Politics of Education and Teachers ’ Support for High-Stakes Teacher Accountability Policies

Although educators are at the center of contentious high-stakes teacher accountability policies, we know very little about their attitudes toward these policies. This research gap is unfortunate because teachers are considered key actors in successful implementation of educational reforms. To what extent do the politics that accompany the introduction of high-stakes teacher accountability policies affect teachers’ support for the policies themselves? To address this gap, we used data from an experimental survey of teachers in New Jersey (n=444), where a new reform— Teacher Effectiveness and Accountability for Children of New Jersey Act (TEACHNJ)—was signed into law in 2012 and implemented shortly thereafter. The cornerstone of the reform is a new Education Policy Analysis Archives Vol. 25 No. 87 SPECIAL ISSUE 2 evaluation system that ties student performance on standardized tests to teachers’ evaluations. We found that the majority of teachers in our study oppose the new evaluation system. Teachers’ attitudes were shaped by the politics of the key actors advocating for the policy, perceptions of implementation efforts, and beliefs in the potential outcome of the policy. Open-ended responses indicated that teachers question the validity of the evaluation system and are concerned about the negative intended and unintended consequences of the system. We conclude this paper by discussing the implications of these findings for policy studies and policymaking.


Introduction
High-stakes teacher accountability policies are the new frontier in the application of private-sector discourses and practices to the field of education (often called New Public Management; Gewirtz & Ball, 2000).Major components of these policies include (a) teacher evaluation that is based on how much academic growth students experience over the course of the school year, and (b) teacher evaluation that has implications for professional development, compensation and benefits, and tenure (see reports by Educational Testing Service [Braun, 2005] and RAND Corporation [Steele, Hamilton & Stecher, 2010], and see review by Baker, Oluwole & Green, 2013).To some degree, these policies draw on opportunities created by international large-scale assessments (ILSA) as well as national assessments (Elmore, Abelmann, & Fuhrman, 1996).In many countries, poor performance on ILSA was blamed on teachers and teacher preparation programs (Figazzolo, 2009;Pizmony-Levy, under review).In turn, this interpretation generated a discursive opportunity structure (Koopmans, 1999) that is open for high-stakes teacher accountability policies as a panacea.Furthermore, the creation of an accountability infrastructure with an abundance of student-level data created the illusion that high-stakes teacher accountability policies are not only needed, but also possible (e.g., through recycling and reapportioning data collected for another purpose).
High-stakes teacher accountability policies are contentious.On the one hand, policymakers and practice communities (e.g.., testing agencies) advocate enthusiastically for the implementation of teacher evaluation systems that can bring about improvements in instructional practices (e.g., Tucker & Stronge, 2005).On the other hand, scholars still debate whether these evaluation systems are scientifically and technically sound (AERA Statement on Use of Value-Added Models (VAM) for the Evaluation of Educators and Educator Preparation Programs, 2015).This is not only a debate among the "elite," but the broader American public is also split about the question of basing part of teachers' salaries on how much their students learn.The 2016 Ed Next survey shows that slightly more than half of Americans (52%) favored this approach, with more support among Republicans (57%) than Democrats (50%).Further, Phi Delta Kappa (PDK) surveys show a significant decrease in public support for state polices requiring that teacher evaluations include how well a teacher's students perform on standardized tests.The 2015 PDK survey shows slightly more than two-fifth of Americans (43%) favored this approach (Richardson & Bushaw, 2015).
The opt-out movement in the US, in which parents refuse to have their school-aged children take federally mandated educational assessments, provides additional evidence to the contentious nature of high-stakes teacher accountability policies (Mitra, Mann & Hlavacik, 2016).A recent survey of activists in the opt-out movement estimated that teachers represent about half of the movement's social base (Pizmony- Levy & Green Saraisky, 2016).When asked to rank the main reason for their participation in the movement, more than one-third of activists (36.9%) indicated that they "oppose using students' performance on standardized tests to evaluate teachers."Moreover, teachers were more likely than others to mention this reason as a motivation for activism (44% versus 27%).
Although educators are at the center of high-stakes teacher accountability policies, we know very little about how they actually feel about these policies (Ballou & Podgursky, 1993;Goldhaber, DeArmond, & DeBurgomaster, 2010).The 2016 Ed Next survey, for example, shows that less than

