Research paper on Primary Education Reform Plan for Afghanistan

CHAPTER 3

Thinking about the Contexts and Processes of Change in Education

Previous chapters have considered what might be termed the macro- contexts of education change: chapter 1 discussed dominant traditions of education that, institutionalized during formative periods, continue to influence the thinking and activities of education actors and their institutions long after their introduction. From this analysis, we draw several insights. First, despite underlying commonalities among forms of the ‘‘modern school,’’ there are important differences across sys- tems, differences that have important impacts on the lives of education actors and that frame the possibilities for reform of education institu- tions. Despite convergence at some levels, there are, at other levels of analysis, influential, and persistent, institutional differences in the organization and operation of education systems. These differences suggest the possibility of policy space for change. In addition, the dis- cussion of traditions of education reminds us that education reform is nowhere undertaken on a blank slate. Any effort at change takes place in context, among people with history, traditions, institutions, and memory. In this way, reform would appear to involve a substantially deeper level of institutional change than is considered in many policy and planning exercises. At the same time, national education systems are not unique unto themselves. They tend to share features with other systems, based in part on common background and the resilience of institutions.

In the context of these influential models, Chapter 2 then elaborated a roughly historical typology of approaches to the education develop- ment process. Chapter 2 ends with the assertion that, given the prob- lems of current approaches, the profession lacks a coherent paradigm for managing systematic change in education. At the same time, the

PAGE 59

59

................. 11172$ $CH3 12-16-04 14:29:17 PS

60 CHAPTER 3

expectations for reform have grown: reform is charged not only with putting in place a new curriculum, for example, but also with develop- ing the capacity of the system to develop, implement, and evaluate cur- riculum on its own on a continuing basis. Not only is reform called upon to complete the job of the reform in a technically effective man- ner, but it is also tasked with collaborating in that process with new and various groups of stakeholders and managing different actors in the system. Reform is charged not only with making needed improve- ments in the delivery of education services but also with transformation of its own system and ultimately of society.

Working to achieve such objectives—indeed any changes to the deep structure of education system—requires critical attention to ‘‘local’’ as well as macro contexts. This is not to deny the salience of macro condi- tions, especially in the face of powerful external forces and organiza- tions. But the effects of the external on reform take place locally. Regardless of technical rationality, local political, organizational, and social factors and local implementers can decide the fate of an initia- tive. And so, this chapter discusses some of the issues of context and process that affect the change process. In particular, the chapter dis- cusses:

• multiple contexts of education change—the policy and the organi- zational contexts in which change is decided and enacted; the larger social contexts within the country; and the external political economy.

• processes of education change—surfacing and testing implicit change theories-in-use; development hypotheses, program theo- ries, and their empirical bases; factors affecting organizational learning; and (more and less useful) models of the change process.

Looking ahead, chapter 4 turns to the outputs and outcomes of edu- cation in different national systems, particularly in light of interna- tional efforts to mobilize education systems to meet targets associated with Education for All (EFA) and the Millennium Development Goals. The remainder of the book elaborates strategic options for change agents seeking to bring about reform.

PAGE 60................. 11172$ $CH3 12-16-04 14:29:18 PS

THINKING ABOUT THE CONTEXTS AND PROCESSES OF CHANGE IN EDUCATION 61

MULTIPLE CONTEXTS

Contexts of Change

The first insight we would offer is a simple one, though relatively infrequently observed: the effectiveness of reform must be understood (and plans made) not primarily in terms of intent, but in terms of what is implemented and then what is sustained or institutionalized. Accord- ingly, the change agent’s job is not simply to develop a technically sound and financially viable plan but to plan for implementation and (help) develop a process that will sustain and institutionalize the desired change. At a minimum, then, at least three contexts are directly relevant to this view of change:

• the context of policy formulation and planning; • the context of implementation; and • the context of institutionalization.

However well designed a policy is from the perspective of actors in one context, it is likely to look quite different to actors in another. A decentralization initiative, for example, may be intended by those in the formulation/planning context primarily as a means of enhancing the authority of actors closer to the ‘‘action.’’ However, to those imple- menting the reform, the same initiative might appear, for example, to be an unfunded mandate, a ‘‘decentralization’’ of problems the center is incapable of handling, even an unwanted increase in workload. Thus, to enhance chances of effective formulation, implementation, and insti- tutionalization, a policy should ‘‘make sense,’’ initially and on an ongo- ing basis, in all three contexts. We use the term naı̈ve planning to refer to change initiatives that are uninformed by the multiple contexts of policy, for example, policy developed with only the policy-making context in mind.

How important is awareness of context? That depends. The greater the need for economies of scale or national mobilization of resources, the greater the reliance on central expertise or coordination, the greater the importance of systemic linkages or of national political leadership or law, the greater the importance of attention to the contexts of formu- lation and planning. In terms of attention to the ‘‘bottom,’’ or ‘‘enacting

PAGE 61................. 11172$ $CH3 12-16-04 14:29:19 PS

62 CHAPTER 3

contexts’’ of reform, the more the reform relies on buy-in on the part of implementers and sustainers—immediately, or over time—the more the reform requires professional judgment on the part of implementers and/or sustainers, the more change agents need to consider the differ- ential contexts of implementation and institutionalization in their plans. Similarly, the more local contexts vary in important or unknown ways, or the more unpredictable the course of reform in local contexts, the greater the need for awareness, and consideration, of local context in planning.

