Current theorizing in education, like that in industry, is largely devoted to explaining trial-and-error, failure-tolerant, low-reliability organizations. This has been appropriate for the industries and schools of the past. However, the current generation of demands placed on education are, at heart, that all students must succeed in school. This is a demand that is novel in the history of U.S. education.
In some areas of modern culture, institutions and technologies have emerged that have great productive as well as destructive powers. Quoting LaPorte and Consolini (1991): Increasingly, any failure of these technologies is perceived by both their operators and the public to have such potentially grave consequences as to warrant the absolute avoidance of failure (p. 19). The issue is not whether the technology can usually function correctly; that is a given. Rather, the requirement is that the technology never totally malfunction. The issue is not validity, which is assumed, but extraordinarily high reliability. In order to educate those U.S. students who are now deemed "at risk," in order for those students to achieve the educational accomplishments that are already being experienced by many students, we will need a similar increase in the reliability of schooling for U.S. children. We will need to create school systems that exhibit an absolute avoidance of student failure.
In several areas deemed critical to the public interest, such as the operation of nuclear powerplants and aircraft control towers, new types of organizational structures have evolved to meet the requirements of virtual 100 percent reliability. These organizations are required to engage not in trial-and-error improvement, as is common both to much of industry and management. Rather, the organizations are expected to operate "trials without errors" (LaPorte and Consolini, 1991, p. 20). Researchers investigating the characteristics of these systems have referred to them as "high reliability organizations (HROs)" (Roberts, 1990).
If one accepts the National Education Goals, then it is no longer acceptable for significant numbers of students to not learn "the basics", or for only many students to learn them well. It is no longer acceptable for large numbers of students to drop out of school or to remain and be provided with a substandard curriculum. The costs to individuals and to society, once low, have become too high. Schools are no longer afforded the luxury of blaming the students and their families for students' failures. Schools are now seen as accountable for the successes and failures of virtually all their students. In order to respond to these new realities, schools and school districts will have to abandon industrial efficiency models and take on the operating characteristics of HROs. Failure to do so will cripple any education reform efforts designed to aid students at risk.
An overview of the primary characteristics of HROs (adapted from Pfeiffer, 1989; Roberts, 1990; and LaPorte & Consolini, 1991) follows. Each characteristic is followed by this author's impressions of the status of U.S. education on the dimension at present.
Establishing and maintaining clear goals has been one of the most frequently cited characteristics of the school-effects research base (Edmonds, 1979; Good & Brophy, 1986; Levine & Lezotte, 1990). Stringfield and Teddlie (1988,1991) find that a cacophony of most important goals often resulted in a lack of clear, unifying goals for the faculties and students in low-achieving schools. Creating a strong set of agreed upon goals across a school faculty takes time, a lot of listening, and occasional firmness on the part of school leadership.
HROs extend formal, logical decision analysis, based on standard operating procedures (SOPs), as far as extant knowledge allows. At the time of the publication of the first Handbook of Research on Teaching (Gage, 1963), there was not a sufficient body of educational research to guide the development of standard operating procedures for schooling in any rational way. In the absence of such knowledge, and in the absence of clear national or state goals, districts developed local procedures. Districts, local superintendents, and principals often abdicated these responsibilities to teachers, with virtually no monitoring of resulting classroom practices. By the publication of the third Handbook (Wittrock, 1986), the research situation had changed considerably. The rudiments of a science of education now exist. However, practice has been slow to follow.
An important link between school-effects research and HRO literature is provided by two studies of teachers behavior in more and less effective schools. Both Mortimore et al. (1988) and Teddlie et al. (1989) find that teachers in high-outlier schools (schools in which students performed much higher on achievement tests than did most schools serving communities with similar economic and ethnic characteristics) behaved in manners more like those predicted by the teacher-effects literature (Brophy & Good, 1986). Perhaps as importantly, these studies found greater consistency among teachers in more effective schools. Teachers were more likely to be moving students through their lessons at a good pace, and fewer teachers allowed high rates of time off task in effective schools. Stringfield and Teddlie (1991) find, for example, that schools in which many students are performing poorly, some teachers are getting no more than half way through their math texts by the end of the school year. This poor performance means that the next year's teacher faces the extraordinarily difficult task of bringing many students up to grade level. The authors find that this higher rate of consistency on the part of teachers was clearly related to the behaviors of the principals, and not related to the income levels of the communities being served. Some principals are already insisting on, and getting, relatively high reliability in instructional delivery.
HROs recruit and train extensively in order to compel adherence to standard operating procedures. Yet at peak times, professional judgment is valued. Kozol (1991) repeatedly observed that children at risk need the highest quality teachers, yet often do not get them.
