write a summarize (th)

Policy and Practice: The Dilemma Author(s): David K. Cohen, Susan L. Moffitt and Simona Goldin Source: American Journal of Education, Vol. 113, No. 4 (August 2007), pp. 515-548 Published by: The University of Chicago Press Stable URL: http://www.jstor.org/stable/10.1086/518487 Accessed: 05-05-2016 21:40 UTC

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at

http://about.jstor.org/terms

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted

digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about

JSTOR, please contact [email protected].

The University of Chicago Press is collaborating with JSTOR to digitize, preserve and extend access to American Journal of Education

This content downloaded from 192.251.139.3 on Thu, 05 May 2016 21:40:28 UTC All use subject to http://about.jstor.org/terms

AUGUST 2007 515

American Journal of Education 113 (August 2007) � 2007 by The University of Chicago. All rights reserved. 0195-6744/2007/11304-0001$10.00

Policy and Practice: The Dilemma

DAVID K. COHEN University of Michigan

SUSAN L. MOFFITT Harvard University

SIMONA GOLDIN University of Michigan

The relations between policy and practice embody a dilemma. Policies aim to solve problems, yet the key problem solvers are those who have the problem. Governments devise instruments to encourage implementation, but they help only if used well by those with the problem, whose capability is often weak. The realization of policy in practice thus depends on the fit between capabilities that support implementation and aims. The more aims outstrip capabilities, the less likely is effective implementation. When Title 1 of the 1965 Elementary and Secondary Education Act began, capabilities were modest: the program’s in- struments were limited, and neither practitioners nor workers in the environment knew much about school improvement. Yet Title 1 became a viable program in the 1970s because its aims were defined in ways that were achievable, given the capability that was possible. Title 1 later was made the chief lever for national standards-based reform policies, but capability was far from what would be required to eliminate the achievement gap, and the revised Title 1 did not contain means to vastly increase capability. Thus an effort to promote ambitious instruction led to reduced standards of success.

Policy, Practice, and Position

In 1950, the relations between policy and practice in education were not a puzzle of much interest to anyone. There was no federal education policy to puzzle about, and though states made education policy, it seemed not to puzzle anyone, at least not in print. States contributed to local school funds, but the main issues concerned amounts, not effects on practice. It was widely assumed that more money would buy more educational resources, which would create

Electronically published May 22, 2007

This content downloaded from 192.251.139.3 on Thu, 05 May 2016 21:40:28 UTC All use subject to http://about.jstor.org/terms

Policy and Practice

516 American Journal of Education

better education, which would enable more learning. So pervasive were these ideas that, though policy typically was thought to have substantial effects on practice, warrants for the claim were rarely investigated. Discussion centered on how much should be spent on what and for whom.

Money was not the only instrument of policy. States set standards for teacher certification and licensing and regulated school safety and curricula.1 Here, too, the effects on practice were assumed to be plain: setting certification standards would translate into better-educated teachers, which would enable more learning. Generous assumptions about the potency of policy fit the condition of government: the U.S. Office of Education did little but collect data, and most state education agencies were quite weak. If states needed only to appropriate funds and broadly regulate schools in order to influence prac- tice, weak government would do.

Since then, education policy has had five boom decades, but the more policy there has been, the more doubts have grown about its potency. There is more state and federal money, with more mandates and incentives and dozens of policies and programs, and recent policies embody much more ambitious efforts to change practice. Yet the relations between policy and practice are now much more controversial than they were in the 1950s, and there are more doubts about the efficacy of policy.

Research fueled the doubts and controversy. A bit more than a decade after Brown v. Board of Education, researchers showed that funding differences among U.S. schools and differences in the resources that money bought had modest and inconsistent effects on school-to-school differences in learning. Despite large differences in average achievement among schools and especially trou- bling differences between the children of affluent and poor parents, the ed- ucational resources that most people had thought significant were weakly related to school average student performance.

These studies provoked others. Some scholars moved from studying the impact of familiar resources like money, teachers’ degrees, and facilities to

DAVID K. COHEN is John Dewey Professor of Education and professor of public policy at the University of Michigan. SUSAN L. MOFFITT is a Robert Wood Johnson Scholar in health policy research at Harvard University. She received a PhD in political science from the University of Michigan in 2005. Her research considers the effects of government-sponsored knowledge cre- ation and distribution on policy implementation and regulation. SIMONA GOL- DIN holds a master’s degree in management and urban policy analysis from the New School University. She is currently a doctoral student at the University of Michigan in educational foundations and policy. Her research interests include the relationships between policy and practice and the role of students in the implementation of education policy.

This content downloaded from 192.251.139.3 on Thu, 05 May 2016 21:40:28 UTC All use subject to http://about.jstor.org/terms

Cohen, Moffitt, and Goldin

AUGUST 2007 517

focusing on instructional practices, organizational arrangements, and teachers’ knowledge and skill. These studies showed that practice-embedded knowledge and action affect learning. Teachers’ and students’ knowledge and academic practices, as well as their school organizations, mediate between the conven- tional resources that schools deploy, on the one hand, and learning, on the other hand (Cohen et al. 2003).

Those inquiries began to clarify how schools made a difference, and they implied a novel idea: the effects of resources depend on their use. Different schools and teachers did different things with the same resources, with different effects on learning. That seemed to imply that the classical instruments of education policy—funding and regulation—were not as uniformly potent as had been thought.2 The headline news in the early 1970s was that resources had relatively weak and inconsistent effects on practice. That shook inherited ideas about the power and effects of policy and helped to generate new ideas about what it might take for policy to influence practice.

The more fundamental point, which made no headlines, was that if the effects of resource allocation or regulation depend on practitioners’ use, then policy depends on practice. Resource provision is crucial, for students and teachers cannot use books or laboratories they do not have, but books and teaching do not unfold automatically in students’ minds. Resource provision creates opportunities for use, but resources are only effective if they are used well (Hanushek 1996). If that is so, then it follows that the quality of use will be influenced by practitioners’ understanding and by the circumstances that influence that understanding. It will then be reasonable to infer that policies will be more likely to influence practice if they cultivate practitioners’ use of policy and less likely to influence practice if they only allocate and regulate resources.

Studies of implementation in education began in response to Lyndon John- son’s Great Society. Several studies seemed to assume that efficacy resided in policy: well-designed policies would contain the resources, incentives, and oversight that were needed to shape practice; so most early studies focused on compliance. Studies of this sort tended to view things from the vantage point of policy—thus the appellation “top-down”—from which the main issues appeared to concern how control, compliance, and management influenced policy outcomes (Bardach 1977; Mazmanian and Sabatier 1989).3 These au- thors recognized that some discretion or adaptation was unavoidable, but they did not think that added value to the policy; implementers’ discretion would weaken policy. If policies could drive practice, study of implementation would be unnecessary: evidence of resource allocation and regulation would do.4

Doubts about these ideas arose in part from evidence that education policy did not have the expected effects. One case was the NAACP Legal Defense and Education Fund report, mentioned earlier, on state and local misuse of

This content downloaded from 192.251.139.3 on Thu, 05 May 2016 21:40:28 UTC All use subject to http://about.jstor.org/terms

Policy and Practice

518 American Journal of Education

Title 1 funds. Instead of spending the money on poor children, as the act required, some localities spent it on band uniforms and other things that were plainly prohibited (Martin and McClure 1969). Another case was the G. E. Tempo study of Title 1, for it reported that Title 1 had no average effect on learning (Mosbaek et al. 1968). The researchers found that, though some schools spent Title 1 money on reading, others spent it on music instruction or field trips; they concluded that since practitioners put the program to varied noninstructional uses, it could hardly improve achievement.

Both cases showed that practice shaped how policy turned out. Drawing on such research, Richard Elmore titled early ideas about the potency of policy a “noble lie” and argued that policy should be made by “mapping backward” from practice to rest policy design with the needs of practitioners (1979–80, 603).5 Paul Berman and Milbrey McLaughlin (1979, 21) took a similar view in the Rand Change Agent Study, in which they portrayed teachers’ motivation as a key to implementation: “Projects begun with broad-based support were not only more likely to have been implemented in a mutually adaptive way, but they also stood a better chance of attaining a stable con- tinuation.” The sense that implementers’ discretion was unavoidable was taken to mean that policy could not design practice; McLaughlin wrote “policy cannot mandate what matters” (1991a, 187). Practice was too complex and policy making was too remote for policy makers to prescribe for practitioners.

That nicely represented what came to be known, inelegantly, as the “bottom- up” perspective: practitioners’ will was their own, unreachable by policy yet crucial to its success. From the practitioners’ point of view, the key issues are not control and compliance but concern how the work that implementers do and the situations in which they do it influence their response to policy (Berman and McLaughlin 1979; Elmore 1979–80; Lin 2000; Lipsky 1980). Practitioners know things that policy makers could not, and they use the knowledge to modify policy. McLaughlin and other researchers saw teachers’ adaptation of policy as the result of professional judgment about policies’ utility: “one ironic consequence of this street-level policy making and professionals’ priority for client service is that reformers often may not get what they want, but get what they need as policy is put into practice” (1991a, 189).6 Other studies expanded on these ideas, most influentially, Michael Lipsky’s Street Level Bu- reaucracy, which argued that policy was effectively made in practice (1980). In less than a decade, inherited assumptions about the potency of policy had been reversed by ideas about the potency of practice, most of them arising in research on implementation.

In these two lines of work, policy and practice are often portrayed in conflict or opposition, as policy makers attempt to secure compliance from implementers who respond by attending to their situations and often ignore, evade, or attempt to buffer themselves from policy.7 In some cases, the argument touches on

This content downloaded from 192.251.139.3 on Thu, 05 May 2016 21:40:28 UTC All use subject to http://about.jstor.org/terms

Cohen, Moffitt, and Goldin

AUGUST 2007 519

democratic norms: some analysts view policy as the authoritative expression of representative government, while others see it as an imposition on those, close to the people, who do the real work of social service. Knowledge also is a key element in the debate, but it divides policy and practice. Street-level bureaucrats are portrayed as knowing things about their work, their clients, and their cir- cumstances that policy makers do not know but that are crucial in any effort to adapt policy to local practice. Policy makers are portrayed as knowing things about political goals, policy instruments, and the policy-making process that are crucial to policy formation, but these are things that practitioners are unlikely to know.