Literature Review
Scholars have agreed that teachers' engagement is key for successful policy implementation (Fullan, 2001;Lipsky, 2010;Spillane, 2004).It is reasonable to assume that teachers' engagement would be even more important in the context of high-stake teacher accountability policies.After all, these policies pose sanctions and rewards on teachers based on student performance on standardized tests.To date, however, scholars have mostly overlooked teachers' views toward these policies (for research on teachers' support of merit pay, see Ballou & Podgursky, 1993;Goldhaber, DeArmond & DeBurgomaster, 2010).
Our study is informed by two lines of research.The first is concerned with teachers' engagement with policy and policy implementation.Scholars have pointed to four factors that enhance teachers' engagement with top-down educational change: policy "buy-in" and consensus over the principles of policy (Datnow, 2000;Datnow & Castellano, 2000;Fullan, 1991Fullan, , 1993;;Sarason, 1990Sarason, , 1996)), good experience with implementation efforts (method, type, and pace; Desimone, 2002), knowledge and understanding of the policy (Spillane, Reiser, & Reimer, 2002), and beliefs in the efficacy of the policy (Datnow & Castellano, 2000;Desimone, 2002).Most of the research that has been done is based on qualitative methods, which limits the possibility to test these four factors simultaneously.In this study, we use a quantitative research design to address this gap.
The second line of research focuses on teachers' views toward NPM-inspired policies and reforms (e.g.privatization and choice).A review of the factors related to teachers' attitudes toward these policies suggest that only a few personal and organizational characteristics have been studied (Belfield & Wooten, 2003).Hess, Maranto, & Milliman (2000), for example, found that more experienced teachers, those who identified as Democrats, those who majored in education, or those who had never worked in a competitive educational environment were more likely to oppose public school choice.Further, drawing from a micro-political approach, Addi-Raccah (2012) found that Israeli teachers holding internal leadership positions or those who reported feeling empowered by their principals were more supportive of privatization.
A common theme in both lines of research is the role of politics and power in shaping teachers' engagement with policies.Given the heated debate surrounding high-stake teacher accountability policies, we hypothesize that the political environment in which these policies are discussed and introduced will affect teachers' support.Indeed, Mortimore and Mortimore's (1998) analysis of the British education system, for example, concluded that the public discourse characterized with "antiteacher" sentiments and scandalization of student performance on standardized tests had negative consequences on policies intended to improve teaching and learning.
In the context of our case, Governor Christie, known for his harsh, real "New Jersey talk," has repeatedly and publicly disparaged any organized teacher effort to critique the policy.Therefore, in order to further Mortimore and Mortimore's (1998) analysis we examine the potential impact of the political environment on the perception of the TEACHNJ policy.To do this we draw from the literature discussing frames (Benford & Snow, 2000) and framing effects (for summaries, see Levy 2002;Levy 2003).Frames serve as thought organizers and can alter an individual's response in a survey even though the items that are framed differently may be logically equivalent.Equivalency framing usually involves "casting the same information in either a positive or negative light" (Druckman, 2004, p. 671).For example, a frame that mentions 95% employment as opposed to one that refers to 5% unemployment, although substantively equivalent in content, might elicit divergent opinions as to the state of the labor force and economy (example taken from Druckman, 2004).Equivalency frames involve alternative descriptions of a qualitatively similar phenomenon (Druckman, 2001), such as the frames used when describing a new teacher evaluation system as either "proposed by the Legislature" or "proposed by the Governor Christie." In this article, we systematically examine teachers' views toward high-stakes teacher accountability policy (TEACHNJ) and its implementation platform (Achieve NJ).We show a gap between teachers' endorsement of policy principles (i.e., policy alignment) and teachers' view of the specific policy.Drawing on an embedded survey experiment, we demonstrate the important role of the political environment on teachers' support for a new teacher evaluation policy.We also demonstrate the role of teachers' beliefs and attitudes play in the efficacy of the policy and its implementation.

