Educational Evaluation and Policy Analysis http://eepa.aera.net

The Role of Nonsystem Actors in the Relationship Between Policy and Practice: The Case of Reading Instruction in California Cynthia E. Coburn EDUCATIONAL EVALUATION AND POLICY ANALYSIS 2005 27: 23 DOI: 10.3102/01623737027001023

The online version of this article can be found at: http://epa.sagepub.com/content/27/1/23

Published on behalf of

American Educational Research Association

and

http://www.sagepublications.com

Additional services and information for Educational Evaluation and Policy Analysis can be found at:

Email Alerts: http://eepa.aera.net/alerts

Subscriptions: http://eepa.aera.net/subscriptions

Reprints: http://www.aera.net/reprints

Permissions: http://www.aera.net/permissions

Citations: http://epa.sagepub.com/content/27/1/23.refs.html

>> Version of Record - Jan 1, 2005

What is This?

Downloaded from http://eepa.aera.net at NORTHWESTERN UNIV LIBRARY on February 4, 2014 2791-02_Coburn.qxd 4/11/05 13:48 Page 23

Educational Evaluation and Policy Analysis Spring 2005, Vol. 27, No. 1, pp. 23–52

The Role of Nonsystem Actors in the Relationship Between Policy and Practice: The Case of Reading Instruction in California

Cynthia E. Coburn University of California, Berkeley

Studies of the relationship between policy and practice typically focus on the formal policy system alone. Yet, the public policy system does not exist in isolation. A host of nonsystem actors promote, translate, and transform policy ideas as they carry them to teachers. This study draws on neoinstitu- tional theories of organization to investigate the role of nonsystem actors in the relationship between policy and teachers’ classroom practice. A cross-case, historical design was used to investigate how teachers in two California elementary schools responded to changes in state reading policy from 1983 to 1999. The way in which teachers responded depended in part on the nature of their connections to policy messages, which varied substantially across teachers and across policy initiatives. Policy mes- sages from nonsystem actors were more consequential for teachers’ classroom practices. Teachers’ connections to nonsystem actors were influenced by the interrelationship between system and non- system actors as policy ideas emerged, diffused, and were implemented over time.

Keywords: nonsystem actors, policy implementation, reading instruction

MOST studies of the relationship between state (Burch, 2000; Cohen & Hill, 2000; Hill, 2000). instructional policy and classroom practice focus However, these studies provide little guidance as solely or predominantly on how policy ideas move to how or why nonsystem actors are so influen- through the formal public policy system (state, tial or what accounts for the nature of teachers’ county, district) into schools. Yet, the formal pol- differential connections to them. icy system does not exist in isolation. A host of Taking up this challenge, the present study drew nonsystem actors—independent professional on neoinstitutional theories of organizations to development providers, reform organizations, investigate how teachers in two California ele- publishers, and universities—promote, translate, mentary schools learned about and responded to and even transform policy ideas as they carry changes in state reading policy from 1983 to 1999. them to teachers (Burch, 2000; Coburn, 2001a; How teachers responded to changes in reading Cohen, 1982, 1995; Cohen & Hill, 2000; Hill, policy depended, in part, on the nature of their 2000). Although policy studies have investi- connections to policy messages, which varied gated the role of such organizations in the policy- substantially across teachers and across different making process (see, for example, Cusick & policy initiatives. Despite these variations, how- Borman, 2002; Eilers, 2004; Fuhrman, Clune, & ever, policy messages from nonsystem actors Elmore, 1988; Ogawa, 1994), their role in policy tended to be more consequential for teachers’ implementation in general and the relationship classroom practice. And the nature of those con- between state policy and teachers’ classrooms nections to nonsystem actors themselves was in- more specifically has rarely been explored sys- fluenced by the complex interrelationship be- tematically. Studies that have addressed this issue tween system and nonsystem actors as particular provide evidence that nonsystem actors are a policy ideas emerged, diffused, and were imple- key mediating link between policy and practice mented over time.

23

Downloaded from http://eepa.aera.net at NORTHWESTERN UNIV LIBRARY on February 4, 2014 2791-02_Coburn.qxd 4/11/05 13:48 Page 24

Coburn

Reading instruction in California provides a service providers; membership organizations, such fertile context for this study because California is as professional associations, advocacy groups, in the midst of its third major shift in reading pol- and networks; and nonprofit organizations, in- icy since 1980. After the rise of a movement to- cluding universities, research firms, and quasi- ward literature-based instruction in the late 1980s governmental agencies that provide research and and early literacy approaches in the early 1990s, development and technical assistance or act as tremendous activity and controversy since the intermediaries. These nonsystem actors, as I call mid-1990s has led the state and the profession them, play an important role in carrying policy toward varying versions of what is frequently ideas to teachers in the form of information, pro- called a “balanced” approach to reading instruc- fessional development, technical assistance, and tion. The intensity and temporal proximity of these curricular materials. shifting policies, as well as the differences in fun- System and nonsystem actors interact in a damental assumptions about teaching and learn- multiplicity of ways. Nonsystem actors play a ing they imply, throw into relief the pathways by role in the formulation of state policy by partici- which teachers engaged them. By understanding pating in state task forces and commissions and these pathways, we not only gain further insight by providing intellectual resources on which pol- into the complex relationship between policy and icy is sometimes based (Cohen, 1982; Cusick & classroom practice but also uncover potential Borman, 2002; Hill, 2000). But nonsystem actors points of leverage for influencing how teachers also play a crucial role in policy implementation. respond to changes in instructional policy. States often depend on nonsystem actors to carry out instructional policy in the absence of capac- Theoretical Framework ity to do it themselves (Cohen, 1982; Hill, 2000; Observers have noted the complexity and frag- Honig, 2004). Thus, policymakers provide funds mentation of what David Cohen has called the for schools and districts to purchase curricular system of “instructional guidance” in the United materials and contract professional development States (Cohen, 1995; Cohen & Spillane, 1993; with nonsystem actors. In some states, including Rowan, 2001). Teachers learn about new ap- California, state funding is linked to specific cri- proaches to teaching and learning through a di- teria meant to govern content.2 Often, however, verse and at times loosely connected set of poli- the content is left up to local discretion (Cohen, cies, organizations, and actors. The formal system 1982). Beyond this, numerous professional de- of public policy and governance is complex and velopment providers and instructional materials multilayered (Cohen, 1982; Cohen & Spillane, are available to teachers that have no connection 1993; Meyer, Scott, & Deal, 1981; Scott, 1994). to the public policy system but nevertheless con- These system actors, as I call them, include the vey policy ideas to teachers. various individuals and organizations that con- In sharp contrast with this portrait of the mul- stitute state and local governance of schooling, tiple, interdependent paths from instructional including state departments, county offices, school policy to classroom practice, most studies of the districts, and schools.1 Policy decisions related relationship between policy and practice focus to instruction are made at the state, district, and solely or predominantly on the formal policy sys- school levels, and policy decisions made at higher tem (see, for example, Fuhrman & Elmore, 1990; levels of the system must move through multiple McGill-Franzen, 2000; Odden, 1991; Valencia layers of system actors to reach teachers (Cohen & Wixson, 2000). This narrow focus essentially & Spillane, 1993; McLaughlin, 1987; Pressman excerpts policy implementation from its broader & Wildavsky, 1984). social context, failing to take into account or in- In addition to the public policy system, how- vestigate the role of nonsystem actors in the re- ever, teachers also learn about policy ideas from lationship between policy and practice. a range of actors that are not formally part of the In this study, I drew on neoinstitutional orga- system. Rowan (2001) identified three classes nization theory to put forth a broader conceptu- of private organizations that play an important alization of the ways in which policy messages role in instructional improvement in U.S. public about appropriate reading instruction move in schools: for-profit firms, including textbook pub- and through the environment and into schools. lishers, instructional program vendors, and other At root, institutional theory is a cultural approach. 24

Downloaded from http://eepa.aera.net at NORTHWESTERN UNIV LIBRARY on February 4, 2014 2791-02_Coburn.qxd 4/11/05 13:48 Page 25

The Role of Nonsystem Actors

It emphasizes how norms and cultural concep- connected to policy messages through a web of tions are constructed and reconstructed over time, interactive and, at times, reciprocal linkages. The carried by individual and collective actors, and analytic task, then, is to understand the nature of embedded within policy and governance struc- these linkages between system and nonsystem ac- tures (Scott, 2001; Scott, Mendel, & Pollack, tors and how they shape the way in which teachers 1996). Institutional theorists suggest that policy learn about and respond to policy messages from messages shape patterns of action and beliefs the environment. within organizations through regulative means A few studies in education provide initial guid- as they are incorporated into formal policy, but ance in this task. These studies have shown that also through normative means as teachers and nonsystem actors play a mediating role between others feel pressured to adopt certain approaches state instructional policy and teachers’ classroom to maintain legitimacy and through cognitive practice (Coburn, 2001a; Cohen & Hill, 2000; means as conceptions of appropriate practice at- Hill, 2000). To the degree that these individuals, tain taken-for-granted status as natural or com- organizations, and materials provide teachers ac- monsense (Scott, 2001). cess to policy ideas that are grounded in the cur- This conceptualization highlights the ways in ricula, connected to several elements of instruc- which policy is situated in and interacts with a tion, and extended in time, they influence the broad range of actors, organizations, and sets of degree to which teachers make changes in class- ideas. In particular, institutional theorists have room practice in the direction of state policy offered a new unit of analysis to facilitate the (Cohen & Hill, 2000; Hill, 2000). Nonsystem ac- study of these interactive relationships: the or- tors are especially influential when policy ideas ganizational field. Situated between individual are complex and require substantial changes in organizations and populations of organizations, modal teacher practice (Hill, 2000). Finally, non- the organizational field has been defined as “those system actors shape what teachers learn about organizations that, in aggregate, constitute a rec- policy messages. They not only carry policy ideas ognized area of institutional life: key suppliers, to teachers but also translate and transform them resource and product consumers, regulatory as well (Coburn, 2001a; Hill, 2000). In this way, agencies, and other organizations that produce nonsystem actors play a key role in shaping teach- similar services or products” (DiMaggio, 1991, ers’ access to some policy ideas and not others. pp. 64–65). The organizational field is charac- The present study builds on and extends this terized by governance structures and actors that work in three ways. First, I develop a frame- are connected with one another through both hor- work for conceptualizing the nature of teachers’ izontal and vertical ties and share a “common connections to policy ideas in the broader envi- meaning system” that guides action and interac- ronment. To engage the question of teachers’ con- tion (Scott, 1994). nections to messages via nonsystem actors, it is Empirical work outside of education has em- first necessary to develop the conceptual tools phasized how policy emerges from vertical as for describing and analyzing the broader ways well as horizontal interactions between the pub- teachers are connected to policy, of which non- lic policy system and professional organizations, system actors are one route. Second, I use this university researchers, policy entrepreneurs, and framework to provide additional evidence that others in the organizational field (Kingdon, 1984; nonsystem actors play a critical role in shaping Laumann & Knoke, 1987; Moore, 1988; Scott, what policy ideas teachers are connected with Ruef, Mendel, & Caronna, 2000). But it also pro- and how they respond to them. In so doing, I move vides evidence that these nonsystem actors, in beyond documenting the nature of the relation- turn, shape and interpret policy messages as they ship between nonsystem actors and teachers’ move through the system to target organizations change in practice to develop an account of why (Dobbin, Sutton, Meyer, & Scott, 1993; Dowd & these connections are particularly influential. Fi- Dobbin, 1997; Edelman, 1992; Pollack, 1997). nally, I discuss the ways in which the interrela- Rather than proffering a vision of policy imple- tionship between system and nonsystem actors mentation in which policy messages move through in the organizational field influences the nature the system in a top-down, more or less linear man- of teachers’ connections to nonsystem actors in ner, this perspective suggests that teachers are schools. 25

Downloaded from http://eepa.aera.net at NORTHWESTERN UNIV LIBRARY on February 4, 2014 2791-02_Coburn.qxd 4/11/05 13:48 Page 26