There is little evidence that policy makers noticed any of this. While many researchers argued that policy should be understood from the bottom up, state and federal governments increasingly tried to direct practice with policy from the top down. In the 1970s many states mandated basic skills instruction and minimum competency testing. The latter sought, for the first time in U.S. history, to regulate learning and teaching by requiring outcomes and attaching incentives to attain them. Government began to turn away from efforts to shape practice by allocating or regulating resources, moving toward efforts to shape practice by requiring outcomes. The state policies drew mixed reviews: researchers frequently portrayed them as efforts to “de-skill” teachers, but there is evidence that practice responded. National Assessment of Educational Progress (NAEP) scores for African American and Hispanic students improved between the mid-1970s and mid-1980s; one cannot draw clear causal con- nections, for other factors probably were partly responsible, but better NAEP scores were roughly what the state policies sought (Smith and O’Day 1990).

Policy makers took a bolder step in the mid-1980s, with the implementation of standards-based reform. Critics in and out of government complained that academic standards were generally weak, that students were not educated for the complexities of life and work in an “information age,” that more aca- demically demanding instruction was in order, and that educators should be held to account for what schools produced. Standards-based reform began in California and Kentucky in the mid- and late 1980s, and these policies sought to drive practice by building a new framework around schools. The frame’s key elements were content standards, tests aligned to standards, and account- ability to encourage schools to boost test scores. By 1993 more states had adopted versions of the policy, and more fell into line after President Clinton’s Improving America’s Schools Act (IASA) and President Bush’s No Child Left Behind (NCLB).

Researchers had a mixed view of state and federal efforts to shape practice. Some doubted the efficacy of outcome-oriented policies and criticized them: Audrey Amrein and David Berliner (2002) made such claims in a multistate study, and Walt Haney (2000) doubted the Texas policies in several reports.

This content downloaded from 192.251.139.3 on Thu, 05 May 2016 21:40:28 UTC All use subject to http://about.jstor.org/terms

Policy and Practice

520 American Journal of Education

Several other researchers reported constructive effects: Martin Carnoy and Susanna Loeb’s study of state accountability systems (2002) found that states with more stringent systems had stronger student NAEP performance, but a later report by Carnoy (2005) found no effects. David Grissmer and his col- leagues (2000) reported policy effects when comparing student scores in Texas, which had a relatively strong accountability system, to those in California, which had a relatively weak system. These arguments continue, but researchers have been more likely to find effects on student learning on state tests that have been keyed to state accountability schemes than on NAEP, which is national and more difficult and has not thus far been adjusted to any state accountability scheme.

The years in which standards-based reform swept the country also saw a novel school of thought take shape, one that focused on the role of cognition in the relations between policy and practice. If, as we wrote above, the effects of resources depend on their use, and if practitioners are the resource users, then the effects of resources depend partly on practitioners’ knowledge and skill and the circumstances that influence them. Several researchers elaborated on this idea. One of us published an article that focused on the role that one teacher’s knowledge played in her implementation of California’s ambitious new mathematics policy in the late 1980s (Cohen 1990). A few years later, the same author argued that practitioners’ knowledge mediated the relations between policy and practice and that the further educational policy departed from conventional practice, the more opportunities to learn practitioners would need if they were to implement effectively. Implementation was por- trayed as a species of learning and policy as a sort of instruction (Cohen and Barnes 1993a, 1993b).

A string of studies in the years that followed offered related ideas, probably due to the great knowledge demands that standards-based reform placed on many teachers and also to a revived interest in the social features of cognition. These studies portrayed practitioners’ knowledge and understanding as key features in their response to policy. Some called attention to the cognitive chains that mediated the relations between policy and practice (Cohen and Hill 2001; Spillane and Thompson 1997; Spillane and Zeuli 1999). Others focused on the role that communities of practice play in shaping practitioners’ understanding and thereby their apprehension of and response to policy (Coburn and Stein 2006). Another study focused on the role that language played in mediating the comprehension of policy and, through that, practitioners’ response (Hill 2006).8

James Spillane and his colleagues (2006, 63) captured several main threads in this line of work when they wrote that “reforms are complex ideas,” and that “artifacts used to communicate these reforms encode representations of a set of ideas about the reform. . . . Implementation [thus] involves cognition . . . [and] requires thoughtful use of these artifacts.” The relations between

This content downloaded from 192.251.139.3 on Thu, 05 May 2016 21:40:28 UTC All use subject to http://about.jstor.org/terms

Cohen, Moffitt, and Goldin

AUGUST 2007 521

policy and practice are here transformed into thought, and the key influences on implementation are whatever shapes thought. It would not be too much to say that, in these studies, the primacy of thought replaces that of political positions. That seems fitting at a time when policy places unprecedented demands on practitioners’ knowledge and skills.

Reconsidering the Relationship

Given this curious history, how might we understand the relations between policy and practice? In 1941, Carl J. Friedrich wrote: “Public policy, to put it flatly, is a continuous process, the formation of which is inseparable from its execution. Public policy is being formed as it is being executed, and it is likewise being executed as it is being formed” (356). Roughly three decades later, in 1979, Giandomenico Majone and Aaron Wildavsky wrote that policy “shapes implementation by defining the arena in which the process takes place, the identity and role of the principal actors, the range of permissible tools for action, and of course by supplying resources. The underlying theory provides not only the data, information, and hypotheses on which subsequent debate and action will rely, but also, and most importantly, a conceptualization of the policy problem” (reprinted in Pressman and Wildavsky [1984], 174). These ideas imply that policy and practice shape each other in continuing interaction and that they depend on each other. Policies and programs cannot specify practice but depend on practitioners to realize them in varied situations, while practice depends on policy to frame action and offer resources.9

The Dilemma

Our discussion begins from these ideas and from puzzles about the primacy of position that have absorbed researchers since the late 1960s. But where others see positional primacy as the key, we see a dilemma. Governments and private agencies identify problems and propose solutions, whether for welfare, drug use, or weak schools, yet the key problem solvers are the offending, needy, or damaged organizations and people. Governments can devise in- struments—incentives, ideas, money, leadership, rules, and more—to en- courage implementation. Governments typically have more power than the objects of social policy. These things can help but only if attended to and used well by those who are said to be the problem and by others who may assist. This dilemma can be difficult to manage even in seemingly simple situations, as when teachers try to help students improve weak compositions. Teachers may see students as the problem, but students are the key problem solvers.

This content downloaded from 192.251.139.3 on Thu, 05 May 2016 21:40:28 UTC All use subject to http://about.jstor.org/terms

Policy and Practice

522 American Journal of Education

Teachers can offer examples, assign exercises, coax, and make suggestions, but they depend on students to use the advice and examples to write better compositions. Teachers have more knowledge, power, and authority, but stu- dents’ will and ability are their own. Like other higher-ups, teachers depend on the people with problems to solve them, despite the imbalance in power and authority. If students will not do as teachers say, teachers must settle for what students will do.10

The policy makers who define problems and devise remedies are rarely the ultimate problem solvers. They depend on the very people and organizations that have or are the problem to solve it. At the same time, those that have or are the problem depend on policy makers or others for some of the re- sources—ideas, incentives, money, and more—that may enable a solution. As Ann Lin wrote (2000, 164), the incentives to reconcile policy and practice mean that policy is “a joint project.” This formulation calls attention to the tension between official authority and human agency, which can be a source of conflict, and to the potential for mutual dependence in the relations between policy making and practice, which can be a source of cooperation.

A key influence on the part that conflict and cooperation play in the relations between policy and practice is the risks that policies can cause by creating incompetence in practice. On the one hand, legislators and executives devise policies and programs as they try to lead society, solve social problems, main- tain political office, and reinforce the legitimacy of policy and agencies. The design of policies and programs mixes political, professional, and technical concerns. As policy makers define problems and solutions, their work also is shaped by ideology, local or regional interests, and the desire to advance their political interests and legitimate their actions. Yet, on the other hand, the policies and programs that they devise change service delivery, and that often implies change in the practices that deliver services or in the services that clients use or in how they use them or some combination of these. The sharper the departures from conventional practice that policy urges, the more radical are the entailed changes for practitioners and recipients, and the more practice is placed at risk of failure. For more ambitious policies require practitioners to acquire new capabilities and to unlearn present capabilities; such policies create even more incompetence among the practitioners whose incompetence already is implied in the policies’ novel aims (Bardach 1977). The more in- competence policies create, the more difficult it is to mobilize the resources that express mutual dependence and that can enable cooperation. The more that difficulty grows, the more likely is conflict.

Policies create incompetence for a variety of reasons. Policy makers may be ignorant of practice. Or they may believe it essential to set very ambitious goals to stir practitioners to act. Or they may wish to set goals that will cause practice to fail. At the same time, there are incentives for policy makers to

This content downloaded from 192.251.139.3 on Thu, 05 May 2016 21:40:28 UTC All use subject to http://about.jstor.org/terms

Cohen, Moffitt, and Goldin

AUGUST 2007 523

restrict the incompetence that they create in practice. For if practitioners cannot readily acquire new competencies, they are unlikely to deliver on policy. The less practitioners know about what to do, the more there is a risk of implementation failure or service delivery disruption or resistance by practi- tioners or clients or some combination of these. That can erode trust in practice, and because that problem can be traced to policy, that contagion could threaten the legitimacy of policies or programs and even the authors’ political interests. Such things play out in different ways, often in anticipated actions, reactions, and adjustments, before matters spill into public view. But, however they happen, the more policy puts practice at risk of failure, the more it can damage the legitimacy and political interests of practitioners or policy makers or both.

Politicians do not always aim for policy success. When some federal law- makers argued against school aid in the early 1960s, they worried about the threat to local control and about federal intervention on behalf of black Amer- icans in the South. In bargaining about the draft legislation, they wanted to sharply limit federal influence, and the resulting program design placed stiff constraints on what the central government could do by way of school im- provement. That limited what many other supporters of Title 1 hoped for, but the conservatives cared much more about restraining federal influence than improving practice. Policy designs reflect political considerations, and policy outcomes include politics as well as practice.

Yet, to the extent that policy makers do intend success for policy, they depend on practice for it. Practitioners, in turn, depend on policy makers and others in the environment for the resources to assist their work, including policy implementation. Practitioners have incentives to implement policy, for policy makers exercise authority and influence resources. Policy makers often have incentives to adjust to practitioners’ capabilities, because failure in prac- tice can affect policy makers’ interests and even the legitimacy of their en- terprise. Policy and practice contain opportunities for both cooperation and conflict. Although mutual dependence, as Majone and Wildavsky implied, is a stable feature of the relations between policy and practice, it does not lead inexorably to cooperation. Policy and practice can encourage either conflict or the trust, collegiality, cooperation, and communication that can enable effective implementation; which it is depends in good part on the distance that policy creates from practice and also in part on the resources that are deployed—or not—to bridge that distance.

