Saultz, McEachin, & Fusarelli, 2016). The aggressive use of executive power was met by congressional resistance. With the 2012 presidential election approaching, Congress planned to reauthorize ESEA in 2013, more likely in 2014 if a Republican was elected president (Michelman, 2012). After Obama won reelection, Congress passed ESSA with provisions to sharply reduce the role of the executive branch, specifically limiting the ability of the Secretary of Education to grant waivers requiring states to (1) adopt specific academic standards, (2) assessments, (3) accountability systems, or (4) teacher evaluation systems (ESSA, 2015). All four of these components were included in previous waivers granted under Secretary Duncan. Under ESSA, the Secretary of Education can no longer use the waiver process as a policy lever to coerce or provide incentives for states to implement policies set by USDE. If power and control indeed shift substantially back to the state and local levels, it will reverse decades of increasing federal control over education (Fusarelli & Fusarelli, 2015). If so, it is hard to imagine that USDE would cede its power willingly. The implementation of ESSA was expected to be a bumpy road over the power shifting process.

Studying the U.S. Congressional Hearings
One way to understand national policy and politics surrounding ESSA is to study the legislative branch of the government. A congressional hearing is "a meeting or session of a Senate, House, joint, or special committee of Congress, usually open to the public" (U.S. Government Publishing Office, 2019, para. 1). Congressional hearings are an important part of policymaking process in the United States, as they are the primary ways that Congress members "obtain information and opinions on proposed legislation, conduct an investigation, or evaluate/oversee the activities of a government department or the implementation of a Federal law" (U.S. Government Publishing Office, 2019, para. 1). In fact, holding a hearing on an issue itself signals the salience of the issue perceived by a congressional committee (Diermeier & Feddersen, 2000).
Testimonies in congressional hearings are given by a range of policy actors, including members of the U.S. Congress, officials in the executive branch of the government, interest groups, businesses, think tanks, and academic researchers (Gormley, 1998). Being called to testify in hearings indicates a policy actor's influence on an issue (Smith, 1995). Policy actors see hearings as an important policy platform to articulate their policy preferences in front of legislators and the public, as well as garner attention and support of other policy actors (Cmiel, 1999). Some might deem hearings as "stage-managed spectacles" (Burstein & Hirsh 2007, p. 179), when witnesses are chosen for political reasons and what they say are known in advance. Yet policy actors believe their policy preferences provided in the hearings influence many policy changes and policymaking. For instance, in congressional hearings on the Equal Pay Act of 1963, policy actors who advocated for gender equality won support for prohibiting sex discrimination in employment in Title VII of the Civil Rights Act of 1964 (Burstein, 1998). In congressional hearings on hate crimes, policy actors had a significant impact on the content and enactment of federal hate crime legislation (Jenness, 1999). Congressional hearings thus have become a proxy for policy research (Arnold, 1990;Clifton, 2004;Fisher, Leifeld, & Iwaki, 2013;Leifeld, 2013). In research on climate change, data collected from congressional hearings on climate change and global warming were compared with the data collected from media coverage (e.g., New York Times, Los Angeles Times, and Chicago Tribune) to reveal that Republican-controlled Congress appeared to restrict the climate change issue (Liu, Lindquist, & Vedlitz, 2011). Congressional hearings on climate change were also used to examine how climate change issues were framed (Park, Liu, & Vedlitz, 2010). Park et al. reported that when the Democrats held the majority in Congress, hearings on climate change occurred more frequently in Democrat-controlled Congressional sessions, and had more scientists and pro-environment policy actors as testimony witnesses. By contrast, testimony witnesses were mostly from business and industrial sectors in Republicans-controlled Congress, challenging the science of climate change and offering information on the potential negative impact of climate change policy on economic growth and foreign relations (McCright & Dunlap, 2003;Park et al., 2010). Further, to understand interest groups' influence on policymaking process of a variety of public policies (e.g., the savings and loan bailout, the Americans with Disabilities Act, security on college campuses, solid waste disposal, and patents for inventions developed in outer space), congressional hearings were analyzed to conclude that "information-particularly information regarding policy effectiveness-does affect the likelihood that a policy proposal will be enacted" (Burstein & Hirsh 2007, p. 174).
In education policy research, very few studies have capitalized on data collected from congressional hearings. One example of analyzing congressional hearing is a recently published study on the role and contribution of academic researchers in congressional hearings on the reauthorization of the Higher Education Act-the legislation for promoting higher education access, attainment, and affordability (Perna, Orosz, & Kent, 2019). The results of a critical discourse analysis indicated that academic witnesses played an expert role in congressional hearings: "both experts who offer independent knowledge and experts who validate or confirm a legislator's preferences and priorities" (Perna et al., 2019, p. 111).
After the passage of ESSA in 2015, at the ensuing congressional hearing "Next Steps for K-12 Education: Upholding the Letter and Intent of the Every Student Succeeds Act," the thenranking member of House Committee on Education and the Workforce stated, "passing legislation is only one step. … fulfilling the promise of Every Student Succeeds Act rest[s] on successful implementation that honor Congress' long-standing commitment and intent of the ESSA" (Scott, 2016, p. 4), and the role of Congress was to maintain "vigorous oversight and enforcement to ensure that these expectations are met" (Scott, 2016, p. 5). To that end, congressional hearings are a fitting proxy to study policy discourse, as congressional hearings provide direct accounts of policy discourse on ESSA as well as broad education issues. Given the role of Congress in ESSA and the value of congressional hearings, this study takes an important first step in understanding policy actors and their policy coalitions via congressional hearings on ESSA. Building on the earlier work on policy discourse in hearings (Fisher, Waggle, & Leifeld, 2013;McCright & Dunlap, 2003), this study sheds light not only on who has a say in ESSA implementation policy discourse in the U.S. Congress, but also on what they are saying and with whom they are aligning.