There is very little research on teacher, principal, and superintendent recruitment, and essentially none on the effects of recruitment. (For an initial effort that deserves multiple replications, see Wise et al., 1987). Research on the how to of effective training has made much greater progress (Showers, Joyce, & Bennett, 1987; Lewis & Miles, 1990; Fullan, 1991). For staff development to change the instructional patterns of teachers, it must include presentation of theory, modeling, time for practice, and timely, individualized feedback. There are three important points here. First, these findings have face validity. Any adult who has learned to ski, or play tennis, or to use a new computer software package can quickly recognize that these are very nearly the steps they took toward incorporating new skills. Second, these steps are often assumed in other fields. Resident surgeons do not learn new procedures just through reading a book or attending an inservice session after work. They observe their colleagues, and then are repeatedly observed performing the new task and provided with "real time" feedback. Third, we have extraordinarily limited experience with such change-bearing staff development in education. The 1-day workshop is the norm. Multiple-day workshops, spread across a semester or year, are becoming more common; but the large-scale work of coordinated, in-class modeling and feedback must necessarily be the work of principals, program facilitators, and within-school peers. These are not nearly frequent enough events in U.S. education today.
Ward and Tikunoff (1989) and Kirby et al. (1992) have conducted detailed studies of teacher induction programs, both of which found school-level effects. However, much more research on the long-term effects of various induction and staff development programs is needed.
Stringfield and Teddlie (1991) found that principals in positive-outlier schools were more likely to take an intense interest in staff recruitment. They found that, by contrast, principals in low-outlier schools passively accepted what they [central administration] send us. Stated differently, where pairs of schools that were serving economically and ethnically similar families were obtaining markedly differing student achievement rates, the school with the higher achieving students almost invariably paid much greater attention to recruitment and evaluation of teachers. This was particularly true in schools that served high numbers of students at risk.
HROs have initiatives that identify flaws in standard operating procedures, and nominate and validate changes in procedures that prove inadequate. The author is unaware of large-scale research on school- or district-level efforts at systematic, organizational efforts to identify flaws within schools and correct them. However, Mortimore et al. (1988) and Stringfield and Teddlie (1991) describe principal and staff actions in more effective schools that might serve to identify problems within schools and facilitate correction. Stringfield and Yoder (1992) present a case study of an exemplary school serving highly disadvantaged Hispanic children. That school consistently developed standard operating procedures and with equal frequency sought out methods for improving those procedures.
The 1988 Hawkins-Stafford amendments to the U.S. compensatory education laws mandate a self-study and program improvement process for schools in which compensatory education students are not making adequate academic gains (LeTendre, 1991). Thousands of schools have been identified nationwide (Heid, 1991), and in several states those schools have entered into an extended self-study with the eventual goal of school-directed improvement (see Stringfield, Davis, & Billig, 1991).
Most visitors to "light house" schools have seen schools and, less often, districts in which teachers had clear, open invitations to provide corrective feedback regarding flaws in schooling systems. They may have seen clear effects from that feedback. However, it is probably safe to say that the typical U.S. teacher does not view herself as having open access to processes that could change significant school and district procedures. Note that the very existence of "light house" schools within districts speaks to the acceptance of low reliability in the implementation of schooling in U.S. education. If schools and districts are to reliably educate virtually all students at risk, teachers and principals must have more voice in establishing and modifying school procedures.
HROs are sensitive to the areas in which judgment-based, incremental strategies are required. They therefore pay considerable attention to performance, evaluation, and analysis to improve the processes of the organizations. U.S. education is currently undergoing increased emphasis on performance-based evaluation of teachers (e.g., Millman, 1981, 1989). Bridges (1986) has conducted a scholarly analysis of procedures for managing incompetent teachers. Bridges work demonstrates that effective action is possible, though rarely undertaken. The author concedes that the majority of incompetent teachers are allowed to continue to intervene in the educational lives of children. Obviously, this derails all efforts at raising education's reliability.
Stringfield and Teddlie (1991) report that principals in more effective schools took teacher recruitment, development, and evaluation more seriously than did principals in negative-outlier schools. Most of the principals in the positive-outlier schools had counseled-out, forced the transfer, or otherwise removed one or more teachers from their staffs. In some schools, that rate of moving out lower performing teachers who decline repeated opportunities to modify their instructional performances exceeds 50 percent (e.g., C.de Baca, Rinaldi, Billig, & Kinnison, 1991). That was rarely true of principals in negative-outlier schools.
In HROs monitoring is mutual (administrators and line staff) without counterproductive loss of overall autonomy and confidence. In the United States, there is currently minimal evaluation of teachers by administrators or fellow teachers, and virtually no evaluation of administrators by teachers. Both Rumberger and Larson (in press) and Fine (in press) present systems in which peers become much more aware of each other's teaching. Team meetings regarding steps to accelerate an individual student's progress, team planning by teachers, team teaching, and peer observation are all steps toward opening up the process of teaching. Anonymous teacher evaluations of principal performance are rare, but I have visited a few schools which gather and aggregate this data. In almost every case, the principal said that this new procedure was initially painful but ultimately productive. This is a step toward institutional reliability.