No Child Left Behind is a case in point. It requires the elimination of racial and social class differences in student achievement in a dozen years. That implies dramatic change in practice and sets up the possibility of conflict between policy and practice. But the probability of conflict depends on several other factors, including whether the policy deploys instruments that enable

This content downloaded from 192.251.139.3 on Thu, 05 May 2016 21:40:28 UTC All use subject to http://about.jstor.org/terms

Policy and Practice

524 American Journal of Education

change in practice, what resources are available in the environment, and practitioners’ capabilities. The more abundant are these instruments, re- sources, and capabilities, the more policy and practice express mutual de- pendence, and the more likely is cooperation. Yet, as things stand, there is little evidence that most practitioners have the capability to make such change largely on their own, little evidence that NCLB offers the instruments that would enable such change, and abundant evidence that resources in the en- vironment are modest. Situations of this sort, in which programs’ aims are much more ambitious than the resources available to implement them, are likely to trigger conflict. In this case, as in many others, knowledge can play an important role. If policy makers had known more about practice and what it would take to improve it, policy might have been designed differently; if practitioners were more skilled and knowledgeable, they might have responded to the policy more effectively. The lack of knowledge on both counts has encouraged continuing conflict between Washington, DC, and states and localities.

Positional analyses tend to distract attention from the mutual dependence that ties policy and practice, the distance that can separate policy goals from practitioners’ competence, the risks for practice and policy that incompetence can create, and whether resources that help bridge the distance and enable cooperation are present or absent and encourage conflict. In addition, “top” and “bottom” are, as an empirical matter, rubbery categories. For example, Title 1 of the 1965 ESEA was a federal program, so one might assume that viewing its implementation from the top would begin and end in Washington, DC. But this program devolves extensive authority for program decisions to the states; they allocate funds, screen local applications, assist local programs, and evaluate them. The federal government is the “top” agency for states, and they are the federal agency’s “bottom” of first resort. But those agencies in Albany and Sacramento also are the “top” agencies for local education authorities (LEAs), even as the school principals’ “top” in those LEAs is the central office, and the teachers’ top is the principal. In modern societies, in which political organization grows ever more dense and elaborate, one agency’s top is more and more likely to be another’s bottom.

As a result, the problems of control and compliance that researchers identify with some policy-making tops also are problems for organizations and indi- viduals that researchers identify with the bottoms. The problems of control and compliance that state Title 1 offices have, vis-à-vis LEA central offices, closely resemble the problems of control and compliance that LEA central offices have with schools. The LEA central office workers who deal with problems of Title 1 control and compliance vis-à-vis principals, teachers, and students also must manage their work as “street-level” bureaucrats vis-à-vis central and state agencies.

This content downloaded from 192.251.139.3 on Thu, 05 May 2016 21:40:28 UTC All use subject to http://about.jstor.org/terms

Cohen, Moffitt, and Goldin

AUGUST 2007 525

In such cases, many public managers are both central controllers and street- level bureaucrats, trying to influence those for whom they make policy while trying to negotiate with those who make policy for them.11 In such cases, the roles of “policy maker” and “practitioner” converge in the same person and office. The dilemma of which we write occurs in many locations, and many individuals and agencies experience it from both sides at once. They make and implement different parts of a complex policy task, embodying elements of dependence and the specialized knowledge that their roles entail.

There are some limiting cases: the central government is subordinate to no higher domestic agency, though it depends on many, and students exercise no positional authority over others on whom they depend for services, though they can have enormous influence on the teachers and managers who exercise authority over them. But nearly all others who relate policy and practice are at neither extreme: they try to influence those who depend on them for resources and guidance while at the same time attempting to respond to those who seek to influence them by offering guidance and resources. In these cases, the key to the relations between policy and practice is not positions in hier- archies but mutual dependence and the cooperation or conflict that it en- genders.12 Whether it is cooperation or conflict depends less on positional authority than on the distance between policy and practice and on the actions and resources that help to narrow or increase that distance.

Managing the Dilemma

We group those actions and resources into four factors. One is the aims that policies set: some are modest, while others are ambitious; some are relatively simple, while others are ambiguous. The more ambitious aims are and the more they depart from conventional practice, the greater the resources that will be needed to realize them in practice. The more ambiguous these aims are, the more difficult it will be for policy makers to craft suitable instruments and for practitioners to figure out what to do.

Three sorts of resources bear on the relations between policy and practice. One comprises the instruments that policy makers and managers deploy to support change in practice; this is our second factor. Instruments are the resources that policy makers bring to relations with practice; they include money, mandates for action, incentives to comply, flexibility to adapt policy to local conditions, and ideas that inspire or inform implementers’ under- standing and actions or contain systematic evidence about implementation and effects, among other things.13 These and other instruments are socially created tools. They are intended to encourage assent to policy or to help realize aims in practice, or both, but there is no science of implementation,

This content downloaded from 192.251.139.3 on Thu, 05 May 2016 21:40:28 UTC All use subject to http://about.jstor.org/terms

Policy and Practice

526 American Journal of Education

and there often are deep differences of opinion about how best to achieve a given goal or whether it is worth achieving. As a result, instruments sometimes have been designed as much to satisfy the interests and ideas of legislative factions or interest groups as to increase effective implementation (Scholz 1991).14

Another set of resources is commonly referred to as capacity, though we prefer the term capability.15 This is our third factor, and it refers to the resources that practitioners and others in the environment bring to policy. Those re- sources include interests, ideas (including knowledge, values, and skill), personal resources such as will, and such other things as money. These elements of capability mirror several of those that characterize policy instruments; that is no accident, for when instruments are effective, it is partly because they help to mobilize the capabilities that enable practitioners or organizations to solve problems.

A last set of resources is found in the environments of policy and practice, which is our fourth factor. Environments play a variety of roles: policies are designed in them, instruments and capabilities are formed in and operate through them, and practitioners work in them. The term refers to two things. One is the broad circumstances in which schools and education policy subsist. For example, one prominent feature of the educational environment in the United States is the central role that schools play in social policy. Several European nations directly address social inequality with health care, education, incomes, and tax policies, but the United States restrained social policy to modest welfare, health, unemployment insurance, and old-age pensions (Katz- nelson and Weir 1985). Public schools were thought to guarantee equal op- portunity and reduce the need for other interventions (Katznelson and Weir 1985; Weir et al. 1988). Another key feature of the environment understood this way is that economic liberalism and skepticism about government kept schools and governments weak. States delegated central functions such as test construction, curriculum making, and text writing to private firms, effectively outsourcing technical and professional capabilities that most other school sys- tems retained in-house. A third feature of the environment understood in this sense is that the combination of state and local control with growing residential segregation by race and class created enormous economic and social inequality among local and state school systems.

The result is that schooling was central to social policy, on the one hand, but was offered modestly and very unequally by weak state and federal gov- ernments, on the other hand. Schooling is central to the ideas that have guided social policy in the United States, but it has had a limited capability to deliver on those ideas. It has had a paradoxical role in U.S. social policy: strong in principle, popular belief, and policy but relatively weak in practice. Access to schools was seen as the key to opportunity, and the United States opened

This content downloaded from 192.251.139.3 on Thu, 05 May 2016 21:40:28 UTC All use subject to http://about.jstor.org/terms

Cohen, Moffitt, and Goldin

AUGUST 2007 527

access earlier and more broadly than most other nations; yet schools were modestly and quite unequally supported. There was less to access than met the eye, since schools’ capabilities to equalize opportunity were very uneven— and incommensurate with the rhetoric of policy and popular belief.

Environment also refers to the circumstances in which specific policies and programs subsist. In this sense, the environment is a variable subset of extant social organization, political structures, economic arrangements, fiscal prac- tices, and culture. The boundaries of any particular subset, and thus the salient environment, are determined by the actions under consideration. For example, teacher education is part of the larger environment of education in the United States, but it is not part of the environment that is salient to a policy that changes Title 1 fund allocation.

These environments enable and constrain policy. If we consider the first meaning of the term, the environment affects policy and practice in a variety of ways. It can, for example, influence educational capability through teacher education: most U.S. elementary school teachers learn little about the subjects that they will teach, owing to the very limited professional education that higher education and state government have supported. That limits the ca- pabilities that teachers bring to academically ambitious curricula or to de- manding academic standards, and that limits implementation. It also can shape policy makers’ and practitioners’ ability to communicate and understand what is at stake between them by shaping their familiarity with each other’s work. Centralized systems reduce the fragmentation that is endemic in the United States and increases the chance of understanding between policy makers and practitioners. Size also matters: communication between policy makers and practitioners would be less difficult in Rhode Island than California. Size and fragmentation increase the costs of communication, while coherence and com- pactness reduce them. Less understanding between policy makers and prac- titioners reduces the chances that they will be well informed about each other’s work, thereby reducing the opportunities to see common interests and inter- dependence. That increases the chances of conflict.

If we consider the second meaning of the term, environments regularly affect the design, content, and implementation of policies. Goals 2000 and IASA were written in extensive consultation with chief state school officers and several other education groups. One result was that the legislation had broad support from these groups, but another was that the bills delegated most key decisions about implementation to state and local education agencies. The education groups’ support was probably essential to legislative success, but one cost of that support was a policy that they found acceptable. Private organizations can enable or constrain relations between policy and practice, or, quite often, they do both at once, as they influence policy formation and continue to contest the ends and means of policy during implementation.

This content downloaded from 192.251.139.3 on Thu, 05 May 2016 21:40:28 UTC All use subject to http://about.jstor.org/terms

Policy and Practice

528 American Journal of Education

Environments of both sorts influence policy and practice as they act on aims, instruments, and capabilities. Hence, we discuss them below in con- nection with those influences.

Policy Aims: Ambition and Ambiguity

Other things being equal, the more ambitious the change in practice that policies seek, the more difficult it is to manage the dilemma. Some policies have no ambitious aims, for they seek only modest extension of current prac- tices: they add five minutes a day to the 10 already spent on composition, they include a new wetland in existing shoreline preservation, or they increase the income threshold for Medicaid eligibility. Because they call for only modest change, they create little incompetence in practice. Other policies call for change that has little technical content; they shift school schedules or require patients to communicate with nurses rather than doctors. Their difficulty varies with the extent to which the changes depart from current practice or encounter resistance. Still other policies have much more ambitious aims for practice, like replacing computation with conceptual mathematics or adding new en- vironmental regulation. Because they depart sharply from current practice, policies of this sort create more incompetence in practice and require more capability to enable the revision of practice. That makes implementation more difficult (Spillane 2000).