Conceptual Framework
This study is conceptually grounded in the advocacy coalition framework, a prominent conceptual lens to investigate coalition formation and its impact on policymaking (Ingold, 2011). The advocacy coalition framework (Sabatier, 1988;Sabatier & Weible, 2007) argues that advocacy coalitions are forged by policy actors who have similar policy preferences. Take climate change as an example. The coalitions around the topics of climate change emerged from the actors' consensus view of economic implications of regulating greenhouse gases and the policy instrument for regulation . In education, the opposition coalition of the Common Core State Standards forged around (1) views about the former President Obama, (2) testing, and (3) misconceptions and negative conceptions about the standards (Polikoff, Hardaway, Marsh, & Plank, 2016). These policy preferences demonstrated the mounting dissatisfaction with NCLB among state education agencies, teacher unions, and interest groups (McDonnell & Weatherford, 2013). Their similar policy preferences function as the "glue" holding coalitions together (Sabatier & Weible, 2007). By contrast, the differences in policy preferences are manifested in the discourse that serves to defend or propose coherent arguments as justifications for policy preferences held by the opposition coalitions (Hajer, 1993(Hajer, , 1995. For example, in the opt-out movement-the grassroots, organized efforts to refuse to take high-stakes standardized tests, on the one hand, movement supporters (pro-opt-out teachers, opt-out parents, opt-out advocacy groups, teachers' unions, parent teacher associations, and the opt-out students) and movement opponents (e.g., the New York State Education Department, the Common Core Task Force, the U.S. Department of Education, and the testing providers) formed competing coalitions through different perceptions of contested goals of standardized testing, different framings of the movement, and different preferred policy solutions (Author, 2017). To that end, the advocacy coalition framework views the policy process as a competition between coalitions of policy actors who advocate their beliefs about policy problems and policy solutions.
Policy actors make public statements about a policy in order to signal their policy preferences to potential allies, convince other actors to adopt their policy preferences, or reduce their uncertainty by learning from other actors. A coalition thus emerges as policy actors share their policy preferences (Sabatier, 1998). Over time, similar policy preferences prompt policy actors to engage in a coordinated activity (Zafonte & Sabatier, 1998). For instance, in the social networks accounting for the rise of civil society in Mexico, the coalition was established as the actors (i.e., workers, peasants, students, civic associations, and non-governmental organizations) shared their policy preferences and then co-participated in protest campaigns together (Wada, 2014). In the networks of reading curriculum policymaking, the coalition emerged as the actors (e.g., educational organizations, school districts, business or businesses associations, private reading consultants, and philanthropic foundations) shared their policy preferences, and then collaborated over the policymaking process (Song & Miskel, 2005Yong, Wang, & Lewis, 2016). In the case of the opt-out movement, the New York State Education Department and the U.S. Department of Education shared a consensus view of sanctioning districts with a high opt-out rate. Their shared view led to the coordination that the U.S. Department of Education urged the New York State Education Department to sanction local education agencies with a high opt-out rate (i.e., exceeding 5%), and the New York State Education Department later punished the schools with the high optout rate by withholding grants (The Lockport Union-Sun & Journal, 2016). Moreover, the teachers' union and the Long Island Opt Out group shared a consensus view that standardized testing does not accurately reflect learning or student achievement. This shared view prompted them to engage in the coordinated activity in which the teachers' union joined a Long Island Opt Out group's event and distributed the fliers about opting out (Ferrette, 2016).
Applying the advocacy coalition framework to study ESSA congressional hearings, the interdependent relationships among ESSA policy actors and their policy claims can be conceptualized as networks. In such networks, policy actors are connected by their articulated policy claims about ESSA. Policy actors within the coalition tend to hold similar views and present similar arguments; those who hold divergent views and present dissimilar arguments tend to forge competing coalitions. If we can detect the clusters of policy actors based on their policy claims, we can then identify the policy coalitions at ESSA congressional hearings. Thus, the advocacy coalition framework is the conceptual root of this study's analytical approach of discourse network analysis.

Methods
To uncover policy coalitions, this study uses discourse network analysis to analyze the discourse data of ESSA congressional hearings. Discourse network analysis, built on content analysis and network analysis, is an emerging analytical approach to investigate policy discourse (Leifeld, 2017). Here I present how the data on the discourse of ESSA congressional hearings were collected and analyzed.

Data Sources
Data for this study consist of 30 testimonies from congressional hearings on ESSA during the 114 th -116 th sessions of the U.S. Congress (see Table 1). The 114 th session (January 3, 2015-January 3, 2017) was during the second term of Barack Obama's Presidency. Arne Duncan was the U.S. Secretary of Education till January 1, 2016, succeeded by John King Jr. (January 1, 2016-January 20, 2017). The 115 th session (January 3, 2017-January 3, 2019) was during Donald J. Trump's Presidency, and Betsy DeVos has been the Secretary of Education since February 7, 2017. The 116 th session (January 3, 2019-present) is the current session of the U.S. Congress, during the third and fourth year of Trump's Presidency. There was a Republican majority in both bodies of Congress during the 114 th and 115 th Congresses. In the 2018 midterm election, the Democratic Party won a majority in the House, while the Republican Party increased its majority in the Senate. However, no congressional hearings on ESSA had been held during the 116 th session of the U.S. Congress, at the time of writing this article. As a result, all ESSA congressional hearings analyzed in this study were conducted in the 114 th and 115 th Congresses with a Republican majority in both House and Senate.
A search for all congressional hearings that discussed ESSA was conducted through the website of the U.S. Government Printing Office (GPO), which makes hearing transcripts available for the public record (U.S. Government Publishing Office, 2019). Using the search terms education, I first identified all of the hearings on education during the 114 th -116 th sessions of the U.S. Congress (2015-present). A total of 28 hearings on education at the Senate Committee on Health, Education, Labor & Pensions and 44 hearings at the Education & Labor Committee at the House were identified. Next, the content of each hearing was reviewed to confirm that the focus of each hearing was on ESSA. After this review, a total of 30 testimonies from eight hearings in the House (10 testimonies) and the Senate (20 testimonies) were found to focus on ESSA implementation (see Table 1). The transcripts of these 30 testimonies and statements were then retrieved from the GPO website and compiled as the ESSA Dataset for analysis.