Method Within the two schools, I focused attention This study involved the use of a cross-case, on early-grade classroom teachers (first and historical, and cross-sectional design. Within a second grade) because early-grade reading in- common state policy environment, I compared struction was the main focus of the policy de- the experiences of three teachers in two Califor- bate in the environment. In each school, I began nia elementary schools in two districts. By in- with breadth, interviewing nearly all teachers in cluding more than one district and more than one the first and second grades as well as resource school, I was able to gain insight into how teach- teachers, members of the leadership teams, cur- ers’ embedded contexts mediated their connec- rent and former principals, and other teachers tions to shifting policy ideas. Focusing on three about their connections to policy messages and teachers allowed for the depth of investigation the nature of their classroom instruction. In all, necessary to capture the nature of teachers’ con- I interviewed 12 current classroom teachers, nections to messages from the environment over 4 current resource teachers, and 2 retired teach- time. Although not generalizable, the in-depth ers at Stadele Elementary and 9 current class- investigation made possible by the small number room teachers, 1 current resource teacher, and 1 of cases provides the opportunity to generate retired teacher at Baldwin Elementary. Nearly new hypotheses or build theory about sets of re- all of the interviews were taped and transcribed. lationships that would otherwise have remained To capture depth, I used information garnered invisible (Hartley, 1994). from the first round of interviews to select three The study was both historical and cross sec- teachers for further study. The focal teachers— tional. I studied the nature of teachers’ connec- Sharon, Marisa, and Deanna—represented the tions to the policy environment from 1983 until range of approaches to reading instruction in the 1999 using document analysis, oral histories of schools, the range of years of experience, and informants at multiple levels of the system, and both first and second grades (see Table A2 in Ap- secondary sources. I also investigated the nature pendix for background information on the focal of teachers’ connections to the environment as teachers). I then conducted additional interviews they unfolded in real time during the 1998–1999 with and classroom observations of each focal school year, relying primarily on in-depth inter- teacher. A significant portion of these interviews viewing (Spradley, 1979) and sustained observa- was devoted to developing oral histories of tion (Barley, 1990). teachers’ classroom practices by investigating I used purposive sampling to select two urban links between teachers’ connections to the envi- elementary schools that represented strategic con- ronment (elicited during initial interviews) and trasts along dimensions that previous research changes in their reading instruction over time. In and theory suggested would be useful for under- all, I interviewed focal teachers 28 times in in- standing the relationship between classrooms and terviews ranging from 45 minutes to 3 hours. I the environment. In particular, the theoretical lit- erature suggests that the nature of an organiza- also spent 89 hours observing focal teachers’ tion’s connections to the environment (D’Aunno, classrooms. Finally, to capture teachers’ connec- Sutton, & Price, 1991; Popkewitz, Tabachnick, tions to messages from the environment during & Wehlage, 1982; Scott, 2001) and prior history the study year, I conducted extensive observa- of practice (Vaughan, 1996; Weick, 1995) shapes tions of teachers’ interactions in formal and in- how it responds to pressures from the institutional formal meetings. In all, I spent more than 130 environment. Thus, I selected schools that had hours in Stadele Elementary and 21.5 hours in contrasting connections to the environment (in Baldwin Elementary observing teacher conver- different districts and involved in different reform sations during formal meetings and professional efforts) and contrasting histories of involvement development activities.4 In both schools, I also in reading reform efforts. The two schools ulti- observed countless hours of informal conversa- mately included in the study—Stadele Elemen- tions during lunch, before and after school, and tary and Baldwin Elementary3—are racially and in the hallways. While the majority of the analy- ethnically diverse, with a substantial portion of stu- sis presented in this article draws on data from dents living in poverty (see Table A1 in Appendix focal teachers, I draw on data from the larger for demographic information on the two schools). sample of teachers when appropriate. 26

Downloaded from http://eepa.aera.net at NORTHWESTERN UNIV LIBRARY on February 4, 2014 2791-02_Coburn.qxd 4/11/05 13:48 Page 27

The Role of Nonsystem Actors

To understand the dynamics of state-level I used NUD.IST qualitative data analysis soft- policy changes in California, I engaged in a strat- ware to analyze documents, meeting observations, egy of snowball sampling (Hornby & Symon, interviews, and classroom observations. I ana- 1994). That is, I started by asking key players in lyzed the data chronologically, first charting how the field for recommendations, asked their rec- policy ideas moved through the environment and ommendations for recommendations, and so on then coding and mapping teachers’ connections until I had a list of 23 potential people to inter- to these movements and changes in their reading view. I selected 12 individuals to interview who beliefs and practices. To capture teachers’ con- represented different aspects of the field (aca- nections to policy ideas, I identified key “mes- demics, representatives of the state department of sages” about reading that teachers came into con- education, representatives of the state board of tact with, which then became the key analytic education, professional development providers) unit for this part of the study. Messages included and who were active in different state policy specific statements or exhortations about how eras. In addition to interviews, I traced policy teachers should or must teach reading that were changes by performing a content analysis of delivered through such things as professional 129 state documents, state legislation, task force development promoting a particular approach, reports, media accounts, and documents from new classroom materials, policy documents, state and regional professional development statements from principals or district officials providers. I situated this state-level analysis in about what they expected to see in reading in- the broader national context by consulting sec- struction, or media reports concerning the “read- ondary sources that charted aspects of the his- ing crisis.” The three focal teachers encountered tory of reading more broadly (see Table A3 in 223 messages from the environment.5 Ninety-five Appendix for a description of sources of infor- of these messages came from system actors, mation for the state policy analysis). and 128 came from nonsystem actors (see Finally, I used a bottom-up strategy to under- Table A4 in Appendix for more specific infor- stand the mechanisms by which teachers were mation on the mechanisms by which focal connected to changing state policy. Rather than teachers encountered policy messages). hypothesizing the likely ways that teachers might To analyze focal teachers’ encounters with become connected with policy and investigating messages, I created longitudinal records of teacher those mechanisms (as has been the practice in interaction with and response to a given message previous studies exploring the role of nonsystem over time. I then used a coding scheme rooted in actors), I started with teachers and identified the earlier theoretical work in institutional theory and actual ways in which they were connected to elaborated through an iterative process of data policy over time through system and nonsys- collection and analysis (Miles & Huberman, 1994; tem routes. I then interviewed as many system Strauss & Corbin, 1990) to identify and code for and nonsystem actors as were possible to locate: factors that influenced the nature of teachers’ six individuals in each district who held positions connections to nonsystem actors. That is, I began related to reading instruction, several of whom with codes that described, with little interpreta- were directly involved in either Stadele or Bald- tion, the dimensions that characterized teachers’ win Elementary, and nine representatives of connections to policy messages. By grouping professional development organizations who had together categories and using systematic com- provided professional development on reading parisons (Strauss & Corbin, 1990), I moved to instruction to the two schools during the period progressively higher levels of abstraction until I covered by the study. I supplemented these in- ended up with the following codes: mechanism terviews with content analyses of professional of connection, intensity, content of messages, development, policy, and curriculum documents. and voluntariness. I then coded each response as In all, I analyzed 12 district policy and profes- high, medium, or low along the dimensions of in- sional development documents, professional tensity and content (see Table A5 in Appendix development materials from 18 local providers, for the definitions used in coding). Finally, I used 10 curricula, and 9 school-level reports identified descriptive statistics to analyze the relationships through this process. between these factors and teachers’ responses. 27

Downloaded from http://eepa.aera.net at NORTHWESTERN UNIV LIBRARY on February 4, 2014 2791-02_Coburn.qxd 4/11/05 13:48 Page 28

Reading Instruction in California Soon many of the pedagogical approaches as- Since the early 1980s, the California public sociated with Reading Recovery moved beyond school system has been the site of tremendous re- this infrastructure as independent professional form energy focused on changing the way chil- development organizations began to offer train- dren are taught to read. Three successive move- ing that adapted pedagogical principles of Read- ments seeking to redefine what constitutes “good” ing Recovery for the classroom (Literacy Initia- reading instruction have gained prominence in tive, n.d.; Ohio State University, 1998; Rigby the profession, become part of state policy, and Education, 1997; St. Mary’s College, n.d.; Swartz, been carried into schools by professional devel- Shook, & Klein, 2000). Finally, these ideas opment providers and instructional materials. This began to work their way into state policy. They tumultuous period of change began when a move- were carried by a state leadership network that ment within universities, the teaching profession, had brought early literacy ideas to nearly 20,000 and the policy community began to challenge teachers and administrators by the time it was widespread and deeply institutionalized ways of disbanded in 1995. They played a heavy role in teaching reading known collectively as the basic the state’s criteria for textbook adoption in 1994 skills approach. Advocates of what would be- (Curriculum Development and Supplemental come known as literature-based instruction put Materials Commission, 1994). They also formed forth a vision of reading instruction rooted in the centerpiece of state position statements in- epistemological assumptions about the nature of tended to provide guidance on early reading in- teaching and learning fundamentally different struction (California State Department of Edu- from the conventional wisdom at the time. The cation, 1995b). movement for literature-based instruction gained Then, in the mid-1990s, questions about what a particularly high profile in California in 1987, constituted “good” reading instruction exploded when it became embedded in the state-level frame- onto the public stage after the release of test work for English-language arts (California State scores that placed California last in the country in Department of Education, 1987), which, in turn, reading, tied with Louisiana and Guam (Carlos & was linked to state textbook adoptions, profes- Kirst, 1997). In what was dubbed “the reading sional development, and, for a short time, stan- wars” by the popular press, controversy raged dardized testing (Brandt, 1989). about the root causes of low test scores. Critics Shortly after literature-based approaches began began to call for a return to “basic skills.” In 1995, to gain prominence in the professional and pol- the state responded to the controversy by pub- icy world, a second movement—what I call “early lishing a task force report calling for a “balanced literacy”—quietly began to put forth an alterna- approach” to reading instruction (California State tive vision of early reading instruction as it moved Department of Education, 1995a), launching the from district to district throughout the state. third policy shift in two decades. For the next sev- Rooted in the pedagogical principles of Reading eral years, different groups of actors put forth dif- Recovery (Askew, Fountas, Lyons, Pinnell, & fering constructions of just what a “balanced ap- Schmitt, 1998; Breneman & Parker, 1991), this proach” is or should be, constructions rooted in set of approaches spread throughout the state, divergent assumptions about teaching and learn- ultimately making its way into state policy and ing. Over time, a coalition of state actors and re- spawning a host of professional development searchers was able to define and embed in state providers focused on applying Reading Recov- policy a conception of “balanced” instruction. ery principles to early-grade classrooms. Ini- During the second half of the 1990s, the state tially, this set of ideas spread from district to dis- legislature passed 12 bills allocating nearly half trict as the infrastructure for Reading Recovery a billion dollars toward reform efforts promot- developed in the state. From 1990 to 1996, the in- ing this approach (California State Board of Ed- frastructure grew to include four regional Read- ucation, 1999). With this legislation, the state ing Recovery training sites at universities linked systematically undid linkages between state pol- with 70 districts, 135 trained teacher leaders, and icy and literature-based and early literacy ap- thousands of trained Reading Recovery teachers proaches, putting in their place policies that de- in California (Neal, Kelly, Klein, & Schubert, fined good reading instruction in terms of this 1997; St. Mary’s College, 1998). new construction of balance.6 28

Downloaded from http://eepa.aera.net at NORTHWESTERN UNIV LIBRARY on February 4, 2014 2791-02_Coburn.qxd 4/11/05 13:48 Page 29

Teachers’ Connections to sages about appropriate reading instruction as Shifting Reading Policy represented by standardized tests. But teachers Given the multifaceted and multilayered nature also learned about state policy ideas through a of the reading environment, these shifts in reading diverse array of nonsystem actors as they at- policy interpenetrated schools and classrooms to tended professional development sessions with varying degrees. The ways in which teachers re- independent professional development providers sponded to reading policy depended in part on the or school reform organizations, used textbooks or nature of their connections to a given policy ini- other curricular materials in their classrooms, or tiative. Teachers’ connections varied both across learned about reading instruction in preservice teachers and across policy initiatives, at times sub- teacher education or continuing coursework. In stantially, shaping teachers’ access to some policy addition, they encountered policy ideas through ideas and not others. Beyond access, however, such informal routes as conversations with their the nature of teachers’ connections—their mech- colleagues and media coverage of ongoing pol- anisms, intensity, and content—also influenced icy debates. how they responded to the policy messages they As can be seen in Table 1, the three teachers encountered. In this regard, teachers’ connections varied considerably in their connections with with nonsystem actors were particularly impor- different policy initiatives. For example, teach- tant, as they tended to be more consequential for ers varied in the type and frequency of their con- classroom practice than those with system actors. nections to ideas associated with early literacy In this section, I begin by providing an overall por- in the mid-1990s. Deanna was connected in trait of the nature of teachers’ connections to pol- multiple, overlapping ways that involved both icy messages about reading through both system system and nonsystem actors. She taught in a and nonsystem routes. In so doing, I identify key district that placed an enormous emphasis dimensions along which teachers’ connections to on early literacy approaches, embedding ap- policy vary. proaches into multiple aspects of district policy Teachers were connected to messages about and investing a great deal of resources in reading both when messages pressed in through district-sponsored professional development. policy and school- or district-sponsored profes- These district-level expectations were rein- sional development and as teachers themselves forced by the school leadership, which both reached out to new resources, training, and ma- adopted the approach in school policy and terials in their proximal environment. Teachers’ brought in additional independent professional connections to messages can be described in development providers to work with early-grade terms of their mechanism (how teachers learned teachers in the school over the course of a num- about policy messages), their content (what they ber of years. In contrast, Marisa was connected learned), and their intensity (their level of en- to early literacy ideas in less intensive ways via gagement with policy ideas). These variables— system actors but reached out on her own to col- the how, what, and how much of teachers’ expe- leagues, curricular materials, and independent riences with policy—involved both the formal professional development providers who carried policy system and nonsystem actors. this set of ideas. Finally, Sharon represented a third pattern. Her primary connection to early literacy ideas was through participation in dis- Mechanism trict-sponsored professional development. Al- Teachers were connected to messages in mul- though she had connections to ideas through tiple and at times overlapping ways. All three nonsystem actors, these connections tended to teachers were connected to policy ideas through be short-term and disconnected experiences. formal policy system channels. They received The nature of teachers’ connections also varied copies of state and district frameworks, stan- across different policy initiatives. While teachers dards, and position papers; attended district pro- encountered state policy focused on literature- fessional development sessions; were connected based instruction through both system and non- through specific district and school policies that system means (predominantly textbooks and pro- promoted particular approaches to reading in- fessional development), by late 1999 all three struction and assessment; and encountered mes- teachers had learned of state policy promoting 29