Study Setting: New Jersey
With over 8.5 million people, New Jersey is the country's most densely populated state per square mile, with 17.3% of the population's school-age children being 5 to 18 years old (U.S. Census Bureau, 2010).As of 2008, the New Jersey educational system was the eleventh largest educational system in the United States (National Education Association, 2011).Various reports have suggested that New Jersey offers high-quality public education.The Editorial Projects in Education Research Center (2010), for example, ranked New Jersey seventh overall in the United States based on students' chance for success, overall K-12 achievement, standards, assessments and accountability, and school finances.New Jersey has traditionally outranked its peers with its average Advanced Placement exam test score (O'Neill, 2011).New Jersey has consistently outperformed 90% of the states in fourth-and eighth-grade reading and mathematics assessments as determined by the National Assessment of Educational Progress (NAEP; National Center for Educational Statistics, 2011).This positive view of the New Jersey educational system was challenged in 2009 with the release of a twelfth-grade assessment by NAEP.While New Jersey outscored all other participating states (only 11 states volunteered to participate in this assessment) in mathematics, the state matched the national reading scores.In addition, this assessment pointed-again-to a wide achievement gap between whites and minority students.The release of the report sparked a local policy "shock."In an interview with local media, then Acting Education Commissioner Rochelle Hendricks said, "While these results show that our seniors score well when compared to other participating states in math, the bigger picture is clear in that we must do better" (Mascarenhas, 2010).
Despite New Jersey's comparatively successful educational system, the 2009 gubernatorial candidate Christopher Christie, a Republican, ran on a platform arguing that "Nationally, academic achievement has been flat for many years... New Jersey is suffering from the same stagnation when we should be pushing for greater student achievement and academic excellence" (Rundquist, 2009).1With that belief, upon taking office in January 2010, Governor Christie began a series of education reforms, including "merit pay, changes to tenure, completing a statewide data system that tracks student achievement, forming of a teacher evaluation task force, creating the designation of master teacher and allowing alternate route certification for principals" (Fleisher, 2010).On the issue of tenure and merit, Governor Christie was quoted in the same Fleisher (2010) article: "Tenure has become a job guarantee regardless of performance or success.Tenure has become the sclerosis that coats the veins of our school system […] Any type of compensation that allows for anything but merit-gone."This, coupled with contentious interactions with his Democratic opponent who supported the New Jersey National Education Association (NJEA), created a hostile political environment for education reform.Governor Christie has consistently criticized the NJEA, calling the union "a group of political thugs" (Blackburn, 2011) and refusing to meet with the NJEA, arguing that "Frequently, the leadership of the NJEA has been a strong advocate for the status quo, whether the status quo is succeeding or failing" (Springer, 2009).The NJEA has not hesitated in launching media campaigns against the governor.In the same news article, Fleisher quoted NJEA President, Barbara Keshishian, who said, "'He is proposing reforms that are not based on good educational research or practice.What he proposes-an over-reliance on student test scores to make critical decisions from compensation to employment-is fatally flawed'" (Fleisher, 2010).
In his first year in office, Governor Christie reformed public employee pensions and benefits to reduce the state deficit.In addition, the governor placed a two percent cap on property taxes, which in turn capped school budgets that are funded through town property taxes.At the same time, 475 million dollars was cut in school aid, and further budgets were redirected to expand school voucher programs and support charter schools, forcing districts to reduce budgets and staff.All these reforms were introduced with little-if any-consultation with the NJEA.These actions and the overall approach were interpreted by NJAE as a personal attack from Christie against educators and public education (NJEA, 2013).
Subsequently, Governor Christie introduced the flagship of his educational reform: Teacher Effectiveness and Accountability for Children of New Jersey Act (TEACHNJ).TEACHNJ's theory of action is that improving educator effectiveness through evaluation and feedback will improve student performance (New Jersey Department of Education, 2010).The cornerstone of the TEACHNJ reform is a new teacher evaluation system, which is currently in its third year of implementation in schools throughout the state.Despite the reform's positive language of improving teacher effectiveness and student learning, the NJEA has met it with resistance and animosity (2013).Further, although the Act passed the state legislature unanimously, public discourse highlighted the important role of the governor in advancing this agenda (see for example, Rizzo, 2012).The question remains: To what extent did the politics of introducing and advocating for TEACHNJ affect teachers' support?TEACHNJ was developed as part of a federal grant called Race To The Top (RTTT).The funding received by state from the RTTT grant was based on the ability of the New Jersey Department of Education (NJDOE) to meet the deadlines established in its proposal, which placed TEACHNJ implementation on a fast track.The newly developed evaluation system consists of four categories of effectiveness (ineffective, partially effective, effective, and highly effective) based on multiple measures that are used to determine tenure.This drastically altered the former system of evaluation and tenure in the state, which used a binary measure of effectiveness, based on a single measure of educator practice and only required observations of non-tenured staff.Prior to TEACHNJ, teachers with tenure were not required to be observed by administrators and the removal of ineffective teachers often took years and hundreds of thousands of dollars, which discouraged firings.The Act made the revocation of tenure easier with automatic firing after two consecutive years of ineffective ratings and a streamlined tenure due process.TEACHNJ features numerous changes in evaluation structure, expectation, and support and is part of an educational reform movement in New Jersey focused on increasing student performance through effective educator evaluations that provide recognition, feedback, and support to education professionals (New Jersey Department of Education, 2013).
To implement TEACHNJ, the state developed AchieveNJ-a teacher evaluation and support system.Under AchieveNJ the state DOE has made significant efforts to provide communication and support to teachers through the creation of informational guides, presentations, forms, templates, etc. AchieveNJ was designed according to the following guiding principles: 1. Educator effectiveness can and should be measured to ensure our students have the best teachers in the classroom.2. Evaluations should always be based on multiple measures that include both learning outcomes and effective practice.3. Timely feedback and high-quality professional development, tied to evaluations, are essential to help educators improve.4. Evaluation and support systems should be developed with significant input from educators.5. Tenure and other forms of recognition should be based on effectiveness.
(NJDOE, 2013) As we discuss in the data and methods section, we draw on these statements to evaluate teachers' alignment with the underlying principles of the policy.
The new teacher evaluation system, mandated by TEACHNJ, is based on up to three components of assessment used to evaluate teacher performance.First, the median Student Growth Percentile (SGP) represents the median score of the growth that the individual students make from one year to the next on a state-wide assessment.New Jersey adopted the Partnership for Assessment of Readiness for College and Careers PARCC.The SGP only applies to fourth-to eighth-grade language arts teachers and fourth-to seventh-grade mathematics teachers (hereafter "tested subject teachers").This design might lead to differences in support for TEACHNJ between those subject teachers who were tested and others.Second, Student Growth Objectives (SGO) are academic goals with measurable assessments designed by the teacher, with assistance from the principal and/or the supervisor.All teachers are required to develop SGOs.Third, Teacher Practice is based on observations of instruction.
The state of New Jersey has approved numerous observation instruments (e.g., the Charlotte Danielson's Framework for Teaching, and the McREL Teacher Evaluation System).Taken together, these three components of the evaluation system are weighted and calculated to reflect a teacher's overall effectiveness on a scale from one to four, ineffective (1), partially ineffective (2), effective (3), and highly effective (4) (NJDOE, 2013).The goal of this evaluation system is to redefine the teaching profession based on quantitative terms with the hopes of improving student outcomes by improving teacher effectiveness.