Discourse Network Analysis
Discourse network analysis was performed to examine the ESSA Dataset. All of the 30 testimonies were organized according to when the hearings were held and which witnesses testified. Within each testimony, a policy claim is coded by three variables. First, the actor a  A = { 1 , 2 , … , } is a person or an organization who testified in the hearings. In this study, the policy actors were those who spoke in the hearings on ESSA. According to the ESSA Dataset, policy actors of ESSA fall into eight categories: (1) federal administrative and executive offices (e.g., the Department of Education and the Government Accountability Office), (2) state administrative and executive offices (e.g., governors and state superintendents), (3) teachers unions (e.g., the American Federation of Teachers and the National Education Association), (4) education interest groups (e.g., the Education Trust, the Alliance for Excellent Education, and the Migration Policy Institute), (5) district superintendents, (6) teacher, (7) education professor, and (8) Congress members. It is worth noting that many policy actors belong to more than one group. For instance, Gail Pletnick, who testified at the Senate and House hearings, was both the superintendent of the Dysart Unified School District in Arizona and president of the School Superintendents Association. Gary Herbert, who testified at the Senate hearing, was both the governor of Utah and the National Governors Association. In his testimony, Herbert stated explicitly that he "testify today on behalf of the National Governors Association and the people of Utah" (Herbert, 2016, p. 8). David Steiner, who testified at the Senate hearing, was the executive director of Johns Hopkins Institute for Education Policy, and he also served on the Maryland State Board of Education. Among all witnesses, the lone academic Linda Darling-Hammond was both a professor of education emeritus at Stanford University and president and CEO at Learning Policy Institute.
Second, the claim c  C = { 1 , 2 , … , } is an abstract representation of a policy claim articulated by the policy actors (Leifeld & Eawag, 2017;Hajer, 1995;Koompmans, 1996;Koompmans & Statham, 1999). To identify the claims, I began by reading the hearing transcripts, familiarizing myself with arguments about ESSA implementation. I then identified themes of the arguments, and categorized similar arguments that repeated across the hearing documents. Whenever an argument fell under a claim, the statement was coded; therefore, the coding was deductive. For example, in the testimony on state accountability system, the claims can be "Defining 'consistent' in the 'consistent underperformance' as no more than two years" and "States hold schools accountable for assessing at least 95% of their students and subgroups." The coding was performed by the researcher. While there is no measure of inter-coder reliability in this study, the coded claims were explicitly articulated by actors without much room for speculation.
The third variable in each policy claim is the agreement relation r  R = { 1 , 2 } which is a dichotomous variable of the policy actors' sentiment toward a policy claim (Leifeld, 2017). The agreement relation is "positive" ( 1 ) if the actor refers to the claim in an affirmative way, and is "negative" ( 2 ) if the actor rejects the claim or uses a negative connotation. For example, regarding the claim of "A single summative score of school performance is misleading", some policy actors (e.g., Kentucky Commissioner of Education, district superintendents, teacher, and education professor) agreed with the claim; whereas other policy actors (e.g., U.S. Department of Education and Civil and Human Rights and Education groups) disagreed as they wanted school ratings to be summative. The distinction between the positive and negative sentiment towards a claim is particularly valuable, because it reveals competing coalitions around a policy claim.
After coding the above three variables for all 30 testimonies, I first created an actor-claim bipartite network = (A, C, ) to connect policy actors and their claims. Each tie can be understood as an affiliation tie connecting a policy actor to a claim in a positive or negative way, in short (a, c)  , where denotes the set of ties in relation r. With the 30 testimonies, I constructed the actor-claim bipartite network which had 29 actors and 47 claims. Next, the actor-claim bipartite network was then converted into an actor coalition network in Figure 1 (Borgatti, Everett, & Johnson, 2013;Newman, 2013). In the actor coalition network, the nodes represent the policy actors, and the ties represent the co-occurrences of claims expressed by the actors. For example, if both policy actors-1 and 2 -make the same claim, then there is a tie connecting 1 and 2 in the actor coalition network.
To reveal the structure of the policy actors, the Girvan and Newman's (2002) clustering algorithm was applied by using NodeXL-a social network analysis and visualization software package-to visually map the actor coalitions by illustrating how policy actors clustered in the discourse network. The Girvan-Newman algorithm is a hierarchical agglomeration approach to detect tightly knit clusters in the network so that the nodes within the clusters are densely connected, and the connections between clusters are relatively loose. By using the Girvan-Newman algorithm, clusters of policy actors would suggest policy coalitions through the actors' similar policy claims.

