Politics of Policy: Assessing the Implementation, Impact, and Evolution of the Performance Assessment for California Teachers (PACT) and edTPA

Summative performance assessments in teacher education, such as the Performance Assessment for California Teachers (PACT) and the edTPA, have been heralded through polices intended to enhance the quality of the teaching profession and raise its stature among other professions. However, the development and implementation of the PACT, and subsequently the edTPA, have not been without controversy and debate. The purpose of this article is to assess the implementation, impact, and evolution of the PACT and edTPA. To do so, we review the growing body of literature on the impact and implementation of the PACT and critically analyze the state policies surrounding the edTPA. We raise questions about the practical and policy implications of the PACT’s evolution, as a statewide assessment, to the edTPA, a nationally available assessment.


Introduction
Over the past twenty years, a "remarkable amount" of teacher education reform policy has been proposed, piloted, and implemented at the national, state, and local levels, causing what Darling-Hammond (2010) has called "perhaps the best of times and worst of times" (p.35) for teacher education.Many of these reform policies have promoted increased accountability for teacher education as well as rigorous standards for teacher performance.In particular, performance-based assessments of preservice teachers have been heralded as one policy intended to enhance the quality of the teaching profession and raise its stature among other professions (Darling-Hammond & Hyler, 2013;Wei & Pecheone, 2010).These calls are based on evidence of the potential of performance assessments to improve teacher learning, teacher quality, and, ultimately, student achievement (Boyd et al., 2006;Darling-Hammond, 2010b;Wei & Pecheone, 2010)  1 .
As a result, many state departments of education and national accrediting agencies (e.g., CAEP, 2013) have passed and implemented policies in support of performance-based assessments in teacher education (Arends, 2006).Of particular note is the Performance Assessment of California Teachers (PACT) that was developed as part of a consortium of teacher education institutions in response to California legislation SB 2042 (Merino & Pecheone, 2013;Pecheone & Chung, 2006).The PACT 2 is a subject-specific portfolio that assesses teacher candidates' ability to plan, teach, assess, and reflect on 3-5 hours of instruction across four integrated tasks (planning, instruction, assessment, and reflection) through multiple sources of evidence including lesson plans, teacher artifacts, student work samples, video clips of teaching, and personal reflections (Lombardi, 2011;Pecheone & Chung, 2006).The PACT represented a historic collaboration among teacher educators across more than 12 California institutions to design, develop, and implement a performance assessment that met the AERA/APA/NCME (1999) standards for testing (Merino & Pecheone, 2013).
Based, in part, on the PACT's success, professors and researchers at Stanford University developed an "updated," nationally accessible version of the PACT, called the edTPA 3 (Darling-Hammond, 2010, p. 44).Together, they formed a partnership among the Stanford Center for Assessment, Learning, and Equity (SCALE), national organizations such as the American Associate for Colleges of Teacher Education (AACTE), and Pearson Education, a for-profit publishing company, to jointly field test and implement the edTPA on a national scale (AACTE, 2014;Sato, 2014).Rapid program and state adoption and implementation of the edTPA followed.As of 2014 more than "500 educator preparation programs in 34 states and the District of Columbia" (AACTE, 2014) have implemented the edTPA.
The development and implementation of the PACT, and subsequently the edTPA, have not been without debate.Specifically, scholars have criticized the content of the assessments (what is included, what is left out) (e.g., Cochran-Smith, Piazza, & Power, 2013) and the implementation of the assessments (e.g., use of external scorers) (Au, 2013).This debate has extended across teacher education faculty, programs, local and national organizations and into the public media and blogosphere (e.g., Ravitch, 2013).
The purpose of this article is to assess the implementation, impact, and evolution of the PACT and edTPA.To do so, we synthesize the growing body of literature surrounding the PACT and critically analyze state policies involving the edTPA to address three overarching questions: 1. What does the literature say about the implementation and impact of the PACT? 2. How do state policies on the edTPA define the problem of teacher education?3. What are the policy and practical implications of the evolution of the PACT and edTPA?
1 For a comprehensive review of the literature on performance assessments in teacher education, see Wei & Pecheone (2010).In this article, we analyze the implementation, impact, and evolution of two prevalent performance assessments in teacher education, the Performance Assessment for California Teachers and the edTPA, not specifically documented in Wei & Pecheone (2010). 2 For purposes of this article, we operationalize the PACT assessment as the "teaching event" or the set of tasks included in the capstone assessment.Unless noted, we do not include the "embedded signature assessments," in our description of the PACT. 3 Building on the PACT model, the edTPA "requires candidates to use teaching materials (e.g.lesson plans, assignments, video and classroom assessments of learning) to evaluate their teaching and student learning" ("edTPA Minnesota," 2015).The edTPA includes additional subject specific assessments, as well as modified tasks and rubrics.

Conceptual Framework
To guide our work, we draw on Cochran-Smith, Piazza, and Power's (2013) "politics of policy" framework.As Cochran-Smith and colleagues argue, "policies governing teacher education are not developed and enacted at a single level by a single agency, but at multiple levels and by many actors" (p.8), often with competing agendas.These actors include federal, state, and local agencies, national and state accrediting bodies, professional and research organizations, higher education institutions, advocacy groups, centers, and, in some cases, for-profit companies.Furthermore, they operate in what Cochran-Smith and colleagues (2013) describe as a "policy web" (Joshee & Johnson, 2005;Joshee, 2007Joshee, , 2009)), which emphasizes the interconnected roles of those involved in developing and implementing policy, and a "policy cycle" (Bowe, Ball, & Gold, 1992) of competing contexts in which policies are remade and reworked.
The "politics of policy" framework presents four interrelated components: 1) "discourses and influences," defined as broader political, economic and social conditions that shape teacher education policy; 2) "constructions of the problem of teacher education," that articulates the major arguments and actors behind policy; 3) "policy in practice," that represents how policy is written and interpreted; and 4) "impact and implementation," that explores the broader outcomes of policy.Together, these components highlight the various "interactions of…state control, professional influence, public advocacy and local agency" (Cochran-Smith et al., 2013, p. 9) within and across teacher education policy.