We would argue that three questions are particularly salient:

• Are enactors able to? Are implementers and sustainers able—at individual and organizational levels—to implement/sustain the policy or change? Ability includes resources, know-how, and organizational wherewithal.

• Do they have to? How free are implementers and sustainers not to implement/sustain the policy?

• Do they want to? Does it make sense for implementers/sustain- ers—on balance, given other choices, everything else going on, history, and the costs and benefits from the perspective of implem- enters and sustainers—to implement/sustain the policy or change? Put another way, given the values, goals, competing claims on attention and time, and the relative ability of authority to coerce, and the past, is it rational for implementers or sustainers to imple- ment/sustain this policy? This cost–benefit calculus is likely to include organizational, economic, political, social, cultural, histor- ical, and perhaps ideological considerations.

Participation of implementers and sustainers in the formulation and planning of policy is one useful way to work toward responsiveness to these contexts. Participation, of course is a fuzzy term, referring to a variety of particular strategies and begging for problematization. Indeed, participation can refer to very casual to quite meaningful involvement in decisions that affect one (see Shaeffer, 1992). Partici- pation, whichever meaning is used, is not a panacea, but we would argue that participatory approaches likely offer a better possibility of contextually informed planning than nonparticipatory approaches.1

PAGE 62................. 11172$ $CH3 12-16-04 14:29:20 PS

THINKING ABOUT THE CONTEXTS AND PROCESSES OF CHANGE IN EDUCATION 63

In addition, the organizational contexts in which policy maker, implementer, and sustainer work are likely to play particularly influen- tial roles in shaping response to a reform initiative—the extent to which an organization’s survival and well-being is threatened or assured; internal systems of accountability, reward, and support; the extent of hierarchy and cohesion within the organization; an organization’s openness to change and learning; and so forth.

In thinking about context, it is important to be clear about the point or points at which implementation decisions are made. Thus, for exam- ple, if a policy aims to increase girls’ enrollment in primary school, an obvious context of implementation is the family. (Local officials may promulgate, even promote, the policy, but the family is the critical implementer.) Does it make sense—given the family’s economic con- ditions; other children; the availability, safety, and attractiveness of the school; and the family’s cultural values; and in light of the govern- ment’s ability and willingness to coerce—to send a particular girl to school? Regardless of the rationality of girls’ education from a national perspective, or even from the long-range perspective of the family, a family is unlikely to enroll a girl in school if the perceived costs (physi- cal, economic, social, cultural, etc.) of enrollment outweigh the per- ceived benefits and if no one compels enrollment. Many well-intended but centrally focused diagnoses of girls’ education, for example, con- strue the problems and solutions only from the perspective of policy makers and planners, not, critically, from the perspective of those who will decide whether or not girls attend school.

If part of the girls’ education initiative involves a pregnancy policy, for another example, whereby adolescent mothers are permitted to return to school after childbirth, then the school leadership is likely to be a critical context for implementation. If a school director has sub- stantial discretion in implementing the policy, for example, his or her internal calculus of costs and benefits is likely to determine implemen- tation (see Wolf Lang, Mount, & VanBelle-Prouty, 1999).

A policy that enjoys widespread support is also more likely to be sustained, whereas one lacking such support, whether through opposi- tion or the noise of competing claims, is likely to suffer from benign neglect. In contexts of choice, we have argued, the social, cultural,

PAGE 63................. 11172$ $CH3 12-16-04 14:29:21 PS

64 CHAPTER 3

political, and organizational viabilities of a policy or change are as important to its sustainability as its financial feasibility.

We would argue that implementers and sustainers almost always act in ways that are rational, and can be understood, but these ways are rational in their own contexts. What might appear to be irrational on the part of an implementer and sustainer is often a misplaced projection of the rationality of the context of the planner/policy maker onto the context of implementer/sustainer.

Planners working under the assumptions of ‘‘rational planning’’ and hierarchical authority are particularly prone to missing the diverse rationalities of the multiple contexts and players in policy reform. In an effort to solve education problems expeditiously and under the con- sistently tight constraints of funding and attention, overly rational plan- ners tend to downplay the history and context of a problem, the meaning it has to different players. They tend to apply the technical rationality of a solving a problem to the different rationalities of imple- mentation and institutionalization, whereas solving a technical problem is only part of the process of implementing and institutionalizing change. In this way, outsiders may be particularly challenged, as com- pared with knowledgeable insiders, in developing implementable, socially sustainable plans, as they tend to lack a sufficient understand- ing of the particular local contexts.

At the same time, administrative authorities are prone to confusing legitimate or positional power with enforcement power. They are likely to assume a policy or plan will be implemented in the specified way because authorities, who are in charge, have made the policy decision. In this way, administrative authorities may be more challenged than outsiders to develop implementable, socially sustainable policies and plans, for they may be too close and too convinced of the power in their authority. We remember a regional decentralization initiative that the central ministry announced as fully implemented the afternoon it was promulgated by central authorities.