Results
The ESSA actor coalition network, as seen in Figure 1, has 20 policy actors. The policy actors are connected by ties if a pair of actors articulated the same policy claims. The thickness of ties represents the number of shared policy claims. The thicker the ties are, the more policy claims are shared between a pair of policy actors. The actor coalition network includes only the actors whose testimony had at least a policy claim that was shared by another actor. This is because if a policy claim was shared by only one actor, then there would be no tie connecting two policy actors who articulated the same policy claim. In the actor coalition network, the node size represents how central an actor is in the network. The bigger the node is, the more central the actor is in the network. Each actor's degree centrality is displayed in Table 2. Superintendent Pletnick had the highest degree (26), indicating he was the most central actor whose policy claims were articulated by many other actors (e.g., both House and Senate Education Committee Chairmen John Kline and Lamar Alexander, Congressional Asian Pacific American, Black, and Hispanic Caucuses, state education commissioner, other district superintendents, teacher, teachers union, interest groups, and education professor). Specifically, superintendent Pletnick (Arizona) and superintendent Schuler (Illinois) had the strongest tie (tie strength = 5) in the network, represented by the thickest tie in Figure 1. That is, the two superintendents concurrently made the largest number (5) of the same policy claims at ESSA congressional hearings: (1) A single summative score of school performance is misleading; (2) The USDE's proposed timeline for rating schools is rushed; (3) States need to justify if n-size is above 30, (4) USDE's "supplement, not supplant" regulation exceeds its statutory authority, and (5) child welfare agencies and local education agencies should collaboratively cover the transportation cost of foster children.
After applying the Girvan and Newman's (2002) clustering algorithm to the actor coalition network, four clusters of policy actors were detected. Among them, two clusters (represented in green and blue colors) are in one big component-a subset of the nodes connected by at least one tie (Newman, 2013). Moreover, there are two components (represented in black and orange colors), each of which has only two nodes. For the sake of consistency, I termed the clusters and components as groups in the pages that follow. Each group's density is displayed in Table 2. Density is a value ranging from zero to one, calculated by the total number of observed ties divided by the total number of possible ties in the same network. The higher value of a group's density, the higher percentage of observed ties among all possible ties is. In the actor coalition network, Group 3 and 4 Note: The policy actors are represented as nodes. The node size is proportional to degree centrality. The policy actors are connected by ties if a pair of actors articulated the same policy claims. The thickness of ties represents the number of shared policy claims. The thicker the ties are, the more policy claims were shared between a pair of ties. Four colors represent four groups detected by using the Girvan and Newman's (2002) clustering algorithm.
had the highest density (density = 1), meaning all possible ties (one tie connecting the only two actors in Group 3 and 4, respectively) are observed in the groups. Moreover, Group 1 has more actors and higher density than Group 2.
A close examination of policy claims articulated by the policy actors in different groups found that each group appears to have a dominant theme. The policy claims shared by policy actors in Group 1 center around equity issues; the policy claims in Group 2 focus on assessment and accountability; the policy claims in Group 3 are mostly about states have changed/passed legislation to align the state accountability systems with ESSA goals; the policy claims in Group 4 focus on the USDE's state plan approval was inconsistent with the ESSA statutory provisions. Below I detail how the same policy claims shared by different actors have shaped the ESSA policy actor coalition network.

Group 1: Equity
In Group 1 (green nodes on the left in Figure 1), 10 policy actors were densely connected through their shared policy claims. The 10 actors include: a superintendent, a state administrative and executive office (Mississippi Department of Education), USDE, Congress members (e.g., House Committee on Education and the Workforce Chairman John Kline, Congressional Asian Pacific American Caucus, Congressional Black Caucus, and Congressional Hispanic Caucus), interest groups (the Education Trust and Alliance for Excellent Education), and Civil and Human Rights and Education Groups. These actors were connected by their policy claims on equity issues. The policy claim that held most actors together was that the use of super subgroups could not take the place of individual student groups in state accountability systems. A super subgroup refers to an aggregated subgroup of students that combines subgroups of students (e.g., black, Hispanic and Native students). The then-Secretary of Education John King asserted that each performance indicator needed to be measured "separately for each subgroup of students" [ESSA Sec. 111(c)(4)(B)] for accountability purposes. House Education Committee Chairman specifically noted the importance of disaggregating data on the subgroup of English language learners who were also students with disabilities (Kline, 2016). Likewise, Congressional Asian Pacific American, Black, and Hispanic Caucuses stated that prohibiting the use of super subgroups would help ensure that the academic performance of disadvantaged students could not be masked by average. Interim Vice President for Government Affairs and Communications of the Education Trust, Daria Hall, agreed that "all indicators in the accountability system [should] be disaggregated by each group of students, so schools can't sweep the performance of some students under the rug" and "super [sub]groups can't take the place of individual student groups, so progress among one group can't mask stagnation or declines for another" (Hall, 2016, p. 70). This policy claim was echoed by another interest group Alliance for Excellent Education. Its Vice President of Policy Development and Government Relations, Phillip Lovell, testified that aggregating subgroups of students together violated ESSA and ran the risk of "inadequately supporting historically underserved students" (Lovell, 2017, p. 5). A total of 31 civil and human rights and education groups unequivocally stated that super subgroups must not be allowed (U.S. House Committee on Education and the Workforce, 2016a).
Moreover, the House of Representatives in Congressional Asian Pacific American, Black, and Hispanic Caucuses formed a strengthened coalition through four additional claims: (1) States and districts take timely action in any school with at least one low-performing student subgroups; (2) States hold schools accountable for assessing at least 95% of their students and subgroups; (3) USDE needs to clarify the "supplement, not supplant" provision, and (4) States need to justify if nsize of a subgroup of students is above 30. However, the last claim ran against the policy claim articulated by the Civil and Human Rights Groups who believed that a maximum n-size of 30 was too high (U.S. House Committee on Education and the Workforce, 2016a). Despite the differences in the claim about the n-size, the Civil and Human Rights Groups was still in the same group with Congressional Asian Pacific American, Black, and Hispanic Caucuses, because they agreed on more claims than their disagreement on the n-size.

