Neither top-down nor bottom-up strategies for educational reform work. What is required is a more sophisticated blend of the two. In this paper I examine the problem in three ways. First, I review briefly some evidence that corroborates the proposition that neither centralized nor decentralized change strategies work. Second, I present the conceptional and empirical case that a blend of the two strategies is essential. Finally, I consider two levels of the problem - school-district, and school/district-state - to illustrate how simultaneous centralized-decentralized forces can be combined for more effective results. Thus, centralized and decentralized are relative terms that can be applied at any two adjacent levels of hierarchical systems.
As dissatisfaction with failed implementation grew in the 1970s, states and districts turned more and more to mandatory solutions. Corbett and Wilson's (1990) study of the impact of compulsory statewide testing in Maryland and Pennsylvania is a case in point. They found that new statements testing requirements caused action at the local level, but in ways that narrowed the curriculum and created conditions adverse to reforms:
Corbett and Wilson also identified other unintended consequences including the diversion of attention and energy from more basic reforms in the structure and practice of schools, and reduced teacher motivation, morale, and collegial interaction necessary to bring about reform. They conclude: "when the modal response to statewide testing by professional educators is typified by practices that even the educators acknowledge are counterproductive to improving learning over the long term, then the issue is a 'policy making problem'" (p. 321).
One a more sweeping scale, Sarason (1990) argues that billions of dollars have been spent on top-down reform with little to show for it. Sarason observes that such reform efforts do have an implicit theory of change:
Political impatience and expediency are as understandable as motivators, as they are ineffectual as strategies for educational reform. Governments can't mandate what mattes, because what matters most is local motivation, skill, know-how and commitment. As Goodlad (1992) observes: "top-down, politically driven education reform movements are addressed primarily to restructuring. The have little to say about educating" (p. 238).
In short, centralized reform mandates have a poor track record as instruments for educational improvement.
The failure of centralized reform has led some to conclude that only driven reform can succeed. Site-based management is the most prominent current manifestation of this emphasis. So far, however, the claim of superiority of grass-roots initiatives is primarily theoretical. In reviewing evidence on site-based management in The New Meaning of Educational Change, I concluded that restructuring reforms that devolved decision making to schools may have altered governance procedures but did not affect the teaching-learning core of schools (Fullan, 1991, p. 201). The evidence continues to mount.
Taylor and Teddlie (1992) draw similar conclusions in their study of the extent of classroom change in "a district widely acclaimed as a model of restructuring" (p. 4). They examined classrooms in 33 schools (16 from pilot schools that had established school-based management [SBMI programs and 17 from non-pilot schools in the same district). Taylor and Teddlie did find that teachers in the pilot schools reported higher levels of participation in decision making, but they found no differences in teaching strategies used (teacher directed instruction, low student involvement in both sets of cases dominated). Further, there was little evidence of teacher-teacher collaboration. Extensive collaboration was reported in only 2 of the 33 schools and both were from the low participation (non-pilot) schools. Taylor and Teddlie (1992) observe: "Teachers in this study did not alter their practice ... increasing their participation in decision-making did not overcome norms of autonomy so that teachers would feel empowered to collaborate with their colleagues" (p. 10).
Other evidence from classroom observations failed to indicate changes in classroom environment and student learning activities. Despite considerable rhetoric and what the authors saw as "a genuine desire to professionalize teaching ... the core mission of school seemed ancillary to the SBM project" (p. 19). Substantive changes in pedagogy (teaching strategies and assessment), and in the way teachers worked together on instructional matters, proved to be elusive. These findings would not be as noteworthy, claim the authors, except for the fact that "the study occurred in a district recognized nationally as a leader in implementing restructuring reforms" (p. 16). Similarly, Hallinger, Murphy, and Hausman (1991) found that teachers and principals in their sample were highly in favor of restructuring but did not make connections "between new governance structures and the teaching-learning process" (p.11).
Virtually identical findings arise in Weiss's (1992) investigation of shared decision making (SDM) in 12 high schools in 11 states (half were selected because they had implemented SDM; the other half were run in a traditional principal-led manner). Weiss did find that teachers in SDM schools were more likely to mention decision about the decision-making process (e.e., composition of committees, procedures, and so on), but "schools with SDM did not pay more attention to issues of curriculum that traditionally managed schools, and pedagogical issues and student concerns were low on the list for both sets of schools" (p.2).
Similar findings were obtained in the implementation of the Chicago Reform Act of 1989. In essence, this legislation shifted responsibility from the Central Board of Education to Local School Councils (LSCs) for each of the city's 540 public schools and mandated that each school develop School Improvement Plans (SIPs). The LSCs by law consist of 11 or 12 members (six parents, two teachers, tow community representatives, the school principal - and, in the case of high schools, a student). Easton (1991) reports that the majority of elementary teachers said that "their instructional practices had not changed as a result of school reform and will not change as result of SIP" (p.41).
In sum, decentralized initiatives, as far as the evidence is concerned, are not faring any better.
Given the absence of any clear superiority of top-down or bottom-up strategies, two patterns, both ineffective, persist. One pattern resolves the dilemma through the false clarity of ideological preference. Many of those in positions of authority opt for centralized reform - "almost always egregiously indifferent to the role of obstacles", says Goodlad (1992:238). Advocates of decentralization, similarly (although from a different ideological perspective), push ahead with site-based management as an end in itself.
The other pattern, of course, rests on ambivalence about which way to go, usually resulting in flip-flops or swings from top-down to bottom-up emphasis. Both strategies are often pursued simultaneously, but in a completely disconnected manner. Rowley's (1992) case study of school district restructuring covering a 12-year period is instructive and I expect represents a familiar story. Sequoia Valley School District in California engaged in a major restructuring effort in the early 1980s following the appointment of a new superintendent in 1979. By 1985, the district had created a mission statement and a comprehensive strategic plan. The superintendent and the board "adopted a philosophy of school-based management". Over time, however, the board became dissatisfied with the uneven development and fragmentation of effort. The superintendent and board began to establish a number of particular programs with external funds and consultants and small groups of teachers and administrators in such areas as "whole language", "early childhood centers", "cooperative learning", and so on and so on. Observes Rowley (1992):
The result, not uncommon, was the appointment in 1990 of a new superintendent known for advocacy of "tight school-based accountability and multiple methods of assessing student performance" (p. 28). Outcome-driven education became just the latest in the ebb and flow of district approaches. Because of its imposition and seeming incompatibility with preferred instructional approaches in many schools, conflict increased sharply. Within a year, doubts were being expressed by the board about the new superintendent's leadership style and its adverse effect on the morale of teachers and administrators.
In conclusion, the whole matter of the relative roles and relationships of centralized and decentralized strategies for educational reform is a morass, badly in need of conceptual and strategic clarification.