The difficulty of implementing ambitious policies is eased or compounded by their relative clarity or ambiguity. A policy that calls for more conceptual work in mathematics could do so by briefly stating that objective, contrasting it with computational work, and justifying the change with reference to the nation’s need for more able scientists and engineers. Alternatively, such a policy could reduce ambiguity by spelling out the mathematical content that should be covered in detail and by offering examples of conceptual work in that content. Such a policy would still be difficult to implement, but that difficulty could not be attributed to ambiguous aims.

In some cases, policies are ambiguous because the implementing organi- zations are quite varied, so that very specific aims would be unworkable. In such cases general aims are associated with local discretion, because the ca- pability that is required for implementation lies with practice. Some other policies are ambiguous because they envision states of affairs so novel that they can be imagined but not specified in much detail. The capability that is required for implementation lies nowhere, and implementation depends on its invention.

In many cases, ambiguous aims are due to coalitions using vague phrases to paper over differences. Majone and Wildavsky (1984, 168–69) write that

This content downloaded from 192.251.139.3 on Thu, 05 May 2016 21:40:28 UTC All use subject to http://about.jstor.org/terms

Cohen, Moffitt, and Goldin

AUGUST 2007 529

“in most policies of interest, objectives are characteristically . . . vague (be- cause that is how we can agree to proceed without having agreed also on exactly what to do).” Federal education policy often has been ambiguous so as to manage both state and local variability and differences within coalitions. Yet such ambiguity affects relations with practice, for “the more general an idea and the more adaptable it is to a range of circumstances, the more likely it is to be realized in some form, but the less likely it is to emerge as intended in practice” (Majone and Wildavsky 1984, 176). The ambiguity that attracts policy makers makes it difficult for practitioners to know what is expected of them (Brodkin 1990). Although ambiguity can enable policy, it inhibits the capability that practitioners need if they are to respond in ways that are consistent with policy aims.

Ambiguity of this sort has marked federal education policy since the 1960s. Reformers sought policies that would solve educational and social problems, while conservatives and many educators resisted change. The aims of Title 1 of the 1965 Elementary and Secondary Education Act (ESEA) were so am- biguous that it could be seen as both a radical departure and a modest ini- tiative. On the radical side, it aimed to improve schools for poor children, and that seemed to imply more federal influence on states and schools. Federal officials tried to encourage school improvement, to advise states and localities, and to evaluate their progress.16 Yet everyone was poorly informed; there was little knowledge of how to improve schools, teaching and learning were weakly understood, and state and federal intervention had been rare. Practitioners at all levels of the federal system faced large new tasks with small capabilities.

These problems were multiplied by the structure of Title 1. Its formula grant came close to being general aid, for it was designed in part to appease opponents of federal school aid by restraining central influence. The federal government had to decide how much to spend, to apportion it, and to mail checks, while localities had only to spend the money on a broad category of school resources for a particular purpose and to fill out applications to that effect. These things were more familiar than school improvement. Congress knew how to appropriate funds, and educators knew how to spend on schools. Distributing funds was a bit novel for the U.S. Office of Education and required learning, yet it was familiar federal work. It pressed everyone a bit beyond extant skill and knowledge but not so far as to make it very difficult to manage.

The implementation of Title 1 has been an exercise in managing ambiguity about its aims.17 For three decades there were cautious federal efforts to im- prove schools with something close to general aid, and some modest improve- ments were made. Then Presidents Clinton and Bush made Title 1 the core of federal efforts to reshape state and local practice with standards-based reform. The aims of Title 1 changed: ambiguous pronouncements about im-

This content downloaded from 192.251.139.3 on Thu, 05 May 2016 21:40:28 UTC All use subject to http://about.jstor.org/terms

Policy and Practice

530 American Journal of Education

proved schools for poor children were replaced with what seemed more specific requirements for students’ academic progress. The focusing of IASA and NCLB on school outcomes seemed to sweep ambiguity out of federal education policy.

Yet great ambiguity remained. One source is familiar from the origins of Title 1: many important decisions about the actual content of the new aims were delegated to states and localities. Although IASA and NCLB did contain a design for unprecedented federal influence, states and localities still were the operating agencies, and they had great political and educational influence. States have most political authority in public education, and that was displayed in 1965 in Title 1 aims that mixed centrally inspired school improvement and support for state and local control. The former was expressed in the purposes of the legislation and the targeting of funds, while the latter was expressed most plainly in the delegation of most operating authority to states and lo- calities. In 1994 and 2001, the ambiguity was displayed in aims that mixed much more ambitious and centrally inspired school improvement with con- tinued delegation of most operating authority and many major decisions to states. For example, the criterion by which schools are judged to have achieved “proficiency” varies among states because the legislation permits states to choose their own tests and set their own proficiency standards. What appears to be unprecedented clarity in the new Title 1’s aims dissolves into a blur as states decide (Olson 2002; Saulny 2005).

A second source of ambiguity in the new Title 1 arises, ironically, from the apparently more precise aims. For, as policy pressed policy further into prac- tice, uncertainty increased because improved learning is very complex and weakly understood. Some sources of this ambiguity are little different today from what they were in 1965: the uncertain validity of tests to measure students’ achievement was known then, but it was not salient to policy because Title 1 did not aim explicitly at tests or use them for accountability. Doubts about the use of tests in Title 1 only surfaced intermittently, especially when every- five-years evaluations sought to estimate the effects of Title 1 and Congress debated the program. By 2003, however, tests were in the foreground because they had become the way to judge accomplishment of Title 1’s central aim. All of the ambiguities that some researchers had known about for decades became highly salient to policy and blew into public debate. Knowledge about tests and their use, which had not been a serious problem for policy or practice, suddenly became a central issue for both as disputes erupted about whether tests can measure learning with sufficient validity to decide students’ profi- ciency or schools’ progress.

Other sources of ambiguity for the new Title 1 are more novel. Teaching, learning, and school improvement are not well enough understood to specify the most effective ways to teach and learn multiplication or composition to

This content downloaded from 192.251.139.3 on Thu, 05 May 2016 21:40:28 UTC All use subject to http://about.jstor.org/terms

Cohen, Moffitt, and Goldin

AUGUST 2007 531

a great range of students in a great range of circumstances. These uncertainties were familiar to many researchers and educators, but they had not been salient to policy because no one had set out explicitly to improve teaching and learning and no policy had made school improvement the key to its success. When standards-based reform moved aggressively into the weakly understood and contested domains of instruction and its outcomes, it opened up large domains of ambiguity, both about how to improve teaching and learning and about how to improve weakly performing schools. The seemingly more precise aims of these policies masked enormous uncertainty and ambiguity, precisely be- cause so little was understood about how to improve teaching, learning, and weak schools. Whether policy aims are ambiguous or not depends at least as much on knowledge of how the aims might be achieved as it does on how clearly they are stated.

The expanded ambiguity might have been less of a problem if pressure for improved test scores and schools had been modest, but standards-based re- forms coupled ambiguous aims to growing pressure for results. That turned the lack of knowledge about teaching, learning, and school improvement into a major problem for policy and practice. As Herbert Simon and his colleagues observed nearly half a century ago, “If the people to be regulated do not know what they are supposed to do, they cannot do it” (Simon et al. 1950, 415).

The further policy aims depart from conventional practice, the more acute the dilemma becomes, for such aims reduce practitioners’ capability and in- crease the risk of practice failure. The key mechanism is the capability deficits in practice—the incompetence—that ambitious aims can create. Ambiguous aims also can influence the dilemma, for ambiguous aims make it difficult for practitioners to know what policy proposes or how to do it. The recent changes to Title 1 combine ambition and ambiguity. Ambiguity is often essential to policy formation, but it can be crippling for practice. Whatever the aims of policy, though, the means to achieve them lie elsewhere, that is, in practitioners’ capabilities, in the instruments that policies deploy, and in the environments of policy and practice.

Instruments: Resources That Policy Brings to Practice

Policies create a gap with practice by moving at least a bit beyond it; imple- mentation depends on closing that gap. That depends in part on the instru- ments that policies deploy. They offer funds, mandate or forbid actions, create incentives to comply, offer flexibility to adapt to local conditions, and deploy ideas to inform practice. These and other instruments are socially created tools. Because they are created or deployed politically, they reflect the views and interests of those who devise policy. Although policy instruments often

This content downloaded from 192.251.139.3 on Thu, 05 May 2016 21:40:28 UTC All use subject to http://about.jstor.org/terms

Policy and Practice

532 American Journal of Education

are crafted to encourage assent to policy and to deploy resources to change practice, they also have been devised so as to diminish or constrain policy.

Instruments vary in strength, or their influence in practice, and in salience, or how closely they connect with what must happen in practice to achieve policy aims. Strong but not salient instruments are likely to be less effective, for they create capability problems for practitioners by telling them to do something but leaving them to figure out what and how. Potent incentives to achieve aims that are far beyond most practitioners’ capabilities are a case in point. Weak instruments that are salient for practice also are unlikely to be effective, for they offer help that bears on practice but exert little influence because they offer little or no encouragement to use the instruments. Technical assistance to produce performances that implementers had no incentive to produce is a case in point (Fuhrman and Elmore 1992, 1995). Instruments that are strong and salient are more likely to help, but they are more difficult to craft, in part because they typically require more knowledge in both policy design and practitioners’ use. That problem grows as policy departs further from conventional practice.

Strength and salience are not inherent in instruments but vary with the ambitions and ambiguities of aims, the extent of capabilities, and the resources available in environments. Instruments that are strong and salient for policies that depart only a bit from conventional practice are not equally strong and salient for policies that depart much further. For the further policies depart from conventional practice, the more incompetence they create and thus the more capability they require of practitioners. To illustrate, consider the in- struments that four federal programs deploy. Impact Aid helps schools that serve families at federal installations, and three instruments suffice: money, flexibility to spend it, and occasional oversight with audits. Government com- pensates for a burden imposed, but it seeks no change in behavior and creates no incompetence. Flexibility was essential, noncontroversial, and feasible, and local authorities needed little more knowledge to make effective use of the instruments.

Flexibility delegates responsibility to practitioners, either because policy makers want to avoid conflict, because some interested parties want to weaken policy, or because circumstances require local adjustment. In principle, flex- ibility promises to encourage useful adaptation and to mobilize support for policy. The assumption typically is that the knowledge required for effective implementation is local and variable and that it cannot be mobilized at the center in which policy is made; flexibility thus is partly a means to bring knowledge for practice to bear. Flexibility can help to manage the dilemma by delegating agency for policy to practice, inviting practitioners to engage policy on terms that suit their values, interests, and knowledge.