Data & Methods
Our analysis is based on the New Jersey Teacher Engagement with Education Reform Study, an original survey conducted by the second author (see Woolsey, 2014).The primary goal of this survey was to collect information from teachers about their evaluation of the implementation of TEACHNJ and to document their professional experiences in the context of a large education reform.The survey was conducted in the winter of 2014, during the first academic year after the implementation of TEACHNJ.Because of restricted resources, the study is based on a nonprobability sample of 481 teachers.
To capture the wide swath of teachers across the state, data collection efforts included three recruitment strategies.First, we asked all public-school superintendents in the state of New Jersey (with correct contact information in the state DOE directory) to share the link to the survey with teachers.This strategy yielded the majority of responses (78%).Second, we asked all building administrators and supervisors in the Fair Lawn School District, where the second author has taught for the eight years, to share the survey link with teachers.Third, we posted a link to the survey on Facebook, a common social media platform, with a request to complete and share the survey.Following the initial contact, focal contacts-public-school superintendents as well as Fair Lawn School District building administrators, and supervisors-received one reminder about the survey.
The survey instrument was designed as a web-based, self-administered questionnaire.It included questions gauging respondents' attitudes toward TEACHNJ and its implementation schema, AchieveNJ.We developed these items based on a thorough examination of documents published by the New Jersey Department of Education about TEACHNJ and AchieveNJ (e.g., policy documents and training materials).The survey also contains detailed information on respondents' sociodemographic and professional characteristics.Most of these items were adapted from the General Social Survey and the Trends in International Mathematics and Science Study (TIMSS) Teacher Questionnaire.To evaluate the reliability of the instrument, the second author administered the survey to five teachers and conducted a short follow-up cognitive interview about the survey format and item wordings.The instrument was revised to incorporate their feedback.
Despite the richness of the data, one caveat should be mentioned: the study does not involve a random selection of respondents.This design might have implications on (a) the extent to which our sample is representative of the teacher population, and (b) the extent to which we can generalize patterns about the teacher population.Comparing the sample to the general population of New Jersey public-school teachers, we find that our sample includes more women (81% vs. 73%) and more white teachers (93% versus 85%).While our sample includes teachers from all 21 counties in the state, it overrepresents some of the counties and underrepresents other counties.To ameliorate this issue, we calculated and applied sample weights based on gender, race/ethnicity, and county.2Although the generalizability is compromised, the study still sheds light on important and understudied perspectives on new teacher evaluation systems.
Measures Support for AchieveNJ.The dependent variable in this study is the level of support for the new teacher evaluation system.As discussed earlier, teacher support or "buy-in" is crucial for successful implementation of any educational reform effort (Akiba, 2013;Fullan, 1993;Spillane, 2004).The item, which was the final question in the survey, included five response categories: (1) strongly oppose, (2) oppose, (3) I'm not sure; (4) favor, and (5) strongly favor.
Teachers' support for AchieveNJ can be shaped by various factors.The following measures capture the possible effects of the political environment, support for the underlying principles of the reform, self-reported policy knowledge, evaluation of implementation efforts, and perceived policy impact.
Political environment.To test whether teachers' support was affected by the political environment in which the new teacher evaluation system was proposed and implemented,, we used an embedded survey experiment.This design is useful when scholars seek to infer causal relationships from a cross-sectional survey (i.e., the survey experiment clearly distinguishes cause and effect).For example, Pizmony-Levy and Ponce (2012) used a survey experiment to examine whether terminological changes in the debate over same-sex marriage affected public support for this cause.
In another study, O'Brien and Pizmony-Levy ( 2016) used a survey experiment to examine whether a description of a university professor as an engaged scholar affected students' perception of the professor.
We manipulated the question stem by varying the terms describing the proposer of the new system, AchieveNJ.We used "Governor Christie" to signal a politicized sponsorship and "NJ Legislature" to signal a more neutral sponsorship.Recall that the TEACHNJ bill was proposed by Governor Christie and unanimously passed into law by the NJ Legislature.This manipulation yielded two versions of the item: "Do you oppose or favor AchieveNJ, the new teacher evaluation system, as proposed by [the NJ Legislature / Governor Christie]?"Each version was randomly assigned to respondents.In the analysis to follow, the version of the question is captured by a dummy variable coded 1 for the "Governor Christie" frame and 0 for the "NJ Legislature."If mentioning the name of Governor Christie reduced teachers' support for AchieveNJ in the context of a survey, then this means that the political environment shaped teachers' attitudes in the real-world.
Reform principles.Scholars have demonstrated that teachers' endorsement of reform principles is important for their support of the policy (e.g., Deci, 2009).Thus, the survey asked respondents to indicate the extent to which they agree or disagree with five statements describing the underlying principles and assumptions of the new teacher evaluation system.These statements were derived from the AchieveNJ communication materials (e.g., "evaluation and support system should be developed with significant input from educators" and "tenure should be based on effectiveness," etc.).These items included four response categories: (1) strongly disagree, (2) disagree, (3) agree, and (4) strongly agree.Drawing on these items, we calculated an index where higher values reflect higher levels of alignment with the policy intention of the new teacher evaluation system (Cronbach's alpha = .79).
Self-reported policy knowledge.Successful implementation of policy requires that stakeholders "on the ground" demonstrate knowledge of different aspects of the reform (e.g., Healey & DeStefano, 1997).To assess policy knowledge, the survey asked respondents to indicate the extent to which they are knowledgeable about nine key concepts related to the reform.These concepts included the law itself (TEACHNJ), the implementation platform (AchieveNJ), and state-mandated components of the reform (e.g., student growth objectives, teacher observation instrument, etc.).These items included four response categories: (1) not at all knowledgeable, (2) not very knowledgeable, (3) somewhat knowledgeable, and (4) very knowledgeable.Drawing on these items, we calculated an index where higher values reflect higher levels of perceived knowledge about key concepts in the new teacher evaluation system (Cronbach's alpha = .83).