Group 2: Assessment and Accountability
Group 2 (blue nodes in Figure 1) is the second-largest group, composing of six policy actors. They include a state administrative and executive office (e.g., Kentucky Department of Education), a district superintendent, a math teacher, an education professor, Senate Committee on Health, Education, Labor & Pension Chairman Lamar Alexander, and teachers union (American Federation of Teacher). Most of the policy claims in this group center around assessment and accountability. Specifically, the policy claims that shaped the formation of Group 2 include (1) a summative score for school performance, (2) supplement, not supplant regulation, (3) assessment participation rate, (4) rushed timeline to rate schools, (5) multiple indicators in the accountability systems, and (6) identification of consistently underperforming schools.
A summative score for school performance. One policy claim that shaped the coalition formation was whether a summative score was needed for school performance ratings. ESSA requires that a state, on an annual basis, meaningfully differentiate its schools using all the indicators in the state accountability system (ESSA, Section 1111(c)(4)(C)). Yet in the USDE's proposed regulation (Section 200.18 (4)), states were required to provide a single summative rating from multiple measures of school performance (U.S. Department of Education, 2016, p. 34601). The then-Secretary of Education John King justified it by arguing that a summative rating for a school was needed, because it would be used by states to identify the schools at the bottom 5% for intervention (King, 2016b). However, many policy actors testified against such a summative score of school performance. The education professor argued that a single summative score would "limit a state's ability provide the data needed for schools and states to act wisely and well on behalf of the students and families, while hindering the ability of parents and community members to advocate wisely and well on behalf of their children" (Darling-Hammond, 2016, p. 3). Superintendent Pletnick testified that reliance on a summative score provides a distorted view of the strengths and weaknesses of programs and practices in our schools, and can be misleading. In reading a report card, many readers look at the summative indicator and move on, and that one score does not provide a complete picture (Pletnick, 2016, p. 1-2).
Likewise, Kentucky Commissioner of Education asserted that While the proposed regulations claim to replace NCLB's narrow definition of school success with a more comprehensive picture of school performance, the requirement of a single summative score seems to go well beyond what the statute calls for and would limit states' ability to leave data at a dashboard level, which is a broader, fairer and more accurate representation of school performance (Pruitt, 2016, p. 53).

Supplement, not supplant regulation.
Another policy issue that unified many policy actors (e.g., superintendents, Kentucky Commissioner of Education, and Congress members) was about the "supplement, not supplant" provision. The USDE has a long-standing supplement, not supplant rule that requires Title I funds not to be used to replace the state and local funds a local education agency would have spent in a Title I school if it did not participate in Title I. ESSA retains the supplement, not supplant rule, but prohibits USDE from prescribing the specific methodology a local education agency uses to allocate state and local funds. Congressional Asian Pacific American, Black, and Hispanic Caucuses asked USDE to clarify the compliance with ESSA's supplement, not supplant provision to ensure that states and districts are fulfilling the intent of Title 1: "to provide all children significant opportunity to receive a fair, equitable, and high-quality education, and to close educational achievement gaps" (U.S. House Committee on Education and the Workforce, 2016a, p. 96).
In comparison, the responses from both Senate and House Education Committee Chairmen Lamar Alexander and John Kline were much stronger. They referred to a non-partisan report which stated that a "legal argument could be raised that [USDE] exceeded its statutory authority if it promulgates the proposed [supplement, not supplant] rules in their current form" (Congressional Research Service, 2016, p. 9). This is because ESSA states that nothing in Title I shall be construed to mandate equalized spending per-pupil for a state, local education agency, or school. Yet the USDE's proposed regulation required that the methodology for allocating state and local funds to schools would result in the local education agency spending an equal or greater amount per-pupil in its Title I schools than the average amount it spends per-pupil in its non-Title I schools. Kentucky Commissioner of Education shared his concern that the USDE's regulatory proposal on Title I's supplement, not supplant requirement exceeded the scope of ESSA (Pruitt, 2016). Pruitt further argued that such regulation would force districts to transfer teachers to comply, "place existing state and local initiatives to promote diverse public schools at risk of noncompliance, and penalize states and districts that use a weighted funding methodology" (p. 57). Superintendent Pletnick specified that With teacher salaries[as] the largest expenditure in a school district, it is a false premise to require schools to use teacher salaries in evaluating compliance with supplement, not supplant provisions, as it is a policy built on the false assumption that teacher salaries are a single indicator that can meaningfully and reliably be used in an undisputable manner to indicate effectiveness and quality in programs within Title I schools. That thinking is flawed on many levels. Perhaps most importantly, it assumes states and schools across the nation employ one single approach to determining teacher salaries. This is not the reality (Pletnick, 2016, p. 4-5).
Superintendent Schuler shared the concern that the USDE's proposed regulation blurs[ed] the line between two distinct and equally important statutory provisions, both targeted at maintaining the integrity of Title 1 dollars. … a push to equalize spending assumes that a school can control for all factors that teachers consider when deciding where to apply (Schuler, 2016, p. 77).
Assessment participation rate. Many policy actors (e.g., teacher, teachers union, superintendents, and academics) formed a coalition around the policy claim that USDE should not dictate how states factor the 95% of participation rate in statewide assessment into a school evaluation. While ESSA does require at least 95% of public school students participating in annual state assessments of student achievement, it grants control to each state over how the 95% participation rate is factored into the statewide accountability system (ESSA, 2015). Cassie Harrelson, a math teacher, stated that the proposed regulations go beyond by dictating the actual consequences that schools must face. Instead of punishing districts, we should be helping districts find solutions to solve the lower participation rate and our assessments. And again, how to handle lower test participation rates was supposed to be determined at the State level and once again the Department is taking away that opportunity (Harrelson, 2016, p. 60).
Randi Weingarten, president of American Federation of Teachers, also argued that USDE should not threaten states of losing funding if they cannot meet the test participation rate (Weingarten, 2016). Superintendent Pletnick stated that the USDE's proposed regulations did not allow meaningful input from the local level (e.g., state and community) to address the assessment participation rate. Instead, the USDE's proposed regulations were the equivalent of saying you can paint the house any color you wish, as long as it is green. So what option do you choose? This is not true flexibility and runs counter to ESSA's framing principles, empowering state and local education agencies in their work to provide all students with educational opportunities of ESSA (Pletnick, 2016, p. 3).
Darling-Hammond also testified that The proposed regulations outline very specific consequences to be applied when there is a participation rate of less than 95% for any group in any school. … A number of officials and educators have indicated that these approaches could backfire and cause greater challenges for them as they seek to build a culture of engagement in new assessments and systems. Encouraging states to determine and clearly articulate how they will factor the requirement for 95% participation in assessments without federally prescribed sanctions will likely better help address the previous misuse of and current responses to high-stakes testing (Darling-Hammond, 2016, p. 3-4).
A rushed timeline to rate schools. Another policy claim that held the coalition together was that the USDE"s proposed timeline for rating schools is rushed. This claim was articulated by superintendents, a state education commissioner, an education researcher, and an interest group (the Education Trust). Under ESSA (Section 5(e)(1)(B)), states which receive Title I funding must develop and implement a single, statewide accountability system beginning with the school year 2017-18. However, the USDE's proposed regulation (Section 200.19(d)(1)) required states to use data available in 2016-17 that was generated under the previous accountability system to identify schools for comprehensive and targeted support and improvement (U.S. Department of Education, 2016). This proposed regulation was criticized because "if these schools are re-identified for 2017-18 and the subsequent two years, other schools with [the] chronically low performance for students overall may not get comprehensive support and improvement until 2020-21" (Hall, 2016, p. 71). The Kentucky Commissioner of Education also testified that "Implementing a new accountability system in 2017-18 is already a monumental task on an aggressive timeline, and I have concern that states will be able to implement new systems that take full advantage of ESSA by the 2017-18 school year" (Pruitt, 2016, p. 52-53). Superintendent Pletnick argued that "The proposed timeline will rush the implementation of accountability system decisions and may result in some schools, in the first year of ESSA label implementation, being identified as failing based on 2016-17 data. That data may be more aligned to NCLB mandates" (Pletnick, 2016, p. 3). The education professor also asserted that it was important to "ensure sufficient time to implement thoughtful and effective accountability systems which incorporate stakeholder feedback and have the capacity to drive effective strategies for improvement in schools" (Darling-Hammond, 2016, p. 3).