Like other instruments, however, the effects of flexibility depend on how

This content downloaded from 192.251.139.3 on Thu, 05 May 2016 21:40:28 UTC All use subject to http://about.jstor.org/terms

Cohen, Moffitt, and Goldin

AUGUST 2007 533

practitioners use it. That, in turn, depends on how other policy instruments and practitioners’ capabilities relate to policy aims. The 1994 reauthorization of Title 1 (IASA) set very ambitious new aims, including greatly improved academic work for disadvantaged students. It required states to devise and use strong academic standards and tests that were keyed to the standards and to make schools accountable for students’ performance. At the same time, the legislation gave the states great flexibility to devise standards and tests and to set the student performance criteria for which they would be accountable. That flexibility did not produce deep engagement with the policy: a few states continued with reforms they already had under way, but most implemented Title 1 superficially. State and local capabilities to devise standards and improve instruction were far from commensurate with IASA’s ambitious aims, and flexibility was not coupled to other instruments, such as help to improve practitioners’ capability or incentives to set strong and usable standards. The Clinton administration faced a hostile House of Representatives and opposition in many states, and top officials did not think they could push states to take a more aggressive approach. They expected political and legislative losses elsewhere if they pressed harder.18 Flexibility combined with weak instruments, modest capability, and political conflict to promote weak responses in practice.19

The salience and strength of policy instruments vary not only with other instruments and capability but also with policy aims and environments. Title 6 of the 1964 Civil Rights Act (CRA) sought to change behavior on America’s most painful social problem by withholding federal funds from agencies that discriminated on the basis of race. This instrument was salient because money is essential to schools’ operation; it also was relatively strong, for once Title 1 began, federal funds became a part of school budgets: jobs and operations depended on them. But federal officials used Title 6 cautiously, for they “soon learned that the withholding of funds is a two-edged sword. The recalcitrant state agency is punished, but the Federal program is also wounded through loss of vital local support, political attacks, and general disruption of coop- erative relationships needed for the successful operation of other programs” (Orfield 1969, 7). The threat of fund loss was salient, but it was only inter- mittently strong because of federal dependence on practitioners for success and political support. That limited how far federal officials would push prac- tice, which resistant southern officials quickly sensed and used to their ad- vantage. Implementation continued the political struggles that had marked civil rights and legislative debates.

The salience and strength of policy instruments also vary with the actions that policy aims entail. Title 6 was relatively effective for such acts as denying African American children access to “white” schools, but it was less so for segregated classes or teachers’ treatment of students. For the latter had some-

This content downloaded from 192.251.139.3 on Thu, 05 May 2016 21:40:28 UTC All use subject to http://about.jstor.org/terms

Policy and Practice

534 American Journal of Education

thing to do with teachers’ beliefs and attitudes, and with common, sometimes nonracial, ability grouping. Fund loss was not as salient to teachers’ beliefs and attitudes as to system budgets. In addition, withholding funds was blunt: it was relatively strong in dealing with problems that infected entire jurisdic- tions, in part because the knowledge required to apply and respond to it was neither complex nor difficult to acquire. Yet that same instrument was less salient when dealing with complex problems like racially imbalanced class- rooms, in good measure because the knowledge required to apply and respond to the instrument was both complex and difficult to acquire. A tool that was salient and strong for one situation was not for another that seemed similar, in part because the knowledge demands of the two situations were so different.

The more ambitious or ambiguous a policy’s aims, the more difficult it is to devise salient and strong instruments. The curriculum reforms of the late 1950s and early 1960s urged ambitious change yet offered little assistance to practice, apart from limited summer workshops. Some teachers could apply and take the time, but there was room for only a few of those who needed help. Most teachers had no opportunity to learn about how students might respond to the materials, how teachers might learn how to use them, and how they might teach. If one reason was the lack of many workshops, another was the lack of suitable instruments to enable teachers’ learning. For example, there was no curriculum for teachers that built on the student curriculum to help teachers learn how to use the student curriculum well. If it had existed, it might have included examples of students’ work, analyses of the thinking involved, and how accomplished teachers used the materials and students’ work to improve practice. It would have been difficult to create such an instrument for considerable expertise in teaching and learning would have been needed, along with professional and political influence to gain support and encourage use. No government or private agencies tried to do such work, and those who built student curricula could not imagine what to do; they wrote the materials and let teachers cope.20 That sharpened the dilemma for practitioners, inviting them to produce performances that were far from cur- rent practice, yet provided little help and few incentives.

Here again, knowledge was crucial, but in this case it was the lack of knowledge to craft effective instruments. To do that policy makers need knowl- edge of practice and the likely needs of practitioners and knowledge of what instruments are likely to be effective. We noted earlier that Herbert Simon and his colleagues had written: “If the people to be regulated do not know what they are supposed to do, they cannot do it.” They went on to write that in such cases “agencies must engage in great amounts of public educational activities” (Simon et al. 1950, 415). True enough, but devising effective ed- ucational activities requires knowledge that both policy makers and practi- tioners are likely to lack in the case of innovative policies.

This content downloaded from 192.251.139.3 on Thu, 05 May 2016 21:40:28 UTC All use subject to http://about.jstor.org/terms

Cohen, Moffitt, and Goldin

AUGUST 2007 535

The strength and salience of policy instruments are relative for they depend on how capable practitioners are, on policy aims, and on the environments of policy and practice. Hence, the alignment of these things bears directly on instruments’ likely effectiveness. Yet that alignment can be difficult to achieve. The instruments that IASA and NCLB deploy seem strong, historically con- sidered: clear academic standards, tests aligned with the standards, and schools’ accountability for students’ performance all go far beyond Title 1 in 1965, or 1985 for that matter. The key instrument in 1965 was money, which was targeted to education for poor children in particular schools. In its time that was a sharp departure, but the political influence of conservatives in the federal legislature meant that the formula grant was a virtual entitlement; once the formula was written and funds appropriated, state and local grants were set. States and LEAs had to write proposals and offer assurances, but their grants depended less on the proposals’ merits than on the incidence of poverty and a legislative intent that nearly all districts should receive funds. There were few incentives to change practice and little help. The formula grant was aligned with a legislative intent to offer something like general aid but not with the intent to improve schools. Funds alone were unlikely to be strong or salient for that purpose, since the effects of resources depend on their use.

The NCLB’s standards, tests, and accountability appear much stronger than Title 1’s 1965 instruments, but their effects depend no less on practitioners’ use. That is a problem for NCLB because its instruments are not very salient to the practices at which they aim. Teaching and learning must change if Title 1 is to succeed, yet the NCLB’s instruments do not reach these practices very well. Academic standards can be a useful frame for practice, but few teachers can turn them into lessons that advance learning. They can exemplify policy goals, but not in the detail or with the salience to instruction that would enable most teachers and students to use them effectively. That makes the dilemma more acute and decreases the likelihood that managers and prac- titioners will respond in ways that have educational value. Many states adopted tests of basic skills to create relatively easy outcome measures that would enable at least the appearance of success. In these cases the tests may be used as a sort of curriculum, but it is likely to leave many students with narrow and superficial knowledge.

Incentives also are essential to learning, but these are more likely to be effective if they are salient to instruction. The incentives attached to account- ability in standards-based reform, however, do not operate in instruction; they operate on states, and, through them, on districts, and through them, on schools, and through them, on teachers. There are no direct incentives for teachers, and none for students, even though they are at least as important to outcomes as teachers. The target of these accountability incentives is not instruction but students’ test scores, and the incentives tend to operate outside

This content downloaded from 192.251.139.3 on Thu, 05 May 2016 21:40:28 UTC All use subject to http://about.jstor.org/terms

Policy and Practice

536 American Journal of Education

instruction. They affect teachers through many layers of government; if they operate on students at all, they do so through teachers and schools. These features reduce the incentives’ salience for instruction, and that reduces the probability that they will greatly improve instruction. In such multilevel ac- countability, it would be impossible to attach incentives for performance to a particular instructional design unless the design was a part of the accountability scheme itself. That might have made more sense educationally, but it would have been politically impossible and technically quite difficult, at least on the legislation’s rapid compliance schedule. Some difficulties of such an initiative have been on display in the Reading First program, which attempts to marry accountability for schools and systems to a particular instructional design.

The NCLB’s strong but not salient instruments sharpen the dilemma, and they are likely to encourage responses that do not help to improve instruction. Teachers are likely to reduce the risk of failure by boosting outcomes without improving learning, by prepping students for tests, helping them to cheat, altering scores, or inflating average scores by keeping low-performing students away. Such responses are to be expected if policies set implausible aims or if they set plausible aims but fail to deploy instruments that are strong and salient enough to support their achievement or if instruments are strong but not salient or strong and salient but operate on incapable practitioners. In such cases, teachers can be expected to reduce risk rather than improve perfor- mance because they lack the knowledge required to improve performance, and instruments do not provide it. Under these circumstances, if policies attach consequences to performance, they increase the chance that practitioners will try to reduce risk by complying without improving practice.21 One should expect practitioners to act roughly rationally, given any set of aims and in- struments, and it is rational to reduce risk if one is required to do things that one cannot, especially if one may be penalized for failing to perform.

Instruments are the capability that policy brings to its relations with practice. They vary in strength, or their influence in practice, and in salience, or how closely they connect with what must happen in practice to achieve policy aims. The further those aims depart from conventional practice, the more acute the dilemma becomes and the more likely it is that strong and salient instruments will be required to enable practitioners to respond constructively. Knowledge is a key element in such policy instruments. One reason it is important is that policy instruments often either convey knowledge directly or indirectly enable practitioners to acquire knowledge themselves. Another reason knowledge is crucial is that policy makers require it in order to devise instruments that can enable practitioners to respond constructively.

What practitioners can and cannot do, however, is not a simple function of policy aims and instruments. It also depends on their capabilities.

This content downloaded from 192.251.139.3 on Thu, 05 May 2016 21:40:28 UTC All use subject to http://about.jstor.org/terms

Cohen, Moffitt, and Goldin

AUGUST 2007 537

Capability: Resources That Practice Brings to Policy

Capability is essential to implementation because it consists of the resources that implementers bring to policy. Yet capability is relative: what is sufficient for a policy that departs only a bit from conventional practice is unlikely to be sufficient for a policy that departs much more dramatically. Capability varies not only with policy aims but also with instruments, environments, and prevailing practice.

Capability has individual and social sources. Individual practitioners bring values, interests, dispositions, and skill and knowledge to their encounters with policy. Values bear on capability because policy aims or instruments often attract or offend practitioners’ values or those of actors in the environment. Teaching evolution deeply offends some educators, parents, and politicians who believe that it introduces incorrect or evil ideas, while not teaching evo- lution deeply offends others who accept science. Such values influence ca- pability by enhancing or weakening the will to implement policy or by im- peding or enhancing acquisition of the skills and knowledge needed to implement. Values vary with environments. Aversion to evolution is more likely in the U.S. South, while support is more likely in cosmopolitan urban areas. Similarly, policies that reduce speed limits are more likely to be accepted in urbanized states than in the American West, where they are thought to infringe on basic liberties. Even when programs seem to have no prominent value content, they can come into play. Adding ten minutes of daily writing may reduce recess or music, requiring practitioners to adjust their priorities, not in principle but in the daily conduct of valued work. Such conflict can reduce practitioners’ will to do as policies propose (Lin 2000).