Evaluation of implementation.
Another factor that might play a role in teachers' support is their interaction with and evaluation of the implementation process.To measure teachers' evaluation of the implementation efforts of the new teacher evaluation system, the survey asked respondents to indicate the extent to which they agree or disagree with five statements describing different facets of policy implementation: communication, pacing, resources, support, and monitoring.These items included four response categories: (1) strongly disagree, (2) disagree, (3) agree, and (4) strongly agree.
Drawing on these items, we calculated an index where higher values reflect a more positive view of the implementation process (Cronbach's alpha = .89).
Perceived policy impact.Finally, we determined whether teachers' belief in the efficacy of the new teacher evaluation system is associated with their support for the system.The survey asked respondents to indicate the extent to which they agree or disagree with the following two statements: "The new evaluation system will improve teaching" and "The new evaluation system will improve student learning."The item included four response categories: (1) strongly disagree, (2) disagree, (3) agree, and (4) strongly agree.These survey items are strongly correlated (r=.88, p<.001).

Controls
In addition to examining the effect of different aspects of engagement and framing on support for AchieveNJ, we controlled for a series of sociodemographic and professional characteristics, including gender, race/ethnicity, age, urban/rural status of school community, tenure, and tested subject teachers.In the 2013-14 academic year, when data for this study were collected, the tested subjects included mathematics and English/language arts in grades four through eight.Table 1 presents definitions, metrics, and descriptive statistics for all variables included in the analysis.

Open-ended Responses
The final section of the survey instrument included an open-ended question which prompted respondents to share their thoughts on issues raised in the survey.This qualitative dimension allowed us to preserve the teachers' voices and emphases, and provides a window into how teachers make sense of the new evaluation system.Furthermore, we used the responses to open-ended questions to help corroborate statistical findings.Approximately two-fifths (39.3%) of the sample responded to the open-ended item.Analyses of cases missing open-ended responses revealed several differences in the likelihood of responding.Older and tenured teachers who work in urban contexts were more likely than others to respond to the open-ended item.

Analytical Technique
After removing missing data, the final sample included 444 teachers.We used Ordinary Least Squares (OLS) regression modeling to examine teachers' support for AchieveNJ.Model 1 includes the randomized condition from the survey experiment.Model 2 adds sociodemographic characteristics (gender, race/ethnicity, and age) and professional characteristics (school community, teaching experience, tenure status, and tested subject teachers).The next four models introduce different measures of engagement with the reform: policy alignment (model 3), self-reported knowledge (model 4), perceived impact (model 5), and evaluation of implementation (model 6).Finally, model 7 introduces simultaneously all measures of engagement with the reform.Because the sample includes teachers nested in counties, the usual regression assumption of the stochastic independence of error terms underlying tests of statistical significance is violated.Therefore, in all regression analyses, we corrected for this non-independence of observations using the cluster procedure available in Stata to compute robust standard errors.

Descriptive Statistics
This section presents frequency distribution and descriptive statistics for the factors that might shape teachers' support for AchieveNJ.In our analysis, we began by exploring teachers' support for the underlying principles of TEACHNJ and AchieveNJ.That is, we examined the alignment of teachers' views with the principles behind the policy reform.As seen in figure 1 below, a large majority of teachers (more than four-fifths) indicated that they agree with the five statements describing the new teacher evaluation system.For example, almost all teachers (96.5%) endorsed the notion that "timely feedback and high-quality professional development are essential to help educators improve."Similarly, almost all teachers (95%) endorsed the notion that "evaluation and support systems should be developed with significant input from educators."Even the link between tenure and effectiveness was validated by more than four-fifths of the sample (82.5%), although at a notably lower percentage than other items.

FIGURE 1 -ABOUT HERE
The majority of teachers (94.9%) participated in trainings that were offered as part of the implementation of TEACHNJ and AchieveNJ.When we examined their self-reported knowledge, however, we found much variation across key concepts.That is, the teachers in this study were more knowledgeable of the most immediate and concrete aspects of the new evaluation system and less knowledgeable of the law TEACHNJ and its implementation platform AchieveNJ.For example, using a rating scale, a majority of teachers said they were knowledgeable (combination of "very knowledgeable" and "somewhat knowledgeable") about "student growth objectives" (84.6%), the "teacher evaluation system" (82.5%), and the "teacher observation instrument" (71.9%).In contrast, less than half of teachers said they were knowledgeable about "AchieveNJ" (46.8%), "TEACHNJ" (43.1%), the "school improvement panel" (34.4%), and the "District Evaluation Advisory Committee" (29.1%).

FIGURE 2 -ABOUT HERE
Next, we examined teachers' evaluation of implementation efforts of the new evaluation system.As seen in Figure 3, teachers are critical of these efforts.A clear majority (more than two-thirds) of the sample disagreed to some extent with statements describing the implementation efforts.Slightly more than four-fifths (81.8%) disagreed with the statement "the rollout of the new teacher evaluation system, TEACHNJ, was done at a comfortable pace that allowed for effective implementation," and slightly more than three-quarters (77.9%) disagreed with the statement, "The new laws and changes in teacher evaluation are effectively communicated by NJ Department of Education."