Cochran-Smith and colleagues (2013) highlight overarching discourses and influences prevalent in current reform initiatives in teacher education, particularly noting the discourses of neoliberalism and human capital (e.g., Apple, 2006;Hill, 2007;Leyva, 2009), as well as the discourse of accountability in teacher education that, as many have noted, has shifted from accountability of inputs to accountability of outcomes (Cuban, 2004).As Cochran-Smith and colleagues (2013) note, the discourses and influences are used to frame the conversation and debate around policy in teacher education and have been "pervasive and completely normalized both within the professional/ university teacher education community and outside it" (p.12).
While these normalized discourses and influences serve to frame our work, they are not the focus of our analysis.Following Cochran-Smith and colleagues' (2013) interrogation of teacher education policies, in our analysis, we focus on three components of the "politics of policy" framework to analyze the implementation, impact, and evolution of the PACT and edTPA.First, we explore the "policies in practice" and "impact and implementation" of the PACT, through a review of the literature on the PACT.Then, we unpack various "constructions of the problem of teacher education" through an examination of state-level policies pertaining to the edTPA.In doing so, we extend Cochran-Smith and colleagues' work, critically analyzing the evolution of the PACT and the edTPA.We analyze how these assessments are positioned "within the complex and contested space of teacher education accountability writ large" (Cochran-Smith et al., 2013, p. 10), and we investigate in what ways the spirit of and lessons learned from the PACT, as a statewide assessment, were taken up in the implementation of the edTPA, as a nationally available assessment.

Methods and Data Sources
To examine the "policies in practice" and "impact and implementation" of the PACT, we reviewed the published literature on the PACT between 2005 and 2014.This period of time was selected because it follows initial piloting of the PACT in California (2002-2003, 2003-2004), when the first empirical studies were published (i.e., Chung, 2005).This review covers publications from a variety of theoretical and methodological perspectives from peer-reviewed journal articles, dissertations, technical and policy reports from established centers (e.g., SCALE, 2013), and edited chapters.Specifically, we conducted a search of:  Major scholarly databases including Educational Resource Information Center (ERIC) and Google Scholar using the search terms "performance assessment" or "PACT" or "edTPA" or "teacher education" or "teacher preparation";  Publications that were available on relevant center and organization websites, including the PACT website (PACT, 2014), the SCALE website (SCALE, 2014), and the AACTE edTPA website (AACTE, 2014);  Publications (identified via electronic hand search) from peer-reviewed journals in teacher education including the Journal of Teacher Education, Teacher Education Quarterly, Teaching and Teacher Education, New Educator, Action in Teacher Education, and Issues in Teacher Education;  Relevant papers presented at the American Educational Research Association national conferences that were available in the repository or by permission from the authors; and  More than 120 articles and publications that reference Pecheone & Chung (2006), the most widely-cited article on the PACT.
We began by examining more than 80 abstracts from publications on performance assessments in teacher education.From there, we closely examined publications that specifically identified the PACT, and we included in our review more than 50 publications that include descriptive, conceptual, and empirical pieces that directly address the PACT.
To analyze states' policies on the edTPA, we identified the six states-California, Iowa, Minnesota, New York, Tennessee, and Washington-that adopted and enacted policies at the state level explicitly incorporating the edTPA as of 2014.These policies are currently in practice and demonstrate a range of how, for whom and why PACT and/or edTPA are used in state decisionmaking.We examined state policy documents that were publicly available (e.g., California SB 2042), state websites (e.g., Iowa Department of Education, 2015), as well as official state and national edTPA websites (e.g., AACTE, 2014;"edTPA Minnesota," 2015;SCALE, 2014) that described state policies around the edTPA.
We used a discourse approach to policy analysis (Bacchi, 2000;Sharp & Richardson, 2001;Vidovich, 2001Vidovich, , 2007)), highlighting the iterative, interactive, overlapping, and multi-level nature of policy development and implementation (Cochran-Smith et al., 2013).Within this approach, we frame policy as an "argument-making process" (Stone, 2001) through which multiple actors construct various policy discourses.We recognize the newness of the implementation of these policies and the shifting nature of the state-level policies on teacher education, particularly around cut scores and consequential decisions made as a result of these assessments.Given these limitations, we believe this survey of policies to be relevant and significant because it sheds light on the broader contexts surrounding the emergence of state policies around pre-service performance assessments.Our analysis provides a critical lens with which to view the enactment of state policies and may give states that do not have policies in practice a way to understand what is taking place nationally to contextualize their own response.

Findings: Emerging Tensions in Politics of PACT and edTPA Policies
We organize our findings by focusing on key aspects of the implementation, impact, and evolution of the PACT and edTPA.Acknowledging the existing literature that has examined the policy context of the PACT (e.g., Pecheone & Chung, 2006), in this article we highlight the evolution of the PACT, structured in terms of the "policies in practice" and "impact and implementation" of the assessment.In other words, we explore how the PACT has played out for various actors, including teacher candidates, supervisors and cooperating teachers, and university faculty and programs.
We then turn to an analysis of state policies that explicitly include the edTPA to understand the ways in which the policies and actors involved have constructed the problem of teacher education.In doing so, we explore the purposes of the edTPA at the state level, the rationales behind the assessment, the actors involved in policy adoption, and the initial implementation and enactment of policies surrounding the edTPA.This analysis highlights the ways in which the problem of teacher education has been defined at the state level through the adoption and implementation of the edTPA.
We conclude by looking across the two assessments to analyze factors that may have contributed to the drift from the original PACT spirit of by educators, for educators as they occur in different state contexts involving the edTPA.In particular, we raise questions about how much room each state appears to have allowed for local interpretation and control of the results of the performance assessment system, and to what extent this allowance still results in a valid system.This analysis outlines tensions involved in the implementation, impact, and evolution from the PACT, as a state-level performance assessment, to the edTPA as a nationally available assessment.