Downloaded from http://eepa.aera.net at NORTHWESTERN UNIV LIBRARY on February 4, 2014 2791-02_Coburn.qxd 4/11/05 13:48 Page 30

TABLE 1 Mechanisms of Teachers’ Connections to Policy Initiatives Basic skills Literature-based instruction Early literacy “Balanced” approach Teacher System Nonsystem System Nonsystem System Nonsystem System Nonsystem Sharon State Preservice School policy Textbook Multiple district Textbook State policy Awareness of de- standardized education Principal Multiple professional Professional documents bate in the test Textbooks expectations professional development development State media

Downloaded from District Curricular State policy development School from standardized standardized materials State test from professional independent tests test Multiple District independent development provider District policy District policy professional professional providers District policy District

http://eepa.aera.net School policy development development Joint work with District workshop Principal with colleagues professional expectations independent University development providers coursework School policy Colleagues at NORTHWESTERN UNIV LIBRARY on February 4, 2014 Marisa State Professional School policy Preservice District Textbook State policy Awareness of de- standardized development District education professional Joint work with documents bate in the test with professional Curricular development colleagues State media District independent development materials School Curricular standardized Professional standardized provider Principal Multiple professional materials tests development test Interaction with expectations professional development Multiple District policy with senior development District policy professional independent colleagues from School policy development provider Curricular independent from materials providers independent Conversations providers with colleagues Deanna State Preservice District Textbook District Textbook State Awareness of de- standardized education professional Professional professional Curricular standardized bate in the test Textbook development development development materials tests media District Curricular State policy from District policy Multiple Professional standardized materials independent School policy professional development test Multiple provider State policy development with District policy professional with independent School policy development independent provider Principal with providers expectations independent University providers coursework 2791-02_Coburn.qxd 4/11/05 13:48 Page 31

The Role of Nonsystem Actors

the “balanced approach” to reading instruction lated to literature-based instruction at a medium largely through system actors (especially via or high degree of intensity. In contrast, teachers’ copies of policy documents). In instances in which encounters with the state’s rendition of a bal- teachers did learn about the balanced approach anced approach tended to be fleeting: a casual via nonsystem actors, it was through their aware- look through a policy document, momentary ness of the debate in the media and when these mention in professional development, or reading ideas were alluded to as small parts of profes- about it in the newspaper. sional development more representative of earlier Content state policy. Finally, teachers’ connections with policy mes- Intensity sages varied greatly. Teachers with different The intensity of teachers’ connections, or the connections often had access to substantively dif- degree to which teachers had opportunities to en- ferent sets of ideas, tools, and instructional ap- gage with policy ideas in a sustained way, also proaches, as messages were not only carried but varied enormously. At one end of the spectrum, also reinterpreted and, at times, transformed as teachers encountered messages in brief and fleet- they moved into schools. Previous researchers ing ways such as a momentary mention in a fac- have noted that state policy is frequently recon- ulty meeting, casual conversations with a col- structed at lower levels of the formal policy sys- league, or a brief look through a policy document; tem (McLaughlin, 1987; Pressman & Wildavsky, that is, they encountered messages with a low de- 1984), especially the district level (Hill, 2001; gree of intensity. At the other end, teachers at Spillane, 1996, 1998, 2000). The present study times were connected to policy messages in sus- suggests that this phenomenon characterizes the tained, iterative ways including professional de- activity of nonsystem actors as well. For exam- velopment efforts stretching over many weeks or ple, although textbooks in California are adopted months or the day-to-day use of curricular ma- at the state level according to a set of criteria terials; that is, they encountered messages with a linked to state policy, the degree to which text- high degree of intensity. books reflected underlying shifts in pedagogical To characterize the overall level of intensity of principles depended greatly on the particular se- individual teachers’ encounters, Table 2 presents ries. Some publishers responded to policy shifts a description of the multiple ways in which with major reorganizations in structure and ap- teachers were connected to messages associated proach. However, many responded by layering with a particular policy. Here, as with the mech- new approaches and sometimes new language anism of connection, overall intensity varied on top of the structure and approach of previous both among teachers and among policy initia- editions.7 In this way, these and other nonsystem tives. For example, the teachers’ encounters with actors played an important role in shaping what literature-based instruction tended to come in the teachers learned about state policy ideas. form of curricular materials and professional de- Here I characterize the content of teachers’ velopment that was sustained and intensive. For connections in terms of two dimensions: degree Sharon and Marisa, these opportunities were also of depth and closeness to the classroom. Degree linked with intense and ongoing conversations of depth refers to the extent to which policy mes- with colleagues about policy ideas, adding more sages addressed surface structures (such as spe- salience and longevity to the encounters. Thus, cific activities or classroom organization) or un- all three teachers were connected with ideas re- derlying pedagogical principles. Closeness to the

TABLE 2 Intensity of Teachers’ Connections to Policy Initiatives Teacher Basic skills Literature-based instruction Early literacy “Balanced” approach Sharon High High Low Low Marisa Medium High Medium Low Deanna High Medium High Low Note. Intensity is defined as the degree to which teachers had opportunities to engage with a message in sustained, iterative ways. 31

Downloaded from http://eepa.aera.net at NORTHWESTERN UNIV LIBRARY on February 4, 2014 2791-02_Coburn.qxd 4/11/05 13:48 Page 32

Coburn

classroom refers to the degree to which policy struction advocated by state policy. At the other messages came in forms that teachers found to be end of the spectrum, all three teachers were con- easily translatable into classroom practice. For nected to messages associated with early literacy example, policy documents outlining overarch- approaches via curricular materials or in the form ing visions had little proximity to the classroom; of particular pedagogical tools that allowed them teachers often experienced these documents as to make clear connections to their classrooms. abstract, with little guidance for classroom prac- tice. In contrast, textbooks and other curricular Nonsystem Actors and Teachers’ materials had a high degree of proximity to the Response to Policy classroom, as they often were created to be eas- Teachers responded to policy ideas about read- ily integrated into classroom practice. ing in ways that varied tremendously, from reject- As with the other dimensions, individual teach- ing ideas outright to bringing new approaches into ers varied (see Table 3). For example, the teach- their classrooms in ways that caused them to re- ers varied in the degree of depth with which they construct their program and rethink assumptions were connected to literature-based instruction. about how children learn to read. Here nonsystem Both Deanna and Sharon were strongly con- actors played a crucial role. Teachers’ connections nected to messages associated with policy ideas to policy ideas via nonsystem actors tended to be that emphasized using heterogeneous groups for more consequential for classroom practice than reading instruction. But neither was connected to those coming through the formal policy system. pedagogical principles underlying this recom- To provide evidence for this claim, I first charac- mendation or specific pedagogical approaches terize the nature of teachers’ responses to policy for teaching reading differently using this orga- ideas. I then show that teachers were more likely nizational structure. In contrast, Marisa not only to respond to connections via nonsystem actors by was familiar with heterogeneous grouping but making substantive changes in classroom practice also had access to philosophical arguments un- and less likely to reject, ignore, or respond sym- derlying the approach and pedagogical principles bolically to them. Finally, I argue that the nature of consistent with teaching reading using this orga- teachers’ connections to nonsystem versus system nizational structure. actors helps account for their different responses. The content of teachers’ connections also var- Teachers’ Response to Policy Messages ied with varying policy movements. For exam- ple, all three teachers were connected to mes- I identified five ways that the three teachers re- sages about the balanced approach with a low sponded to their connections to policy ideas: re- degree of closeness to the classroom. They were jection, symbolic response, parallel structures, primarily connected via policy documents and assimilation, and accommodation.8 media coverage of the public debate that were quite general and abstract. Consequently, all three Rejection teachers were aware of the call for an increased Given extensive reform activity related to read- emphasis on phonics instruction and of the call ing during the time period covered by the study, for “balance.” But they had only limited connec- teachers encountered multiple messages. Often, tion with the specific approaches to reading in- teachers rejected or simply ignored policy ideas

TABLE 3 Content of Teachers’ Connections to Policy Initiatives Basic skills Literature-based instruction Early literacy “Balanced” approach Closeness to Closeness to Closeness to Closeness to Teacher Depth classroom Depth classroom Depth classroom Depth classroom Sharon High High Low/medium High Low Medium Low Low Marisa Medium Medium High Medium Medium Medium Low Low Deanna High High Low Medium High Medium Low Low Note. Depth is defined as the degree to which policy messages address underlying pedagogical principles. Closeness to classroom is defined as the degree to which policy messages come in forms that teachers find to be easily translatable into classroom practice. 32

Downloaded from http://eepa.aera.net at NORTHWESTERN UNIV LIBRARY on February 4, 2014 2791-02_Coburn.qxd 4/11/05 13:48 Page 33

The Role of Nonsystem Actors

they encountered. For example, early-grade teach- the day to provide opportunities for students to ers at Stadele Elementary attended a district pro- learn “at their level.” Thus, she essentially con- fessional development day in which the presenter structed two parallel structures for reading in- suggested that teachers use individualized instruc- struction that were rooted in different instruc- tion as a way to meet children’s diverse needs in tional approaches and met different goals. Just reading. Later, Sharon declared to her colleagues over 8% of the 223 teacher responses involved that individualized instruction was entirely in- creating parallel structures. appropriate. She said: “How can you possibly teach reading without putting kids in groups? Assimilation That’s crazy!” Having rejected the idea out of Because teachers drew on their tacit beliefs and hand, Sharon continued using reading groups. assumptions to construct their understanding of Overall, of the 223 responses to messages, 27% in- messages from the environment (Coburn, 2001a; volved rejection. Guthrie, 1990; Jennings, 1996; Spillane, 1999; Spillane & Jennings, 1997; Spillane, Reiser, & Symbolic response Reimer, 2002), they often interpreted and enacted Teachers responded to some policy messages messages in ways that transformed them to fit by making changes in the appearance but not the with their underlying assumptions. Or, in the lan- substance of their work. For example, in response guage of cognitive learning theorists, they assim- to pressure from the district for schools to use a ilated new knowledge or experiences into exist- district rubric to score performance assessments, ing schemas or ways of doing things (Fosnot, Deanna and other teachers at Baldwin Elemen- 1996; Piaget, 1978). In so doing, teachers tended tary posted the rubric on their classroom walls to respond to messages by making changes in ma- but did not use it in their assessment practice. terials or classroom organization but not the un- derlying pedagogical or epistemological assump- Sharon and some of her colleagues at Stadele El- tions that guided their approach. ementary exhibited a similar response to word For example, Sharon participated in several walls, an instructional approach that engages stu- professional development activities that empha- dents in interactive word play intended to help sized using thematic approaches to encourage them learn high-frequency words and letter-sound reading comprehension and put forth a view of correspondence (Cunningham, 1995). Sharon student learning as the active process of making put lists of words up on the wall but neither re- connections across subject matter rather than ferred to them in class nor used them in instruc- mastering isolated skills. However, Sharon drew tional activities with her students. Just over 7% on her view of reading comprehension as mas- of teachers’ responses to policy messages were tering a sequence of skills and came to under- symbolic. stand thematic instruction not as a fundamentally different approach to student learning, as pro- Parallel structures moted in the professional development offerings, At times, teachers balanced multiple and con- but as a way to make sequenced skills instruction flicting priorities by creating two or more paral- more interesting to children. She rearranged her lel approaches that corresponded with different room into centers and created a series of activi- pressures or priorities. For example, Deanna and ties for students to do at the centers that related other teachers at her school balanced conflicting to the story. However, these activities remained pressures from the principal to use a textbook rooted in sequence skills instruction, with a heavy that was above most children’s reading level with reliance on worksheets. In this way, Sharon and pressures from district and independent profes- other teachers brought messages into their class- sional development providers to teach with read- rooms in ways that altered surface manifestations ing materials at children’s instructional level by of their practice (classroom organization or ma- creating two parallel systems for reading instruc- terials) without affecting the underlying beliefs tion. In Deanna’s case, she taught using the text- and assumptions about teaching and learning book in the morning to “expose all children to the that guided their work. Echoing findings from the curriculum and what they need to know” and con- broader literature on teacher change (Guthrie, ducted guided reading groups at another point in 1990; Jennings, 1996; Smith, 2000; Spillane, 33

