An Educational Experiment We Need
Although the performance of public schools continues to disappoint, scarcely a week goes by that does not see the introduction of some innovative educational idea. The abundance and variety of ideas is astounding. Among those I have read about in these pages and elsewhere are: philosophy for children, school-site budgeting, video-disk science instruction, cooperative learning, writing across the curriculum, institution of a traditional dress code, simulated economies in schools, Afro-centered curricula, differentiated staffing, mentoring by community members, reciprocal teaching, public choice and other forms of deregulation, performance testing, whole-language instruction, corporation-school partnerships, assertive discipline, new curricular frameworks, new approaches to teacher education.
I am far from suggesting that these ideas are worthless. Some in fact sound extremely promising. Nevertheless, the disappointing performance of many of our students--low-income and minority children especially--remains. We do not know why this is so. Perhaps the schools' efforts are simply no match for the hours the young spend watching television, the absence of parental guides, or our society's traditional disdain for intellectual performance. Perhaps the ideas are fine, but degenerate in the hands of the ordinary teachers ultimately responsible for their implementation. Perhaps the variety of efforts attempted in the same school system or the same school work against each other and cancel each other out. Perhaps successful ideas are simply not given enough time to develop and grow roots. Perhaps the successful spread of good ideas is impeded by lack of adequate funding. Perhaps underlying structural features of our society, racism and inequality of social and cultural capital, doom efforts to improve schools for all children.
The limited success of creative, often heroic efforts prompts three responses: the proposal of ever bolder ideas, such as "bankrolling educational entrepreneurs," mentioned in a Dec. 5, 1990 Commentary; renewed calls to return to the educational ways of our ancestors; or simply a sense of defeat and resignation. Is such resignation warranted? I do not know but I propose that we try to find out.
Almost all school-improvement efforts are bedeviled by one constraint that, while not without reason, ought to be removed in order to conduct an experiment with impact. This is the requirement that the level of funding per child in the schools that innovate not exceed by very much the level of funding at the neighboring school or school system. Any proposal that would, let's suppose, require that spending per child in an experimental school be increased by a factor of five would be summarily dismissed. The reason is obvious; it would be patently unfair, not to say politically infeasible, to permit some children to receive so disproportionate a share of educational resources. Moreover, some would argue that such a huge investment in a few children, even if successful, would accomplish little since it would stand slight chance of replication.
I believe, to the contrary, that such an experiment would yield an answer to the single most important question that those committed to educational improvement need to confront: to what extent is the continuing mediocre performance of our children, "at risk" children especially, a result of our unwillingness to invest in education? In other words, are our schools failing because we can't make them better or because we won't make the necessary investment?
Let me try to weaken resistance to the idea of spending a large amount on a few children, even in an experiment, by pointing to a setting in which huge amounts are already being spent on only a few children. Consider intensive care for low-birth-weight infants. The cost of such hospital intensive care for infants weighing less than 1,000 grams at birth varied between $50,000 and $144,000 in 1985, according to one study conducted at a Denver hospital. (About two-thirds of the infants survived.) In one cost-benefit analysis of such care, the total costs per surviving baby exceeded their projected average lifetime earnings. These astronomical costs may be compared to the $2,000 it costs for delivery and neonatal care of a healthy newborn in one of our local hospitals. That is, the cost of care for premature newborns may exceed that of healthy newborns by a factor of 50.
It is not difficult to identify children whose likelihood of thriving educationally is very low. What would happen if we invested even five times as much in their education as in that of their more-likely-to-thrive peers--if we invested $25,000 per year instead of the U.S. average of about $5,000? The answer is that we simply have no idea. But we ought to find out.
One of the difficulties of attempting to compare the relative effectiveness of different schools or kinds of schools is the differences in their respective populations. The study I propose, to be carried out in one city, would begin by trying to identify as large a pool as possible of at-risk 2-year olds. From it, 50 matched pairs of children would be chosen by lot for participation in the experiment. One child from each pair, chosen again by lot, would serve as control, attend conventional day-care facilities and local public schools. The other half would, for 12 years, receive an education in which budgetary constraints were virtually eliminated.
Such a program would include an intensive early-education component for the children themselves as well as for their parents. For ages 6 through 13, the program would extend beyond the conventional school day to include various kinds of after-school activities. Throughout the study, but at the end of 8th grade especially, the children's development would be compared. We would then have some idea of what might be done to salvage educationally at-risk children if resources were ample. I emphasize "some" because no such experiment is likely to yield definitive conclusions.
Without going into any details of the experimental design, let me answer several obvious questions that consideration of such a study brings to mind:
- (1) Is it ethical to provide benefits to the children in the experimental group that are not offered to those in the control group? I believe it is not. Therefore, I propose that the control-group children receive an equivalent sum of money at the end of the 12-year experiment, with the proviso that that money be spent only for their education. In exchange for the benefits provided, the parents in both experimental and control groups would promise not to pull their children out of the experiment.
- (2) Who would teach in the experimental school? To test the limits of our pedagogical prowess, we need to invite the best teachers in the nation to participate. I would hope that the professional challenge as well as substantially higher salaries made possible by increased funding would permit a national recruitment effort. I would hope a team of teachers could be assembled whose classroom effectiveness with at-risk children was acknowledged by their own supervisors and communities.
- (3) Don't different communities and parents want different things for their children? What would the school's philosophy be, and according to what criteria would the children's progress be assessed? This question is harder to answer in theory than in practice. While I would suppose that the school would have as its primary aim the intellectual development of its children, there is no reason to think this must be done at the expense of their development as caring, confident, and socially conscious human beings. A parent and citizen board of overseers would monitor the project.
- (4) Given the plethora of ideas and approaches to teaching, who would decide what methods to use? This is a difficult question. If the teaching staff had sufficient lead time to plan the program, perhaps one year before the children arrived, I would hope that they could design a program they had confidence in. They would have the resources to hire whatever consultants they wished to, but unless the teachers felt ownership of the program, their own commitment would be jeopardized.
- (5) Who would run the school? I would hope that the school district in which the experimental school was located could find an administrator who was respected by teachers as well as community members. Since this administrator would have ample resources as well as an outstanding group of teachers, I would expect him or her to play a largely facilitative role.
The proposed experiment would cost about $2.5 million a year for a 12-year period, no small sum, but the total for all 12 years is less than 1 percent of the research budget of the U.S. Education Department for the current year.
Before the experiment got under way, two constituencies would need to be satisfied. The educational-research community would need to be convinced that the research design met the highest standards for such research. And the community of advocates for poor children would need to be satisfied that the project would make use of every available resource for improving the chances of children in the experimental group, given current knowledge about education practices.
What might the experiment tell us? If, despite a substantial investment, the children in the experimental program failed to develop their capacities to a significantly greater extent than their peers in the control group, we would need to lower our expectations about the redemptive power of good schooling. If a significantly greater proportion of the experimental children were to thrive educationally, we could begin to try to change people's attitudes toward educational investment.
When a premature infant is born, even if the prematurity is the result of the mother's abuse of drugs or alcohol, even if the child's chances of a normal life are not assured, we say with one voice: "Save the child; we'll worry about how to pay for it later." If the experiment I propose shows that "intensive educational care" yields similar benefits, shouldn't we adopt the same attitude towards expensive educational intervention?
Vol. 10, Issue 36, Pages 25, 32Published in Print: May 29, 1991, as An Educational Experiment We Need