The most important shift in considering HROs, however, is not in the specific characteristics of the organizations. Rather, it is the intellectual shift that must precede the evolution of the characteristics. The most important characteristic of HROs is a perception, held by the public and the employees, that failures within the organization would be disastrous. Bhopal was disastrous. Three Mile Island was very nearly disastrous. Schools serving children at risk will increase the reliability of their work as the public and the professionals working within the schools come to view children's not learning to read as constituting individual travesties and public disasters. HROs evolve; they do not arrive complete. They evolve because the public demands them and pays for them.
Throughout the history of public education, the responsibility of schools has been to provide instructional processes to those deemed by the schools to be willing and able to benefit from schooling. It was presumed that others, including students who would now be described as at risk," would drop out. The schools systems became efficient at achieving these demands, but the goal was efficiency, not reliability.
Gradually, over the past several decades and rapidly today, the mandate to educators has changed. The new mandate is that schools be responsible for producing high achievement levels among all students. Given that schools have previously demonstrated competence at educating many students, the new issue is one of much higher reliability.
To date, the field of school-effectiveness research has focused much of its energy on identifying variables that explained differences between schools in which students scored higher or lower on achievement tests, attended school with greater or lesser regularity, and so on. The practical implications of this research were assumed to follow naturally: If researchers could determine what the best schools had, and give it to all schools, all students achievement levels would rise.
The high reliability organization literature offers support for some of the findings of previous research. In particular, it supports the familiar characteristics of clear goals, attention to evaluations, and close coordination. However, the HRO literature suggests several other areas of needed changes that may be necessary for schools to serve all students well. Those areas clearly merit practical attention and academic study. The development and maintenance of standard operating procedures where appropriate, the extensive use of staff development, the importance of open lines of communication, mutual monitoring, alertness to surprises or lapses, and the maintenance of equipment may be among the most important factors affecting the quality of educational services provided to students at risk.
In addressing these areas, schools, districts, and the federal government would simultaneously address several of the issues raised in the preceding chapter. By opening communication and increasing two-way accountability, schools would get more accurate pictures of their unique needs.
By loosening regulations and increasing a focus on staff development regarding the use of $6 billion per year in Chapter 1 funding, the federal government would increase the value of human capital serving students at risk in Chapter 1 schools. By modifying the "supplement not supplant" rule so as to require coordination among service providers, Chapter 1 could reduce professional isolation, particularly as it relates to service for individual children at risk.
By entering into a greatly expanded program of national research on teacher, school, and program effects, the federal government could greatly assist local districts that seek proven options for improving services to children at risk. By funding the continued development and expansion of programs that have been proven to work, and by providing independent evaluations and cost-analyses for implementation and institutionalization of the programs in various contexts, the federal government could provide an invaluable tool as schools seek not just "programs that work," but programs that are well-suited to the particular conditions faced by a specific community and school.
Two areas suggested by HRO researchers seem particularly worthy of future research on the education of at-risk students. These are clear evidence of valuing of schools by central offices, and the primacy of high reliability over short-term efficiency.
There is no systematic literature on the role of central offices in creating school effects. Wimpelberg (1989) makes a first approach at defining a central role for the central office and Fullan (1991) notes several features of districts and higher levels of government that facilitate innovation. The HRO literature would suggest that unless the central office places considerable value on the performance of students within each school, long-term school effectiveness is unlikely.
If schools are being increasingly counted on to produce virtually universal high rates of student achievement, if, in the words of LaPorte and Consolini (1991), increasingly, any failure of these technologies is perceived both by their operators and by the public to have such potentially grave consequences as to warrant the absolute avoidance of failure," (p. 19) then over time all schools will be asked to become HROs. This will be most difficult for schools serving large numbers of students at risk. Research on school and program effects now demonstrates overlap with the HRO literature. This suggests a systemic path toward progress.
An important policy-relevant conclusion concerns fiscal support for public schooling. HROs cost more than traditional industrial plants. They cost more because they are operating under different demands. Efficiency assumes a certain level of fault tolerance. The greater the tolerance for occasional failures, the greater the possible efficiencies. HROs are designed to achieve trials without errors. If the rhetoric of the National Education Goals is to be taken seriously, then the supporters of that rhetoric need to understand that higher reliability is, in the middle term, more expensive. It will help if our political leaders remember that in the long run it would have cost less to have operated the Three Mile Island nuclear facility reliably than it is costing to clean it. It will cost less to provide highly reliable schools for students at risk than to pay for continued expansion of welfare, police, and prison programs.
We now know enough to improve the reliability of schools for students at risk. Many of the systemic, school-, and classroom-level changes can be described and undertaken. Whether as one nation and as a conglomeration of over 15,000 school districts we can muster the will to fully implement what we know, fund the search for new knowledge, and achieve universally high quality schooling for students remains an unanswered question.