Downloaded from http://eepa.aera.net at NORTHWESTERN UNIV LIBRARY on February 4, 2014 2791-02_Coburn.qxd 4/11/05 13:48 Page 34

Coburn

1999; Spillane & Jennings, 1997), this response This revised understanding of reading com- was widespread in the present study. Forty-nine prehension was accompanied by new instruc- percent of the 223 teacher responses resulted in tional approaches. Classroom observations later assimilation. in the year provided evidence that she had begun to adopt instructional strategies to encourage Accommodation children to make connections among their prior In rare instances, teachers in this study en- knowledge and experiences, other texts, and the gaged with policy ideas in ways that caused them selections they were reading. For example, in to restructure fundamental assumptions about the one representative observation, Marisa began nature of reading instruction or student learning. the lesson by talking with the children about how Or, in the language of cognitive learning theorists, good readers make connections between what teachers transformed their preexisting knowl- they know and the story. They then brainstormed edge structures to accommodate new information what they knew about gardens (drawing on work or experiences (Fosnot, 1996; Piaget, 1978; Smith, they were doing in a thematic unit on gardens in 2000). For example, Marisa grappled with new science), read a story about gardens, and jointly messages about reading comprehension in ways worked to create a chart comparing their experi- that caused her to restructure her conception of ence growing vegetables with the experience of the cognitive processes involved in comprehen- the main character in the story. In this way, Marisa sion and effective pedagogical strategies to help moved from a pedagogical strategy of asking students develop in that area. Earlier in her career, questions with correct and incorrect answers to a Marisa, like Sharon, saw reading comprehension strategy of encouraging students to develop com- as a collection of discrete skills that students prehension by making connections between what needed to master to capture the meaning inherent they know and what they are learning in the text. in the text. Thus, her instruction involved read- Thus, Marisa responded to messages about ing stories and answering a series of disconnected reading comprehension in a way that went much comprehension questions at the end that usually beyond the introduction of a new classroom or- focused on recall and low-level inference. For ganization or materials. Instead, encounters with example, in one observation that took place at the messages caused Marisa to challenge and re- start of her fourth year of teaching, Marisa had the structure the underlying pedagogical approach children read the story chorally and then asked that guided her interaction with children in the questions that tested their literal comprehension classroom. Overall, accommodation accounted of the events that had unfolded in the story. for only 9% of teacher responses. During the course of that year, however, Marisa Teachers’ Response to Nonsystem participated with her colleagues in professional Versus System Actors development that put forth a view of reading com- prehension as an active process wherein students Significantly, the three teachers in this study make meaning by drawing on their beliefs and were more likely to respond to connections with experiences as well as the text. After repeated nonsystem actors in ways that brought messages engagement and experimentation with this set of into the classroom in substantive ways. In this messages through participation in an ongoing sense, teachers’ connections with nonsystem ac- study group with her colleagues, Marisa gradually tors were more consequential for their classroom began to restructure her view of reading com- practice. As can be seen in Table 4, of the 128 prehension. At the end of the year, she declared: connections to policy messages via nonstate ac- tors included in this analysis, teachers responded Teaching reading comprehension is not just to just over 75% by bringing approaches into about giving them a paragraph and asking lit- their classroom in a way that influenced class- tle questions about it. . . . It’s not just the words room practice (parallel structures, assimilation, on the page are what you comprehend. You accommodation). Furthermore, of those they comprehend because of the experiences that you’ve had, the other books that you’ve read, brought into their classroom, they responded to what your neighbor just said. . . . It’s just the nearly 11% in ways that caused them to recon- whole inner web that we have of thinking and struct their underlying pedagogical or epistemo- knowledge. logical assumptions (accommodation). In terms 34

Downloaded from http://eepa.aera.net at NORTHWESTERN UNIV LIBRARY on February 4, 2014 2791-02_Coburn.qxd 4/11/05 13:48 Page 35

TABLE 4 Teachers’ Responses to Connections via System Versus Nonsystem Actors Response type System (n = 95) (%) Nonsystem (n = 128) (%) Rejection 33.7 21.8 Symbolic 13.7 2.3 Parallel structures 9.5 7.0 Assimilation 36.8 57.8 Accommodation 6.3 10.9

of nonsystem actors, the teachers’ connections documents and 11 of 22 messages encountered with independent professional development through district policy documents. They were providers, university coursework, and curricula most likely to respond symbolically to school were most likely to lead to substantive incorpor- policy messages, doing so in the case of 7 of the ation. Teachers incorporated 67.3% of 49 mes- 11 messages encountered in such a way. sages from independent professional development providers (with 12.2% leading to accommodation), Accounting for the Influence 8 of 10 messages encountered via university of Nonsystem Actors coursework (with 2 leading to accommodation), Why are nonsystem actors so consequential in and 72.7% of 55 messages carried by curricula the relationship between policy and practice? (but with only 1.8% leading to accommoda- Nonsystem actors connected teachers with pol- tion). At the same time, teachers responded to icy messages in ways that tended to have greater only 25% of messages from nonsystem actors intensity, greater proximity to the classroom, by ignoring them or responding symbolically. greater depth, and were more likely to be volun- In contrast, of the 95 connections made through tary. When teachers’ connections had these char- formal system channels, teachers responded to acteristics, teachers tended to respond in more just over 50% by bringing new approaches and consequential ways. Here I discuss each factor in ideas substantively into their classroom. Further- turn, providing an explanation of why it appears to encourage more substantive responses and more, teachers responded to only 6% of mes- comparing teachers’ connections to system ver- sages in ways that caused them to rethink some sus nonsystem actors along the dimension in aspect of their practice. Again, professional de- question. velopment from various system actors was most likely to lead to incorporation. Teachers incor- Intensity porated 51.8% of 27 messages from district pro- Teachers were more likely to respond to pol- fessional development providers (with 14.8% icy messages in ways that influenced their beliefs leading to accommodation) and 4 of 5 messages and classroom practice when their connections encountered through school professional devel- with them were of higher intensity, such as opment providers (with 1 message leading to sustained and ongoing interactions with high- accommodation). However, professional devel- quality professional development or one-on-one opment accounted for a smaller percentage of interaction over time with a key carrier.9 High- and teachers’ connections to system than nonsystem medium-intensity connections—especially to the actors. At the same time, teachers responded to a degree that they involved interaction with col- much higher percentage of messages from sys- leagues—provided the opportunity for teachers to tem actors by rejecting or ignoring them than engage in a process in which they experimented messages from nonsystem actors. Teachers re- with practice and then reflected on their experi- sponded to more than 47% of the messages from ence in ways that helped them shift practice over system actors by rejecting or ignoring them or by time. Greater intensity thus allowed teachers to responding symbolically. Teachers were most work with messages in an iterative fashion, con- likely to reject messages they encountered through structing understandings of new ideas and ap- state and district policy documents, rejecting 8 of proaches that often started with assimilation but, 13 messages encountered through state policy at times, pushed toward accommodation. 35

Downloaded from http://eepa.aera.net at NORTHWESTERN UNIV LIBRARY on February 4, 2014 2791-02_Coburn.qxd 4/11/05 13:48 Page 36

Coburn

For example, in the mid-1990s Deanna partic- incremental shifts in her practice. She began ipated with her colleagues in 18 months of on- teaching phonics in the context of stories but did site professional development on early literacy not shift her understanding of the reading process offered by an independent professional develop- in any significant way. But as she continued to ment provider—a connection with a high degree work with the professional development provider, of intensity. This professional development of- Deanna extended her experimentation, discussed fered a model of reading instruction that differed her experiences with colleagues, and received significantly from Deanna’s underlying beliefs ongoing feedback from the professional devel- about how children learn to decode. At that time, oper that guided her experimentation toward a Deanna taught decoding by using a structured deeper and more substantive response. She began phonics program that taught skills in a system- to see decoding as a complex process involving atic and largely decontextualized way. She also more than attention to phonics and began prompt- worked on decoding when she listened to chil- ing children to pay attention to semantic and syn- dren read in round robin reading groups, taking tactic cues in addition to phonemic information. care to correct them when they made mistakes. And, perhaps more significantly, she reported that The professional development program, in con- she learned a more facilitative way to interact with trast, put forth a model of decoding that empha- students while they were reading stories. Rather sized using multiple sources of information (or than interrupting students when they made mis- cueing systems) to decode, rather than a sole or takes and correcting them, she began to ask ques- predominant emphasis on using phonics. It also tions to help children figure out words on their emphasized teaching decoding in the context of own. For Deanna, this change represented a fun- stories rather than teaching it separately. Finally, damental shift in pedagogical strategy from a it emphasized a facilitative approach to teaching. transmission model wherein teachers tell students Rather than correcting children when they made the answer to a teacher-as-coach model wherein a mistake, the professional development encour- teachers ask students questions to help them aged teachers to ask children questions about their develop strategies of meta-cognition and self- reading (Does that look right? Does it sound right? correction—a change encouraged by long-term Does that make sense?) so that children would participation in professional development activ- develop the ability to be self-monitoring and self- ities over time. correcting readers (Fountas & Pinnell, 1996; Neal As can be seen in Table 5, connections with et al., 1997). nonsystem actors were more likely to be at a high When Deanna began working with the profes- or medium level of intensity than were those with sional development provider, she started to make system actors. Just under one third of teachers’

TABLE 5 Nature of Teachers’ Connections to System Versus Nonsystem Actors Nature of connection System (n = 95) (%) Nonsystem (n = 128) (%) Degree of intensity Low 51.6 17.2 Medium 34.7 50.0 High 13.6 32.8 Closeness to the classroom Low 55.8 7.8 Medium 26.3 42.2 High 17.9 50.0 Degree of depth Low 64.2 42.2 Medium 29.5 37.5 High 6.3 20.3 Voluntariness Normative 24.2 1.0 Regulative 75.8 99.0

36

Downloaded from http://eepa.aera.net at NORTHWESTERN UNIV LIBRARY on February 4, 2014 2791-02_Coburn.qxd 4/11/05 13:48 Page 37

The Role of Nonsystem Actors

connections via nonsystem actors were of high ment designed to provide formative information intensity, and half were of medium intensity. In that could be used to adjust instruction. Prior to contrast, just under 14% of connections via sys- the professional development, Sharon engaged in tem actors were of high intensity, and just over a formal assessment only at the beginning of the third were of medium intensity. In part, these dif- year to place children in reading groups and at ferences in levels of intensity were due to the fact the end of the year to fulfill district requirements. that a greater percentage of the teachers’ con- However, the assessment instrument and the ac- nections with nonsystem actors came in the form companying protocols for planning based on re- of ongoing professional development. But even sults of the assessment brought the idea of for- when teachers participated in professional devel- mative assessment into her classroom in a opment with system actors, these experiences concrete way for the first time. By the end of the tended not to be as sustained as those involving year, she was assessing all children on a monthly nonsystem actors. basis and adjusting the composition of her read- ing groups on the basis of the information pro- Closeness to the classroom vided by the assessment. The closer the connection was to the class- Teachers’ connections with nonsystem actors room, the more likely it was to influence teach- were influential in this regard. Connections with ers’ beliefs and practices. Policy documents, re- nonsystem actors were more likely to be close to source guides, standards documents, and even the classroom than connections with system ac- discussions taking place at professional devel- tors. Fifty percent of messages received via non- opment sessions at times felt very abstract to system actors involved a high degree of close- teachers. Without mechanisms to translate these ness to the classroom, as compared with just abstract ideas into the concrete realities of the under 18% of those received via system actors classroom, teachers were not always able to find (see Table 5). In part, this disparity was due to ways to bring them into the classroom in mean- the fact that teachers’ connections to messages ingful ways. For example, at Stadele Elementary, about reading from system actors often came in school leadership engaged teachers in a series of the form of policy documents—standards, task activities designed to better familiarize them with force reports, frameworks—and exhortations the new district standards. They engaged in a from school, district, and state leaders. Teachers process of writing grade-level indicators for dis- often experienced these documents and state- trict standards on decoding and assessed them- ments as abstract, with few mechanisms allow- selves on progress toward the meeting the stan- ing connections to be made to the classroom. dards. Yet, the conversations remained on an abstract level, and teachers found it difficult to Depth envision how some of the key ideas in the stan- Teachers were more likely to make consequen- dards would be reflected in classroom instruction. tial changes in their classroom practice in response Ultimately, early-grade teachers simply copied to policy messages when the messages presented district indicators as their own, with minimal dis- policy ideas with a high degree of depth. When cussion or consideration. Perhaps it is not sur- teachers were connected to policy messages that prising, then, that observations of teachers’ class- went beyond surface structures (e.g., specific ac- rooms showed little evidence that teachers were tivities or ways to organize the classroom for influenced by the work they did writing decoding reading instruction) to address underlying peda- standards. Approaches to decoding remained as gogical principles, they were more likely to ex- they had prior to standard writing and, in many amine and refine their assumptions about teach- cases, were quite different from the approaches ing and learning, leading to accommodation. emphasized in the standards. Deanna’s engagement with early literacy ideas In contrast, curricular materials, classroom as- about decoding, discussed earlier, provides a good sessments, and professional development activi- example of this phenomenon. In the course of ties that were close to the classroom brought pol- this professional development, Deanna not only icy messages into teachers’ classrooms in concrete learned new activity structures for teaching read- ways.10 For example, Sharon participated in pro- ing (guided reading) and incorporated new mate- fessional development on an assessment instru- rials within these structures (leveled texts), she 37

