Interrelations between Policymakers ’ Intentions and School Agents ’ Interpretation of Accountability Policy in Israel

The study examined the interrelations between policymakers’ intentions for test-based accountability policy, and school agents’ perceptions and actions with regard to this policy. Mixedmethods were used and encompassed 24 policymakers, 80 school principals, 168 teachers and case studies of four schools. New institutional theory, including the concept of “environmental shift” (Powell & Di Maggio, 1991) and the metaphor of “coupling” (Weick, 1976), served as the conceptual framework. Findings indicate that the interrelations between intentions, perceptions and actions are mainly tightly coupled. These are discussed by invoking three types of institutional isomorphism (DiMaggio & Powell, 1983): coercive, mimetic and normative.


Introduction
Within the literature on public policy, there is extensive research on both policy determination and policy implementation; however, there is less research with regard to the translation of policy from intention to practice.Within the field of educational policy, accountability, the process of evaluating schools based on student performance (Figlio & Loeb, 2011), is prominent in education systems across the world.Indeed, it is believe that accountability extends down into the system and increases the influence of political authorities in shaping classroom behavior (McDonnell, 2013).Thus, the purpose of the current study is to examine the relationship between the policy intentions of policymakers for accountability, and the interpretation of that policy by school agents.An understanding of this relationship may shed light on the extent to which policy implementation is consistent (or inconsistent) with policy aims (Spillane, Gomez & Mesler, 2009) with the primary focus being on the intentions of policymakers vis-à-vis the perceptions, interpretations and actions of school principals and teachers.
This study situates test-based accountability at the intersection between policy studies and institutional theory.The former views the determination and implementation of test-based accountability as a contextualized process involving two main agents: policymakers, who determine policy, and school agents (principals and teachers) who implement it.The latter positions policy implementation within the broader context of complex organizational settings and relates to how implementers perceive and interpret policy, as well as what actions they take to address it (McDonnell, 1994a(McDonnell, , 1994b)).The theoretical framework utilized for an analysis of this relationship is new institutional theory (DiMaggio & Powell, 1983;Powell & DiMaggio, 2012).

Test-based Accountability
Accountability systems involve the collection of information regarding student performance, which is usually carried out through state-mandated tests, considered by some to be the "centerpiece" of accountability systems (Hanushek & Raymond, 2002;Hanushek & Woessmann, 2014).Policymakers, perceiving tests as being an effective policy instrument to influence education (McDonnell, 1994b), are attracted to them for a number of reasons.They perceive tests as being authoritative, objective, and fair (Shohamy, 2001;Shohamy, 2017).The visibility of tests provides an obvious indication of action (Linn, 2000) and thus policymakers are perceived to be promoting a sound educational agenda.Furthermore, tests are not only visible to the public, but the public also believes in tests.
Due to the relatively low cost, at least in comparison to other means of reform, testing is best suited to the constraints of government budget.Finally, tests can be implemented quickly which may be most significant to policymakers, considering their often short terms of office.

Intentions of Policymakers for Test-based Accountability
Policy intentions relate to the purpose and expectation of a policy (Shohamy, 2001).Intentions not only determine actual policy but may also shape policy adoption and implementation (Dunn, Jaafar, Earl, & Katz, 2013;May, 2012;Unger, 2008).Policy has even been considered as a "transformation of intentions" (Placier, Hager & Hull, 2005, p. 96).More often than not, a range of intentions drives a given policy and this appears to be the case for test-based accountability, for which policymakers have multiple intentions.
One of the most prominent intentions for test-based accountability is school improvement and the enhancement of school performance (Cohen, Moffitt, & Goldin, 2007;Malen & Rice, 2008).Another intention is to raise student achievement (Hamilton, Stecher, & Klein, 2002).When studying policy intentions through interviews with policymakers McDonnell (1994a), a leading scholar in the field of public policy, detected an additional intention for test-based accountability whereby policymakers use tests with the intent to heighten the awareness of teachers and administrators regarding instructional goals and expectations.Another intention has been found to be motivating teachers to work harder and make more effort (Firestone, 2014;Linn, 2004).Some scholars have argued that it is a means for top-down command and for wielding control (Brindley, 2008;Moe, 2003).

Policy Presentation
Policymakers are aware of the importance of gaining the attention of the implementing agents (Spillane, 2002).This is accomplished in many ways, primarily through laws, directives and regulations.However, the policy message may also be transmitted through other channels, for example through the provision of inducements (Bardach, 1977).Apart from gaining the attention of future implementers, policy presentation is designed to encourage implementing agents to think differently about their behavior in order for policies to work as levers for change and reform (Weiss, 1990).Thus the present study focuses not only on intentions but also on ways through which policymakers have presented their test-based accountability policy to school agents.

Policy Implementation
Policy implementation can be perceived as "the connection between the expression of governmental intention and actual results" (O'Toole, 1995, p. 43).Thus, implementation should correspond to intent and those who implement policy are required to interpret and understand policy and take action accordingly.However, implementation does not always play out according to the original blueprint set out by policymakers and there may be what Marshall (1998) has coined a "slippage between policy intent and outcomes" (p.103).
Furthermore, policy implementation is a contextualized process, typically occurring in complex organizational settings wherein implementers occupy different levels of hierarchy (O'Toole, 2000).In the case of educational policy, its implementation depends specifically on principals and teachers (McLaughlin, 1991;Russell & Bray, 2013) who are working within the complexity of a school organization and who are expected to have the required capabilities (Cohen, Moffitt, & Goldin, 2007).It is often adapted to the school and classroom contexts in different ways than may have been intended (Hill, 2001).
Even though the responsibilities and professional demands of principals and teachers differ, both groups face the same policy expectations; both must interpret policy intentions and expectations (Heineke, Ryan & Tocci, 2015;Singh-Pillay & Alant, 2015) and must respond by setting a course of action.Thus, in understanding the implementation of test-based accountability it is necessary to focus on these individuals, specifically paying attention to their perceptions of policy and resultant actions as well as their educational context.