PACT: Policies in Practice
The PACT presents a unique example of the development and implementation of a performance assessment within one state's policy context.The impetus for the PACT was teacher educators' dissatisfaction with the performance assessment options outlined in response to California SB 2042 (Pecheone & Chung, 2006).As a result, the PACT was an opportunity for teacher educators who were "motivated by a desire to develop an integrated, authentic, and subjectspecific assessment that is consistent with the core values of member institutions while meeting the assessment standards required by the state" (Pecheone & Chung, 2006, p. 22-23) to proactively take part in the policy-making process at the state and local levels.
The "problem of teacher education" (Cochran-Smith et al., 2013) around performance assessments has been constructed in terms of the professionalization agenda in teacher education (Cochran-Smith et al., 2013;Cochran-Smith, 2001;Darling-Hammond & Hyler, 2013;Zeichner, 2003Zeichner, , 2006) that seeks to elevate the teaching profession through rigorous standards and the involvement of teacher educators (and teachers) as active participants in the making of policy (Peck, Gallucci, & Sloan, 2010).The "problem of teacher education" as taken up by the PACT, included direct involvement and collaboration of teacher educators across the state of California.These developers sought to elevate the status of the profession through a common assessment, in response to a state legislative requirement (Pecheone & Chung, 2006).
From this construction of the problem of teacher education, our review of the literature on the PACT explores ways in which the interpretation and enactment of policies around the PACT affected the experiences of key actors in the California context.In the sections that follow, we parse the experiences of teacher candidates, supervisors and cooperating teachers, as well as university faculty and programs, and we explore how the PACT policy was remade in local contexts.The tensions revealed through the ways these stakeholders interpreted and remade these policies within their local contexts direct attention to key questions around acceptance and resistance, the dynamics of program change, and the (un)intended privileging or marginalizing of perspectives and populations.These tensions provide lessons for the implementation of the edTPA within and across state policy contexts.
Teacher candidate experiences with the PACT.At a general level, studies that addressed teacher candidate experiences with the PACT support the assertion that candidates acquire knowledge about their teaching practice from completing the performance assessment, but that the value of this learning opportunity is mediated in significant ways by the nature and structure of their student teaching experience (Chung, 2007(Chung, , 2008;;Okhremtchouk, Newell, & Rosa, 2013;Okhremtchouk et al., 2009).Candidates, for example, highlight the PACT's positive influence on their ability to reflect deeply on their teaching (Okhremtchouk et al., 2009) and to assess student learning (Chung, 2005(Chung, , 2007(Chung, , 2008;;Pecheone & Chung, 2006).At the same time, these contributions are unevenly mediated by multiple variables, including: when teacher candidates obtain targeted supports to complete the tasks of the PACT (Chung, 2007;Pecheone & Chung, 2006); when and how tasks are scaffolded (e.g., training candidates in use of video to enhance reflection, embedding PACT-oriented tasks into previous coursework) (Okhremtchouk et al., 2009); the extent to which candidates are given feedback on the tasks of the PACT (Chung, 2008;Okhremtchouk et al., 2009); and the extent to which teacher candidates are given opportunities to independently plan, teach lessons, and reflect on their classroom experiences apart from the PACT (Bunch, Aguirre, & Téllez, 2009;Chung, 2007Chung, , 2008;;Liu & Milman, 2013).
Further complicating these conditions are issues beyond the control of teacher candidates, including the freedom to make independent teaching decisions in their student teaching placements in order to optimize what they learn from the PACT teaching tasks (Chung, 2007(Chung, , 2008)).For example, Chung (2007) found that "candidates who reported higher levels of constraints on their teaching decisions (e.g., constraints related to their mentor/cooperating teacher's expectations, district mandated curriculum, required pacing, or required district/departmental tests) were significantly less likely to report learning important teaching skills from the PACT" (p.34) than those with greater independence in their student teaching placements.Other studies highlight the timing of PACT completion as a critical factor that influences candidates' performance on the assessment as well as their overall learning experiences (Chung, 2007(Chung, , 2008;;Lit & Lotan, 2013;Okhremtchouk et al., 2009;Sandholtz & Shea, 2012).Researchers generally agree that the PACT assessment should not reside at the very end of the program or student teaching experience, as this would result in candidates having little or no opportunity for adjustment, adjudication (should the need arise), or continued learning as a result of the PACT.
At issue across most of these studies is the underlying question of how to determine preparedness to teach and at the same time facilitate novices' learning as teachers.Of particular concern, studies highlighted that student teachers perceived the PACT as overwhelming their student teaching experience (Pecheone & Chung, 2006).The risk highlighted by these researchers is that teacher candidates might have to choose between being a "good student" or "task completer" in order to achieve a high score on the PACT versus being successfully and critically engaged in learning to teach during the process of completing the performance assessment (Liu & Milman, 2013;Sandholtz, 2012;Sandholtz & Shea, 2012).
Supervisor and cooperating teacher experiences with the PACT.Findings across a number of studies direct attention to tensions that stem from lack of knowledge about the PACT assessment on the part of cooperating teachers and supervisors and, in turn, their potential to support teacher candidates (Lit & Lotan, 2013;Okhremtchouk et al., 2009).This issue is further complicated by the nature of their roles: supervisors and cooperating teachers are positioned to observe, support, and gauge teacher candidates' progress over time and, in this formative stance, may be inclined to anticipate professional growth in the candidates over the course of their student teaching experience (Sandholtz & Shea, 2012;Sandholtz, 2012).Among other considerations, this sets up potential misalignments between their assessments and those of PACT scorers who make a single summative judgment at one point in time.