FIGURE 3 -ABOUT HERE
Not only were the teachers in this study critical of the implementation efforts, but the majority of respondents reported pessimistic views of the impact of the new evaluation system on teaching and learning.About three-quarters (74.0%) of respondents said it will not improve teaching (30.1% strongly disagreed and 43.9% disagreed with a positive statement).Slightly more than four-fifths (81.1%) of respondents said the new system will not improve student learning (30.1% strongly disagreed and 51.0% disagreed with a positive statement).
Finally, we examined the distribution of the dependent variable: Support for AchieveNJ.As expected, two-thirds of respondents (66.5%) indicated they oppose AchieveNJ (29.5% strongly opposed and 37percent opposed).Only one-in-ten respondents (9.5%) said they favor AchieveNJ.The remainder, one-fourth of the sample (24%) said that they are not sure whether they favor or oppose AchieveNJ.

Multivariate Analysis
Earlier we noted that teachers' support for educational reforms is often shaped by the political environment in which the reform was introduced.Furthermore, we found that the process in which TEACHNJ and AchieveNJ were developed was highly contentious.Consequently, we now turn to discuss the effects of different sponsor frames on support for AchieveNJ, as illustrated in Table 2.

TABLE 2 -ABOUT HERE
Model 1 presents the bivariate correlation between the sponsor framing of the new evaluation system and the respondent's support for the system.We found that when the system was presented as an initiative of Governor Chris Christie (rather than an initiative of the New Jersey Legislature), the new evaluation system received less support from teachers (model 1; b=-.356, p<.001).This pattern persisted after we controlled for sociodemographic and professional characteristics (model 2).This should not be a surprise given the fact that the sponsor framing was randomly assigned to respondents.The significant effect of sponsor framing holds across all other models, regardless of the controls added to the model.
Model 2 introduces the control variables.Sociodemographic characteristics and most professional characteristics had no significant effect on support for the new evaluation system.The only exception was the coefficient for tested subject teachers (those teaching mathematics and English/language arts in grades four through eight).In the first two models (models 2 and 3), the coefficient's effect size is very small and not significant.However, in the final three models, the coefficient increases and becomes statistically significant (model 6; b=.191, p<.01).In other words, the effect of tested subject teachers on the outcome variable was suppressed by teachers' evaluation of implementation and their expectations of policy impact.Compared to other teachers, tested subject teachers showed more support for AchieveNJ, but they were also more critical about the implementation efforts and the impact of the new evaluation system.3Thus, without these engagement measures in the model, the contribution of tested subject teachers would be suppressed.
Finally, we examined the effect of four engagement measures on support for the new teacher evaluation system.Not surprisingly, teachers' endorsement of the underlying principles of the new teacher evaluation system had no significant effect on support for the new system (model 3).In contrast, teachers' self-reported policy knowledge had positive and significant effects on support for the new evaluation system (model 4).Teachers who perceived themselves as knowledgeable of various components of the reform expressed greater support for the new system.This coefficient, however, reduced in size and loses significance when we accounted for other engagement measures (model 7).This might be a result of the strong correlation (or overlap) between self-reported policy knowledge and evaluation of implementation (r=.366, p<.001).
Both teachers' evaluation of implementation efforts and teachers' expectations of policy impact had positive and significant effects on support for the new evaluation system (models 5 and 6).Teachers who expressed a more positive view of the implementation efforts also showed more support for the new evaluation system.Teachers who perceived the new policy as efficacious also showed more support for the new evaluation system.This pattern was consistent, though the coefficients' effect size shrunk, after we accounted for other engagement measures (model 7).This might be a result of the strong correlation (or overlap) between the two variables (r=.486, p<.001).These variables accounted for much of the variation in the outcome variable, as reflected in the adjusted R-squared.