HROs are alert to surprises or lapses. The experience of HROs is that small failures can cascade into major system failures, and are hence monitored carefully. The author is unaware of studies focused specifically on schools responses to surprises or lapses. Mortimore et al.'s (1988) purposeful leadership, maximum communication between teachers and students, and record keeping, all found to be characteristics of more effective schools, would logically lead to quicker responses to surprises or lapses. Similarly, Stringfield and Teddlie's (1991) attention to daily academic functioning, as a school-level predictor of school effects, would logically lead to quicker discovery of and response to lapses. The teaming in Philadelphia's Charter Schools or in several Coalition of Essential Schools projects almost certainly leads to quicker identification of problems and responses to lapses. The quarterly, curriculum-specific testing is a component of Success for All (Slavin et al., 1992) and provides an example of steps that can greatly decrease opportunities for cascading errors in students' educations.
In many schools serving students at risk, significant percentages of students arrive in second and even third grade unable to read. It is not plausible that previous teachers have not noticed these deficiencies. Yet, it is obvious that adequate remediation had not been instituted. Such programs as Success for All (Slavin et al., 1992) and Reading Recovery (Pinnell, 1988), when successfully implemented, show that these situations are in no way inevitable, rather they are allowed to happen. For the error of not teaching a child to read to be allowed to cascade through several grades, a school must be operating at a low reliability level. We will not achieve the National Education Goals with such performances. High reliability organizations point to a different structure and results.
HROs are hierarchically structured, but during times of "peak" loads, they emphasize a second layer of behavior that emphasizes collegial decisionmaking regardless of rank. This second layer is characterized by cooperation and coordination. At times of peak activity, line staff are expected to exercise considerable discretion. U.S. schools are hierarchically structured. Expectations during peak times have not been systematically studied, but probably vary greatly. The author's impression is that during peak times at some schools, accommodations are made, everyone pitches in, and important functions continue to operate. By contrast in some schools, temporarily heavy loads become excuses for nonperformance by key staff members. Things break down. To educate all students at risk, systems must be in place that rarely break down and invariably catch the breakdowns and correct them. HROs rely on the competence of all professionals in the facility and distribute responsibility accordingly.
HROs regularly respond to potentially disastrous situations as being far too important to trust to rules alone. Authority patterns shift from hierarchical to functional-skill based authority, as needs arise. Large U.S. school districts generally, and special education programs in particular, are rule-focused and rule-driven. In some schools and districts, exceptions to rules are almost never tolerated. This is often to the short- and long-term detriment of specific children. Schools and districts must be responsive to the needs of students who might fall through the cracks in a rule-driven system. Madden et al. (1991) offer examples of systems designed so that professionals "catch" students who might fall behind.
Especially during times of peak performance, staff are able to assume a close interdependence. Relationships are complex, coupled, and sometimes urgent. A high level of coordination between compensatory and regular classrooms was found to be a characteristic of more effective compensatory education programs in the United States (Griswold, Cotton, & Hansen, 1985; Allington & Johnston, 1989). Mortimore et al. (1988) find that involvement of teachers in decisionmaking and consistent teacher inservice programs were both related to school effectiveness. Stringfield and Teddlie (1991) also report that high levels of cross-classroom and cross-grade coordination were positive predictors of school effects. Professional isolation is the enemy of any program trying to truly serve students at risk.
Equipment is maintained and kept in the highest working order. Responsibility for checking the readiness of key equipment is shared equally by all who come in contact with it. This is not true in most U.S. schools, where nonfunctioning equipment often sits unrepaired for months or years. A striking feature of equipment maintenance in many schools is the extent to which unnamed persons beyond the school are often viewed as being responsible for maintenance. This situation may in part be a result of many educational administrators viewing no equipment, beyond heaters and an office telephone, as key and deserving of their personal attention.
HROs are invariably valued by their supervising organizations. There is some evidence (e.g., Wimpelberg et al., 1989) that school districts provide more attention and support to some schools than others, and that it is often the schools in the least advantaged neighborhoods that receive the least attention and local support. Stringfield and Teddlie (1988) observe that some school districts, rather than retraining or terminating manifestly unskilled teachers and principals, assign them to schools serving the least advantaged families. Such actions do not indicate high valuing of all students and schools by central offices.
The accountability requirements of the Hawkins-Stafford Chapter 1 Amendments of 1988 have led some districts to pay increased attention to schools serving higher percentages of students at risk. Winfield (1991) reports that in order to facilitate a shift to schoolwide- project use of federal compensatory education funds for schools serving students 75 percent or more of whom receive free lunches, one large school district discontinued its practice of requiring some schools to accept the involuntary transfer of probationary teachers. This would appear to be evidence of increased valuing by central administration of the schools serving high percentages of students at risk.
Short-term efficiency takes a back seat to high reliability. U.S. education has spent much of the last 30 years attempting to become more efficient, and much public dialogue concerns ridding education of wasteful management. For the Charter Schools to succeed, Fine notes that teachers need to work together over time. The school district undermines these efforts when it bounce[s] teachers out of the Charter because the school has a momentary drop in enrollment (1992, p. 10). If the goal is restructuring, some short-term efficiencies must be temporarily set aside.