School Principals and Test-based Accountability
Studies have shown that school principals pay attention to state policy signals in general and focus on accountability policies in particular (Herman, 2007).When principals are confronted with the demands of test-based accountability, they make their own interpretations regarding the implications for them and their schools (Louis, Leithwood, Wahlstrom, & Anderson, 2010;Seashore & Robinson, 2012).In general, principals who perceive testing policy to be supportive of their work and valuable to the school are more inclined to support and implement it (Lyons & Algozzine, 2006).
The actions of principals may be divided into those on the level of the school and those on the level of the classroom.Actions on the level of the school include: planning resource allocation according to state testing demands (Dorn, 2007;Jones & Egley, 2004;Nichols & Berliner, 2007;Smith, Edelsky, Draper, Rottenberg, & Cherland, 1989), increasing oversight of teachers' work (Anagnostopoulos & Rutledge, 2007;Spillane, 2013), changing student classifications so as not to include them in the statistics (Figlio & Getzler, 2002;Glass 1990;McGill-Franzen & Allington 1993;Smith 1991), and analyzing test results and developing appropriate school plans (Herman, 2007).

Teachers and Test-based Accountability
Teachers perceive both the positive and negative aspects of test-based accountability (Louis, Febey, & Schroeder, 2005).Those who perceive the positive aspects of test-based accountability believe that a system of checks for maintaining standards is necessary (Pedulla et al., 2003) and that there is value in testing as it ensures personnel will take responsibility for their actions (Robertson, 2003).Teachers perceive the tests as contributing to school effectiveness (Louis, et al., 2005), making them more aware of educational outcomes (DeBard & Kubow, 2002) and directing them to planned and focused instruction (Hamilton, 2003).On the other hand, teachers perceive test results as reflecting test-taking skills rather than language ability (Tahmasbi & Yamini, 2012).Furthermore many believe that test-based accountability decreases morale (Abrams, 2004), limits autonomy and harms professionalism (Crocco & Costigan, 2007;Leithwood, Steinbach & Jantzi, 2002;Rustique-Forrester, 2005).

Theoretical Framework: "New Institutionalism"
The theoretical framework of new institutionalism has been used in a variety of organizational and institutional settings, not least being schools, as "public schools display all of the features addressed by institutional theory" (Mitchell, 1996, p. 174).This framework can provide the perspective through which to examine interactions between educational policies on the one hand, and school and classroom practices on the other (Burch, 2007, Burch & Spillane, 2005;Cerna, 2013;Coburn, 2004;Hasse & Krücken, 2014;Metz, 1990;Ogawa, 1992;Spillane & Burch, 2006).
Schools are considered to be interrelated organizations, which are open to external influence from their environment, or institutional field (DiMaggio & Powell, 1983).As such, schools must cope with diverse and sometimes contradictory demands and pressures filtering in from the institutional field.Of particular influence is an "environmental shift" that occurs when a component of the institutional field modifies an expectation or requirement of the organization within the field (Powell & DiMaggio, 1991;Rowan, 2002).One may consider state-mandated tests as influential environmental shifts that critically affect schools (Spillane & Burch, 2006).
New institutionalism has shown that schools generally seem to respond to environmental shifts in similar ways (Burch, 2007;Ogawa, 1992;Powell & DiMaggio, 1991;Rowan, 1982Rowan, , 1995;;Rowan & Miskel, 1999).For example, Hanson (2001) suggests that "the pressures on schools from organizations and agencies in their environmental fields...are quite similar across the country, and in consequence, public schools in one region of the country tend to act like schools in other regions" (p.648).Furthermore, school personnel develop their understanding of what is expected of them through participation in the same institutional environment (Meyer, Scott & Deal, 1981;Rusch, 2005).
Following from the idea that the institutional field exerts pressure on schools to which they respond, test-based accountability policy may exert pressure on the perceptions and actions of school agents (McLaughlin & Talbert, 1993;Rustique-Forrester, 2005;Spillane, 1999).These will be investigated in the context of a national test in Israel.

The Context
The Israeli education system is centralized and as such is characterized by the control of education by a central government body and a clear hierarchical structure.All public primary and secondary education, including school personnel, is government funded.Administrative and curricular decisions cascade down from the Ministry of Education and through the different levels of officialdom via reliance on inspectors who ensure the implementation of central regulations.Schools are required to follow national curricula that specify the standards and objectives to be pursued by all schools in all subjects and textbooks used in schools are to be approved by the Ministry.The Ministry also defines exit level achievement through a national matriculation exam (known as Bagrut).
In 2002, the Ministry of Education introduced a new, compulsory test battery, known as the Growth and Efficiency Measures (GEMs), into elementary and junior high schools.It was designed to provide information about three aspects of schooling: academic achievement in four core subjects -Hebrew/Arabic (as native language), Mathematics, Science and English as a foreign language (EFL); school climate; and pedagogical work environment.The information relevant to each of these aspects of schooling is collected through the following: achievement tests in grades 5 and 8; questionnaires completed by all students from fifth through to ninth grade; and interviews with school principals and teachers.The GEMs is designed and administered by the Ministry of Education.The Ministry also analyzes the results and provides individual reports to each school.
The school principal is responsible for making all the necessary arrangements for test administration, and for interpreting the findings and implementing any necessary changes based on the results.As for teachers, as it has been suggested that teachers of subjects tested respond differently to their institutional field than non-tested subjects (Spillane & Burch, 2006;Spillane & Hopkins, 2013) the focus of the present study is on EFL teachers.English is a compulsory subject taught in all public schools from fourth to 12 th grade and as such has been included among the core subjects to be tested on the GEMs test battery.
The research questions studied within the context of test-based accountability (as exemplified in the GEMs tests) were: 1. What are the intentions of policymakers related to test-based accountability, and how are policy intentions presented to school agents (principals and teachers)?2. How do school agents interpret the test-based accountability policy?3. How can the interrelations between policy intention and implementation in the context of test-based accountability be characterized?