Larger Social Contexts of Reform

Given the pressures of time, attention, and resources, and the desire to improve the system, a conscientious change agent is likely to focus

PAGE 64................. 11172$ $CH3 12-16-04 14:29:22 PS

THINKING ABOUT THE CONTEXTS AND PROCESSES OF CHANGE IN EDUCATION 65

on the problem and solution, to the exclusion of context. If a rational planner, she or he will tend to focus quite intently on the best technical solution to the problem. This makes sense given the logistical and orga- nization challenges of getting policy on the agenda, marshalling sup- port, and gaining approval, and the technical challenges of defining problems, gathering data, devising affordable and hopefully effective strategies to deal with the problem, all fast. Moreover, it is difficult for technical staff, sitting in a planning office, for example, to do much about the context, whereas the change agent can work on the proposals for change. From the longer-range perspective of implementation and institutionalization, however, and of working to bring about deep change in education systems, it makes sense to think about context.

Nowhere, as we have asserted, does reform start with a blank slate. Change agents who miss the larger context run the risk of proposing naı̈ve, underinformed strategies for change. Failure to see the larger context may result in a misunderstanding of the full origin and dynamic of the problem, leading, for example, to treatment of the symptom, to use a medical metaphor, rather than the underlying cause. Still, under- standing context is an enormously complex undertaking, and it may be useful to highlight some of the dimensions of context of likely rele- vance.

Social Context: How does an initiative look not only in terms of the problem and solution, but also in terms of the larger social con- text?

Historical Context: What is the history of reform in this context? What historical experiences are likely to affect development, implementation, and success of the reform? How does an initiative look, not only in terms of current problems and possible solutions, but also in light of the history of problems and solutions?

Political Economic Context: Which groups stand to gain and lose by the reform?Economic Context: How will the reform be financed? Over the short and the long term? What are the relative costs and benefits? How will costs and benefits accrue to different groups?

Organizational Context: Does the ministry have the capacity to implement and institutionalize a reform such as this?

PAGE 65................. 11172$ $CH3 12-16-04 14:29:22 PS

66 CHAPTER 3

Sociological Context: How will the reform affect predominant divi- sions in society?

Technical Context: How is the solution going to solve the problem? Do the parties responsible for implementation have the general knowledge, the specific understanding of tasks and responsibili- ties, and the resources to carry out their responsibilities?

Cultural Context: Is the reform compatible with local culture(s)? What meaning will the reform acquire in the context of local cul- ture?

Political Context: How will the reform affect the distribution of power in society? Are dominant coalitions in favor? Why is this issue coming to the fore now?

Processes of Change

In addition to mindfulness of context, we highlight the importance of the change process, in five dimensions: (1) the nature of the problem to be addressed (different problems call for different processes); (2) the social mechanisms and potential for social resolution of policy issues (the greater the social capital and the more developed social institu- tions, the easier the process, all else being equal); (3) the empirical and theoretical foundations of development hypotheses and program theo- ries (relatively few proposals are well-grounded in theory and research); (4) the surfacing and testing implicit theories of change; and (5) more and less useful models of the change process.

Nature of the Problem

Reimers and McGinn (1997) list four questions that provide useful distinctions among policy problems. Participatory processes become important to the extent that a problem is divergent, involves multiple disciplines and levels of knowledge, involves negotiation or multiple actors, and operates in a democratic policy arena. From this perspec- tive, most important problems in education require a more or less par- ticipatory process.

• Convergent or divergent? Convergent problems have single solu- tions as compared with divergent problems, which have many pos-

PAGE 66................. 11172$ $CH3 12-16-04 14:29:23 PS

THINKING ABOUT THE CONTEXTS AND PROCESSES OF CHANGE IN EDUCATION 67

sible good answers. Convergent questions tend to focus on technical issues, such as: What is the cost of educating a student through secondary school? A related divergent question might be: Should the government invest more in secondary education? There is a tendency to answer divergent problems with convergent answers, rather than to use answers to divergent questions to inform a larger process of decision making.

• Single or multidisciplinary/level knowledge? Some problems can be addressed with knowledge from a particular profession or disci- pline or from one level of the system, whereas other problems require knowledge from multiple disciplines and different system levels. Testing specialists, for example, could well design a contin- uous assessment system on their own. However, the knowledge that teachers, administrators, parents, and others have of condi- tions in schools, classrooms, and communities would be needed to design a continuous assessment system that is likely to be imple- mented, and widely used as intended.

• Single decision maker or multiple actors? Obviously, the fewer the number of people and levels of an organization involved in a deci- sion, the less negotiation is required. A simple policy decision about administrative matters, for example, might be made by a minister or her designee, whereas use of a new instructional approach in teaching mathematics would require a much more complex set of negotiations to be successfully implemented.

• Soluble with technical knowledge? Some problems can be addressed using technical knowledge, whereas other problems can only be resolved by reference to values and preferences, which often vary greatly. Compare two questions: Will the elimination of school fees result in more girls enrolling and completing school? Is the purpose of education to prepare students for the workforce or to transform society and redress past inequities?

Authoritarian or Democratic Decision Making?

Reimers and McGinn raise a fifth question about the context of deci- sion making—top-down or participatory? On the one hand, policies are more easily developed and implemented in a more authoritative sys-

PAGE 67................. 11172$ $CH3 12-16-04 14:29:24 PS

68 CHAPTER 3

tem. At the same time, we would argue, more participatory policy proc- esses, which often become necessary in addressing complex policy problems such as those outlined above, are facilitated by a successful history of participatory decision making in a system; the existence of social and political structures outside education in which public partici- pation plays an important role; and by social cohesion or ‘‘social capi- tal’’ and an operational level of trust among important groups of stakeholders and social groups. At the same time, the need to make important decisions in education can provide an occasion for initiating new forms of participation, as there is near universal agreement as to the importance and value of education, if not the means. Regardless, a society’s predominant approach to decision making plays an important role in determining appropriate process.