Multiple indicators in the accountability systems.
How to weigh multiple indictors in the accountability system is another fissure between USDE and other policy actors. ESSA Section 1111(c)(4)(C)(ii) specifies that each academic indicator has to receive "substantial" weight in the state's accountability system, and that "much greater weight" than the school quality indicators in the aggregate. USDE acknowledged that ESSA allowed states to "build accountability systems that build in questions like are students progressing in science or social studies, are students getting access to art and music" (King, 2016a, p. 37). However, both Kentucky Commissioner of Education and the education professor testified that the USDE's proposed regulations went beyond the scope of the statute, negating "a state's ability to determine the impact that 'substantial' and 'much greater' weights have in the overall accountability system" (Pruitt, 2016, p. 54). Moreover, these regulations should not restrict state options for weighting and using these additional measures in meaningful ways to add to the information that is used to examine school success. This should include the meaningful use of extended-year graduation rates in state accountability systems, which incentivize schools to keep in, rather than pushing out, students who cannot graduate in 4 years and to re-attract those who have left (Darling-Hammond, 2016, p. 3).

Identification of consistently underperforming schools.
Other claims that frequently occurred in Group 2 centered around how to identify underperforming schools to provide comprehensive support and improvement. ESSA Section 1111(c)(4)(C)(iii) provides that each state must establish and describe in its state plan a methodology to identify schools for targeted support and improvement and leaves the determination of consistently underperforming up to states. The USDE's proposed regulation (Section 200.19(c)(1)), however, defined consistently underperforming as failing to make progress for two years. The Civil and Human Rights and Education Groups agreed with this definition (Civil and Human Rights and Education Groups, 2016, p. 92). However, Kentucky Commissioner of Education testified that "the regulation oversteps the bounds of the statutory language which leaves the definition of consistently underperforming up to the states" (Pruitt, 2016, p. 55).

Group 3: From ESSA to State Legislation
Group 3 (black nodes in Figure 1) has a relatively small size. The limited number of hearings with a focus at the state level led to a rather small number of policy actors in Group 3 (From ESSA to State Legislation) and 4 (USDE's State Plan Approval) in the discourse network. There were only two actors-a dyad, to be precise-in Group 3 and 4. Nevertheless, Group 3 is composed of the heads of two states' administrative and executive offices, including Superintendent of South Carolina Department of Education Molly Spearman and Commissioner of Nebraska Department of Education Matthew Blomstedt. They both testified at the Senate's hearing on "The Every Student Succeeds Act: States Leading the Way" on September 25, 2018. At this hearing, they both claimed that their states have changed and/or passed legislation to align the state accountability systems with ESSA goals. To align ESSA with state statute, in April 2018, Nebraska passed LB 1081 (Change education provisions regarding reporting, penalties, residency, boundaries, priority schools, subpoena authority, poverty, and limited English proficiency and adopt the Nebraska Reading Improvement Act; Nebraska Legislature, 2018). South Caroline General Assembly put forward an outcome-based system, including the Profile of the South Caroline Graduate-a document that outlines the knowledge, skills, life, and career characteristics that every graduate should possess (Spearman, 2018).