Method
A mixed-methods approach was chosen as it generates most useful results (Johnson, Onwuegbuzie, & Turner, 2007) particularly in accountability studies (Stecher & Borko, 2002) and within institutional theory (Weick, 1976).A qualitative approach was used to collect data from policymakers and school personnel as it provides more specific information and leads to deeper understanding (Firestone & Mayrowetz, 2000;LeCompte & Preissle, 1993).Explicit intentions appear clearly as written directives and articles of law, but alongside these are implicit intentions, which can only be tapped through interviews with the policymakers themselves.A case study approach allowed for a more in-depth study of the beliefs and actions of school agents, but it was supported by a quantitative approach, which allowed for the collection of information from a large number of participants and allowed for a description of general trends.

Participants
The participants included policymakers, school principals, EFL teachers and four schools.

Policymakers.
A purposeful sampling technique served to select and recruit participants, with the intent of reaching all those policymakers who were directly or indirectly involved in the policymaking process (Creswell, 2007).This led to 18 policymakers who were holding or had held key government positions in education and were involved in setting testing policy in general and GEMs policy in particular.Six additional participants were recruited utilizing a snowball sampling technique (Warren & Karner, 2010) where we asked our initial 18 participants to connect us to any other policy actors who could serve as informants on the policy design.The final sample was thus comprised of 24 participants who reflected various levels of position in the hierarchy of the Ministry of Education, held roles with a range of responsibilities, and had participated in the preparation of policy documents and directives regarding testing policy.
The policymakers who participated in the study included three Director Generals of the Ministry of Education (past and present), the head of the Pedagogical Directorate who also serves as the Deputy Director General, the head of the Pedagogical Secretariat, two chief scientists (past), two subject chief inspectors, three district inspectors, heads of various divisions and departments in the Ministry of Education, including the Division for Elementary Education, the Division for Secondary Education, the Department of Evaluation, the Curriculum Division, the Department of Teacher Training, and District Offices.
Principals.Principals were randomly chosen from the official list of schools participating in the GEMs test during the year of data collection.Three-hundred principals were asked to complete a questionnaire of which 80 fully completed the questionnaire, representing a 27% response rate.Among the final sample of principals there was almost equal representation of female and male principals (56.2% female; 43.8% male); two-thirds administered elementary schools and one-third secondary schools.Half served in the secular sector, a quarter in the religious sector and a quarter in the Arab sector.Only a fifth of the principals had experience of three years or less and more than half had over nine years' experience.
EFL teachers.A purposive sampling procedure was employed by which questionnaires were distributed to EFL teachers attending compulsory training courses held at Centers for Teacher Professional Development.An effort was made to ensure that the participating teachers represented a valid sample of EFL teachers working in schools around the country.A total of 168 teachers participated in the study, the majority being female but with equal representation of elementary and junior high school teachers.Half of the teachers taught in the secular sector, a quarter in the religious sector and a quarter in the Arab sector.Two-thirds of the teachers had over nine years teaching experience, and 80% of teachers held at least a bachelor degree.

Schools.
The inclusion of a case study of schools was intended to provide additional information about the perceptions and actions of principals and teachers within their working context and thus enrich the data collected through quantitative means.Four schools participated in the case study.Those interviewed and observed within the case study of schools included all the school principals alongside administrative staff, and the EFL teachers in each of the schools.Schools were selected based on typical case sampling which is based on the rationale that these schools are illustrative as they are in no way extreme, deviant, exceptional or irregular (Patton, 2015).

Data Collection and Analysis
Data was collected over a one-year period, from spring 2005 to spring 2006 using a wide range of instruments.Data collection began with the collection of qualitative data from policymakers and continued throughout the year in the case studies of schools, through interviews and conversations with principals, administrative staff and EFL teachers.All the qualitative data was coded using preassigned theory-based and inductive data-driven codes (Bogdan & Biklen, 1998).All qualitative data was coded using Atlas.ti.The quantitative data was collected from principals and teachers during the school year through the use of specifically designed questionnaires.
Interviews.Three semi-structured interview protocols-for policymakers, school principals and EFL teachers-were developed and were based on the conceptual framework presented in the literature review.The semi-structured interview protocols included open-ended questions which were designed to allow for freedom of speech and encourage participants to pursue various directions of thought as they saw fit.The protocols underwent content validation and qualitative evaluation by four experts consisting of an expert in policy studies, an expert in assessment, an expert in educational administration and a teacher trainer.Interview questions were altered systematically as new information became available and themes emerged from on-going data analysis.Policymaker interviews included such topics as goals and priorities, views on testing, performance and achievement.Topics for school principals included impact of the GEMs prior to and following administration, performance expectations and consequences and general impressions of accountability system.Topics for EFL teachers included impact of the GEMs prior to and following administration, curriculum design, and classroom instruction and testing.
Observations.Observations of EFL teachers were carried out in a number of classrooms mainly, but not exclusively, in the grades participating in the GEMs (i.e.fifth and eighth grades).The main purpose of these observations was to complement the data collected through the interviews.The observations concentrated on the topics focused on by teachers in the classroom and on the types of materials that teachers used in the classroom, such as worksheets and practice tests used in test preparation.In addition to the observation of the classroom, observations of staff meetings were also carried out, particularly of those held prior to GEMs administration and upon receipt of the GEMs results.The foci of such observations were issues related to the GEMs, such as the division of labor in preparing materials for test preparation, and the discussion of test results.