Downloaded from http://eepa.aera.net at NORTHWESTERN UNIV LIBRARY on February 4, 2014 2791-02_Coburn.qxd 4/11/05 13:48 Page 38

Coburn

also learned about underlying principles of student gree of intensity (usually district or school pro- learning (that students learn to read by relying on fessional development actors) were much less multiple cueing systems) and pedagogical prin- likely to be associated with a high or medium de- ciples (teacher as facilitator to develop self- gree of depth. monitoring, self-correcting readers). That is, the professional development provided a connection Voluntariness to the approach with a high degree of depth. Finally, degree of voluntariness influenced In contrast, when Deanna and other teachers teachers’ responses to policy messages. The vast encountered instructional approaches or organi- majority of messages teachers were connected zational structures without underlying princi- with via both system and nonsystem actors (just ples, they tended to do one of two things. In the under 90%) came in the form of normative pres- first scenario, they tried to incorporate these ap- sure. That is, these messages put forth ideas proaches or structures into their classroom but about what teachers should do, specifying valued transform them significantly through the lens of ends for instruction or the appropriate means to preexisting practice. In these cases, underlying get there (Scott, 2001). Thus, while teachers classroom norms and assumptions about teach- were encouraged to change their practice in a ing reading remained unaltered, leading to as- certain direction, ultimately it was a voluntary similation. In the second scenario, they experi- decision. However, teachers’ connections to 23 mented with approaches, found them difficult to out of 223 messages—all but one from system implement absent a deeper understanding of the actors—were accompanied by regulative pres- pedagogical principles involved, and failed to sus- sures. That is, they involved rule setting, moni- tain them. For example, in the early 1990s, both toring, or sanctioning (Scott, 2001). Teachers Deanna and Sharon experienced pressure to use were mandated or required to teach in a particu- whole-class instruction. However, both teachers lar way, to particular ends, or using particular were connected to these ideas with a low degree of curricular materials, and as a result the messages depth. That is, they learned about the organiza- were seen as involuntary. Regulative pressures tional structure but had virtually no professional most often accompanied school policy and, to a development on instructional approaches to use in lesser extent, district and state policy messages. this structure or the underlying theory of student As illustrated in Table 5, teachers were more learning at its root. After initially reorganizing their likely to respond to normative messages than reading programs to do whole-group instruction, regulative messages by incorporating them into Sharon abandoned the approach when she could their classroom and doing so in ways that altered not figure out a way to make it work with the ped- preexisting practice. Two thirds of teacher re- agogical approaches she used, and Deanna succes- sponses to normative pressure involved changes sively modified whole-class instruction over time in classroom practice, while just over half of until it resembled her original approach. teacher responses to regulative pressures did. Teachers’ connections with nonsystem actors However, when teachers did respond to regula- tended to have a higher degree of depth than their tive pressure by making changes in classroom connections with system actors. Just over 57% of practice, they did so by developing parallel struc- teachers’ connections to nonsystem actors were tures when the messages had a medium degree at a medium or high degree of depth, as com- of congruence with their preexisting practices pared with just over 35% of their connections (seven messages) and assimilating messages that with system actors (see Table 5). In part, degree had a high degree of congruence (two messages). of depth was linked with degree of intensity, in (Teachers did not incorporate regulative messages that greater intensity seemed to allow teachers to at all when they had a low degree of congruence.) encounter policy messages that pushed beyond Thus, when teachers responded to mandates by superficial manifestations to include underlying bringing them into their classroom, they did so in pedagogical principles. Interestingly, however, ways that left their underlying pedagogical and in comparison with nonsystem actors associated epistemological assumptions intact.11 Regulative with a high or medium degree of intensity, sys- messages accounted for nearly one quarter of all tem actors associated with a high or medium de- of the teachers’ connections to policy messages

38

Downloaded from http://eepa.aera.net at NORTHWESTERN UNIV LIBRARY on February 4, 2014 2791-02_Coburn.qxd 4/11/05 13:48 Page 39

The Role of Nonsystem Actors

via system actors, hence accounting for some of policy over time, so too did the content and focus the failure of messages related to system actors of the professional development and instructional to influence teacher practice. materials each teacher encountered via nonsystem actors. Figure 1 shows that, after the shift in state Interrelationship Between System policy from basic skills to literature-based instruc- and Nonsystem Actors tion beginning in late 1987, Sharon’s connections Thus far, I have argued that (a) teachers’ con- to policy ideas via nonsystem actors began to shift nections to policy messages vary considerably from an exclusive basic skill orientation toward in their mechanism, intensity, and content and literature-based instruction—a trend that contin- (b) messages carried by nonsystem actors tend ued throughout the first half of the 1990s. After a to be more influential for classroom practice brief period with limited connections via nonsys- than those carried by system actors. But if teach- tem actors, Sharon began to be connected to mes- ers’ connections with nonsystem actors are so sages associated with early literacy and, to a lesser consequential, what factors shape these connec- extent, “balanced” instruction via nonsystem ac- tions? In this section, I argue that the interrela- tors after policy shifts in the early and mid-1990s. tionship between system and nonsystem actors Deanna exhibited a similar pattern in her con- plays a key role in structuring teachers’ access nection to nonsystem actors. When Deanna reen- to and experience with nonsystem actors. There tered teaching in 1990, she experienced a major are three key elements of this interrelationship. shift from what had been exclusively basic skill First, at the state level, changes in policy influence messages when she started teaching in the early the content of teachers’ connections to messages 1970s (not represented in Figure 1) to a via nonsystem actors by helping to define and heavy connection with messages associated with create legitimacy for particular conceptions of literature-based instruction that continued through “good” reading instruction. Second, within the the mid-1990s. Starting in 1993, her connections to context of these state-level shifts in conceptions nonsystem actors began to shift to early literacy, of good reading instruction, local system actors a trend that continued through the late 1990s. play a mediating role between teachers and non- After the emergence of the state’s version of “bal- system actors. This accounts in part for the vari- anced” instruction in 1995, she was connected to ation observed among teachers. Third, the spe- a few messages carrying these ideas via nonsys- cific ways in which system and nonsystem actors tem actors. Finally, Marisa entered teaching interact as particular policies emerge—the dif- during an especially active period of state policy- ferent roles played by actors at different histori- making; literature-based instruction and early cal moments—have an important impact on the literacy messages were being challenged at the mechanisms by which teachers come into con- state level by the movement toward balanced tact with nonsystem actors during policy imple- basic skills. Thus, she experienced these shifts in mentation. This last factor accounts for variation a highly compressed fashion. Initially, Marisa’s in teachers’ connections with nonsystem actors connections to messages via nonsystem actors during different policy initiatives. heavily favored literature-based instruction, but shortly thereafter she became increasingly Influence of State Policy Shifts on connected to early literacy ideas. Finally, near Content of Connections the end of the decade, as state policy-making fo- As discussed earlier, previous research has doc- cused on balanced instruction increased, she umented the way in which policy emerges from began to be connected to these ideas via non- and is affected by interaction with nonsystem ac- system actors as well. tors such as professional organizations, university This pattern whereby teachers’ connections to researchers, policy entrepreneurs, and others (Hill, nonsystem actors shifted with changes in state 2000; Kingdon, 1984; Laumann & Knoke, 1987; policy was especially pronounced in the case of Moore, 1988; Scott et al., 2000). But state policy nonsystem actors with direct connections to the also seemed to influence the content of messages state policy system. For example, given that carried by nonsystem actors to schools. As con- textbooks are adopted according to a set of state- ceptions of good reading instruction shifted in state developed criteria, it is not surprising that the

39

Downloaded from http://eepa.aera.net at NORTHWESTERN UNIV LIBRARY on February 4, 2014 1=basic skills 2=literature-based instruction 3=early literacy 40 4=“balanced” instruction

35 = Sharon

30 = Deanna

= Marisa 25

20 Number of connections Number

15

10

5

1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4

1983-1985 1986-1988 1989-1991 1992-1995 1996-1998 FIGURE 1. Teachers’ connections to policy ideas via nonsystem actors.

content and format of textbooks changed with from 1983 to 1985, 60% of curricula and pro- these criteria, albeit to varying degrees. How- fessional development activities promoted basic ever, this phenomenon also occurred when non- skills, and 40% promoted literature-based instruc- system actors had no such direct connections, as tion. By 1989–1991, this balance had shifted, was the case for the majority of teachers’ con- with 40% emphasizing basic skills and 60% em- nections to messages via nonsystem actors (75 phasizing literature-based instruction. Finally, of 128 messages). during 1996 through 1999, the balance shifted Changes in state policy seemed to influence yet again, as 3% of curricula and professional de- the content of teachers’ connections to nonsys- velopment offerings from nonsystem actors were tem actors on both the supply and the demand literature based, 68% promoted early literacy, side. On the supply side, state policy played an and 29% promoted balanced instruction. Thus, it important role in influencing the nature of what appears that there were different resources avail- nonsystem actors offered. Analyses of profes- able to teachers from nonsystem actors at different sional development and instructional documents historical moments. In a particularly vivid exam- available at the local, regional, and state levels ple, the same independent provider of professional indicate that the content of nonsystem actors’ work development on literature-based approaches to evolved (to varying degrees) with the shifting ideas both Baldwin and Stadele Elementary in the about what constituted good reading instruction in late 1980s provided professional development to the policy and professional environment. Thus, teachers in the region on systemic approaches to

40 Downloaded from http://eepa.aera.net at NORTHWESTERN UNIV LIBRARY on February 4, 2014 2791-02_Coburn.qxd 4/11/05 13:48 Page 41

The Role of Nonsystem Actors

phonics and phonemic awareness associated define what constituted good instruction. This, in with the state’s version of balanced instruction in turn, influenced both the content of professional the late 1990s. development, teaching materials, and reform ef- On the demand side, the nature of teachers’ forts and teachers’ decisions to reach out to new connections to nonsystem actors changed as ideas and approaches. teachers, schools, and districts began seeking out different kinds of messages about reading in Mediation of Access by Local System Actors ways that coincided with changing state policy. Within the broader context of shifting content, As messages linked with a particular policy be- teachers’ access to policy messages via nonsys- came increasingly salient and legitimate in the tem actors was mediated at the local level by dis- environment, teachers, schools, and districts began trict and school leadership. Local leaders influ- to be aware of their presence and began to hear enced what nonsystem resources teachers had about them from their colleagues. This salience, access to by making decisions about contract- coupled with recommendations from trusted ing with independent professional development others, contributed to the sense that particular ap- providers, purchasing instructional materials, pro- proaches represented “best practices” in reading viding resources for teachers to attend profes- instruction. For example, at the district level, sional development activities on their own, and, both districts adopted Reading Recovery after in recent years, working with local universities to district officials learned about the program from provide coursework for uncredentialed teachers. officials in other districts at conferences or meet- By choosing to allocate resources for curricular ings. At the school level, school leaders often materials and professional development, and by made decisions about professional development contracting with some nonsystem actors and not or materials on the basis of recommendations others, district and school leaders influenced not from trusted colleagues or an emerging aware- only the quantity and variety of connections teach- ness of the pervasiveness of a given approach. ers had to policy messages but the content and in- For example, the school leaders at Stadele Ele- tensity of these connections as well. As a result, mentary provided training on a particular ap- teachers in different districts and schools at times proach to reading comprehension as they became had access to substantively different messages increasingly aware that many other schools in the about reading policy. This local mediation in part region were using it. The principal explained, accounts for the variation observed among the “Everybody had always been talking about [the teachers in the areas of mechanism, content, and approach] and how helpful it is.” Similarly, all intensity (see Tables 1–3). three teachers repeatedly explained their deci- At the district level, this phenomenon can be sions to participate in a particular type of profes- illustrated by comparing teachers’ connections sional development or coursework with refer- with policy messages concerning early literacy. ence to recommendations from colleagues or While both districts in the study were strongly their sense that everyone was doing it. connected with and supportive of policy messages Thus, as suggested by institutional theorists, linked to early literacy, differences in approach shifts in state policy, themselves often part of shaped how and to what aspects teachers were broader movements whereby particular sets of connected. Deanna’s district not only invested ideas were legitimized, furthered the legitimiza- enormous resources to bring policy messages tion of these ideas. Nonsystem actors—both about early literacy to the school but relied on those with direct ties to system actors and those a strategy that simultaneously involved using without—shifted their content and approach district personnel to provide instructional guid- in subtle and not-so-subtle ways in response. ance to schools (e.g., hiring district literacy And decision makers at each level of the sys- coaches to work with elementary schools to im- tem reached out to nonsystem actors in ways that prove reading instruction) and contracting with brought teachers into contact with favored mes- a range of independent professional develop- sages. In this way, state policies shaped mes- ment providers to provide an enormous amount sages teachers encountered not only through of professional development to K–3 teachers in organizational means, via the formal policy sys- the district. Deanna and the 11 other teachers tem, but through normative means, by helping to interviewed at Baldwin Elementary were thus 41