Group 4: USDE's State Plan Approval
Like Group 3, Group 4 (orange nodes in Figure 1) is also composed of two policy actors, including Secretary of Delaware Department of Education Susan Bunting and President of Education Reform Now Shavar Jeffries. They both claimed that USDE's state plan approval was inconsistent with the ESSA statutory provisions. USDE did not initially approve Delaware ESSA plan which included science and social studies proficiency as indicators of academic achievement in the state accountability system, because USDE's understanding of ESSA was that only reading and math test results should count as academic indicators, and should have much greater weight in rating schools. To have its ESSA plan approved, Delaware was forced to relegate science and social studies proficiency to another part-School Quality section, instead of originally proposed sections on academic achievement and progress-of its accountability system (Bunting, 2018). Moreover, president of Education Reform Now presented in the hearings a list of more than 30 sections of ESSA that was not followed by state ESSA plans approved by the USDE (Jeffries, 2018). One of the inconsistencies between ESSA and USDE's approved state plans was the ESSA's requirement that all academic indicators used in state accountability systems are the same ones used statewide for all children. However, in Arizona and New Hampshire, individual school districts were allowed to violate ESSA by choosing which assessments to administer (Kline, 2017).

Discussion
Applying the discourse network analysis, this study analyzed the ESSA discourse in congressional hearings. For the first time, discourse network analysis was applied to map and identify coalitions of policy actors in ESSA in congressional hearings. Many coalitions center around the two goals of ESSA-accountability and autonomy (U.S. Committee on Education & Labor Republicans, 2017). The salience of accountability and autonomy issues in congressional hearings can be attested by the coalitions identified in this study. The findings of this study have substantial policy and methodological implications.

Power Shift from USDE to State and Local Education Agencies
One policy implication is the unfolding tug of war as the power shifts from USDE to state and local education agencies. ESSA aims to restore states' and local leaders' control over and increase flexibility on school accountability, teacher quality, and school improvements (U.S. House Committee on Education and the Workforce, 2016). The shift of power from USDE to state and local education agencies has been a bumpy road since the passage of ESSA. When the then-Secretary of Education John King testified in Congress in 2016, he was reminded multiple times by House Education Committee Chairman about his predecessor Arne Duncan's "smart-lawyer remarks" (Kline, 2016, p. 15). When Duncan was asked about ESSA narrowing the Secretary's authority in an interview with POLITICO Pro in December 2015, Duncan responded, the final thing is we have every ability to implement, to regulate the law. Philosophically, I agree with a bunch of the stuff, and candidly, our lawyers are much smarter than many of the folks who were working on this bill. There is some face-saving things you give up, some talking points that you give up, which we always do because we are focused on substance, and we have every ability to implement. That is all I have ever wanted (Duncan, 2016, p. 15).
In addition to Duncan's initial resistance, although USDE could use regulations to provide clarity on issues that were not clearly specified in ESSA (King, 2016), USDE used its proposed regulations as an instrument of power grip. Many of the policy claims in Group 2 of assessment and accountability manifest the tension between USDE and other policy actors. For example, the USDE's proposed supplement, not supplant regulation was criticized by both Senate and House Education Committee Chairmen who believed that USDE exceeded its statutory authority (Congressional Research Service, 2016). Another example is how to hold states and districts accountable if less than 95% of students participate in statewide assessment. ESSA allows states to decide how the 95% participation rate is factored into the statewide accountability system (ESSA, 2015). However, USDE's proposed regulations had specific consequences of low participation rate in state accountability systems. Moreover, the approval of state ESSA plans has been another instrument used by USDE to retain its power. ESSA allows states to use "student access to and completion of advanced coursework" as an indicator of school quality and student success [Section 1111(c)(4)(B)(v)(V)]. However, USDE criticized a state's proposal to incorporate performance in Advanced Placement (AP) and International Baccalaureate (IB) exams in this indicator (Lovell, 2017). Another example is that Secretary DeVos strongly encouraged states to "think out of the box" (Bunting, 2018, p. 4), However, when Delaware thought of the box and proposed to include science and social studies proficiency as part of student achievement, its state ESSA plan failed to receive USDE's initial approval.
Amidst the power shifts from USDE to state and local education agencies, Congress has played a critical role. On the one hand, House Education Committee Chairman stated that USDE should behave like "a partner, not a dictator" (Kline, 2016, p. 3). The Senior Republican of the House Committee on Education and Labor, Virginia Foxx, stated that "ESSA has stripped away powers of the Department of Education, such as the ability of the Secretary of Education to legislate through executive fiat, or the ability of the Department's bureaucrats to substitute their judgment for states" (U.S. Committee on Education & Labor Republicans, 2017, para. 8). On the other hand, the incumbent House Education Committee Chairman Bobby Scott told state school chiefs that ESSA's flexibility is not "a blank check" (Klein, 2019, para. 1). With the ongoing implementation of ESSA at state and local district level, it remains to be seen how the power shift from USDE to state and local education agencies will continue to unfold. Over the years, federal policymakers and presidents have increasingly discussed education as a national priority. The findings in this study suggest that most of the bumpy ride of power shift took place when Obama was in the White House with Arne Duncan and John King as the Secretaries of Education. By contrast, Betsy DeVos, as the Secretary of Education in the Trump administration, encouraged states to embrace the opportunities and flexibility afforded by ESSA in an innovative way. Unlike USDE under the leadership of Duncan who used NCLB waiver to exert power over states, DeVos mostly declined to aggressively use the USDE's power of approving of states' proposed ESSA plan as an opportunity to assert more federal power (Ujifusa, Blad, & Burnette II, 2019). Yet in the early days of the 2020 presidential campaigns, Pete Buttigieg, the mayor of South Bend, Indiana, described education as a national security issue (Stoll, 2019). Is this a signal of gaining more control and power over education policy? It remains to be seen how the 2020 election will influence ESSA implementation in the future.