Documents.
Published documents related to official legislation were collected, such as the Director General Code of By-Law which is disseminated to all schools and available to the public through the Internet.In addition, the documentation of the official protocols of the Education Committee of the Knesset (Parliament) which are freely available on the Knesset Internet site was also collected.Documentation also included any written material prepared by the Department of Evaluation, such as PowerPoint presentations, meeting agendas, and written correspondence.
Questionnaires.Two questionnaires were developed to elicit information regarding the principals' and teachers' perceptions of and actions taken in response to the GEMs.The questionnaires were based on other surveys used in previous studies related to principals' and teachers' perceptions of state testing programs (Kennedy, Ball & McDiarmid, 1993) and on the literature.The questionnaires underwent content validation and qualitative evaluation by eight experts consisting of one national English inspector, two teacher trainers, an expert in language test design, an expert in educational evaluation, a school principal and two teachers.

Results
Test-based accountability is a prominent fixture in the centralized education system of Israel and one of its foremost policies.Its central feature, the state-mandated GEMs test administered to fifth and eighth graders in all public schools, has impacted those schools in various ways.The results of the current study relate to the GEMs test during its third year of implementation.

Intentions of Policymakers
The first research question related to policymakers' intentions for the test and its presentation.It was found that policymakers attach multiple intentions to the testing policy.One of the main intentions was to address the decline in student achievement as highlighted in the low ranking of Israeli students on international comparison studies.Many of the policymakers mentioned that "the beginnings of the problems were the disastrous results on the PEARLS and PISSA tests".Moreover, within a single year prior to the GEMs, three separate meetings held by the Education Committee of the Israeli Parliament were exclusively devoted to the topic.This problem was perceived as being particularly harsh as, in the words of one school principal, "the results are in the media and policymakers don't like bad publicity."A policy of additional testing to ensure success on future international tests seemed logical, especially as policymakers were aware of the advantages of testing.Moreover, there had been national tests in the past, thus policymakers saw the GEMs as a natural progression from previous state-mandated tests.
Intentions can be divided into those relevant to the level of the school and those relevant to the level of the state.With regard to the school, policymakers believed that "there was no accountability, kids go from grade to grade, whether they know the material or not, with no accountability."Thus, the GEMs was necessary in order to ensure school accountability including towards the state: There must be accountability toward …the provider of the resources [the Ministry].The bottom line is that the state gives money to the school principal so he can do something, not so he can waste the money, and he is required to report what he did with it.
Further, policymakers believed in the need to set minimum standards and common goals for all."The number one goal of the GEMs is to reach certain standards, or agreed basic achievement that is clear to principals, teachers, and inspectors, what is the minimum expected from the pupil."The belief expressed by many policymakers who were interviewed was that, by getting schools not only to work toward common standards but testing schools to make sure that these common standards are achieved, there would be an automatic narrowing of the education gap.This sentiment is clearly expressed in the words of one policymaker: "The GEMs is a good infrastructure…it is an equalizing instrument that we can say all students need to reach." A consistent finding and commonly cited intention of policymakers in the context of statemandated tests is that the test ensured that certain subjects as well as specific content within a particular subject be taught and this is considered "the power of assessment…something that is assessed is of value."Another policymaker supported this by stating that "if there is no test in a subject then it doesn't get taught…in our culture tests are a defining factor."With regard to defining content, a declared intention of policymakers was for EFL teachers to implement the national curriculum in their classrooms.It should be mentioned here that a new EFL curriculum had been introduced to schools just prior to the implementation of the GEMs and policymakers apparently thought to "take advantage of the GEMs as a trigger to get teachers working according to curriculum." Besides the above intentions that related to the level of the school, policymakers mentioned two additional intentions that related to the level of the state.One was the intention of collecting data that would contribute to a more effective allocation of resources.The other was for evaluation purposes, though policymakers were aware that this may "be hard for principals to accept…that their work is checked".
Policymakers not only determine policy, but they are also aware of the need to gain the attention of policy implementers and motivate them to think about policy requirements.The findings indicate that policymakers did consider ways in which to present their policy to school agents: "You need to know how to market it correctly…so that the use will be right." Policymakers presented policy by various means and through a range of channels.Policy was portrayed formally through government regulations and directives.Official documents included the Director General Code of By-Law which is binding for all public schools and the official test specifications.Informal communication was also harnessed as "sometimes when wanting to bring about change it is better to operate through channels that have the least resistance." Informal communications were either written or oral, were more responsive to inquiries from the field than initiated by the policymakers, and were usually aimed at individual schools or school agents.Policy trickled down through the Ministry hierarchy thus: "The general inspector guides the district inspectors who guide the principals and the subject matter inspectors guide the subject matter coordinators in the schools."In-service training was another informal means of conveying policy.The Ministry organized "many conferences and study days for principals to learn about the GEMs".In the words of one policymaker "it's enough the school sees the syllabus of the in-service training given around the country to know what is expected of them." Having discussed the policymakers' intentions for and presentation of policy, we turn now to the second research question and describe school agents' policy interpretation, considering their perceptions and their actions.