Interests are another source of capability. A policy that requires firms to cease dumping PCBs (polychlorinated biphenyls) in rivers may cost them money or market share, and requiring educators to test students and track their progress is costly. Even if manufacturers or educators agree with the policies, implementation can damage their interests by reducing profits, cutting products, or changing work. That can diminish the will to implement, just as policies that enhance practitioners’ interests can increase it. Rewarding cor- porations for protecting the environment, by trading pollution-reducing prac- tices, is intended to enhance their will to advance policy by enhancing their interest, just as paying teachers to learn more about their subjects could increase their will to improve instruction by showing them how more knowl- edge could advance their interest in student learning. Such interests can vary with environments. Towns that are on polluted rivers far downstream from the pollution source are likely to have a greater interest in the creation and implementation of remedial policies than those situated at the source, where

This content downloaded from 192.251.139.3 on Thu, 05 May 2016 21:40:28 UTC All use subject to http://about.jstor.org/terms

Policy and Practice

538 American Journal of Education

families depend on polluters for jobs and towns depend on them for taxes and economic development.

Knowledge and skill are a third source of individual capability. Motorists use their driving skills and knowledge of speeding fines to deal with speed limits, and judges use their knowledge of federal and state sentencing guidelines and their skill at weighing evidence to decide criminal cases. When policies urge change in practice, they often create incompetence by requiring prac- titioners to do things that they do not know how to do (Bardach 1977). Some policies support education for practitioners so as to deal with the incompetence that the policies create.

Dispositions are a fourth source of capability. Teachers with similar values, interests, and knowledge often approach their work quite differently: some are warmly engaged and eager to work, while others are diffident. Similarly, some managers are aggressive in pursuit of their objectives, while others are more cautious. Such differences contribute to capability by influencing the will to work, the motivation to learn, and engagement with practice.

The social sources of capability are found in organizations and environ- ments. In some cases the salient considerations are economic. Affluent school districts typically employ more qualified teachers, who have more resources and often use them more effectively than colleagues in poorer districts. Schools with better leadership, which also is more often found in affluent communities, make better use of teachers and other resources. Schools in less affluent com- munities, by contrast, have more staff turnover, which reduces capability, in part by impeding the development of stable knowledge among practitioners. Since policies operate in environments, such economic differences count. Title 1 is a modest federal supplement to state and local operations: the entire federal contribution is now a bit less than 10 percent of total local revenues. Title 1’s influence in practice thus depends greatly on what local school or- ganizations bring to bear, yet states and local school systems have very unequal capabilities, owing to very large differences in finance and income. The states and localities that need Title 1 the most therefore bring the least capability to implementation, so that the program’s operation and effects are shaped by the very inequalities that it aims to ameliorate.

Some social sources of capability are cultural and educational. Title 1’s influence in practice depends both on what local school organizations bring to bear and on what families and communities bring to children’s schooling. Yet states and local school systems have very unequal capabilities, owing in part to large differences in parents’ education and communities’ resources. Here again, the communities and families that need Title 1 the most bring the least capability to implementation; here, too, the program’s operation and effects are shaped by the very inequalities that it aims to ameliorate.

Organizations and environments also influence capability by shaping the

This content downloaded from 192.251.139.3 on Thu, 05 May 2016 21:40:28 UTC All use subject to http://about.jstor.org/terms

Cohen, Moffitt, and Goldin

AUGUST 2007 539

exchange of knowledge about both practice and the interests, values, and knowledge that inform practice. For instance, the size and complexity of service-delivery organizations affect practice by affecting the exchange of knowledge about practice; the larger and more complex they are, the more difficult it is to exchange knowledge. Other things being equal, it is likely to be less difficult to implement a school improvement policy in a medium-sized subdistrict of New York City than in the entire city, for communication, learn- ing, and oversight would be more difficult for the entire city (Cohen and Spillane 1992; Elmore and Burney 1997).

Organizations can structure themselves to enhance or constrain capability. Some school systems use inspectors to monitor and improve instruction: teach- ers’ promotion to tenure and school headships depends on inspectors’ judg- ment about candidates’ demonstrated quality of practice, and that can help to sustain or improve individuals’ performance. The key point, though, is that inspectorates can be repositories of capability if they institutionalize knowledge, skill, and the disposition to maintain or improve practice. In such school systems, inspectorates can build and sustain social capital. Organizations also can inhibit the development of capability: in the United States, excellence in teaching is not required to maintain teaching or administrative posts, and there are no inspectorates. To the extent that it exists, instructional expertise is concentrated among classroom teachers. The lack of such capability outside classrooms makes it more difficult to improve performance inside them.

Our discussion implies that capability is specific to tasks or sorts of tasks; there appears to be little generic capability. When policies propose to change practice, they create incompetence, but the “incompetence is not, of course, a trait like having brown eyes. It is a description of a relationship between a task or function in the situation up to a given, though ultimately arbitrary, standard of some sort” (Bardach 1977, 126). Some policies set relatively narrow tasks for practitioners: federal impact aid required practitioners only to accept money and spend it on educational activities. In this case, the capabilities in question centered on accounting and had little or nothing to do with teaching or learning. Other policies, however, set very broad tasks for practitioners: Presidents Clinton and Bush revised Title 1 to set ambitious new tasks for educators, for they required the elimination of race, class, and gender-based inequality in student performance. In this case, the capabilities in question spanned teaching, learning, school management, school improvement, as- sessment, and the use of scientific knowledge in practice. Both impact aid and standards-based reform focus on education, but they require quite different capabilities from practitioners.

Additionally, practitioners’ capability arises not from a single attribute— skill and knowledge are most often mistaken for the whole—but from com- binations of individual knowledge, values, interests, and dispositions, and social

This content downloaded from 192.251.139.3 on Thu, 05 May 2016 21:40:28 UTC All use subject to http://about.jstor.org/terms

Policy and Practice

540 American Journal of Education

capabilities. Practitioners are likely to find it more difficult to acquire or use knowledge when policies negatively affect their interests or values or when they are not disposed to engage with practice. Values also can help to create interests, as when lobbying for charter schools yields legislation that creates networks of charter sponsors, educators, and parents, which became significant in state politics and support for charter schools. Capability subsists in such interactions and is maintained, enhanced, or degraded by them.

Organizations can enhance capability by enabling and institutionalizing the exchange of knowledge about practice and the interests, values, and knowledge that inform practice. They also can constrain capability by inhibiting such exchange. This sort of capability seems to be a form of social capital that grows if organizations enable and institutionalize collective practice (Bryk and Schneider 2002).

Finally, because the elements of capability interact, policy makers may have more than one point of leverage. Policies often try to mobilize knowledge and skill indirectly by offering incentives for results and leaving it to practitioners to exert or acquire the required capability. That can work if practitioners have or can easily acquire the capability, but policies that offer incentives alone are less likely to work if practitioners cannot easily learn what they need to know. Recent state policies make schools accountable for student performance and reward or penalize them accordingly. Most states set modest goals, but a few set tougher targets. Teachers in both sorts of states acquire an enhanced interest in students’ performance, but those in states with less tough targets require only a modest increase in skill and knowledge. As policies depart further from current practice, interest alone is less likely to provoke improved knowledge, skill, or practice, for tougher requirements create more incompetence and imply a need for more capability, which teachers are less likely to create on their own. Policies that press practitioners to produce performances for which they lack capability sharpen the dilemma and increase the chances that practitioners will attempt to comply without actually affecting the announced outcome.

Capability is relational, like the other resources that we have discussed. It waxes and wanes in interaction with the aims that policies set, the instruments that they deploy, and the environments in which policy and practice subsist. One can speak accurately of capability only if one speaks in relational terms, and one can shape capability only by shaping it in relation to those other resources.

Conclusion

A dilemma lies at the heart of relations between policy and practice: many policies seek to correct problems in the definition or delivery of social and

This content downloaded from 192.251.139.3 on Thu, 05 May 2016 21:40:28 UTC All use subject to http://about.jstor.org/terms

Cohen, Moffitt, and Goldin

AUGUST 2007 541

educational services, yet the key problem solvers are the failing schools, welfare clients, or drug addicts that policy identifies as the problem. The success of policy depends on the flawed clients, practitioners, and organizations that policy would correct. The central puzzle for policy and practice is how to enable these people and organizations to do as policies propose. Policies may offer incentives, ideas, money, and more, yet they will help only if they are used well by those who are said to be the problem and by others who may assist.

Our discussion has highlighted the parts that cooperation and conflict play in the relations between policy and practice. On the one hand, policy makers rely on practitioners for their success, and, on the other hand, practitioners rely on policy makers for resources and guidance. Both contain incentives to cooperate. Policy makers’ legitimacy and political interests depend in part on policies’ fate in practice, so policy makers have reasons to accommodate prac- titioners’ concerns and capabilities. Practitioners have incentives to execute policy, for it arises in legitimate authority, it can offer valuable resources, and failure could weaken the interests and legitimacy of practice. But these in- centives to cooperate often exist alongside sources of conflict. Most generally, conflict presents itself in the relationship between the tasks that policies set for practice, on the one hand, and the resources available to accomplish those tasks, on the other hand. The further policy departs from extant practice and the less instruments, capabilities, and environments are able to bridge the distance, the more likely is conflict between policy makers and practitioners. Inconsistency in these relationships arises from several sources, including re- source constraints, conflicts in policy making that are expressed in vague aims and weak instruments, and lack of knowledge.

Yet the dilemma itself implies a fundamental tension between policy and practice, which often generates conflict. Consider an example: recent edu- cation policies assume that state and local school systems served poor children badly, and some prescribe forceful measures to get them to correct the problem. From this angle, it made sense for NCLB to use tough instruments to mobilize educators’ will to improve, to identify weak professionals, and to spur per- formance. Yet if educators are to improve weak schools, it also makes sense for policy to offer the assistance, encouragement, and trust that they need to do the job. Advocates of the first view tend to see advocates of the second as apologists for those who failed poor children and would defeat policies like NCLB by allowing weak performance. Advocates of the second view tend to see partisans of the first as inflexible enforcers who are likely to defeat school improvement by driving practitioners into mechanistic compliance. The dif- ference is built into the dilemma that frames relations between policy and practice, and it was expressed in conflict during legislative drafting over how aggressive or helpful NCLB should be. It was expressed during implementation

This content downloaded from 192.251.139.3 on Thu, 05 May 2016 21:40:28 UTC All use subject to http://about.jstor.org/terms

Policy and Practice

542 American Journal of Education

in conflict over how aggressive or helpful governments should be. It was expressed in continuing debate and conflict on these points among those interested in and engaged with implementation. And it will be expressed in debates concerning the reauthorization of Title 1 in the coming years.