Empirical Basis for Development Hypotheses and Program Theories

While advocating appropriate structuring of participation in the process of policy-making, such processes should also be as informed as possible by a solid empirical base. As Reimers and McGinn argue, virtually all important decisions in education involve both technical and normative dimensions. Unfortunately, education facts are rarely cut and dried. Answers tend not to be right or wrong so much as right if one wants to achieve a certain goal. Or wrong when certain condi- tions prevail. In cases of divergent problems, or where the answer depends on the differing values of stakeholders, information can serve to inform trade-offs and suggest likely consequences of different courses of action.

We have argued that the empirical basis for most reform efforts in education is extremely weak. Most of what is done, we would argue, even by the most rational of technical planners, is based on something akin to faith. Elmore (1996), for example, argues that U.S. educators do not know how to improve education on a system-wide basis. Lacking knowledge of what is sufficient, we develop intervention after interven- tion of necessary measures that may, or may not, achieve targeted goals. It is worth inquiring, at appropriate points in the policy process, how grounded our underlying theories of change are in research and/or

PAGE 68................. 11172$ $CH3 12-16-04 14:29:24 PS

THINKING ABOUT THE CONTEXTS AND PROCESSES OF CHANGE IN EDUCATION 69

practice. Will holding teachers accountable for their students’ scores on standardized tests result in improved instruction for all students? It seems logical that it would, but what does the research say? Do stu- dents who are forced to repeat a grade learn more than similar students who are promoted? We don’t really know. What are the effects of school choice, for example, on equity? In a particular community? Across a system? We don’t know. Given such a murky theoretical foun- dation, it is useful to make as explicit as possible and to test the under- lying hypotheses on which we expect change initiatives to work.

Surfacing and Testing Theories-in-Use

Deep learning becomes possible as change agents become aware of their own, and others’ (implicit, and often inaccurate), theories of change and work to surface and test them. Such an approach places the change agent in a learning as opposed to a knowing stance, useful given how little is known.

At this point, it may be useful to digress a bit into some basic points in organizational theory. Argyris and other constructivist theorists argue that people use ‘‘mental maps’’ to guide their plans, actions, and assessments of situations (Argyris and Schön, 1978).2 These mental maps can be considered theories; they have the characteristics of the- ory—generality, consistency, completeness, testability, centrality, and simplicity. They may explain a particular event: for example, why a class scored poorly on a test. Or they may explain how the class, or the world, works in a more general way: for example, why students in rural areas score lower than students in urban schools. Theories are simpli- fications. They are more or less accurate in explaining how things work and more or less useful in helping people make sense of the world around them. People take action on the basis of theories they hold. Arg- yris and Schön call theories that individuals, or organizations, use to guide their actions as ‘‘theories-in-use.’’

Most theories-in-use are tacit, that is, they do not operate at a con- scious level. Tacit theories enable people to deal efficiently with a host of daily events without questioning causality at each step of the way. Tacit theories determine (and limit) how we see the world (Howley et al., forthcoming). Though most people are generally unaware of their

PAGE 69................. 11172$ $CH3 12-16-04 14:29:24 PS

70 CHAPTER 3

own theories-in-use, theories guide behavior. Tacitly held theories can create difficulties, for they often contain unexamined assumptions. They can color understanding of what’s happening and restrict one’s ability to respond flexibly. Of course, being unconscious makes them difficult to identify, surface, and change. Theories-in-use can be deduced by observation of behavior and critical analysis. If a teacher only asks boys to answer questions in a classroom of mixed gender, one can infer that she does not think much of girls’ ability to respond. There may be other reasons, of course, but she has a theory-in-use that causes her not to call on girls, whatever she says.

At a conscious level, people are aware of their ‘‘espoused theories,’’ the theories they use to explain their behavior, and the explicit values they support. People’s behavior is generally guided by theories-in-use, while explanations and self-understanding are usually governed by their espoused theories. Our teacher, for example, may state, and firmly believe, that she sees girls and boys as equally capable of learning. There is usually a certain amount of discrepancy between people’s espoused theories and their theories-in-use, sometimes more than a lit- tle discrepancy. Argyris and Schön believe that the closer espoused theories are to theories-in-use, the more effective the organization or individual (1978).

Like Argyris, we believe that this ‘‘theory’’ of theories-in-use is use- fully applied to organizations as well as individuals. We would further apply theories-in-use to practices of education reform in general; to the approaches to change of particular education organizations such as schools, ministries, and funding agencies; and to particular initiatives. In the case of general approaches to reform, we might use the term development hypotheses to describe the developmental effects expected to follow from a particular initiative. In the case of a particular inter- vention, we use the term project theory to describe the chain of causes and effects expected to result from project activities. (See chapter 9 for discussion of program theory in the context of monitoring implementa- tion.)