Downloaded from http://eepa.aera.net at NORTHWESTERN UNIV LIBRARY on February 4, 2014 2791-02_Coburn.qxd 4/11/05 13:48 Page 42

Coburn

connected to policy messages about early liter- nected to literature-based approaches through acy in multiple ways involving both system and their credential program, most of the seven teach- nonsystem actors. ers interviewed who were present at the school at In contrast, Sharon and Marisa’s district de- that time, including Deanna, had relatively few voted fewer resources to learning opportunities connections to such approaches. As illustrated in for teachers related to early literacy in general Tables 1–3, Deanna was not connected to mes- and had a philosophy of providing opportunities sages about literature-based instruction in as for professional growth using district staff rather many ways, or with as much intensity, depth, and than contracting out with independent profes- proximity to the classroom, as was Sharon in the sional development providers. Thus, while the early 1990s. district personnel were for a time deeply com- It is important to note that while local system mitted to early literacy ideas, and these ideas in- actors such as districts and schools played an im- fluenced many aspects of district policy and pro- portant role in mediating teachers’ access to non- fessional development, Sharon, Marisa, and the system resources, they were not the only influ- 16 other teachers interviewed at Stadele Elemen- ences. Teachers became connected to messages tary had far fewer connections with early literacy via nonsystem actors as they reached out on their policy messages via nonsystem actors. As can be own to materials, coursework, and professional seen in Tables 2 and 3, Marisa and Sharon’s con- development available in their environment. In nections to early literacy were less intense and addition, they encountered messages as they in- less deep than were Deanna’s. teracted with their colleagues in their proximal School leadership also played an important role teacher communities. Teachers in different micro- in mediating teachers’ connections to nonsystem communities of practice, even at the same school, actors. This phenomenon can be illustrated by at times were connected in different ways to non- comparing Sharon and Deanna’s connections to system actors in the environment. policy ideas related to literature-based instruction. In the early 1990s, the principal of Sharon’s school Process of Policy Emergence was a key advocate of literature-based approaches Finally, the ways in which system and non- to reading instruction. She engaged the staff in system actors interacted in the environment as a whole-school experiences with independent pro- given policy emerged influenced the mecha- fessional development providers on portfolio nisms by which teachers ultimately learned about assessment and thematic teaching, provided re- new policy ideas. Nonsystem actors played dis- sources for early-grade staff to attend professional tinctly different roles in bringing about changes development activities involving shared reading, in state-level policy during different historical and devoted enormous resources to purchasing periods. To simplify greatly, the movement to- children’s literature. As a result, Sharon and the ward state policy emphasizing literature-based 11 other teachers interviewed who were present at instruction involved the confluence of top-down the school at that time were connected to multiple state policy and a grass-roots teachers’ move- nonsystem actors promoting literature-based ap- ment. The attempts of Bill Honig, the state su- proaches. Many of these connections were char- perintendent of instruction, to bring high-quality acterized by high degrees of depth and intensity. literature to children through an early form of In contrast, the principal of Deanna’s school state systemic reform intersected with a bur- was skeptical of approaches associated with geoning grass-roots movement promoting whole- literature-based instruction. In the early 1990s, language approaches to instruction spreading she spent funds to purchase packaged phonics ap- through the state via local teacher study groups. proaches and train teachers on them to supple- The early literacy movement spread more at the ment what she saw as a lack of attention to phon- middle level of the policy system. Reform ideas ics in the state-adopted reading series. She also spread as nonsystem actors (primarily universi- brought in a researcher from a local university to ties connected with Reading Recovery and pro- provide professional development on approaches fessional development organizations) formed al- to decoding that she saw as moving away from liances with local districts. It was only after the literature-based instruction. Thus, while there movement had begun to spread from district to were teachers at Deanna’s school who were con- district that it became embedded in state policy. 42

Downloaded from http://eepa.aera.net at NORTHWESTERN UNIV LIBRARY on February 4, 2014 2791-02_Coburn.qxd 4/11/05 13:48 Page 43

The Role of Nonsystem Actors

In a third pattern, the movement toward a “bal- were not connected with policy messages either anced” approach was predominantly top down. through the district or through nonsystem actors While members of professional associations such as professional development providers to were involved in task force commissions in the the same extent as in previous movements.14 early years, the particular configuration of “bal- Rather, at least during the first 4 years of this pol- ance” that was built into policy came from a icy movement, the main ways in which these coalition of researchers and policymakers, and teachers learned about ideas associated with the participation on the part of professional associ- state’s version of “balanced” instruction were ations and local educators was more limited. through widespread media coverage and stan- The state did interact with some professional dardized testing, routes that are long on reach but development providers, but it did so by provid- short on detailed information regarding the spe- ing funds to local districts that could be used cific approach to reading instruction involved in only to contract with a relatively small number the policy.15 of professional development providers that Thus, the process by which particular sets of passed state certification for the content of their ideas about reading instruction gained legitimacy approaches.12 and became embedded in policy influenced the Each of these sets of interrelationships had mechanisms by which teachers ultimately were implications for the nature of teachers’ connec- connected to these ideas. The nature of teachers’ tions to policy ideas via nonsystem actors. Per- connections, in turn, influenced whether and how haps because state policy promoting literature- teachers made changes in their classroom practice based instruction grew out of a broader movement in response to shifting state reading policy. Policy involving professional organizations and profes- changes that came about in partnership with non- sional development providers, Deanna, Sharon, system actors, who had greater capacity to bring and other teachers who completed interviews policy ideas close to the classroom in more sus- and were present at the school at that time were tained ways, tended to result in connections for more likely to be linked to policy ideas through teachers that had more depth, intensity, and prox- independent professional development providers imity to the classroom. This was the case with lit- (and the associated school-level decisions) than erature-based instruction and early literacy. In during later policy movements.13 Twenty-six contrast, the policy change that emerged with lim- percent of teachers’ connections during the era ited engagement of local districts and the multi- of literature-based instruction were via indepen- faceted set of nonsystem actors—the state’s ver- dent professional development providers, while sion of the balanced approach—seemed to result only 16% of connections during the era of early in fewer overall connections for teachers with literacy and 7% during the era of balanced in- state policy ideas and connections that were less struction involved such providers. sustained and less close to the classroom. In contrast, given the key role of the district and its relationship to statewide reform orga- Conclusions nizations during the early literacy movement, Scholarship on policy implementation often Deanna, Sharon, and other teachers at both schools excises policy from its broader social and cultural who taught during all three periods were much environment. Scholars focus primary attention more likely to be connected to policy ideas via on the way policy ideas move through the for- the district than via other means during the early mal policy system and into schools, paying little literacy movement. Just over 30% of teachers’ if any explicit attention to the myriad nonsystem connections to messages during the early literacy actors that interact with the system throughout era came via the district, while only 17% of mes- the implementation process. A few researchers sages during the era of literature-based instruction have argued that nonsystem actors are a key me- and 5% during the era of balanced instruction diating link between instructional policy and came via the district. classroom practice (Cohen & Hill, 2000; Hill, Finally, perhaps because of the limited in- 2000). The present study echoes this point, pro- volvement of either districts or professional or- viding evidence that teachers’ connections with ganizations with the state’s version of “balance” nonsystem actors tend to be more consequen- in the early stages of the movement, teachers tial for changes in classroom practice as well. 43

Downloaded from http://eepa.aera.net at NORTHWESTERN UNIV LIBRARY on February 4, 2014 2791-02_Coburn.qxd 4/11/05 13:48 Page 44

Coburn

Because teachers tend to have connections with wide range of system and nonsystem routes. nonsystem actors that are more intense and in- While the strategy of linking state reading frame- volve greater depth, proximity to the classroom, works to textbooks, professional development, and voluntariness, they are more likely to re- and assessment both in 1987 and then again in spond to policy messages carried by nonsystem the late 1990s created a great deal of normative actors by bringing approaches into the class- pressure around each set of policy messages and room, at times rethinking assumptions and re- shifted the focus of nonsystem actors to some de- configuring instruction in substantive ways. gree, teachers continued to be connected to a But in this study I go further, arguing that the range of messages about reading instruction as- interaction between system and nonsystem ac- sociated with earlier policy initiatives. The no- tors at both the state and local levels influences tion of congruence becomes even more compli- the nature and content of teachers’ connections cated when the temporal element is introduced. to nonsystem actors. Policy emerges out of move- At any given time, teachers are not only con- ments in the environment whereby particular sets nected to current policy but also connected to of ideas gain legitimacy over time in ways that previous policies as they have become embedded are influenced by nonsystem actors. Policy, in in teaching materials, local teaching practices, turn, creates further legitimacy for these ideas, and routines in their proximal community. Shifts influencing what is available to teachers via non- in the professional and policy world introduce system actors. The particular ways in which sys- ideas into the system in ways that persist long tem and nonsystem actors interact during the after the policy has gone away. Thus, while ele- emergence of policy influence the mechanisms ments of the state policy system may be coherent by which teachers become connected to policy to varying degrees, teachers continue to be con- during implementation. And, at the local level, nected with ideas related to previous policy in teachers’ connections are further mediated by their day-to-day work in schools. school and district leaders, who shape access for All of this suggests that coherence is more teachers to some nonsystem actors and not others. than a property of the policy system at a fixed These findings have important implications time. Rather, coherence should be seen as a prop- for policy and research. First, this portrait of the erty of the broader environment for instructional role of nonsystem actors complicates notions of guidance. This recasting highlights the enormous policy congruence along several dimensions and challenge in bringing about coherence, especially levels. Calls for increased policy coherence have when the policy change envisioned is a large one become increasingly widespread in the educa- along multiple dimensions, as has been the case tional policy community (Fuhrman, 1993; Smith with recent shifts in reading and other instruc- & O’Day, 1991). The argument is that teachers tional policies. It further suggests that if coher- are more likely to make changes in their practice ence is the goal, policymakers and others must in a particular direction if the policy system pro- employ a much broader range of strategies than motes these approaches along multiple dimen- targeting the formal policy system alone. Coher- sions (Cohen & Hill, 2000; Spillane & Jennings, ence seems to require connections with a broader 1997). And, indeed, California was one of the range of actors at multiple levels and from mul- early states to link aspects of instructional policy tiple sectors of the environment. Chief among together in a move toward congruence (Carlos & these actors are the professional networks, re- Kirst, 1997; Massel et al., 1994). In reading in- sources, and reform organizations that work most struction, the state linked instructional frame- closely with teachers and schools. works (and later standards) to textbook adoption, However, this study also suggests that policy- assessment, and professional development start- makers face trade-offs in relying on nonsystem ing in 1987. actors in this manner. On the one hand, as sug- However, the strategy of policy coherence gested earlier, many nonsystem actors have a seems to assume that policy messages are the pri- greater capacity than policy actors to reach teach- mary or, at least, most influential source of mes- ers in ways that are substantive, sustained, and sages about instruction for teachers. This study situated in their day-to-day work in the class- documents the ways in which teachers are con- room. On the other hand, nonsystem actors, like nected to messages about instruction through a local policy actors (e.g., districts), tend to trans- 44

Downloaded from http://eepa.aera.net at NORTHWESTERN UNIV LIBRARY on February 4, 2014 2791-02_Coburn.qxd 4/11/05 13:48 Page 45