Such conflict is difficult to avoid because each view has a portion of truth. The people and organizations that have been the problem often are complicit in social damage, and, from that perspective, they cannot be trusted to do the right thing; it makes sense to treat them with some skepticism and to call them to account. Yet, because they are the chief available agents of improve- ment, it also makes sense to give them the help and trust that could enable them to improve. In a reasonable world, policy makers and practitioners would choose both, for, short of creating entirely new systems of service provision, policies would be more likely to improve practice if they blended the two. One example is policy that combines instruments that require performance with instruments that are calibrated to practitioners’ capabilities to improve.22

But even if policies did choose both, it is very unlikely that any policy could be written in ways that entirely satisfy the contrary views of truth and fairness that the dilemma embodies.

The key element here is not position in hierarchies but tensions between policy and practice that the dilemma highlights. But if conflict is built into the relationship between social policy and practice, it can be eased or made worse by how the relations among aims, instruments, capability, and envi- ronments are managed. One enduring source of tension between policy and practice is inadequate knowledge; improved knowledge could inform efforts to manage conflict. If evidence were systematically collected on the operation and effects of policies, to illuminate the causes of success and failure, the resulting knowledge could help to improve instruments and capabilities. Sys- tematic inquiry into educational processes and school improvement could have similar effects.

That point has special importance for standards-based reform because knowledge assumes such a crucial role in these policies. They propose to improve schools’ effectiveness, and they can succeed only if schools improve learning. That greatly expands what educators need to know, for if school staff lack knowledge of effective practice, they will not greatly improve learning. If practice fails, it could make standards-based policies a political liability for schools and the policies would lose legitimacy if failure was epidemic. By requiring schools and school systems to be more effective, these policies set a new priority on the creation and use of valid knowledge about teaching, learning, and school management.

Yet neither IASA nor any state policies were written or executed in ways that enabled a great leap forward in knowledge production and use. There is little provision for continuing formative studies. Moreover, though NCLB provides

This content downloaded from 192.251.139.3 on Thu, 05 May 2016 21:40:28 UTC All use subject to http://about.jstor.org/terms

Cohen, Moffitt, and Goldin

AUGUST 2007 543

for more focused research on schooling and school improvement in the Institute of Education Sciences, that agency has a huge agenda and a modest budget. In addition, tight compliance schedules in NCLB mean that, even if the funds had been much more ample and more research was done, much of it would be concluded after the key decisions were made. This situation encouraged practitioners to adjust their interpretation of policy to something closer to con- ventional practice. Although NCLB rhetorically recognized the importance of new knowledge, it placed no high priority on producing it nor did it frame implementation as a means to generate and use it.

Several decades ago, Daniel A. Mazmanian and Paul A. Sabatier (1989) wrote: “[If a] valid causal theory linking target group behavior to policy objectives is not available or is clearly problematic . . . then the proponents of the statute must make a conscious effort to incorporate in it a learning process through experimental projects, extensive research and development, evaluation studies, and an open decision process involving as many different inputs as possible” (273). It is a wise observation, even more apposite today than it was then, but few policy makers appear to have noticed. If punishment is a purpose of policy, generating knowledge to solve problems is beside the point. But punishment alone cannot work. Standards-based reform has created a new frame for education policy, in which both practitioners and policy makers have an unprecedented need for much better knowledge about teach- ing, learning, and school improvement. Without such knowledge, it will be impossible to deliver on the policies’ promises, yet these policies have neither significantly enhanced the capability to produce such knowledge nor have they devised ways in which it might be used by practitioners and policy makers.

Notes

We are indebted to Fritz Mosher and Richard Elmore for comments on an earlier draft.

1. Kenneth Meier defines regulation as “government restriction of individual choice to keep conduct from transcending acceptable bounds” (1993, 82).

2. Eric Hanushek (1996), e.g., argues that the effects of resources depend on their use and, hence, that resources should be tied to incentives in order to maximize effects on student outcomes.

3. Top-down theorists often consider policy makers to be those at higher levels of government, i.e., federal and state officials.

4. The literature on implementation is extensive. Several useful accounts are the following: Elmore (1978, 1979–80); Lindblom (1959); Lipsky (1980); Mazmanian and Sabatier (1989); McLaughlin (1991a, 1991b); Pressman and Wildavsky (1984); and Van Meter and Van Horn (1975).

5. “The notion that policymakers exercise—or ought to exercise—some kind of direct and determinant control over policy implementation might be called the ‘noble lie’ of conventional public administration and policy analysis” (Elmore 1979–80, 603).

This content downloaded from 192.251.139.3 on Thu, 05 May 2016 21:40:28 UTC All use subject to http://about.jstor.org/terms

Policy and Practice

544 American Journal of Education

6. Ten years after the Rand Change Agent Study, though, McLaughlin (1991b, 149) began to revise her account and allowed for the potential for policy to affect practice: “Belief or commitment can follow mandated or ‘coerced’ involvement at both the individual and the system level. . . . We did not see or did not recognize instances in which belief follows practice. Individuals required [by policy] to change routines or take up new practices can [thereby] become ‘believers.’”

7. Commentators often portray this as conflict over goals. 8. Several of these studies and others have been collected in a very useful recent

volume (Honig 2006). 9. The terms policies and programs are equivalent for our purposes, so we use them

interchangeably. 10. Although this point has not been central to most analyses of implementation,

several scholars have called attention to it. Richard Elmore wrote that “unless the initiators of a policy can galvanize the energy, attention, and skills of those affected by it . . . the effects of policy are unlikely to be anything but weak and diffuse” (1979–80, 611). Milbrey (McLaughlin 1991a) wrote that policy implementation is a “problem of the smallest unit” and that policies cannot “mandate what matters” (189 and 187, respectively). Elmore and McLaughlin (1988) wrote that the fate of reforms ultimately depends on those who are the object of distrust. Ann Chih Lin writes that “every grand idea and good wish that policymakers have lies in the hands of others who implement them” (2000, 14).

11. Some political scientists use principal agent theory and its hierarchy to explain intraorganizational issues (see, e.g., Brehm and Gates 1997). It recognizes control and compliance problems at several points along an implementation path, which might imply that a “bottom” can also be a “top.” But it treats these relations as layered and in isolation from each other. In our formulation, individuals or groups of individuals can be simultaneously a top and a bottom.

12. For a helpful and somewhat parallel discussion, see Elmore (2005). 13. On ideas as policy instruments, see Cohen and Hill (2001) and Weiss and Gruber

(1984). 14. We use the term instruments in a sense similar to McDonnell and Elmore’s (1987)

use of the term. They refer to policy instruments as “the mechanisms that translate substantive policy goals (e.g., improved student achievement, higher quality entering teachers) into actions” (134). They do not address either adaptability or instruments’ salience for practice. For a discussion of the former, see McLaughlin (1991a). For a discussion of the latter, see Cohen and Hill (2001, 2–8, 148–51, 166–67, 180, 184).

15. That is because warehouses have capacity, while people and organizations have capabilities. Discussions of the implementation of education policies often mention capacity, but we have found no analysis of the concept in these accounts; typically it is taken to refer to the knowledge and skills of implementers. There is a substantial literature on capability in the philosophy of economics, due largely to Amartya Sen’s use of the concept in his critiques of utilitarian views of the quality of life. Sen’s view and several related discussions can be found in Nussbaum and Sen (1993).

16. Michael Kirst, interview, September 1995. 17. These paradoxical features of Title 1 closely resembled much state school policy,

in which weak school finance equalization policies were seen as powerful means to equalize educational opportunity.

18. Those estimates were mistaken. At the end of the second Clinton administration, Michael Cohen was appointed Assistant Secretary of Education for Elementary and Secondary Education; he vigorously enforced IASA’s requirements, and more states complied.

This content downloaded from 192.251.139.3 on Thu, 05 May 2016 21:40:28 UTC All use subject to http://about.jstor.org/terms

Cohen, Moffitt, and Goldin

AUGUST 2007 545

19. Some readers may see our analysis as a version of principal agent theory, but we do not. Two chief components of principal agent models, as applied to government agencies, are goal conflict and information asymmetry (Waterman and Meier 1998). Much principal agent theory and application build on those components. Goal conflict is pervasive in the United States, but fundamental implementation problems would persist even if goals or preferences were perfectly aligned throughout the school system and even if all principals were privy to all agents’ information. The problems of aligning implementers’ capabilities, policy goals, and instruments can be exacerbated by goal conflict and/or information asymmetry, but they are independent of goal conflict and/ or information asymmetry. Thus a central problem for Title 1 and many other policies and programs is that they have goals that few, if any, have the capabilities to implement and for which policy or program instruments typically offer tepid and partly salient support at best. In such situations, principals are unlikely to behave differently if they know everything that agents know, for neither know nearly enough to implement policies broadly and effectively.

20. Peter Dow, in Schoolhouse Politics: Lessons from the Sputnik Era (1993), offers an account of these problems in connection with his analysis of one such curriculum— Man, A Course of Study.

21. Traditional notions of information asymmetry do not seem to help here. Al- though gaming is one reaction to incentives sans capability, this does not hinge on agents’ private information and how it enables them to shirk, drift, sabotage, and otherwise pervert democratic processes.

22. Milbrey McLaughlin (1991a) wrote: “Experience shows that some balance of pressure and support is essential. Pressure is required in most settings to focus attention on a reform objective; support is needed to enable implementation” (188).

References

Amrein, Audrey L., and David C. Berliner. 2002. The Impact of High-Stakes Tests on Student Academic Performance: An Analysis of NAEP Results in States with High-Stakes Tests and ACT, SAT, and AP Test Results in States with High School Graduation Exams. Tempe, AZ: Education Policy Studies Laboratory.

Bardach, Eugene. 1977. The Implementation Game: What Happens after a Bill Becomes a Law. Cambridge, MA: MIT Press.

Berman, Paul, and Milbrey W. McLaughlin. 1979. Federal Programs Supporting Educational Change. Vol. 8. Implementing and Sustaining Innovations. Santa Monica, CA: Rand.

Brehm, John, and Scott Gates. 1997. Working, Shirking, and Sabotage: Bureaucratic Response to a Democratic Public. Ann Arbor: University of Michigan Press.