Other studies of supervisor and cooperating teacher experiences with the PACT appear to mirror concerns raised by teacher candidates.In particular, supervisors and cooperating teachers raise concerns about the PACT dominating the student teaching experience (Pecheone & Chung, 2006).Some cooperating teachers viewed the PACT requirements as compromising the actual work of teaching.Other supervisors were frustrated with attempts to extend support and guidance beyond the requirements of the PACT (Lit & Lotan, 2013).A key question emerging from these studies is: To what extent are cooperating teachers' and supervisors' voices valued beyond information provided by the PACT about teacher candidates' mastery of skills required to be novice teachers?
University faculty and program experiences with the PACT.Studies generally reveal faculty perspectives on the PACT as a real or potential catalyst for program change (Darling-Hammond, Newton, & Wei, 2010;Lit & Lotan, 2013;Pecheone & Chung, 2006;Peck & McDonald, 2013;Whittaker & Nelson, 2013), as well as for faculty feeling more accountable for their practice (Sloan, 2013).Across studies, however, we find mixed interpretations of the PACT by program faculty.In some cases, the PACT was viewed by university faculty as an assessment "done to us," which resulted in "cannibalizing" (Whittaker & Nelson, 2013) or "colonizing" (Chung, 2008;Liu & Milman, 2013) of the program's curriculum, shifting the curriculum toward a more limited consideration of perspectives, or more fundamentally contributing to a loss in program identity (Lit & Lotan, 2013;Liu & Milman, 2013;Peck et al., 2010;Sandholtz & Shea, 2012;Sloan, 2013;Whittaker & Nelson, 2013).In these instances, faculty perceived that PACT-related policies were a threat to their autonomy, which contributed to their resistance to what they perceived to be an externally imposed assessment, or, in some cases, an unwelcome force to promote or extend a particular vision of the profession (Peck et al., 2010;Peck & McDonald, 2013).
In other cases, the PACT was viewed as a process of moving faculty and their perspectives along (Lit & Lotan, 2013;Liu & Milman, 2013;Peck et al., 2010;Sandholtz & Shea, 2012;Sloan, 2013;Whittaker & Nelson, 2013).These studies highlighted enhanced faculty collaborations across courses, clear articulation for faculty and students across coursework, and greater shared understanding of the program as a whole (Sloan, 2013).Factors that appeared to contribute to faculty "buy-in" included incorporating all program members as part of the PACT scoring process (Sloan, 2013), designing formative, embedded signature assessments that permitted a "standardized flexibility" across the curriculum and supported academic freedom (Whittaker & Nelson, 2013), and designating program leaders to coordinate efforts and manage private/public space and work loads (Sloan, 2013).
A key question emerging across these studies is: How are the outcomes of policies like those associated with the PACT affected by the ways in which they are interpreted and taken up by program faculty within their local context?

PACT: Implementation and Impact
Our analysis demonstrates how the PACT was taken up at the local level.Drawing on the experiences of multiple stakeholders with the PACT, here we lay out three tensions around the impact and implementation of the PACT: professionalization vs. local control; marginalizing vs. privileging; and formative vs. summative assessment.We discuss the tensions in detail because our review of the literature reveals that they are significant across the experiences of all involved and are critical in understanding how the PACT is influencing and shaping the discourse of teacher education.
Professionalization vs. local control.As others have noted, one of the rationales for the development and use of a performance assessment in teacher education is the potential benefit toward the professionalization of the field (Cochran-Smith et al., 2013;Darling-Hammond, 2010b;Darling-Hammond & Hyler, 2013).Performance assessments have the potential to serve as highly valid assessments that can determine readiness for the classroom (Darling-Hammond, Newton, & Wei, 2012).They also have the potential to hold teacher candidates to rigorous and consistent standards for novice teachers.In other words, performance assessments can offer a common vision of good teaching and a standard of excellence for the teaching profession (Pecheone & Chung, 2006).
Because the PACT is an assessment with high stakes for both teacher candidates and teacher education programs, it is critical, then, that scores are consistent across scorers (i.e., high inter-rater reliability), the results of the assessment are consistent across time, location, and context, and the inferences made about the candidates have a high degree of validity.In general, the PACT is considered to have strong psychometric properties.Researchers have determined that the PACT has a high degree of inter-rater reliability (Pecheone & Chung, 2006), content validity (Duckor, Castellano, Tellez, Wihardini, & Wilson, 2014;Pecheone & Chung, 2006), and some evidence of construct validity (Duckor et al., 2014;Newton, 2010;Pecheone & Chung, 2006).There are concerns that males scored significantly lower than females on the PACT and student teachers in suburban schools scored higher than those in urban or inner-city schools (Pecheone & Chung, 2006).Furthermore, recent Rasch analyses suggest that the tasks themselves do not represent independent constructs (Duckor et al., 2014).In other words, statistical analyses raise questions about what the PACT actually measures.For the PACT to be a device that contributes to the professionalization of teaching, more work will need to be done in order to address these systematic biases as well as further research on the internal structure of the assessment (Duckor et al., 2014) in order to make appropriate inferences about the candidates who complete the PACT.
In the conceptual and empirical literature, there are concerns about how the professionalization of teaching through performance assessments may affect the local control of teacher education (Cochran-Smith et al., 2013).For example, to ensure a high degree of inter-rater reliability across institutions, the PACT is officially scored by external scorers who have received training on the PACT rubrics (Pecheone & Chung, 2006).This means that scorers do not know the teacher candidate whose PACT they are assessing (Sandholtz & Shea, 2012).For example, one supervisor expressed her frustration with this when she stated that "after only two days of training, a calibrated scorer's judgments about candidates' competency to teach English language learners was considered more valid than hers" (Berlak, 2010).The impact around issues of scoring, then, can be disempowering to certain stakeholders, including supervisors and cooperating teachers, because of the way in which their assessments of a teacher candidate are or are not considered in relation to candidates' scores on the PACT.Similarly, the implementation of the PACT could be viewed as a loss of local control to determine the readiness of a teacher to enter the profession, given the shifting roles of university faculty, supervisors, and cooperating teachers in the assessment process.In some ways, this message detracts from the message of professionalization of teaching and raises question about whose voices are valued as experts in the field.