Often, assumptions are not wrong so much as they are too general, in which case they need to be qualified to be useful or the conditions need to be specified under which the assumption will accurately describe a particular situation. Take for example the widespread asser-

PAGE 70................. 11172$ $CH3 12-16-04 14:29:25 PS

THINKING ABOUT THE CONTEXTS AND PROCESSES OF CHANGE IN EDUCATION 71

tion that decentralization will lead to improved school effectiveness. Evidence suggests that this may be the case, or it may not. Decentral- ization may lead to improved school effectiveness, for example, if decentralization provides schools with the resources and authority to make instructionally sound decisions (and if effectiveness is well mea- sured by scores on cognitive achievement tests). However, this desir- able outcome may depend on the capacity of schools to make good instructional decisions. Thus, while the theory that decentralization will lead to higher student performance is not exactly wrong, it is somewhat misleading as a guide to action. Surfaced, it can be exam- ined, and refined into a more useful theory, which can then be used, tested, and further refined. In this process, of course, useful lessons are being learned about how to improve schools. Of course, it is also the case that some theories-in-use are simply, and verifiably, wrong. For example, the theory of the earth being flat limited national policy options in Europe and elsewhere for thousands of years. Due to the high stakes of being wrong, very few people, for centuries, were will- ing to test it.

A second set of associated concepts relates to learning. Argyris and Schön see learning as taking place as a consequence of the detection and correction of error. When sensing an error, most people look for an operational strategy that will work within the ‘‘goal-structure,’’ and ‘‘rule-boundaries’’ they commonly use. This is ‘‘single-loop learning,’’ a simple feedback loop, where outcomes cause adjustment of behav- iors, as with a thermostat. (A similar concept, first-order change, was discussed in the introduction.) A thermostat senses an ‘‘error,’’ or gap between actual and desired temperature, and responds by following the rule that says ‘‘turn on the heat.’’ People involved in single-loop learn- ing accept the framing of the problem (not enough heat) and take steps to address the problem by following the conventional ‘‘rule’’ for such situations (turn up the heat). Single-loop learning takes existing goals, beliefs, values, conceptual frameworks, and strategies for granted with- out critical reflection. It may work to increase the effectiveness of the strategies used to deal with the problem, it may even adopt a slightly different strategy for reaching the same goal (within the same rule boundaries), but it leaves most of the existing ways of doing things, including dysfunctional strategies and assumptions, in place. It works

PAGE 71................. 11172$ $CH3 12-16-04 14:29:25 PS

72 CHAPTER 3

fine, when, as in our example, adequate heat is the goal, lack of heat is the problem, and turning on the heat is an effective solution. However, single-loop learning is not effective when the environment changes, and when heat, to continue the example, is not the problem.

In such cases, double-loop learning (discussed in the introduction as second-order change) is needed to modify the governing variable of the system—in this case, seek heat—to some other goal. Double-loop learning involves surfacing and critical reflecting on goals, beliefs, val- ues, conceptual frameworks, and strategies. As such it leads to deep learning and the possibility of deep system change. Double-loop learn- ing is critical in organizations and individuals who find themselves in rapidly changing and uncertain contexts, when solving problems that are complex, ill structured, or that change over time. Double-loop learning, while arguably essential for organizational effectiveness in uncertain environments, is generally discouraged in organizational life. Double-loop learning, by definition, calls into question the status quo and the ways things are conventionally done.

Generally, of course, people lose awareness of most of the governing variables they rely on, and focus their efforts on implementing their existing theories-in-use—doing what they’ve always done, only harder, or again and again. In many such cases, a more effective approach would be to question the goal or the theory underlying their strategy for reaching that goal. Lacking awareness of their own theories of change, policy makers may continue to utilize change strategies long after they have lost relevance or effectiveness to the problem at hand.

Rondinelli (1993) suggests, perhaps playfully, that most develop- ment work ought to be undertaken as policy experiments. Indeed, a considerable body of research and practitioner literature suggests strat- egies such as these for organizational as well as individual develop- ment. Peter Senge, notably, has written numerous books about learning organizations (see, for example, Senge, 1994).

Some faulty theories-in-use might fall into a category we might call ‘‘Theories Where We Don’t Know Any Better’’:

• Doing something necessary is (implicitly) sufficient to improve a system. Provision of quality inputs will (necessarily) lead to improved learning—a common assumption we have discussed. In

PAGE 72................. 11172$ $CH3 12-16-04 14:29:26 PS

THINKING ABOUT THE CONTEXTS AND PROCESSES OF CHANGE IN EDUCATION 73

the case of schools, good textbooks are generally necessary to high levels of student achievement. However, textbooks must be ade- quately utilized to be effective in promoting learning. Thus, provi- sion of textbooks may not be sufficient to improve learning. Even so, how many development projects have focused on curriculum development and distribution, with the expectation that such steps will necessarily result in measurably higher levels of student achievement?

• Schools and teachers can improve student learning (if they want to)—holding schools and teachers accountable will lead to increased learning. A number of education systems have put accountability systems in place to track student achievement and to sanction schools or teachers when students do not score suffi- ciently well. Implicit in such policies is the assumption that teach- ers or schools are able to improve student learning—that teachers and schools have the know-how, the resources, and the organiza- tional wherewithal to improve student acquisition of the desired material. It may be the case that many schools, especially the weaker ones, may not be able. A policy that simply puts account- ability systems in place, but does not provide support to schools in improving, may run afoul on this theory. Of course, failure to meet the specified standards will be understood as failure on the part of teachers and school rather than failure of theory.