The Role of Nonsystem Actors

form messages as they carry them to teachers. As ment providers, local policymakers indirectly in- a result, nonsystem actors are a powerful yet not fluence teachers’ connections to policy by shap- entirely controllable mechanism for reaching ing the nature and intensity of their connections teachers. to nonsystem actors. This suggests the need for In California, different policy strategies vis-à- policy researchers to expand implementation vis nonsystem actors, during both the emergence studies to include careful attention to the full of policy and its implementation, emphasized range of mechanisms by which teachers are con- these two dimensions to different degrees. As the nected to policy messages, moving beyond sys- formal policy system increasingly worked with tem actors to include the myriad of nonsystem or through nonsystem actors, teachers were in- actors as well. creasingly connected to policy messages in sub- stantive but often somewhat transformed ways. Notes State efforts in the late 1990s to control messages I wish to thank Betty Achinstein, Larry Cuban, from nonsystem actors resulted in more uniform Meredith Honig, Nathan MacBrien, Teresa Mc- connections. Because the state scaled back infor- Caffrey, Milbrey McLaughlin, Janet Schofield, Charles mal relationships with a range of teacher profes- Thompson, and Stacey Rutledge for comments on ear- sional associations and networks, however, the lier versions of this article. Support for data collection teachers in this study were connected to recent was provided by the Center for Research on the Con- reading policy in ways that were less intense and text of Teaching at Stanford University. Data analysis further from the classroom than was the case and writing of the article were supported by a disser- with previous state initiatives. In a policy system tation fellowship from the Spencer Foundation and by as complex and fragmented as that of the United support from the School of Education and the Learn- States, policymakers must think carefully about ing Research and Development Center at the Univer- sity of Pittsburgh. how to manage strategic trade-offs between 1 Following Scott and his colleagues (2000), I use the reaching teachers and controlling their message. term actors to refer not only to individuals but also to Finally, these findings highlight the need for collective actors such as organizations and associations. policy researchers to pay greater attention to the 2 For many years, California has provided funding ways in which the formal policy system is situ- for districts to purchase reading textbooks provided ated in the broader social and cultural environ- that the districts choose textbooks that are approved by ment. Formal policy is one part of a larger social the state according to a set of state criteria. However, and organizational web through which teachers since 1997, the state has extended the use of criteria are connected with ideas about instruction. By beyond textbooks in an attempt to control the content focusing solely or predominantly on the formal of professional development supported by state funds. policy system, most policy research misrepre- Under AB 1086, professional development providers sents the policy implementation process, failing had to submit professional development plans to the to take into account the ways in which the recip- state to be approved in order for districts to contract with them. rocal relationship between system and nonsys- 3 All school and teacher names used here are pseu- tem actors influences what and how teachers donyms. learn about ideas associated with shifting in- 4 The large difference in observation time in Stadele structional policy. versus Baldwin Elementary reflects the fact that Bald- Furthermore, perhaps ironically, focusing on win Elementary had significantly fewer structured op- policy alone may tend to underplay the influence portunities for teachers to meet. In part, this differ- of policy shifts on classroom practice. In empha- ence stems from district-level policy decisions. The sizing policy’s direct influence on classroom state provided both districts 3 pupil-free professional practice, this approach often fails to capture the development days that could be used for teachers’ way policy influences practice indirectly through meetings. However, Stadele’s district funded 6 addi- tional pupil-free days, such that Stadele had 9 pupil- normative means. By playing a key role in defin- free days devoted to professional development during ing what constitutes good reading instruction, the 1998–1999 school year. In contrast, Baldwin’s dis- state policy influences the content of professional trict did not elect to provide funds for any additional development, teaching materials, and reform ef- days to supplement those provided by the state, and forts. And in making decisions about teaching thus Baldwin had only 3 pupil-free days. Beyond dis- materials and independent professional develop- trict decisions, however, the amount of time teachers 45

Downloaded from http://eepa.aera.net at NORTHWESTERN UNIV LIBRARY on February 4, 2014 2791-02_Coburn.qxd 4/11/05 13:48 Page 46

Coburn

had to meet in the school depended on how school were closer to the state’s vision of balanced reading in- leaders decided to allocate building-level resources. struction. Second, as mentioned earlier, the state pro- The school leaders at Stadele devoted a significant vided grants to districts to make available professional portion of their discretionary resources to providing development for K–3 teachers on elements of the bal- opportunities for teachers to meet, paying teachers to anced approach as specified in state legislation. The meet after school once a week and hiring substitutes state attempted to control messages about reading by to release teachers for a half day per month to work on requiring professional development providers to sub- reform-related projects. In contrast, school leaders at mit plans to the state to be approved in order for dis- Baldwin did not choose to devote resources in this tricts to contract with them. However, districts could way. Thus, teachers at Baldwin met for only 1 hour bypass this approval process by creating their own after school three times per month to work together, professional development. Both of the present districts fulfilling the requirements of their contract. took this route and created their own professional de- 5 All three focal teachers encountered roughly the velopment for K–3 teachers in ways that reinterpreted same number of messages. This is somewhat surpris- the state’s rendition of balanced instruction through ing because one of the three focal teachers, a begin- the lens of an early literacy approach. ning teacher, taught for only 4 years of the study. 15 It is important to note that the study was con- However, this disparity can be attributed to the fact ducted relatively soon after state-level policy changes that this teacher was connected to messages about promoting the balanced approach. (The study was reading through multiple state and district programs conducted during the 1998–1999 school year. State and professional development activities targeted to policy promoting the balanced approach began in new teachers, programs that did not exist when the 1995 and continued at an aggressive pace through the other two teachers entered the profession. study year and beyond.) Relative to other approaches, 6 For a more in-depth analysis of changes in read- policy ideas associated with the state’s rendition of ing policy in California from 1983 to 1999, along with balance had not had as much time to work their way characteristics of particular movements, see Coburn through the system and into schools. It is possible that (2001b). teachers’ connections to nonsystem actors carrying 7 See Hill (2000) for a similar conclusion about math messages about the state’s version of the balanced ap- textbooks in Connecticut. See Rowan (2001) for an proach have increased in the time since the present explanation of this practice rooted in the market pres- data were collected. However, especially in the case of sures faced by textbook publishers. literature-based instruction, teachers’ connections with 8 For a more elaborated discussion of this typology, nonsystem actors began in close proximity to the pol- see Coburn (2004). icy change itself. In cases in which there was more par- 9 Cohen and Hill (2000) offered further support for ticipation in the movement toward new sets of ideas this point, providing evidence that greater time spent about reading instruction by nonsystem actors, teach- in professional development results in greater change ers were connected more quickly to these ideas in teachers’ practice. through nonsystem actors than appears to be the case 10 See also Ball and Cohen (1996) on this point con- with more recent policy initiatives. cerning textbooks. 11 This finding should be viewed as tentative given References that it is based on a small number of messages. Askew, B. J., Fountas, I. C., Lyons, C. A., Pinnell, 12 See Coburn (2001b) for a more extended descrip- G. S., & Schmitt, M. C. (1998). Reading Recovery tion of the dynamics of shifting policy. review: Understandings, outcomes, and implications. 13 For the analysis reported in this section, I relied on Columbus, OH: Reading Recovery Council of North data from teachers who taught through all three policy America. movements. Thus, I used data on connections to policy Ball, D. L., & Cohen, D. K. (1996). Reform by the messages from Sharon, Deanna, 11 additional teachers book: What is—or might be—the role of curriculum at Stadele Elementary, and 7 additional teachers at materials in teacher learning and instructional re- Baldwin Elementary. form? Educational Researcher, 25(9), 6–8, 14. 14 In fact, the districts, with their heavy emphasis on Barley, S. R. (1990). Images of imaging: Notes on doing early literacy resulting from their previously strong longitudinal field work. Organizational Science, 1, connection with it, tended to reconstruct state policy 220–247. messages through the lens of early literacy, such that Brandt, R. (1989, November). On curriculum in Cali- increased state policy-making toward balanced in- fornia: A conversation with Bill Honig. Educational struction resulted in increased district activity promot- Leadership, pp. 10–13. ing early literacy. First, both districts adopted text- Breneman, B., & Parker, D. (1991). Reading Recovery: books from the state list of approved texts that tended An idea whose time has come. California Reader, to favor early literacy approaches rather than those that 24(3), 24–28. 46

Downloaded from http://eepa.aera.net at NORTHWESTERN UNIV LIBRARY on February 4, 2014 2791-02_Coburn.qxd 4/11/05 13:48 Page 47

The Role of Nonsystem Actors

Burch, P. E. (2000). Moving from the margins to the at curricular and systemic reform. Teachers College mainstream: Teaching and learning reform in local Record, 104, 765–786. policy context. Unpublished doctoral dissertation, D’Aunno, T., Sutton, R. I., & Price, R. (1991). Iso- Stanford University, Stanford, CA. morphism and external support in conflicting insti- California State Board of Education. (1999). Guide to tutional environments: A study of drug abuse treat- the California Reading Initiative 1996–1999. Sacra- ment units. Academy of Management Journal, 34, mento: California Reading Initiative Center. 636–661. California State Department of Education. (1987). DiMaggio, P. J. (1991). Constructing an organiza- English-language arts framework. Sacramento, CA: tional field as a professional project: U.S. art muse- Author. ums, 1920–1940. In W. W. Powell & P. J. DiMaggio California State Department of Education. (1995a). (Eds.), The new institutionalism in organizational Every child a reader: The report of the California analysis (pp. 63–82). Chicago: University of Chicago Reading Task Force. Sacramento, CA: Author. Press. California State Department of Education. (1995b). Dobbin, F., Sutton, J. R., Meyer, J. W., & Scott, W. R. Ready, set, let’s read!: A helping handbook for com- (1993). Equal opportunity law and the construction prehensive early literacy instruction (draft version). of internal labor markets. American Journal of Sacramento, CA: Author. Sociology, 99, 396–427. Carlos, L., & Kirst, M. (1997). California curriculum Dowd, T. J., & Dobbin, F. (1997). The embedded policy in the 1990s: “We don’t have to be in the front actor and the invention of natural economic law: to lead.” San Francisco: PACE. Policy change and railroader response in early Coburn, C. E. (2001a). Collective sensemaking about America. American Behavioral Scientist, 40, 478– reading: How teachers mediate reading policy in their 489. professional communities. Educational Evaluation Edelman, L. B. (1992). Legal ambiguity and symbolic and Policy Analysis, 23, 145–170. structures: Organizational mediation of civil rights Coburn, C. E. (2001b). Making sense of reading: Log- law. American Journal of Sociology, 97, 1531–1576. ics of reading in the institutional environment and Eilers, A. M. (2004). Below the accountability radar the classroom. Unpublished doctoral dissertation, screen: What does state policy say about school Stanford University, Stanford, CA. counseling? Education Policy Analysis Archives, Coburn, C. E. (2004). Beyond decoupling: Rethinking 12(3), 1–32. the relationship between the institutional environ- Fosnot, C. T. (1996). Constructivism: Theory, perspec- ment and the classroom. Sociology of Education, 77, tives, and practice. New York: Teachers College 211–244. Press. Cohen, D. K. (1982). Policy and organization: The Fountas, I. C., & Pinnell, G. S. (1996). Guided reading: impact of state and federal educational policy on Good first teaching for all children. Portsmouth, NH: school governance. Harvard Educational Review, Heinemann. 52, 474–499. Fuhrman, S. (Ed.). (1993). Designing coherent educa- Cohen, D. K. (1995). What is the system in systemic tion policy: Improving the system. San Francisco: reform? Educational Researcher, 24(9), 11–17, 31. Jossey-Bass. Cohen, D. K., & Hill, H. C. (2000). Instructional pol- Fuhrman, S., Clune, W. H., & Elmore, R. F. (1988). icy and classroom performance: The mathematics Research on educational reform: Lessons on the reform in California. Teachers College Record, implementation of policy. Teachers College Record, 102, 294–343. 90, 237–257. Cohen, D. K., & Spillane, J. P. (1993). Policy and Fuhrman, S., & Elmore, R. F. (1990). Understanding practice: The relations between governance and local control in the wake of state educational reform. instruction. In S. F. Fuhrman (Ed.), Designing co- Educational Evaluation and Policy Analysis, 12, herent education policy: Improving the system 82–96. (pp. 35–95). San Francisco: Jossey-Bass. Guthrie, J. W. (Ed.). (1990). Special issue. Educational Cunningham, P. M. (1995). Phonics they use: Words for Evaluation and Policy Analysis, 12(3). reading and writing (2nd ed.). New York: Harper- Hartley, J. F. (1994). Case studies in organizational Collins. research. In C. Cassell & G. Symon (Eds.), Qual- Curriculum Development and Supplemental Materials itative methods in organizational research: A prac- Commission. (1994). K–8 language arts instruc- tical guide (pp. 208–229). Thousand Oaks, CA: tional resources evaluation form. Retrieved from Sage. http://www.cde.ca.gov/cilbranch/englang/ela_adopt Hill, H. C. (2000). Implementation networks: Non- .htm state resources for getting policy done. Unpublished Cusick, P. A., & Borman, J. A. (2002). Reform of doctoral dissertation, University of Michigan, Ann and by the system: A case study of a state’s effort Arbor. 47