Several authors have argued that the PACT can and should only be one of multiple measures to assess a teacher (Darling-Hammond et al., 2010) and should not replace other measures already in place (Porter & Jelinek, 2011).In this way, the PACT has both the capacity to further the teaching profession through an externally validated assessment, while honoring what is valued at the local context.
A final consideration around the perceived loss of local control is who has a voice about the content of the assessment.While the PACT was developed and revised by teacher educators, the assessment is now in a more permanent state, and cannot be altered at the local university level4 .Some have argued that the common vision of good teaching implicit in the assessment designates a specific view of how teachers ought to be taught to teach and what to value, and therefore, the PACT recognizes not only a common vision of teaching, but a narrowed view of how one ought to teach (Peck & McDonald, 2013).
Marginalizing vs. privileging.Fears around the loss of local control in decisions of teacher education often manifest in the literature as reported concerns about the systematic inclusion or exclusion of specific voices, populations, or points of view.Our review of the literature suggests that two entities are at the heart of the PACT's role in marginalizing and privileging: people and curriculum.
Findings in the empirical and conceptual literature repeatedly highlight four groups of people who can be simultaneously marginalized and privileged as an outcome of implementation policies surrounding the PACT: students, teacher candidates, cooperating teachers, and university supervisors.The conclusions point to the perceptions from these key players that their voices and opinions are not valued in the same way within the context of teacher education as they were prior to the implementation of the PACT.One example of marginalization is the finding that universities have become hesitant to place teacher candidates in rural or urban settings that might negatively impact candidates' ability to complete the PACT (Okhremtchouk et al., 2009).Researchers suggest that this unintended consequence marginalizes certain communities served by rural or urban schools by limiting access to high quality instructors because student teachers may be less likely to apply for positions in these schools upon graduation.As a result, this has the potential to perpetuate cycles of inequity.This finding also highlights the potential to position teacher candidates in a less powerful position to make conscientious decisions about where they teach.
Equally important in understanding the discourse of the PACT's impact is exploring ways in which the curriculum of teacher education is being influenced.Studies found that when the PACT is seen as an add-on, teacher candidates report less satisfaction and learning than when the task is integrated into the existing program (Chung, 2007(Chung, , 2008)).Teacher education programs have started to alter their curricula in such a way as to include content that aligns with the PACT, and faculty have reported that these changes mean that certain courses or priorities of programs, at times, can no longer be taught and are felt to be devalued (Berlak, 2010;Chung, 2007Chung, , 2008;;Lit & Lotan, 2013).This perceived colonization of curriculum is important in several ways.First, the findings suggest that if the PACT privileges a certain curriculum, university faculty may not have the space to make professional decisions about what teacher candidates ought to be taught (Sloan, 2013).Second, some worry that a high stakes performance assessment may shift the focus of teacher learning towards the "right way" to answer questions posed by the PACT because such an assessment may not allow the space for critical thinking or multiple ways of thinking by the teacher candidate (Liu & Milman, 2013).For example, Liu and Milman (2013) found that there is little space in the PACT for multicultural education.Finally, some are concerned about the time, cost, and resources needed to implement the PACT as part of a broader assessment system (Pecheone & Chung, 2006).Despite these concerns, Whittaker and Nelson (2013) found that embedded signature assessments that are locally developed and administered can offer faculty some flexibility and academic freedom to make curricular decisions.
Formative vs. summative assessment.Teacher candidate performance assessments have the potential to contribute to the field in part because of their dual ability to be formative and summative in nature not only for candidates, but teacher preparation programs as well.On the one hand, from the perspective of the teacher candidate, participating in the PACT has been reported to be a powerful learning opportunity (Chung, 2007(Chung, , 2008;;Darling-Hammond et al., 2012;Darling-Hammond, 2010a;Okhremtchouk et al., 2009).On the other hand, the tension with the PACT as a formative tool for teacher candidates persists because it is meant to be a summative assessment to determine readiness to teach.Teacher candidates also report learning less through their completion of the PACT when they felt they had received little support to do so (Chung, 2007(Chung, , 2008).It appears that the successful negotiation of PACT completion for teacher candidates lies in a better understanding of balancing support and autonomy as a means to maximize the learning potential.
The PACT has the capacity to be a powerful formative tool for teacher education programs particularly through analysis of teacher candidates' work (Peck & McDonald, 2013).Several studies have found that the most successful integration of the PACT and its ability to be a learning tool for programs rests on faculty buy-in (Peck et al., 2010;Peck & McDonald, 2013;Sloan, 2013).The transition from understanding the PACT through the lens of compliance towards embracing it as a learning tool exists often under conditions of strong leadership (Sloan, 2013) where faculty (among others) are given ample time and space to work through their anxieties and questions (Peck et al., 2010;Peck & McDonald, 2013).Once the PACT is perceived as having the ability to be informative for teacher education programs, researchers have noted several findings about the ways that programs have incorporated the outcomes for their own improvement (Bunch et al., 2009;Guaglianone, Payne, Kinsey, & Chiero, 2009;Pecheone & Chung, 2006;Peck & McDonald, 2013;Sandholtz, 2012).These include increasing consistency across the program around messages of good teaching, increasing consistency within courses taught by different faculty, and identifying individual faculty's strengths and weaknesses when working with teacher candidates (Chung, 2008;Peck & McDonald, 2013;Sloan, 2013).