Perceptions of School Agents
School agents understood that the GEMs was "intended as a lever for improving achievement" believing that indications of the need for improvement came from the international tests.School principals were aware that one of the driving forces behind the GEMs was the low ranking of Israel on international studies: "if we go for the international standard then I believe that Israel won't want to be left behind, so they will have to have the GEMs."EFL teachers also seemed to be aware of this, as one EFL teacher succinctly put it "it's obvious those on high had to do something after everybody sees how bad we do in comparison to the rest of the world." School principals and EFL teachers alike thought that one of the intentions was to ensure minimum standards for all students.School principals understood that the Ministry was interested in improving achievement and that the means to ensure it was by setting minimum standards.For one school principal it was clear that the Ministry "wanted to ensure a basic level that all the students in the country would reach…and by having the test the teachers would be oriented to this topic and they would work for the GEMs."Another school principal suggested that "there were some very weak populations that they [the policymakers] wanted to raise, so there was a need for a minimum standard that everyone needs to reach."One vice-principal believed that "the GEMs is a lever for improving achievement…we saw that if we teach according to the GEMs we can improve achievement."From among the thirteen EFL teachers in the case study, five mentioned the idea of minimum standards.One EFL teacher believed that it was important that students "everywhere around the country" reach a certain level of English proficiency.This standard could be helpful in that "we're all on the same page."Another teacher believed that this insurance of a "standardized standard" would also ease the transition as "kids can go from junior high to high school without any problems." School agents were aware of what was expected of them in terms of "action on the ground", specifically school principals were to prepare an action plan and teachers were to work according to the curriculum.In the words of one school principal: "I derive my action plan from the test, that was the intention and that is the use I make of the tests."Another principal added that "based on the test results I can set my school's aims and goals."The majority of principals who responded to the questionnaire understood they were expected to prepare an action plan.An EFL teacher explained that she understood policy intention to "make sure we [the teachers] are teaching the curriculum and we're teaching what we need to teach."Some EFL teachers thought that not only was the GEMs intended to ensure implementation of the curriculum but also when preparing tests they "should be based on the format of the GEMs." Finally, some attested to understanding that policymakers were mainly interested in evaluating their performance: "I'm being measured as a principal, and my teachers are being measured as teachers."One junior high principal indicated "it [the GEMs] is to evaluate staff, but from a desire to treat staff favorably, not to entrap the staff, though some may believe that this is the case."However, one principal specifically stated that evaluation "serves as ammunition against the principals and teachers."Some thought that the school is evaluated in comparison with other schools as the "inspector compares schools."In the words of one EFL teacher "the test is per class, or per domain, that shows you it's testing the teachers."

Actions Taken by School Agents
With regard to the actions taken by school agents to address the GEMs it was found that school principals and EFL teachers took specific actions, mainly prior to the test in the context of test preparation, and their actions seemed to be synchronized and well-orchestrated.Even though EFL teachers were not as enamored with the test as school principals were, the actions of both school principals and EFL teachers were similarly directed toward preparing students for the test and ensuring their success on it.School principals and EFL teachers were less synchronized in relation to the use of test results.
School principals were found to support test implementation as they encouraged their staff to cooperate and follow all the directives related to test preparation and administration.Principals also paid particular attention to test preparation activities, allocating additional hours for test preparation and what some have called "GEMs-oriented learning", assigning the best teachers to the classes to be tested and ensuring that teachers could provide individualized attention to students in need.Their budget allocation also reflected the emphasis on test preparation as they provided the necessary funds for the purchase of booklets for test practice and the photocopying of test preparation materials and practice tests.In addition, principals advocated teacher training if and when they considered it necessary or appropriate.As test preparation and administration require extensive work and coordination principals sometimes delegated the administrative responsibility of the test to a pedagogical or assessment coordinator, who served as liaison between the school and the Ministry, dealt with test administration procedures, monitored progress of test preparation, and was also responsible for the action plan subsequent to the receipt of test results.
Upon receipt of the GEMs report principals, together with the pedagogical or assessment coordinators, analyzed the strengths and weaknesses of their school, made changes where necessary, and an action plan was designed based on the data.In addition, they not only shared the results of the GEMs tests with the teachers, but also provided them with guidance in instructional planning and in identifying areas in need of improvement.Some school principals shared the results with parents, and some met with colleagues to compare results and discuss cooperative action.This was particularly evident between feeder elementary schools (in which the fifth grade tests take place) and their receiver junior high school (in which the eighth grade tests take place).
With regard to the actions of EFL teachers that address the GEMs, EFL teachers in both elementary and junior high schools focused on the test requirements and spent inordinate amounts of time exclusively devoted to test preparation.This was not only observed at the four schools participating in the case study, it was reaffirmed through the EFL teacher questionnaire.Based on the quantitative data 33% of the 134 participating EFL teachers reported devoting 11-20 hours to preparing their students for the test and a total of 21-30 hours of preparation time was reported by a quarter of the participant teachers.Taking into account that at the time of data collection the test was administered about eight weeks into the school year, and English lessons took place only three or four hours a week, one may conclude that most of the lessons prior to GEMs administration were almost exclusively devoted to test preparation.
In preparing students for the GEMs EFL teachers focused on the expected content based not only on the test specifications but also on published materials such as practice workbooks.EFL teachers also worked on promoting students' test-taking strategies, teaching the test and its format by giving practice tests made available by the Ministry on the Internet.
Data regarding ways in which teachers made use of the actual test results was limited, mainly due to the fact that the official report for each school only arrived toward the end of the year.Based on the existing data teachers made only limited use of the actual results included in the report that was sent to the schools as grades were reported on the level of the classroom and not on the level of the individual student.Thus, EFL teachers had an idea of the general trends in their classes, and if, for example, they perceived students to be weaker in writing they would include writing skills in their lesson planning.