In addition, a number of other theories-in-use fall into a category we might call, ‘‘Theories Where We Do Know Better.’’ Though rarely stated as such, these implicit theories can be identified in the design of a surprising number of education interventions. (Of course, the prob- lem with theories-in-use is not a matter of explicit knowledge or val- ues, but of implicit beliefs and resulting behavior.)

• Policy will be implemented as intended. Though rarely stated as directly as this, change processes that include no feedback loops to check the implementation of policy are ‘‘in-use’’ of this theory of change.

• Implementation will (necessarily) follow the policy decision. • There will be no unintended consequences.

PAGE 73................. 11172$ $CH3 12-16-04 14:29:26 PS

74 CHAPTER 3

• People will adopt an idea because it is a good idea. • A project will be sustained because it is a good project. • People will act according to the rationality of the planner’s con-

text. • People will do something because they are supposed to (or because

it is their job, or because an authority told them to). • Authorities, by virtue of their authority, can bring about imple-

mentation of a policy. • Policy is made (only) by ‘‘policy makers’’ and officials at the

‘‘top’’ of the system. • Technical considerations can be separated from politics.

Ideas, even inaccurate ones, live long beyond their originators, as John Maynard Keynes famously observed:

The ideas of economists and political philosophers, both when they are right and when they are wrong, are more powerful than is commonly understood. Indeed, the world is ruled by little else. Practical men, who believe themselves to be quite exempt from any intellectual influences, are usually the slaves of some defunct economist. Madmen in authority, who hear voices in the air, are distilling their frenzy from some academic scribbler of a few years back. . . . Sooner or later, it is ideas, not vested interests, which are dangerous for good or evil.

MORE AND LESS USEFUL MODELS OF THE CHANGE PROCESS

Many of these assumptions are legacies of rational and stages models of the policy process, discussed briefly in the introduction, and of com- mand and control models of education governance. Stages models sug- gest that the policy process is usefully understood as a sequential series of steps, distinct in function and time. Haddad and Demsky (1994) present a typical cycle, composed of eight ‘‘policy processes’’:

• Analysis of the existing situation • Generation of policy options • Evaluation of policy options • Making the policy decision

PAGE 74................. 11172$ $CH3 12-16-04 14:29:27 PS

THINKING ABOUT THE CONTEXTS AND PROCESSES OF CHANGE IN EDUCATION 75

• Planning for policy implementation • Policy implementation • Policy impact assessment • Subsequent policy cycles

Other models vary slightly in the naming and specification of stages. Some models, for example, include more explicit stages of agenda set- ting, whereby policy problems are placed on the national decision- making agenda. (See Grindle and Thomas, 1991, for a useful discus- sion of stages models of policy.)

This model of the policy cycle has a number of advantages. It is use- ful heuristically. Indeed, we use it here to organize the second half of this book. It parallels the individual process of cognitive decision mak- ing and permits analysis of a variety of policies across a diversity of institutional contexts and policy domains. It serves as a checklist of necessary steps and their most rational sequence. The wise planner, for example, will base his/her generation of policy options on a thorough analysis of the current situation, else a key element be overlooked. It is good practice to hold evaluation of policy options until after a complete list is generated; typically, there are more and better options than the first ones that come to mind. And so forth. In these ways, the model directs the policy maker to a next logical step in the change process.

Nonetheless, stages models have a number of limitations in describ- ing and prescribing actual policy formation and implementation proc- esses (Porter & Hicks, 1997): real policy-making is rarely as neat as described. Policies can be derailed at any point, by forces having noth- ing to do with the technical quality of analysis. Policy-making does not necessarily proceed through all stages, nor does it necessarily follow the steps in order. Steps may be skipped, or returned to at subsequent times. Rarely do policy makers generate and evaluate a full range of options. Policy agents may use rational policy processes to justify deci- sions already reached rather than to identify the best technical solution. Then too, a full discussion of the pros and cons of different strategies is counterproductive in a contentious political environment where agreement may be more important to the success of education than the technically best solution. Moreover, while the stages model may accu- rately describe an individual’s decision-making process, it does not

PAGE 75................. 11172$ $CH3 12-16-04 14:29:28 PS

76 CHAPTER 3

describe collective decision making well, which is much more diffuse and less linear. Finally, stages models tend to reinforce a false division of labor, between those who formulate the policy and those who imple- ment it, for example. This is associated with one of the pathologies of education reform efforts, the development of ‘‘good’’ plans that are not implemented. Stages models of policy tend to be front-loaded, with much greater attention devoted to the plan than to implementation and beyond.

Porter & Hicks (1997), building on the work of Kingdon (1984), suggest that it is more useful to think of the policy process in terms of multiple ‘‘streams.’’ At any given time, policy elites can focus on only a limited number of issues, hence the useful concept of policy agenda. Given the many potential issues that could be targeted, a policy is much more likely to get onto the agenda and to be acted upon if it represents a viable solution to a politically important problem. (As such, the pol- icy can be seen as lying at the intersection of three somewhat indepen- dent streams, the stream of problems, the stream of solutions, and the stream of politics.) Many social conditions, Porter and Hicks point out, exist, without being considered policy problems. Girls may drop out of school due to pregnancy, for example, and this can be experienced as private tragedy, or it can be defined as a policy issue by being linked to public awareness of the issue as one about which action can and should be taken, and by government.