Overlapping tensions in the PACT.These tensions-professionalization vs. local control, marginalizing vs. privileging, and formative vs. summative-exist in conjunction with one another.This has often meant that key stakeholders are faced with the difficult task of negotiating a balance across all three tensions simultaneously.For example, while trying to promote the professionalization of teaching through the PACT, stakeholders must also be willing to grapple with the implications for changes in local control as well as who and what is being marginalized and privileged through the assessment process.This often entails understanding the PACT not only for what it is and is not but also for what it is capable of doing.Liu and Milman (2013) argue that teacher performance assessments like the PACT "cannot be all things to all people" (p.137; also see Porter & Jelinek, 2011;Sandholtz, 2012).As they note, a performance assessment emerges as a necessary albeit insufficient measurement of preparedness to teach, and may not fully meet the needs of certain stakeholder groups, just as it may not fulfill certain locally contextualized program aims.However, as the reviewed literature suggests, the PACT has in some cases demonstrated its capacity to inform programs about teacher candidates' areas of strength and future growth, as well as areas of need tied to ongoing program improvement.The perceived "forced choice" between focusing on PACT-related activities and actually doing the "work" of student teaching in the classroom frequently positions teacher candidates, as well as their cooperating teachers and supervisors, in a stance that compromises candidates' effectiveness in both roles: student and teacher.Across these tensions, these studies underscore the value of using multiple methods to assess candidates' preparedness to teach and support their learning as preservice and novice teachers.

edTPA: State Policies' Construction of the Problem of Teacher Education
In contrast to the PACT as a state-level assessment, the edTPA operates within and across national, state, and local contexts as a nationally available performance assessment.Given the relatively recent full-scale implementation of the edTPA, the peer-reviewed literature is largely conceptual, descriptive and positional, taking one side or another on the edTPA (see, for example, Au, 2013;Darling-Hammond & Hyler, 2013;Dover, Schultz, Smith & Duggan, 2015;Madeloni & Gorlewski, 2013;Sato, 2014).Although there is emerging empirical research on the edTPA (e.g., edTPA, 2015; Meuwissen & Choppin, 2015), the body of research is limited due to the newness of the assessment.As a result, we turn to an examination of state policies pertaining to the edTPA.
Here we explore the ways in which state policies have adopted and implemented the edTPA.In our analysis, we find that among the six states that adopted and enacted statewide policies involving the edTPA as of 2014 (California, Iowa, Minnesota, New York, Tennessee, Washington), there is significant variation in terms of the purposes and rationale for implementing the edTPA, stakeholder involvement in the development of state policies, and how the edTPA has been initially implemented (see Table 1).For example, California SB 2042 is cited as the impetus for the development of the PACT.As such, the state policy context was the motivating force behind the development of the PACT by teacher educators5 , and ultimately, the edTPA across multiple organizations.This is very different from the other five states that adopted the edTPA existing model for performance assessments in teacher education.For example, New York adopted and implemented the edTPA as the only approved performance assessment for teacher certification.This section explores the evolution of the PACT to the edTPA in terms of how state policies define the problem of teacher education.

2009-2014
Note: MACTE = Minnesota Association of Colleges for Teacher Education; PACT = Performance Assessment for California Teachers; WACTE = Washington Association of Colleges for Teacher Education.
Purposes of the performance assessment.Across the six states in our analysis, the edTPA has been used for multiple purposes at the state and program level, including program completion, state certification/licensure requirements, program approval, program accreditation, and external accountability outside of accreditation.The states also differ in terms of how problems and solutions in teacher education are framed.These differences are illuminated by the varied purposes forwarded by states around the adoption of performance assessments.
In general, looking across states with policies in practice as of 2014, the most common purpose of the edTPA is either for program completion and/or teacher certification/licensure.In all of these states, the edTPA is forwarded as an "authentic" assessment of teacher candidate quality ("edTPA Minnesota," 2015) and, in some cases, an alternative to multiple-choice/constructedresponse assessments such as Praxis II (e.g., Iowa).As such, implicit in the implementation of the edTPA is an assumption that teacher education needs a highly valid and reliable assessment of teacher candidate quality and readiness to teach to make decisions for licensure and certification purposes.
In two states-Iowa and Tennessee-the edTPA is supported as an optional assessment for program completion and recommendation for licensure, with no stakes attached to the assessment at the state level (see Table 1).In contrast, in New York, the edTPA has been framed as an external accountability measure with the purpose of assessing teacher preparation program quality through the overall pass rates of a program's teacher candidates.Specifically, the State Department of Education has recommended that if fewer than 80% of a program's teacher candidates pass the edTPA, the teacher preparation program must submit a professional development plan to the New York State Department of Education (D'Agati, 2014)6 .The NY State Department of Education also recommended the use of each program's edTPA scores in comparative institutional profiles made available to the public (D'Agati, 2014).Our analysis suggests that policymakers in New York perceived a significant problem with teacher preparation programs and the need for further oversight to ensure the preparation of high quality teachers, as measured by pass rates on the edTPA.The solution was to enforce stringent external accountability measures on teacher preparation programs as a lever for reform and enhanced program quality.
California, in contrast, uses state-approved performance assessments including PACT and edTPA to inform program accreditation, however, there does not appear to be a specific requirement for program pass rates tied to program approval or accreditation.The way teacher education has been problematized in New York-in contrast to the way problem of teacher education has been framed in the other states in our analysis-may contribute in some ways to the deprofessionalization of teaching and teacher education, through its use of the assessment for external accountability.
Rationale for the edTPA in state policies.As Cochran-Smith, Piazza, and Power (2013) suggest, constructions of the problem of teacher education include the political strategies used to move the policy forward.In the rationale and making of the state policies, there are several overarching strategies used to adopt the edTPA: institutional credibility of the assessment; common expectations and highly rigorous assessment; and prevalent implementation of the assessments across the nation.These strategies provide some evidence as to how states frame the problem of teacher education, and how these performance assessments can serve as part of the solution.
First, across our analysis of state policies, there is a rationale for enactment that centers on the edTPA's development coming out of Stanford (e.g., "edTPA Minnesota," 2015), a prestigious university with a reputation for high quality teacher preparation.At least in the case of New York, Linda Darling-Hammond is specifically mentioned as one of the designers of the PACT and edTPA to invoke the credibility of the performance assessment (e.g., D 'Agati, 2012;King, 2012).Second, the edTPA has been promoted as an assessment that provides common expectations for what teacher candidates should be able to demonstrate by the end of teacher preparation ("edTPA Minnesota," 2015), and has met measurement standards for validity and reliability.As such, the edTPA is viewed as a rigorous and a seemingly objective measure in determining readiness to teach.Third, and somewhat related, across policies, the edTPA is forwarded because of its prevalence across multiple states and hundreds of teacher education programs (D'Agati, 2012;"edTPA Minnesota," 2015).In other words, the rhetoric behind the edTPA could be interpreted as: everybody is doing it, shouldn't you join in, too?