Discussion
The discussion is based on the third research question with relates to the characterization of the interrelations between policy intention and policy implementation.Based on the findings, there is evidence that policymakers hold multiple intentions for the test-based accountability policy and school agents not only perceived these intentions, but also translated them into actions that were congruent with policymakers' expectations.In response to the test-based mandates, principals mainly believed that the test and accompanying report and action plan could bring order to an otherwise chaotic reality.Principals highlighted many of the positive aspects, including an emphasis on staff cooperation and better-planned budget allocation, as well as more attention to teacher professional training.Another positive outcome was a move toward more distributed leadership as principals, aware of the need to devote extensive time and energy to implementing the test, created a new role that involved responsibility for all aspects related to the GEMs.This role could also be expanded to include all aspects of assessment and testing.
Teachers, in response to test-based mandates and in line with the school policy as set out by their school administration, adapted their teaching.During the period leading up to the test teachers focused mainly on test requirements and on the expected content, and they devoted a lot of the available instruction time to test preparation.This sends a message to students that only succeeding on the test is important.Though teachers were found to be teaching skills that had previously not been taught, this entailed narrowing their syllabus, which meant that other content might have been neglected.Furthermore, to some extent teachers' academic freedom was reduced as content was dictated through the test.
It would seem that the principals in general have a more positive attitude toward the test than the teachers.This may be explained by the fact that the principals have more direct contact with the policymakers and with the channels through which the policy is conveyed.In this way they may have a better understanding of policy intentions.This also provides them with more of an opportunity for interaction with the policymaker in cases that they wish to clarify any of their misunderstandings.In addition, teachers are the ones who bear most of the burden in following the directives of the test and the success of their students on the test is a direct reflection on their performance.These pressures may explain the more negative attitude of the teachers toward the test.Though this study focused on a different education system, one that is centralized and has unique characteristics, these outcomes are not dissimilar to those found in many other educational contexts operating under test-based accountability.
Following the alignment of policy implementation with policy intention, institutional theory is invoked to characterize the interrelations between policy intention and implementation in the context of test-based accountability.Institutional theory offers "important insights for understanding contemporary [educational] policy and practice" (Burch, 2007, p. 93).Within the tradition of new institutional theory, the notion of loose coupling has been applied within organization studies for many years (Scott, 2004) and has been "especially prominent in explorations of educational organizations" (Swanson, 2005, p. 245).Loose coupling describes the match between different elements of an organization and elements in its institutional environment, such as governing agencies, or between the elements within the organization itself (Orton & Weick, 1990;Weick, 1976).Coupling elements refer to anything that may be tied or linked together (Weick, 1976).In the current study the coupling elements are the policymakers' intentions and school agents' perceptions and actions with regard those intentions.
The relations between schools and their environment have traditionally been considered loosely coupled (Weick, 1976).However, Spillane and Burch (2006) have argued that: "the ever increasing use of testing and systems designed to hold teachers and administrators directly accountable for student learning raise doubts about the assumption that schools...are loosely coupled systems" (p.88).Diamond (2007) supports this assertion based on his study of schools in Chicago's accountability system where he found that high stakes testing policies are tightly coupled with classroom practice.Fusarelli (2002) has confirmed this and argued that due to test-based accountability "the tightness of coupling is far greater than at any other time" (p.571).In the words of McDermott (2007), "the tightness takes the form of statewide standards, with mandatory tests…and accountability for results" (p.81).
In line with these scholars, the findings in the current study point to tighter coupling between schools and their institutional environment.The coupling of intentions, perceptions and actions will now be discussed through four dimensions of coupling that characterize the quality of the relationship between coupling elements (Weick, 1982) -dependency, directness, strength and consistency.Dependency refers to the extent of dependence between coupling elements whereby loosely coupled elements are relatively autonomous of each other (Pfeffer & Salancik, 1978;Weick, 1976), and interdependence is characteristic of tight coupling.Interdependency was found between schools and government.School agents, by the nature of the centralized education system, are almost completely dependent on the government for their functioning, including funding, teacher allocation, and curriculum.It would appear that the extent of dependence between government and school is amplified within the context of the GEMs especially in the eyes of policymakers who believe that as they fund schools they have the right to test and monitor schools.Furthermore, the Ministry defined test content and conditions for test administration and schools were obliged to cooperate.This and more, the Ministry provided schools with test results and analyses.In addition, the Ministry approved a school's action plan prepared by school agents and based on test results.
Directness refers to how directly the coupling elements are coupled; schools are loosely coupled with the environment when school personnel communicate with officials through several levels of hierarchy rather than directly (Beekun & Glick, 2001;Weick, 1982) and conversely the more direct the linkages between elements, the tighter the coupling (Weick, 1982).Within the context of the GEMs, policymakers were more directly in touch with school agents, through letter writing, responsiveness to individual school agents who initiate contact, participation in study days and even school visits.This type of direct contact was not customary outside the context of the GEMs.Thus, it would seem that for the dimension of directness policymakers and school agents were tightly coupled.
Strength refers to both the frequency and intensity of contact between the coupling elements, and ranges from frequent and intense contact to infrequent and weak contact (Weick, 1982).The stronger the linkage between elements the tighter is the coupling (Weick, 1982).Besides being direct, contact between policymakers and school agents was also frequent and intense.This was even expressly mentioned when school principals were asked to compare a year with the GEMs to a year without and they decidedly mentioned the frequent contact between them and the Ministry.Training, study days and meetings all indicated the strength of contact and provide evidence for the tightly coupled relations between policymakers and school agents.