Open-ended Responses
As mentioned above, we used open-ended response data to further examine teachers' views toward the new teacher evaluation policy (i.e., TEACHNJ) and its implementation platform (i.e., AchiveNJ).We identified five key themes in the data.First, similar to the pattern we discussed in the quantitative analysis, many teachers simultaneously expressed both support for change and critique of the current direction of policy.For example, a Latina teacher with 11 years of teaching experience wrote, "Although I do believe that there needs to be a change in education, I don't feel that implementing a new evaluation system is the answer.We should be more focused on engaging students and families, once this is done, we then can focus on the teacher's accountability."Echoing the concern that there was a problem that needed to be addressed, a White male teacher with 28 years of teaching experience pointed to problems with both old tenure laws, yet argued that the new rules add a new set of concerns: "Although there was abuse of the old tenure laws, I have seen with the new ones administrators using the new rules as an intimidation tool.It seems to me that you are at the whim of evaluators as to your job status.If they take a dislike to you for any reason the new system is sufficiently complicated that it would be easy to mask the real reason for bringing tenure charges and dismissing a teacher." A second theme that emerged highlighted concerns about the negative consequences of the new teacher evaluation policy.One-in-five teachers (20.9%) mentioned these sentiments in their responses.This theme maps well on responses to survey items about the perceived impact of the policy. 4For example, a relatively new White male teacher, with five years of teaching experience, stated that he is now discouraged about how he will continue to grow as a teacher and is considering a career change.He suggested that the new policy negatively affects teacher's morale: "The new teaching evaluation system, with random unannounced visits that seem to count, while providing little feedback has given a sense of "Big Brother" to the classroom.It has taken the fun out of trying new things and sitting back and reflecting to see if it works […] the haphazard implementation of all of these new systems have turned what was a profession that I loved when I first started 5 years [ago] into something that is turning into a job.
[…] I have recently stopped working on my master's and am currently considering a change in profession due to the direction I have seen our school system take in the past few years." Other teachers addressed common narratives of teaching to the test and wasting resources (e.g., time).A Latina teacher who has being teaching for 25 years wrote, "Too much time is spent teaching to the test.This is just a small measurement of student learning."This concern was echoed by a White female teacher who believes that the new policy does not allow her to use the skills that she has gone to school to learn and has been honing over the last ten years.She commented, "Stop wasting teachers' time and let them teach.With all of the time spent learning about the evaluations, teachers could be using that time to address many other issues with their students." The third theme concerned the DOE's poor implementation efforts.One-in-ten teachers (10.4 %) addressed this issue in their responses.Many teachers commented about the quick pace of the implementation, which did not allow teachers to fully engage.For example, a multiracial female midcareer teacher noted: "I do believe that overall this new system will be beneficial.I am concerned that it was rolled out a bit quickly, not giving educators and administrators time to understand it fully.Because of the speedy implementation, I do believe there will be a learning curve.A slower roll out of this process would have allowed everyone to find and correct the issues without having to hold to a flawed process." In addition to writing about the quick pace of the implementation, teachers also commented about poor training and preparation.For example, a White female teacher with nine years of experience reported that she has seen a number of different policy rollouts and has worked in a district that supports its teachers, in her estimation, well.However, the implementation of the new policies caught even the district off guard.As she reported, "Just to clarify, in my district, we were very prepared for the Charlotte Danielson model and started our training about 4 years ago.However, the training with the SGP's & SGO's was poorly done, not because our district did not provide resources, but because the state was slow in providing the district with needed answers." A fourth theme that emerged concerned the politics of introducing and advocating for the new evaluation system.While our quantitative analysis indicates that teachers' support is affected by the political environment, teachers' open-responses shed further light on possible reasons for this association.Some teachers indicated that they view the new policy as part of a political campaign against public education.For example, a White female teacher with over four decades of teaching experience, indicated that she is suspect of anything that Governor Christie proposes given her assessment of his motives.She wrote, "I would never support anything proposed by this governor who has decimated and vilified teachers to the extent that he has."Similarly, a White male teacher with 25 years of experience opined that this policy was meant to inflict damage."This new system is the biggest waste of time and resources I have witnessed in the last 25 years," the teacher wrote."I guess [Governor] Christie is going to try to kill the public schools with this initiative." Other teachers criticized the fact that the new policy is not informed by relevant knowledge and professional experience.For example, a White male teacher wrote, "I would like to see Educators involved in making the decisions.Governor Christie has no knowledge of how a classroom works."Another White female teacher with 14 years of experience pointed to the de-professionalization of education policy and how teachers are going to be held to standards for which they are not trained."Governor Christie just passed legislation on kids with Dyslexia […] yet teachers are not trained in teaching these kids (I am not).Yet we will be held accountable for these students' growth.This is not fair."She noted that basic educational theory was seemingly missing from the creation and consideration of the policy, writing, "What happened to Piaget's theory of childhood development?It seems to have gone out the window when they expect kids to all be at a certain reading, math level by a certain date.Everything we learned in college about child psychology seems to have vanished." Teachers also linked the new evaluation system to the broader phenomenon in which businesssector ideas are entering schools.A White male teacher with over a quarter-century of teaching experience argued that the policies were created to support testing corporations rather than to improve education: "The use of high stakes standardized test score [sic] in evaluating performance is what makes the new system an absolute farce.It ignores the overwhelming amount of research that indicates that these measures are suspect at best.As far as I am concerned, this new system has nothing to do with children or their teachers and everything to do with the corporate reform movement and the testing-industrial complex and their attempt to privatize public education." Another White female teacher added, "They are trying to make teaching a 'business,' and it is not a business and will never be.Silly me, I thought we were teaching kids, not computers and guess what?Kids make mistakes and should make mistakes, because that's how they learn from them.Kids are not motivated by standardized tests and here's another clue, NEITHER ARE TEACHERS!!!!!" (Emphasis added by respondent).
A fifth theme included questions and doubts about the validity of the new evaluation system.In contrast to the previous themes, this theme was not reflected in the survey instrument.Thus, this theme extends our insights regarding how teachers view the new evaluation policy.A female White teacher with over a decade of experience argued that the important role of out-of-school factors in shaping students' achievement was ignored by policymakers."While I believe it is important for educators to be evaluated, and not just a one-time classroom visit, I do not believe that a teacher should be measured by their students' performance," she said."Some students have isolated situations of which [sic] no teacher will make a difference."Similarly, a White female teacher with 21 years of experience added that the learning environments for certain students were not taken into account: "These tests are not designed to fairly assess all students.Special education students have minimal opportunity to be successful on these tests and the way they are administered is not an accurate assessment of how students are assessed in their normal day […] The new PARCC assessments are computer based, and my students are not proficient with technology." In sum, the open-ended responses have suggested that New Jersey teachers recognize the need for change and endorse the principles of the new teacher evaluation system.However, the teachers in this study expressed concerns about different aspects of the evaluation system, including possible negative consequences, poor implementation, the politics that accompanied the policy process, and the validity of the system.These concerns, in turn, translate into clear opposition to the evaluation system.