Policy actors involved in policy adoption.The major actors and audiences involved in policy development also suggest the ways in which the policies on the edTPA frame the problem of teacher education.Across state policies, these include policymakers, foundations, regional and national teacher education organizations, teacher education programs, teacher educators, and in one case private, for-profit companies.Findings from our analysis reveal that two states adopted policies that include the edTPA in collaboration with or based on recommendation from organizations affiliated with teacher educators/teacher education programs (i.e., Washington, Minnesota).For example, in Minnesota, the Minnesota Board of Teaching and the Minnesota Association of Colleges for Teacher Education (MACTE) (a regional affiliate of AACTE) began exploring performance assessments in teacher education in 2009, formed a steering committee in 2010 funded by a private foundation, and then the state legislature approved a law in 2011 adopting the edTPA ("edTPA Minnesota," 2015)-a "policy web" and "policy cycle" for adopting the edTPA that involved state legislators, teacher educators, and funding from the Bush foundations ("edTPA Minnesota," 2015).
In our review of state policies, New York was the only state that adopted an exclusive policy of using the edTPA as a requirement for certification purposes, with little input from teacher education programs or faculty.The policy in New York was developed and approved by the Commissioner of Education and the NY State Board of Regents in 2012 (King, 2012), in response to both a 2010 directive from the Board of Regents for a performance assessment for initial teacher certification and as part of New York's Race to the Top (RTTT) application (D'Agati, 2012).Interestingly, prior to the adoption of the edTPA, New York had contracted with Pearson to develop a New York-specific assessment, which was then discarded after the edTPA was adopted.In constructing the "problem of teacher education," New York State selected an assessment that was developed for educators/by educators and implemented it with minimal input from (teacher) educators in the state.The "policy cycle" and "policy web" involved broader policy contexts (RTTT) and for-profit companies (Pearson) that already had existing relationships with the state.
How/when the policy was enacted (field-testing, roll-out).The ways in which states field-tested the PACT and/or edTPA before enacting policies can be used to explore another aspect of Cochran-Smith, Piazza, and Power's (2013) construction of the problem of teacher education; namely, "stated and hidden agendas, and political strategies used to forward policies" (p.9).In other words, from examining the length of time that states field-tested the PACT and/or edTPA, our analysis explores how the edTPA was rolled out at the state level.In some cases, the roll-out was intended to inform substantive changes before high-stakes implementation of the assessment.In other cases, the roll-out may have taken place as a pro forma activity.
Five of the six states in our analysis field-tested the edTPA for at least 4 years before the edTPA became fully operational in the fall of 2013.7During these years, the five states were in consultation with Pearson leading to revisions to the edTPA handbooks (e.g., Lambert, 2013).These states also used the field tests as opportunities to consult with teacher educators and other organizations to build "buy in" and resolve any institutional questions and concerns (e.g., Pecheone & Whittaker, 2013;Siera, 2013).In some cases, these conversations resulted in revisions to the assessment policy (e.g., Pecheone & Whittaker, 2013b).
In stark contrast, New York field-tested the edTPA for only one academic year (2013)(2014).This may have been due, in part, to pressure to implement the edTPA as part of New York's commitment in the RTTT application in 2010, in which performance assessments "were scheduled to be implemented in May 2013" (D'Agati, 2014, p. 2).Our analysis suggests that rapid implementation of the edTPA in New York may have resulted in pushback and adjustments to the full-scale consequential implementation of the edTPA.Although a common performance assessment was scheduled to be implemented in New York in May 2013, it was delayed until May 2014 "in order to provide educator preparation programs with an additional year to prepare teaching candidates, while at the same time ensuring that the time frames in the State's RTTT application are met" (D'Agati, 2014, p. 2).However, in May 2014 New York instituted a "safety net" policy which allowed teacher candidates who didn't pass the edTPA by May 2014 to instead "demonstrate their readiness for the classroom by passing an older certification exam" (King, 2014) 8 .Therefore, while there were stated and hidden agendas, as well as political strategies used to forward policies in the other five states, the timeline for the edTPA in New York may have limited meaningful dialogue among the multiple actors involved in the implementation of the assessment.
As Cochran-Smith and colleagues (2013) have argued, "in the context of discussions about the TPA as an accountability mechanism, the role of the profession itself is critical" (p.16).Across these state policies we see varying degrees of teacher educator involvement and participation in the discussion around the implementation of the PACT/edTPA.These policy contexts may have encouraged and facilitated involvement from the (teacher education) profession, or in some cases, created barriers for active participation from the process.In this policy context, the quality and the extent to which teacher educators participate in the state policy-making remains a controversial issue (Cochran-Smith et al., 2013).

Discussion and Implications: Growing Tensions with the edTPA
Across our review of the literature on the PACT and state policy analysis of the edTPA, emerging findings suggest that the tensions highlighted in the PACT literature are magnified with the implementation of the edTPA across the nation.Revisiting the "politics of policy" framework (Cochran-Smith et al., 2013), we see divergent approaches across the PACT and the edTPA in terms of the ways in which teacher education policies surrounding these assessments define the "problem of teacher education."The policy surrounding the PACT constructs the problem of teacher education as one of ownership and direct involvement of teacher educators in California in response to a state legislative mandate.Additionally, within the context of the PACT, there are built-in opportunities to include or embed additional programmatic formative assessments as part of the PACT assessment system.In contrast, the state policies surrounding the edTPA problematize and in some cases appear further removed from the spirit of "by educators."This is exemplified in New York where there was limited involvement of teacher educators in the edTPA policymaking process.