Methodological Implications
On the methodological front, the analytical techniques of discourse network analysis employed in this study add to the methodological repertoire of analyzing policy discourse. This study demonstrates a network analytical approach to examine qualitative data in education policy research. A coalition can be defined in many ways, including, but not limited to, supporting financially, sharing resources, and coordinating strategies and campaigns (Wang, 2020). In this article, grounded in the conceptual framework of the advocacy coalition framework, a coalition was theoretically defined through actors' claims that help to bind coalition together (Sabatier, 1998). There is even evidence to indicate that actors with similar beliefs tend to coordinate with one another (Zafonte & Sabatier, 1998). Building on this theoretical definition, a coalition is then operationalized through discourse network analysis which uses network structures to reveal how similar the claims articulated by policy actors. Conceptualizing policy discourse as network structures, this study considers policy actors as the nodes which are connected by their policy claims on ESSA. Most empirical studies grounded in the advocacy coalition framework focus on beliefs rather than network structures; whereas many network analysis studies on policymaking fail to provide sufficient information about the content of the analyzed policies (Ingold, 2011). In this article, discourse network analysis is a fitting analytical tool that uncovers not only network structure of coalitions in congressional hearings on ESSA, but also reveals how the views and beliefs became congruent and help forge coalitions among policy actors. The discourse network analysis performed in this study provides a network analytical approach to studying policy discourse. The contested policy claims across the coalitions, along with the consensus views within the coalitions, provide insight into the education policymaking on ESSA.

Limitations and Suggestions for Future Inquiry
This study has two major limitations. First, the data collected for this study might not provide a full, comprehensive view of ESSA. This study analyzed text data from testimonies in the U.S. Congress rather than full-text archives of news media or social media. The data sources for this study are congressional hearings, suggesting the policy actors and coalition ties identified in this study were influential enough to be invited by U.S. Congress to testify. In addition to influence, witnesses need to contribute to the policy discourse in a way that is aligned with legislators' preferences (Perna et al., 2019). As a result, the discourse at congressional hearings was different from the one on social media, particularly Twitter. Compared to policy actors in the discourse at ESSA congressional hearings, academic researchers' presence on Twitter was minuscule. None of the education researchers frequently posted tweets about ESSA. Nor were they frequently mentioned or replied to in tweets about ESSA. Meanwhile, other ESSA policy actors were as active on Twitter as they were at congressional hearings. Moreover, the topics in the ESSA discourse on Twitter were broader than those in congressional hearings. For example, in addition to the topics of equity, accountability, teaching and learning, and funding, Twitter users discussed other topics that were not discussed at Congressional hearings, including Common Core, school choice, higher education, bullying, and educational leadership (Wang, 2019). More diverse data sources are therefore recommended for future inquiry.
Second, this study offers only a snapshot of ESSA. It is of paramount importance to longitudinally examine how the coalitions evolve over time as a response to policy changes about ESSA. It is possible that policy actors of ESSA implementation were also involved in the policy development relevant to ESSA. Notably, interest groups allied with powerful congressional committees to push their policy agenda (McGuinn, 2006). All three interest groups in Figure 1-the Education Trust, Alliance for Excellent Education, and Education Reform Now-have been involved in education policy issues and retained registered lobbyists to promote their federal legislative agenda. The two interest groups in Group 1 (i.e., the Education Trust and Alliance for Excellent Education) have strong equity agendas (McDonnell & Weatherford, 2013), and they "convene policymakers and provide briefings and fact sheets to legislators" (Reckhow & Snyder, 2014, p. 189). Moreover, the discourse network of ESSA congressional hearings is still evolving. By the end of 2018, there had been only two hearings focusing on ESSA implementation at the state level. The limited number of hearings with a focus at the state level led to a rather small number of policy actors in Group 3 (From ESSA to State Legislation) and 4 (USDE's State Plan Approval) in the discourse network. If more hearings are held in the future, it is possible that the size of the discourse network will continue to evolve. Technically speaking, ESSA will expire at the end of federal fiscal year of 2020, because the authorization period for ESSA is only four years from 2017 to 2020. To date, Congress has prioritized revamping special education and higher education laws over reauthorizing ESSA. Recently, Senate Committee on Health, Education, Labor & Pension Chairman Lamar Alexander said in an interview that he expected "ESSA to be the law of the land for many years even after its authorization period expires" (Ujifusa et al., 2019, para. 8). It is therefore crucial to investigate how the coalitions change over time. This is a necessary step for understanding longitudinal variation over the ESSA implementation process.

About the Guest Editors
Emily M. Hodge Montclair State University hodgee@montclair.edu Emily M. Hodge, Ph.D., is an assistant professor in the Department of Educational Leadership at Montclair State University. She received her PhD from the Department of Education Policy Studies at the Pennsylvania State University. Her work uses qualitative methods and social network analysis to understand the changing nature of strategies for educational equity. Recent projects have explored how educational systems, schools, and teachers negotiate the tension between standardization and differentiation in the context of the Common Core State Standards, and the varied strategies state education agencies are using to support standards implementation.

Joshua Childs
The University of Texas at Austin joshuachilds@austin.utexas.edu Joshua Childs is an assistant professor of Educational Policy and Planning (EPP) in the Department of Educational Leadership and Policy. Joshua received his PhD in Learning Sciences and Policy at the University of Pittsburgh. Joshua's research focuses on the role of interorganizational networks, cross-sector collaborations, and strategic alliances to address complex educational issues. Specifically, his work examines collaborative approaches involving community organizations and stakeholders that have the potential to improve academic achievement and reduce opportunity gaps for students in urban and rural schools.

Wayne Au
University of Washington, Bothell wayneau@uw.edu Wayne Au is an educator, activist, and scholar who focuses on issues of race, class, and power in schooling. He is a professor in the School of Educational Studies at the University of Washington-Bothell, where he currently serves as dean of diversity and equity. Au is an editor of the social justice teacher magazine Rethinking Schools and the author or editor of numerous other publications, including Readers are free to copy, display, distribute, and adapt this article, as long as the work is attributed to the author(s) and Education Policy Analysis Archives, the changes are identified, and the same license applies to the derivative work. More details of this Creative Commons license are available at https://creativecommons.