Downloaded from http://eepa.aera.net at NORTHWESTERN UNIV LIBRARY on February 4, 2014 2791-02_Coburn.qxd 4/11/05 13:48 Page 48

Coburn

Hill, H. C. (2001). Policy is not enough: Language Ohio State University. (1998). The Early Literacy and the interpretation of state standards. American Learning Initiative at the Ohio State University. Educational Research Journal, 38, 289–318. Columbus: Ohio State University. Honig, M. I. (2004). The new middle management Piaget, J. (1978). The essential Piaget. London: intermediary organizations in education policy im- Routledge. plementation. Educational Evaluation and Policy Pollack, S. (1997). Three decades of service-learning in Analysis, 26, 65–87. higher education (1966–1996): The contested emer- Hornby, P., & Symon, G. (1994). Tracer studies. In gence of an organizational field. Unpublished doc- C. Cassell & G. Symon (Eds.), Qualitative meth- toral dissertation, Stanford University, Stanford, CA. ods in organizational research: A practical guide Popkewitz, T. S., Tabachnick, B. R., & Wehlage, G. (pp. 167–186). London: Sage. (1982). The myth of educational reform: A study of Jennings, N. E. (1996). Interpreting policy in real class- school responses to a program of change. Madison: rooms: Case studies of state reform and teachers’ University of Wisconsin Press. practice. New York: Teachers College Press. Pressman, J. L., & Wildavsky, A. (1984). Implemen- Kingdon, J. W. (1984). Agendas, alternatives, and tation. Berkeley: University of California Press. public policies. Boston: Little, Brown. Rigby Education. (1997). Early literacy inservice Laumann, E. O., & Knoke, D. (1987). The organiza- course (ELIC). Retrieved from http://www.rigby. tional state: Social choice in national policy domains. com/about/background.htm Madison: University of Wisconsin Press. Rowan, B. (2001). The ecology of school improvement: Literacy Initiative. (n.d.). The Literacy Initiative: San Notes on the school improvement industry in the Mateo/Santa Clara counties. Palo Alto, CA: Noyce United States. Philadelphia: Consortium of Policy Foundation. Research in Education. Massel, D., Fuhrman, S., Kirst, M., Odden, A., Scott, W. R. (1994). Conceptualizing organizational Wohlstetter, P., Carver, R., & Yee, G. (1994). Ten fields: Linking organizations and societal systems. years of state education reform, 1983–1993: Over- In U. Gerhardt, H. Derlein, & F. W. Scharpf (Eds.), view with four case studies. New Brunswick, NJ: Systemrationalitat und Partialinteresse [Systems ra- Consortium of Policy Research in Education. tionality and partial interests] (pp. 203–219). Baden- McGill-Franzen, A. (2000). Policy and instruction: Baden, Germany: Nomos Verlagsgesellschaft. What is the relationship? In M. Kamil, R. Barr, Scott, W. R. (2001). Institutions and organizations P. D. Pearson, & P. B. Mosenthal (Eds.), Handbook (2nd ed.). Thousand Oaks, CA: Sage. of reading research (Vol. 3, pp. 889–908). Mah- Scott, W. R., Mendel, P., & Pollack, S. (1996, March). wah, NJ: Erlbaum. Studying “everything else”: Assessing the environ- McLaughlin, M. W. (1987). Learning from experi- ence: Lessons from policy implementation. Educa- ments of an organizational field. Paper presented at tional Evaluation and Policy Analysis, 9, 171–178. the Conference on Institutional Analysis, Tucson, Meyer, J., Scott, W. R., & Deal T. E. (1981). Institu- AZ. tional and technical sources of organizational struc- Scott, W. R., Ruef, M., Mendel, P. J., & Caronna, C. A. ture: Explaining the structure of educational orga- (2000). Institutional change and healthcare organi- nizations. In H. D. Stein (Ed.), Organization and zations: From professional dominance to managed the human services (pp. 151–179). Philadelphia: care. Chicago: University of Chicago Press. Temple University Press. Smith, M. S. (2000). Balancing old and new: An expe- Miles, M. B., & Huberman, A. M. (1994). Qualitative rienced middle school teacher’s learning in the con- data analysis: An expanded sourcebook (2nd ed.). text of mathematics instructional reform. Elementary Thousand Oaks, CA: Sage. School Journal, 100, 351–375. Moore, M. (1988). What sort of ideas become public Smith, M. S., & O’Day, J. (1991). Systemic school re- ideas? In R. B. Reich (Ed.), The power of public ideas form. In S. Fuhrman & B. Malen (Eds.), The politics (pp. 55–83). Cambridge, MA: Ballinger. of curriculum and testing (pp. 233–267). New York: Neal, J., Kelly, P., Klein, A., & Schubert, B. (1997). Falmer Press. Touching the future: Executive summary, Califor- Spillane, J. P. (1996). School districts matter: Local nia Reading Recovery 1991–1997. Fresno: Central educational authorities and state instructional policy. California Reading Recovery Project. Educational Policy, 10, 63–87. Odden, A. (1991). Education policy implementation. Spillane, J. P. (1998). State policy and the non- Albany: State University of New York Press. monolithic nature of the local school district: Organi- Ogawa, R. T. (1994). The institutional sources of edu- zational and professional considerations. American cational reform: The case of school-based manage- Educational Research Journal, 35, 33–63. ment. American Educational Research Journal, 31, Spillane, J. P. (1999). External reform initiatives and 519–548. teachers’ efforts to reconstruct their practice: The 48

Downloaded from http://eepa.aera.net at NORTHWESTERN UNIV LIBRARY on February 4, 2014 2791-02_Coburn.qxd 4/11/05 13:48 Page 49

The Role of Nonsystem Actors

mediating role of teachers’ zones of enactment. Valencia, S. W., & Wixson, K. K. (2000). Policy- Journal of Curriculum Studies, 31, 1–33. oriented research on literacy standards and assess- Spillane, J. P. (2000). Cognition and policy implemen- ment. In M. Kamil, R. Barr, P. D. Pearson, & P. B. tation: District policymakers and the reform of math- Mosenthal (Eds.), Handbook of reading research ematics education. Cognition and Instruction, 18, (Vol. 3, pp. 909–935). Mahwah, NJ: Erlbaum. 141–179. Vaughan, D. (1996). The Challenger launch decision: Spillane, J. P., & Jennings, N. E. (1997). Aligned in- Risky technology, culture, and deviance at NASA. structional policy and ambitious pedagogy: Exploring Chicago: University of Chicago Press. instructional reform from the classroom perspective. Weick, K. E. (1995). Sensemaking in organizations. Teachers College Record, 98, 439–481. Thousand Oaks, CA: Sage. Spillane, J. P., Reiser, G., & Reimer, T. (2002). Policy implementation and cognition: Reframing and re- focusing implementation research. Review of Edu- Author cational Research, 72, 387–431. CYNTHIA E. COBURN is an Assistant Professor, Spradley, J. P. (1979). The ethnographic interview. Graduate School of Education, University of Cali- Fort Worth, TX: Harcourt Brace Jovanovich. fornia, 3643 Tolman Hall, Berkeley, CA 94720; ce- St. Mary’s College. (1998). Saint Mary’s College Read- [email protected]. At the time this article was ing Recovery statistics. Unpublished manuscript. written, she was an Assistant Professor in the School St. Mary’s College. (n.d.). Best practices in early liter- of Education and a Research Scientist at the Learning acy instruction: A long term staff development pro- gram for schools and school districts. Moraga, CA: Research and Development Center, both at the Uni- Author. versity of Pittsburgh. Her areas of specialization in- Strauss, A., & Corbin, J. (1990). Basics of qualitative re- clude the relationship between policy and practice, or- search: Grounded theory procedures and techniques. ganizational theory, urban schools, and qualitative Newbury Park, CA: Sage. research methods. Swartz, S. L., Shook, R. E., & Klein, A. F. (2000). California early literacy learning and extended lit- Manuscript received February 22, 2003 eracy learning. Redlands: Foundation for Califor- Revision received November 9, 2004 nia Early Literacy Learning. Accepted December 17, 2004

49

Downloaded from http://eepa.aera.net at NORTHWESTERN UNIV LIBRARY on February 4, 2014 2791-02_Coburn.qxd 4/11/05 13:48 Page 50

Appendix

TABLE A1 Demographic Profile of Schools Characteristic Stadele Elementary Baldwin Elementary Number of students 692 524 Student race/ethnicity (%) Asian 43.4 2.9 Latino 28.5 11.5 African American 12.9 82.8 Filipino 12.3 0.4 White 2.9 1.3 Pacific Islander 0.1 0.2 Percentage of students who qualify 67.0 88.9 for free/reduced-price lunch Percentage of students who are 47.8 14.3 English-language learners Number of teachers 32 27 Teacher race/ethnicity (%) White 56 36 Asian 24 7 African American 10 36 Filipino 5 4 Latino 5 14 American Indian 0 4 Ranking on California Academic Performance Index (1998–1999) All schools 6a 2a Similar school 7a 7a Note. Data were derived from Ed-Data (1998–1999; http://www.ed-data.k12.ca.us). a Out of a possible 10.

TABLE A2 Characteristics of Focal Teachers and Their Classrooms Classroom student Teacher School Age/race Years teaching Grade level composition Sharon Stadele Late 50s; White 34 First ELD 20 students 7 Chinese 4 Filipino 4 Latino 2 African American 1 Tongan 1 Vietnamese 1 White Marisa Stadele 33; biracial 4 Second ELD 20 students 9 Chinese 4 African American 3 Latino 2 Filipino 1 Arab 1 White Deanna Baldwin Mid-50s; White 13 (4 years in Second 20 students the early 18 African American 1970s and 9 years 1 Latino starting in 1990) 1 Arab Note. ELD = English language development. 50

Downloaded from http://eepa.aera.net at NORTHWESTERN UNIV LIBRARY on February 4, 2014 2791-02_Coburn.qxd 4/11/05 13:48 Page 51

TABLE A3 Documents Analyzed for State Policy Review Type of document Number analyzed State legislation 12 State policy documents and reports 34 State professional development materials 3 Task force reports 7 Materials from state and regional professional development providers 24 Media accounts (newspaper and magazine) 49 Secondary sources on history of reading 11

TABLE A4 Mechanisms by Which Focal Teachers Were Connected to Messages Source Mechanism Number Percentage of total System State policy 13 5.8 State standardized test 8 4.5 District policy 22 9.8 District professional development 27 12.1 District standardized test 3 1.3 School policy 11 4.9 School professional development 5 2.2 Principal expectations 4 1.8 Nonsystem University coursework (preservice) 8 3.6 University coursework (in-service) 10 4.5 Independent professional development provider 49 22.0 Curriculum materials 55 24.7 Colleagues 5 2.2 Media 3 1.3

51

Downloaded from http://eepa.aera.net at NORTHWESTERN UNIV LIBRARY on February 4, 2014 2791-02_Coburn.qxd 4/11/05 13:48 Page 52

TABLE A5 Definitions Used in Coding Factor Definition Intensity Degree to which teachers had opportunities to engage with a message in sustained, iterative ways Low Brief or fleeting interactions, including mentions in faculty meetings or professional development, casual conversations with colleagues, quick look through documents or materials, or participating in a single workshop or event Medium Interaction was sustained over a brief period of time or was intermittent over a long period of time, including professional development experiences that were in-depth but short term, curricular materials that teachers used for short periods of time in class, pressures for standardized tests recurring year after year but (in these schools) confined to a single season of the year, and interaction with a key carrier on an occasional basis High Interaction was sustained over a long period of time, including daily contact with curricular materials, intensive professional development that stretched over at least a year and involved opportunities for dialogue between ideas and experimentation in the classroom, interaction with a key carrier on a daily basis, or incorporation of messages from professional development into ongoing dialogue or planning with workplace colleagues Content: depth Degree to which a message addressed underlying pedagogical principles Low Policy message focused solely on surface structures (i.e., specific activity structures, materials, or classroom organization) without attention to underlying pedagogical principles (i.e., theories of pedagogy, models of the reading process, assumptions about how students learn) Medium Policy message focused primarily on surface structures but included some attention to underlying pedagogical principles High Policy message focused primarily on underlying pedagogical principles alongside attention to surface structures Content: closeness Degree to which policy messages came in forms that teachers found to be easily to classroom translatable into classroom practice Low Teachers reported that the message was difficult to use in the classroom, required significant translation or invention to use in the classroom, or had no connection to their classroom Medium Teachers reported that messages could be translated with moderate effort into classroom practice or that it required some degree of invention to create connections to their classroom High Teachers reported that messages were easy to use and able to be readily incorporated into classroom instruction with little translation or invention

52

Downloaded from http://eepa.aera.net at NORTHWESTERN UNIV LIBRARY on February 4, 2014