To explore the practical and policy implications of the evolution of the PACT and edTPA, we revisit the tensions outlined in the review of the PACT literature to explore they ways in which there may be potential drift from the original spirit of the PACT as an assessment controlled by the profession.Finally, we raise questions and recommend future research that explores how and by whom policies are "taken up," and where the responsibility for policy outcomes is located.
First, as others have noted, when exploring the tension between local control vs. professionalization, the implementation of the edTPA at a national level could forward the professionalization of teaching; it draws on established professional knowledge and standards, and is a "uniform and impartial" assessment that was developed and owned by the profession (AACTE, 2014).As an additional actor and partner, the American Association of Colleges for Teacher Education (AACTE) promotes the goal of widespread edTPA implementation across all states in the nation (AACTE, 2014).To achieve this goal, the organization attempts to create a space for conversation across stakeholders in the teacher education profession by providing resources, coordinating national conferences, and developing advisory boards to support its implementation.
However, as a result of scaling up to the national level, the power for decision-making related to the edTPA is further removed from the local level (Margolis & Doring, 2013) than it was for the PACT in the California context.Furthermore, the edTPA's inclusion of multiple actors outside of the profession raises questions about the extent to which the assessment is still close to the spirit of an "assessment for educators, by educators."Outside of specific state policy contexts, much of the controversy surrounding the edTPA in the professional literature and the media has been about the inclusion of Pearson Education as an operational partner (Au, 2013;Madeloni & Gorlewski, 2013;Ravitch, 2013).Pearson Education, an external private corporation, determines who scores the edTPA, how scorers are trained, how scores are presented, and coordinates where data are housed.Although these decisions are made in collaboration with the profession [i.e., Stanford Center for the Assessment of Learning and Equity (SCALE) and the American Association of Colleges of Teacher Education (AACTE)], the inclusion of another external partner in the assessment process creates opportunities for multiple agendas, strategies, and maneuverings outside of the profession.
In some instances, our findings reveal potential de-professionalization in state policy contexts.We see this possibility in New York, when decisions about adoption, implementation and stakes attached to the edTPA took place with limited input from the teaching profession.As a result, there has been concentrated pushback from teacher educators and prospective teachers, particularly in New York state (e.g., Miletta, 2014;Ravitch, 2013;Sawchuk, 2013).Thus, the involvement of additional actors from the PACT to the edTPA raises questions about the "discourses and influences" prevalent in teacher education including accountability and neoliberalism (Cochran-Smith et al., 2013), which could be forwarded through the use of the edTPA as a state level tool.
Second, the edTPA magnifies the tensions of marginalizing vs. privileging groups and ideas highlighted in the PACT literature.In some ways, these tensions are echoed in recent empirical literature, particularly as it pertains to the role of the cooperating teacher and supervisor in the edTPA.For example, Margolis & Doring (2013) found that the implementation of the new teaching assessment created a dissonance and confusion for cooperating teachers and supervisors who had previously been clear about their role in working with teacher candidates.Excluding such key players in edTPA and PACT implementation risks silencing and marginalizing voices of those who are integral to the field.
Like the PACT, the edTPA has been criticized for its lack of stance on issues of social justice and multicultural education.Although some suggest that there is room for multicultural education and social justice in the edTPA (Robinson, 2014;Sato, 2014), others question the absence of explicit references to multicultural education and social justice (Cochran-Smith et al., 2013;National Association for Multicultural Education, 2014).This raises the question: with no room for modifications and varied stakes attached to the assessment, does the edTPA by necessity drive out space for other conceptions of teaching?
Third, we raise questions about the tensions manifested through the edTPA's purported dual role as both a formative and summative assessment.Drawing on the PACT model, the edTPA has the potential to support teacher candidate learning, assess teacher quality, and inform programmatic improvement.However, the emerging literature highlights the thorny issue around what Margolis and Doring (2013) refer to as the "tipping point" in performance assessments like the edTPA; that is, the point at which, in trying to determine candidates' preparedness to teach the assessment no longer facilitates candidates' learning as teachers.In our analysis, the PACT and edTPA present limited opportunities to do both, effectively, at the same time.Additionally, as a result of "growing pains" or external partnerships, some university faculty have found that they could not use the assessment data for program improvement (Place, 2014).
We ask whether the edTPA can be considered "one assessment" when it is implemented in variety of ways, by different stakeholders, for different purposes, and raise the following questions: Is there flexibility in edTPA policies to preserve "common" elements of the edTPA instrument while providing opportunities for locally-developed signature assessments as part of the assessment system?In developing state policy on the edTPA, could states adopt the assessment as a "living document" with room for change over time?What are the unintended consequences of edTPA policies in terms the inclusion and exclusion of multiple voices and perspectives?Can the edTPA serve the multiples purposes and roles defined by the various stakeholders involved in its implementation?
Research is needed within and across state policy contexts to examine how the edTPA affects teacher candidate learning, program improvement, and faculty buy-in.Further inquiry is needed to examine the ways in which teacher education programs manage the conflicts they encounter when attempting to balance statewide accountability policies with the enhancement of candidate learning and informing program improvement.We recommend empirical research that goes beyond a dichotomous framing of the issues and instead creates a space in which there is room to value and account for the interplay of the policies in practice with the impact and implementation of the PACT, edTPA, and other performance assessments in teacher education.As teacher education programs, state policy makers, and local, state, and national organizations race to implement performance assessments in teacher education, further critical analyses of the policies and impact of the PACT and edTPA are necessary to examine the questions that arise and the conclusions that can be drawn from the politics of these policies.accountability policies on teaching and learning.She is interested in all levels of the system (national, state, district, school, and classroom), and focuses on three policy areas at the forefront of educational reform: comprehensive assessment and accountability systems, performance-based assessments, and teacher/teacher preparation program effectiveness initiatives.

Table 1
States with Policies on the edTPA as of December 2014