Consistency refers to the variation in response of the coupling elements to a given stimulus (Orton & Weick, 1990;Weick, 1982).It is evident that on this dimension too there was tight coupling between policymakers and school agents.Their responses within the context of test-based accountability were limited to being test-specific and goal-specific.With regard to being test-specific, the focus of interactions between policymakers and principals was restricted mainly to issues related to the GEMs test, specifically how to implement it, what is expected to be gained from it and making use of the results.With regard to being goal-specific, policymakers focused solely on the goals set by schools and the measures of success as detailed in a school's action plan.Similarly, meetings held between school agents and officials from the municipality were devoted to discussions related solely to the GEMs.
Taken together these four dimensions of coupling provide evidence of the tight coupling between policy and implementation.This begs the question of whether the tight coupling brings about school improvement, being the main intention of policymakers, or whether schools are only "going through the motions" so to speak.It would seem that some changes were not simply superficial or cosmetic.The tight coupling meant that directives and regulations became embedded in formal structures of schools as they adopted routines of alignment with the curricular content of the test and of monitoring of instruction.The creation of the administrative position of assessment coordinator provides further indication that government regulation had been incorporated into the formal structure of school (Spillane, Parise & Sherer, 2011).Some implications of this include greater distribution of leadership, changes in school assessment policy, and greater focus on methods of testing and assessment, and analysis and interpretation of assessment results, to name just a few.
The tight coupling also meant the promotion of a stronger commitment to professional development.Both principals and teachers were required to participate in various workshops and study days offered by the Ministry.These opportunities for professional developed contributed to school agents' enhanced competences.The participation in common professional development also meant the creation of a "common language", something that in principle could bring about mutual understanding at all levels of the education system and between administration and staff in the schools.
The tight coupling evidenced between policy, administration and teaching, and the resultant conforming of schools to government mandates, may be explained by the process of institutional isomorphism (DiMaggio & Powell, 1983), which, as Fusarelli (2002) has argued, "serves to tighten the loose coupling of educational organizations" (p.568).Indeed, institutional isomorphism has been used by institutional theory to explain why schools generally seem to respond to environmental shifts in similar ways (Burch, 2007;Ogawa, 1992;Powell & DiMaggio, 1991;Rowan & Miskel, 1999;Scott, 1991).
Isomorphism is defined as a "constraining process that forces one unit in a population to resemble other units that face the same set of environmental conditions" (DiMaggio & Powell, 1983, p. 66), in other words, that involves pressures towards conformity or compliance.Isomorphism can "cause whole populations and communities of organizations to act in concert as a result of their shared fate" (Rowan 2002, p. 61) and this seemed to be the case for the school agents in the current study.
Three types of pressures leading to institutional isomorphism (DiMaggio & Powell, 1983;Meyer & Rowan, 1977) include: coercive pressures, often involving political or government pressures and the force of the state to meet expectations and requirements, and including regulatory oversight, incentives, and control; mimetic pressures that ensure success in an uncertain environment by organizations drawing on habitual responses that had previously succeeded, by imitating or emulating other organizations' activities and successful peers, or by copying successful approaches of other organizations; normative pressures based on what is thought to be professional knowledge and how professionals should be or that rely on socialization (through professional organizations and professional media), which leads to viewing certain structures and processes as legitimate.
Coercive isomorphism was evident in the context of the GEMs as the test involved governmental regulation and political pressure exerted by the Ministry on the schools.Policymakers issued governmental directives, exercised control over the mechanisms of test development, administration and analysis and monitored the actions of schools within the context of the test.The actions of school agents reflected their compliance with these policy expectations.However, one may claim that compliance is inherent in a centralized education system, as by definition policy is top-down.Following the guidelines for the test may seem natural in schools that this has been there modus vivendi for all government initiatives.
Mimetic isomorphism was evident as schools tended to compare not only their test results but also the actions they had taken in order to achieve those results.They also turned to practices that had been successful in the past as similar tests had been given before.The sharing across schools may seem a positive action as it may contribute to shared understandings, learning from each other, and contribute to easing students' transition from elementary to junior high schools.However, this sharing could be derived not from a desire to share knowledge and expertise, or a learning community, but rather from competition.Whether this type of competition contributes to school improvement or not is an issue to be further pursued.
Normative isomorphism was evident as the test was considered the norm for school work and for judging school performance.The test was considered by school agents to be a legitimate means of collecting information about the status of education in general and for decision making within the school in particular.The assessment imparts legitimacy to the school; therefore, the test becomes the norm, or at least the normative way of evaluating schools and their performance, to the exclusion or omission of alternative means of assessment that may be more conducive to learning.Norms tend to be the most embedded within school culture, and thus they are not easily changed.Because the test becomes the norm, the school agents may be directed toward it to the exclusion of other, perhaps more authentic and valid forms of assessment.
Viewing test-based accountability through the prism of new institutional theory and harnessing the concept of loose coupling is helpful in explaining the response of different schools to accountability pressures (Burch, 2007;Ogawa, 1992;Powell & DiMaggio, 1991).In the context of the Israeli GEMs test it appears that the tightening of relations between policymakers and school agents has heightened the centralization of an already centralized system.The policy of accountability and the policy instrument of a state-mandated test are powerful regulatory mechanisms that have successfully brought about tighter coupling between the school and its institutional environment and it would appear that breaking away may be almost impossible.
Indeed, some have wondered why teachers have not become more socially responsible about the use of tests (Shohamy, 2001(Shohamy, , 2007)).It would seem that the pressures of the institutional environment may be so great that teachers "are bound to uphold the 'way things are' because other ways of thinking or acting have become largely unthinkable" (Stryker, 1994, p. 867).Scott (2001), a leading organizational theorist, concurs and believes that compliance is a result of a belief that, under certain circumstances, "other types of behavior are inconceivable" (p.134).