Discussion
Through an analysis of a survey of New Jersey teachers, we examined teachers' views of a highstakes teacher accountability policy (TEACHNJ) and its implementation platform (AchieveNJ).Our study points to an interesting paradox.On the one hand, a majority of teachers are aligned with the principles of the policy.For example, they agreed that effectiveness should be measured to ensure students have the best teachers.Some even endorsed linking tenure to effectiveness.On the other hand, teachers were critical of the policy and the implementation efforts.The vast majority (75%) of teachers indicated that they believe that the new policy will not improve their teaching or students' learning.And, only 10% of teachers said they support the new teacher evaluation system.The rest opposed or were undecided.The paradox is illustrated in the closed-ended items and the openended responses.Many of these responses had a similar structure: "I believe in teacher accountability, but…."In fact, Woody et al. (2004) echoed these patterns, based on research in four districts in California.Overall, participants in New Jersey and California saw the need for the reform, but they did not necessarily agree with the efforts being carried out to implement the reform.In other words, teachers tend to agree that the idea of "teacher accountability" is now taken for granted, yet the debate about how teachers should be held accountable remains unsettled.
Our empirical findings also highlight the role of the political environment on teachers.As we describe in the study setting, the process in which TEACHNJ and AchieveNJ were introduced was highly contentious and political.Local media coverage reported this tension, regularly citing disparaging remarks by Governor Christie toward the local teachers' union.Rather than asking teachers about whether the political environment affected them, we used an embedded experiment.This innovative approach enabled us to estimate the causal effect of a treatment in the context of a large survey.Respondents who read about AchieveNJ as an initiative of Governor Christie were significantly less supportive of the new teacher evaluation system than respondents who read about AchieveNJ as an initiative of the New Jersey Legislature.Open-ended responses shed light on the how teachers think about the politics of education.Some of the teachers accused Governor Christie for leading a campaign against public education.Others commented on the de-professionalization of education by individuals who craft policy solutions with limited expert knowledge or experience.
Teachers also pointed to the growing role of corporations in the education and schools (e.g., testing agencies).
Finally, our empirical findings suggest that two additional factors are associated with teachers' support for AchiveNJ.First, similar to previous studies (Desimone, 2002), we found that teachers holding positive views of the implementation efforts also reported higher levels of support for the policy.Both closed-ended items and open-ended responses point to the importance of timing and pace.Many teachers reported that TEACHNJ was implemented very quickly with teachers feeling rather unprepared.Second, as reported by Datnow & Castellano (2000), teachers who believed that the policy will have an impact on teaching and learning demonstrated higher levels of support.
Open-ended responses suggest that teachers were concerned about the negative impact of the new policy on morale, motivation, and engagement.
Our study has three limitations that should be considered.First, the study is based on a small nonrandom sample of teachers, which may raise a question about generalizing about the entire population of teachers.As we discussed earlier, this sampling design is a result of limited resources and support.In an effort to address the sampling design, we calculated and applied sample weights (based on gender, race/ethnicity, and county) to adjust for demographic differences between the sample and the population.The relatively small sample (n=444) prevents us from applying more sophisticated analysis to the data (e.g., multi-level analysis of the nested nature of education systems).Second, the study is based on a sample of teachers in one state, New Jersey.It is possible that teachers in other contexts (states or countries) will have different views and attitudes towards high-stakes teacher accountability policies.Third, data for this study were collected during the first academic year after the implementation of TEACHNJ.It is possible that teachers' views may have changed over time once they had more experience with the new policy.Evaluation of TEACHNJ (Callahan & Sadeghi, 2014), however, has shown that "while teachers indicated that they were observed more often, they also noted the value of the observation was diminished" (p.56).
Given these limitations and in light of the findings of this study, there are at least two possible directions for future research.First, additional systematic research on teachers' attitudes toward high-stakes teacher accountability policies is required.Quantitative studies based on large and random samples will address the limitations and expand our findings.These studies could introduce additional factors that may shape teachers' views toward reforms, such as party identification, political ideology, leadership roles, sense of empowerment, and motivations to join the teaching profession.Furthermore, these studies could examine the role of organizational affiliation (e.g., unions) and social networks (e.g., personal and professional networks).Qualitative studies based on in-depth interviews and observations will help scholars and stakeholders to better understand how teachers interpret and "make sense" of high-stakes teacher accountability policies.Additionally, these studies will provide further insights about why teachers support or oppose these policies.
Second, additional research would explore teachers' views of other applications of NPM models, such as school choice, privatization, corporatization, technology, alternative requirements and training programs (e.g., Teacher For America), and standardized testing (e.g., reliance on ILSA and national assessments to inform policy and practice).Comparing teachers' views across these different policies and practices will provide a more refined understanding of their engagement with these policies.Do teachers differentiate between these policies?If so, what factors shape support for one policy versus another policy?
In conclusion, teachers are important actors in the education system.Current education reform movements, reflected in high-stakes teacher accountability policies, seem to undermine teachers by challenging the building blocks of their professional status (e.g., expert knowledge, specialized training, and autonomy).This study demonstrated that teachers endorse the principles of new teacher evaluation policy, but they oppose the actual policy and its implementation platform for a variety of reasons.In an era of multiple education reform movements, scholars can enrich the policy process by engaging teachers' voices.Careful analysis of teachers' views not only could shed light on policy implementation, but it could also yield innovative ideas for policies aimed at improving education quality.

Finally
, we use open-ended comments to shed light on potential reasons behind the statistical patterns.The next section provides background about the study setting: the state of New Jersey.