Brodkin, Evelyn Z. 1990. “Implementation as Policy Politics.” In Implementation and the Policy Process: Opening Up the Black Box, ed. Dennis J. Palumbo and Donald J. Calista. New York: Greenwood.

Bryk, Anthony S., and Barbara Schneider. 2002. Trust in Schools: A Core Resource for Im- provement. New York: Russell Sage.

Carnoy, Martin. 2005. “Have State Accountability and High-Stakes Tests Influenced Progression Rates in High School?” Educational Measurement: Issues and Practice 24 (4): 19–31.

Carnoy, Martin, and Susanna Loeb. 2002. “Does External Accountability Affect Stu- dent Outcomes? A Cross-State Analysis.” Educational Evaluation and Policy Analysis 24 (4): 305–31.

This content downloaded from 192.251.139.3 on Thu, 05 May 2016 21:40:28 UTC All use subject to http://about.jstor.org/terms

Policy and Practice

546 American Journal of Education

Coburn, Cynthia E., and Mary Kay Stein. 2006. “Communities of Practice Theory and the Role of Teacher Professional Community in Policy Implementation.” In Honig 2006.

Cohen, David K. 1990. “A Revolution in One Classroom: The Case of Mrs. Oublier.” Educational Evaluation and Policy Analysis 12 (3): 311–29.

Cohen, David K., and Carol A. Barnes. 1993a. “A New Pedagogy for Policy?” In Teaching for Understanding: Challenges for Policy and Practice, ed. David K. Cohen, Milbrey W. McLaughlin, and Joan E. Talbert. San Francisco: Jossey-Bass.

Cohen, David K., and Carol A. Barnes. 1993b. “Pedagogy and Policy.” In Teaching for Understanding: Challenges for Policy and Practice, ed. David K. Cohen, Milbrey W. McLaughlin, and Joan E. Talbert. San Francisco: Jossey-Bass.

Cohen, David K., and Heather C. Hill. 2001. Learning Policy: When State Education Reform Works. New Haven, CT: Yale University Press.

Cohen, David K., Steven Raudenbush, and Deborah L. Ball. 2003. “Resources, In- struction, and Research.” Educational Evaluation and Policy Analysis 25 (2): 119–42.

Cohen, David K., and James P. Spillane. 1992. “Policy and Practice: The Relations between Governance and Instruction.” In Review of Research in Education, ed. Gerald Grant. Washington, DC: American Educational Research Association.

Dow, Peter B. 1993. Schoolhouse Politics: Lessons from the Sputnik Era. Cambridge, MA: Harvard University Press.

Elmore, Richard F. 1978. “Organizational Models of Social Program Implementation.” Public Policy 26 (2): 185–228.

Elmore, Richard F. 1979–80. “Backward Mapping: Implementation Research and Policy Decisions.” Political Science Quarterly 94 (4): 601–16.

Elmore, Richard F. 2005. “Agency, Reciprocity, and Accountability in Democratic Education.” In The Public Schools, ed. Susan Fuhrman and Marvin Lazerson. Oxford: Oxford University Press.

Elmore, Richard F., and Deanna Burney. 1997. “School Variation and Systemic In- structional Improvement in Community District #2.” Annual report, High Perfor- mance Learning Communities Project, Learning Research and Development Center, University of Pittsburgh.

Elmore, Richard F., and Milbrey W. McLaughlin. 1988. Steady Work: Policy, Practice, and the Reform of American Education. Santa Monica, CA: Rand.

Friedrich, Carl J. 1941. Constitutional Government and Democracy: Theory and Practice in Europe and America. Boston: Little, Brown.

Fuhrman, Susan H., and Richard F. Elmore. 1992. “Takeover and Deregulation: Working Models of New State and Local Regulatory Relationships.” CPRE Research Report no. RR-24, Consortium for Policy Research in Education, Rutgers University.

Fuhrman, Susan H., and Richard F. Elmore. 1995. “Ruling Out Rules: The Evolution of Deregulation in State Education Policy.” Teachers College Record 97 (2): 279–309.

Grissmer, David W., Ann Flanagan, Jennifer H. Kawata, and Stephanie Williamson. 2000. Improving Student Achievement: What State NAEP Test Scores Tell Us. Santa Monica, CA: Rand.

Haney, Walter. 2000. “The Myth of the Texas Miracle in Education.” Educational Evaluation and Policy Analysis Archives 8 (41), http://www.epaa.asu.edu/epaa/v8n41.

Hanushek, Eric. 1996. “School Resources and Student Performance.” In Does Money Matter? The Effect of School Resources on Student Achievement and Adult Success, ed. Gary Burtless. Washington, DC: Brookings.

Hill, Heather C. 2006. “Language Matters: How Characteristics of Language Com- plicate Policy Implementation.” In Honig 2006.

This content downloaded from 192.251.139.3 on Thu, 05 May 2016 21:40:28 UTC All use subject to http://about.jstor.org/terms

Cohen, Moffitt, and Goldin

AUGUST 2007 547

Honig, Meredith I. 2006. New Directions in Education Policy Implementation: Confronting Complexity. Albany: SUNY Press.

Katznelson, Ira, and Margaret Weir. 1985. Schooling for All: Class, Race, and the Decline of the Democratic Ideal. New York: Basic.

Lin, Ann Chih. 2000. Reform in the Making: The Implementation of Social Policy in Prison. Princeton, NJ: Princeton University Press.

Lindblom, Charles. 1959. “The Science of Muddling Through.” Public Administration Review 19 (2): 79–88.

Lipsky, Michael. 1980. Street Level Bureaucracy: Dilemmas of the Individual in Public Services. New York: Russell Sage.

Majone, Giandomenico, and Aaron B. Wildavsky. 1984. “Implementation as Evolu- tion.” In Pressman and Wildavsky 1984, 177–94.

Martin, Ruby, and Phyllis P. McClure. 1969. Title 1 of ESEA: Is It Helping Poor Children? Washington, DC: Washington Research Project of the Southern Center for Studies in Public Policy and the NAACP Legal Defense of Education Fund.

Mazmanian, Daniel A., and Paul A. Sabatier. 1989. Implementation and Public Policy. Lanham, MD: University Press of America.

McDonnell, Lorraine M., and Richard F. Elmore. 1987. “Getting the Job Done: Alternative Policy Instruments.” Educational Evaluation and Policy Analysis 9 (2): 133–52.

McLaughlin, Milbrey W. 1991a. “Learning from Experience: Lessons from Policy Implementation.” In Education Policy Implementation, ed. Allan R. Odden. Albany: SUNY Press.

McLaughlin, Milbrey W. 1991b. “The Rand Change Agent Study: Ten Years Later.” In Education Policy Implementation, ed. Allan R. Odden. Albany: SUNY Press.

Meier, Kenneth J. 1993. Politics and Bureaucracy: Policymaking in the Fourth Branch of Gov- ernment. Belmont, CA: Wadsworth.

Mosbaek, E. J., F. R. Frola, K. F. Gordon, and J. W. Harrison. 1968. Analysis of Compensatory Education in Five School Districts. Final report from General Electric Co.–TEMPO to Office of Education, U.S. Department of Health, Education, and Welfare, under contracts HEW-05-67-55 (A Survey and Preliminary Cost-Benefit Analysis in Elementary and Secondary Education) and OEC-D-8-08042-3513 (Anal- ysis of Compensatory Education within Schools from Five Major School Districts). Washington, DC: TEMPO.

Nussbaum, Martha C., and Amartya K. Sen, eds. 1993. The Quality of Life. Oxford: Oxford University Press.

Olson, Lynn. 2002. “A ‘Proficient’ Score Depends on Geography.” Education Week, February 20.

Orfield, Gary. 1969. The Reconstruction of Southern Education: The Schools and the 1964 Civil Rights Act. New York: Wiley-Interscience.

Pressman, Jeffrey L., and Aaron B. Wildavsky. 1984. Implementation: How Great Expectations in Washington Are Dashed in Oakland; or, Why It’s Amazing That Federal Programs Work at All, This Being a Saga of the Economic Development Administration as Told to Two Sympathetic Observers Who Seek to Build Morals on a Foundation. 3rd ed. Berkeley: University of California Press.

Saulny, Susan. 2005. “State to State: Varied Ideas of Proficient.” New York Times, January 19, 8.

Scholz, John T. 1991. “Cooperative Regulatory Enforcement and the Politics of Ad- ministrative Effectiveness.” American Political Science Review 85 (1): 115–36.

Simon, Herbert A., Donald W. Smithburg, and Victor A. Thompson. 1950. Public Administration. New York: Knopf.

Smith, Marshall S., and Jennifer O’Day. 1990. “Systemic School Reform.” In Politics

This content downloaded from 192.251.139.3 on Thu, 05 May 2016 21:40:28 UTC All use subject to http://about.jstor.org/terms

Policy and Practice

548 American Journal of Education

of Education Association Yearbook, ed. Susan H. Fuhrman and Betty Malen. London: Taylor & Francis.

Spillane, James P. 2000. “Cognition and Policy Implementation: District Policy-Makers and the Reform of Mathematics Education.” Cognition and Instruction 18 (2): 141–79.

Spillane, James P., Brian J. Reiser, and Louis M. Gomez. 2006. “Policy Implementation and Cognition: The Role of Human, Social, and Distributed Cognition in Framing Policy Implementation.” In Honig 2006.

Spillane, James P., and Charles L. Thompson. 1997. “Reconstructing Conceptions of Local Capacity: The Local Education Agency’s Capacity for Ambitious Instructional Reform.” Educational Evaluation and Policy Analysis 19 (2): 185–203.

Spillane, James P., and John S. Zeuli. 1999. “Reform and Mathematics Teaching: Exploring Patterns of Practice in the Context of National and State Reforms.” Educational Evaluation and Policy Analysis 21 (1): 1–27.

Van Meter, Donald S., and Carl E. Van Horn. 1975. “The Policy Implementation Process: A Conceptual Framework.” Administration and Society 6 (4): 445–88.

Waterman, Richard W., and Kenneth J. Meier. 1998. “Principal-Agent Models: An Expansion?” Journal of Public Administration Research and Theory 8 (2): 173–202.

Weir, Margaret, Ann Shola Orloff, and Theda Skocpol. 1988. The Politics of Social Policy in the United States. Princeton, NJ: Princeton University Press.

Weiss, Janet, and Judith Gruber. 1984. “Using Knowledge for Control in Fragmented Policy Areas.” Journal of Policy Analysis and Management 3:225–47.

This content downloaded from 192.251.139.3 on Thu, 05 May 2016 21:40:28 UTC All use subject to http://about.jstor.org/terms