Like problems, there are many solutions floating around in the solu- tions stream, most of which are not acted upon. Solutions, often gener- ated by policy researchers, attract attention when they are linked with problems. Both problems and solutions, however, are unlikely to assume priority unless linked with salient politics. Like problems and solutions, politics typically moves independently in its own stream according to its own logic and imperatives. Indeed, each stream has a life of its own. However, when the three streams converge on a given policy issue, action is likely.

Like all models, the streams model is a more or less useful simplifi- cation. It accounts well, we feel, for the selective rationality of govern- ment systems in deciding on and implementing policy. At the same time, it allows for the multiple rationalities of politics, policy solutions, and potential policy problems. The streams need not flow in any partic-

PAGE 76................. 11172$ $CH3 12-16-04 14:29:28 PS

THINKING ABOUT THE CONTEXTS AND PROCESSES OF CHANGE IN EDUCATION 77

ular sequence, a realistic portrayal, given the way policies are actually formulated and adopted in organizations. The streams model recog- nizes the fluid yet understandable nature of the policy process. Ideas are important but not determinative of the outcomes of policy proc- esses. Bad ideas as well as good may be adopted. Individuals and groups, inside or outside of government, can play an important role in bringing about policy change, by focusing attention and linking prob- lems with solutions to political issues. Indeed, the authors spend con- siderable time discussing strategies that policy entrepreneurs can adopt to influence the policy process.

Thus while we organize our discussion according to a series of ratio- nal analytic steps, the process of policy change itself may be more use- fully understood as an advocacy process (even if one is advocating a policy one feels, through dispassionate scientific analysis, best meets the needs of the nation). Viewed in this way, bringing out policy change is a matter of identifying and acting on leverage points.

Finally, we see a parallel set of problems with common conceptions of agency in the change process. Command and control models of change underlie many strategies undertaken by government. At their most severe, such models assume complete passiveness, compliance, and capacity, on the part of field implementers. Managers have perfect knowledge of what needs to be done at all levels of the system. Their job is to specify tasks and monitor compliance. Field implementers, for their part, are required (and are both willing and able) to do what they are told. (See Table 3.1.)

In fact, these ‘‘ideal type’’ conditions rarely apply to any human organization, and especially not schools, where the primary task of instruction is one that cannot be specified well from afar. School sys- tems, as Weick famously observed, are ‘‘loosely coupled systems,’’ and orders from the top have only modest direct effects down the line (Weick, 1976). As a result, teachers have considerable discretion in how enthusiastically they implement central directives, especially those where judgment is required.

A further problem is that it is unclear whether central administrators have the information needed to make all education decisions. While central administrators may see the big picture, teachers have the best information about the conditions facing the children in their classroom.

PAGE 77................. 11172$ $CH3 12-16-04 14:29:29 PS

78 CHAPTER 3

Table 3.1 Two Approaches to Implementation of Change in Complex Systems

Aspect/Dimension Command & Control More Participatory

Aliases Top-down; hierarchical Bottom-up; top-down and bot- tom-up

Type of Planning Forward Backward

View of Reform Stages Nonlinear, at least recursive

Metaphor for Plan Blueprint to be followed Structured negotiations

Metaphor for Hierarchy, machine, Forum; herding cats Organization bureaucracy

Ideal for implem- Obedience, capacity to Professional judgment, capacity enters carry out tasks; compliance to move toward goals in local

context; discretion

What implementers Clarity & training Discretion & support need from central authorities

Role of Center Enforcement Support; Direction vis-à-vis Plan & Implementers

Causal effect of Deterministic Influential, but contingent on factors affecting transactions implementation

Approach to Develop a good plan and Transactional; develop good implementation follow it process

Approach to Monitor for compliance and Monitor expected inputs, activi- Monitoring adherence to details of plan ties, intermediate outcomes,

unexpected outcomes; monitor critical transactions; monitor to see inputs have arrived, if activi- ties are taking place as expected, if intermediate out- comes are resulting as expected, what unexpected out- comes have resulted, if theory holds; monitor progress toward goals of reform; monitor ‘‘ongo- ing intelligence’’

Understanding of Mistakes; Errors Possibly adaptive, possibly not; Deviations from plan possible innovations,

potential source of knowledge about what works where & when; possibly problem with people, possibly a problem with plan, process, or in context

PAGE 78................. 11172$ $CH3 12-16-04 14:29:29 PS

THINKING ABOUT THE CONTEXTS AND PROCESSES OF CHANGE IN EDUCATION 79

Table 3.1 (Continued)

Aspect/Dimension Command & Control More Participatory

Dealing with devia- Tighten accountability, See what can be learned; check tions from plan measures of central con- theory and actors; balance

trol; reporting, oversight accountability with support, par- ticipation, discretion

Possible (Sometimes) Possible continued deviations consequences of Defensiveness; denial, from plan (due to weaker strategies of dealing unwillingness to admit fail- accountability) with deviations ure; diminished capacity to

learn from mistakes

Management Funding formulas, formal Knowledge, problem solving on Strategies organizational structures, part of implementers, incentive

Authority, relationships, structures, bargaining among regulations, administrative key groups, strategic use of controls—budget, planning, funds (Elmore, 1982: 22) evaluation (Elmore, 1982: 22)