This study has not presented an exhaustive range of responses to the test-based accountability system in place in Israel.However, the results do provide insight into ways in which the policy context of state-mandated tests affects the perceptions and actions of principals and teachers.Studies on policy implementation within accountability systems have highlighted patterns of tight and loose coupling between policy, school and classroom, and have shown that teachers and administrators are attending to policies and working hard to implement them (Firestone et al., 1999;Hill, 2001).The tightly coupled relationship between the policy environment and schools has resulted in schools focusing on the tests as an indicator of school quality (Mintrop & Trujillo, 2007, p. 319).The tight coupling has also meant that policy is exerting control over the curriculum (Au, 2007).This tight coupling between schools and their governing agencies has been confirmed in the current study and it is not surprising when one considers that the system studied is a centralized one and that the general intention of test-based accountability is to increase external control over what happens in schools and classrooms (Fusarelli, 2002).
Policymakers in the centralized education system, focusing mainly on school improvement, should be aware of both the positive and negative aspects of the different types of isomorphism, minimizing coercive isomorphism and highlighting the positive aspects of mimetic and normative isomorphism.This has the potential to reduce the pressure brought to bear by themselves on schools and allow for more school autonomy.It could also have a more positive effect on school agents' morale (Ladd, 2017).In keeping with normative isomorphism, policymakers may also wish to explore other avenues of action, such as allowing schools to implement their own accountability system (Polikoff, McEachin, Wrabel, & Duque, 2015).In such a case, policymakers concerned with school progress, could design a range of tools for monitoring schools.This is especially relevant for policymakers who may have too narrow a view of schooling (Brighouse, et al., 2016).Another suggestion that has arisen from the literature is moving away from achievement gaps across the state to a focus on reducing within-school or within-district gaps (Ladd (2017;Neal & Schanzenbach, 2010).This would divert the focus from low-performing subgroups to reducing the gap within a school or district.It would also encourage teachers to make better use of assessment data, making the information gleaned from the test more meaningful and relevant.Focusing on the district rather than the state is particularly feasible in Israel as the system is divided into a number of districts and schools in the same district could pay more attention to all their students rather than pointing out a selected few.
It is necessary for school agents to not only understand policy messages and translate them into action; they should be aware of the possible sources of pressures on their work and interpret policy in a way that is not simply a direct response to such pressures (Dulude, Spillane & Dumay, 2017).School principals, being more aware of institutional pressures, may enhance the leadership skills necessary when faced with environmental shifts and their possible impact on schools.Programs designed for training future principals and teachers may benefit from insights gained in this study as to how school agents respond to policy mandates.Teachers may also be more empowered as they, too, become more aware of environmental shifts and their impact on their classroom practice.Pressures that seem to them as being unwarranted may be handled in ways that empower them rather than causing them to act in contradiction to their beliefs.
In the current study there was no attempt to analyze causative relations.However, there is a feeling that the test actually led to the response that was exhibited by schools both on the administrative level and on the level of the classroom.Of course, there may be any number of intervening variables that have contributed to the response of school agents to the policy.It is suggested that future research provide more insight into the causal relations between policy and policy implementation, perhaps through the design of a longitudinal study that closely follows schools over an extended period.A longitudinal design could also capture the long term effects of the policy.The local and more immediate responses of school personnel to accountability mandates, such as narrowing the curriculum, teaching to the test, and changing resource allocation, are clear.Indeed teachers and principals attest to the fact that they "get back to normal" after the test.The question remains to what extent the test has more far reaching consequences for the school.
This study focused on the main policymakers within the Ministry of Education as well as a number of officials external to the Ministry of Education, such as those in municipalities.However, other stakeholders holding influential positions may also be a source of important information, for example, officials from teachers' unions.It is recommended for future studies that officials from these unions be included as their response may provide more insight into the position of teacher representatives with regard to the accountability policy.It could also provide insight into the support teachers may expect if they were to consider a more activist stance as it is these unions that protect the rights of teachers.With regard teachers, the study focused on EFL teachers faced with the dictates of the GEMs.It is suggested that future studies focus on additional groups of teachers who are also faced with the same dictates, such as Math or Science teachers.
Another recommendation for future research is to focus on other sets of couplings.While a focus on couplings between educational organizations and their governing agencies was the center of the current study, there are additional coupling elements that are worth investigating as they are related to the effective functioning of schools.Under this heading would be included, for example, the relationship between teachers and their disciplinary societies or the coupling of schools with publishing houses and testing agencies.These issues of coupling are important not only to understand the work performed in schools as organizations in a given institutional field, but also to shed light on prospects for changing the nature of that work (Coburn, Hill, & Spillane, 2016).
In conclusion, a state-mandated test, conceived as an environmental shift through which isomorphic processes are invoked, has the power to bring about tight coupling between policymakers and school agents, between policy intention and policy implementation.This may be attributed to the joint and concerted effort on the part of the policymakers to ensure that schools comply with the accountability requirements.It may also be attributed to the fact that schools are strongly and directly dependent on governmental agencies.Whatever the case, school agents were aware of most of the intentions of policymakers for the test, and pragmatically translated policy into action, particularly in terms of preparing an action plan and following a national curriculum.Perhaps an understanding of policy intentions can temper the negative effects of top-down policy initiatives.Intentions being more transparent and implementers having a better understanding of policy intentions may allow for more mindful implementation of policy that is accompanied by a deep understanding of purpose and objectives.Having a better understanding of test-based accountability in the context of institutional theory could also contribute to school improvement.Taken together these could bring about more meaningful and lasting change.