Works well when Goals are clear, consistent, There is some area of (poten- enjoy broad support; pro- tial) agreement, a goal to be gram theory is accurate; achieved; different groups are implementers are commit- willing to negotiate; skillful facili- ted to change, capable; tator having credibility with dif- organizational conditions ferent groups are favorable; reform is supported by interest groups; contextual condi- tions remain favorable; when central government has power, and little con- sensus possible among powerful groups of stake- holders

Source: Elmore, 1982

The technologies of instruction are not well enough understood to spec- ify with any precision. The improvement of instruction, from this point of view, is one of enabling teachers to make the best decisions, based on their professional experience and knowledge, according to the changing classroom conditions. Finally, given the frontline nature of teachers’ work, the ‘‘bottom’’ of the school system would seem a potentially fertile source of policy ideas. So, given the professional nature of teaching and the loosely coupled nature of schools, central

PAGE 79................. 11172$ $CH3 12-16-04 14:29:30 PS

80 CHAPTER 3

administrators need a more sophisticated system of acquiring teachers’ cooperation, their compliance, and their ideas.

A number of observers see changes in the role of government, as governments decentralize and work with ‘‘partners’’ outside the minis- try. Adams (2000) characterizes these shifts in several ways. The role of government is shifting from financier, designer, and evaluator of education development to catalyst and partner with local organizations. There is a corresponding shift from control and compliance to policy coordination, management, and monitoring of regulations. The purpose of policy planning shifts accordingly, from a primary focus on the allo- cation of public resources to public/private coordination, and the coor- dination and mobilization of interest groups. The focus of strategic planning shifts from control to setting a national vision, and providing leadership in equalizing services. Management support and adminis- tration shift from detailed monitoring of administrative processes to monitoring of national standards, and support of research and develop- ment. Government’s role shifts from provider of (all) services to facili- tator of good services.

All of this suggests that a more participatory, professional process of policy development and implementation may better help education systems reach many of their goals. The complex goals and technologies associated with education suggest that teachers are much more useful as active, intelligent agents working toward and helping define national goals than they are as passive implementers of policy decisions reached at the center. School-based reform models suggest that such a learning process is best fostered by neither the directiveness of centralized and regionalization models of reform, nor the lack of guidance implicit in decentralized-sectoral reform strategies, which provide the incentives but not the technical assistance. What is needed is a more engaged process of instructional coaching. Such a perspective suggests that central ministries shift their orientation from administration (enforcing rules and regulations) to management (maximizing resource use to achieve goals) to leadership (mobilization to define and achieve higher- order goals).

Finally, dynamic systems theories such as complexity theories or chaos theory caution against too optimistic an approach to change (Howley et al., forthcoming; Wheatley, 1994). Schools and school sys-

PAGE 80................. 11172$ $CH3 12-16-04 14:29:30 PS

THINKING ABOUT THE CONTEXTS AND PROCESSES OF CHANGE IN EDUCATION 81

tems, as ‘‘complex adaptive organizations’’ are inherently unpredict- able. Causality is difficult to trace, as direct linear relationships are rare. Adaptive systems cannot be controlled. However, would-be change agents can take advantage of ‘‘leverage points’’ to attempt to influence the direction of change. The Internet, as a common example, cannot be directed, but it can be influenced. In such a system, a blue- print approach to planning is much less useful than a more interactive strategy, with broad goals set but with flexible strategies that shift according to changing conditions. Everything we know about the reform process suggests that no plan can anticipate every development. Too much is unknown. There are too many variables outside the control of ministries of education. More useful than a specification of every detail is general agreement on broad goals, the political will to change, and, as we have argued elsewhere, a process that structures information and participation in negotiating, understanding, and moving toward goals.

SUMMARY: REFORM AS LEARNING

This chapter has discussed a series of micro-contextual factors likely to affect the reform process. We have discussed the importance of understanding ‘‘local context,’’ specifically, the contexts of implemen- tation and of institutionalization, as well as that of policy formulation and planning. Planners eager to institute system reform should pay attention to the sociopolitical context in which the reform will live, attending to the internal and external political, historical, cultural, and organizational factors likely to be relevant. Policy makers have also been urged to surface and test their theories of change and causality in education, to adopt a learning, as opposed to a knowing, or performing, stance. We have discussed ways the change process is commonly, and can more usefully be, understood, both in terms of how change takes place and how it is directed. Rather than a series of sequential, rational steps, policy adoption is better understood as the convergence of prob- lems, solutions, and politics. Change agents can identify leverage points to bring about such convergence.

Rather than a top-down bureaucratic system of command and con-

PAGE 81................. 11172$ $CH3 12-16-04 14:29:31 PS

82 CHAPTER 3

trol, education systems ought to work better as more participatory organizations supporting the work of professional teachers, using an appropriate mix of carrot and stick, accountability and support, identi- fication and mobilization around common goals. Education systems might operate more effectively as learning organizations rather than as knowing, declaring, or performing systems. Given the unpredictability of reform, front-loaded designs are likely less useful than agreement on the need to change and on the broad outlines of reform, and a legitimate process to organize discussion by stakeholders around values, the state of the system, the ways change can be promoted, and the desired direc- tion and nature of reform.

NOTES

1. See Current Issues in Comparative Education (vol. 6, no. 2: Participa- tory Development: A Promise Revisited) for a critical discussion of this issue.

2. See also the relevant discussion of frame-shifting in Bolman and Deal, 1997.

PAGE 82................. 11172$ $CH3 12-16-04 14:29:31 PS