source of persuasion as a rhetorical claim. By saying that one has compared the CE of different approaches and found a particular one to be most cost-effective, one can often disarm opponents. Unfortunately, this is the principal way in which the terminology is used in the educational sector. This volume will go beyond such banality. CE analysis should be a topic of concern because it can lead to a more efficient use of educational resources—it can reduce the costs of reaching particular objectives, and it can expand what can be accomplished for any particular budget or other resource constraint. Similarly, BC analysis can identify which investments make the most sense and how much society should invest in education. But these results must be derived from application of rigorous methods.
This book emphasizes that a proper assessment of costs and effectiveness or benefits is a necessary element of a serious evaluation. All too often, either costs or effects are considered separately, and any combined inferences may be misleading. For the most part, educational evaluators attempt to establish whether an intervention is effective in attaining some goal, such as raising academic achievement. Business managers and administrators are often concerned with lowering educational costs in order to fit school or district expenditures within a fixed budget. Nevertheless, both parties should share a common goal: They wish to attain maximal school effectiveness for a given budget, or conversely, they wish to attain a given level of effectiveness at a minimal cost. They are unlikely to accomplish this goal if higher effectiveness and lower costs are pursued as independent goals. To illustrate this point, it is useful to consider three examples in sequence.
First, there is the peril of ignoring costs. This peril is nicely illustrated by looking at the development of policies related to smaller classes. During the 1990s, reducing class size was among the most popular and intuitively appealing options for improving schools; it seemed to provide a rare instance of agreement among parents, teachers, administrators, and even politicians. It led to a number of initiatives to reduce class size, especially in the primary grades, with the largest being in California’s effort in 1996 to reduce class size in the early primary grades to 20 students (Brewer, Krop, Gill, & Reichardt, 1999).
Much of the support for class size reduction was based on the promising results of the Project STAR experiment in Tennessee (Krueger, 1999; Mosteller, 1995). This experiment compared the achievement of students who were randomly assigned to three kinds of classes: students in regular classes (22 to 25 students) without teacher aides, students in regular classes with teacher aides, and students in small classes (13 to 17 students). Performance on standardized tests increased by 4 percentile points during the first year in which students were assigned to smaller classes (Krueger, 1999). Each subsequent year of participation in a smaller class further increased achievement by about 1 percentile point. (Teacher aides were found to have little effect on student achievement.) Overall, the results provided strong evidence that reducing primary class size has important effects on student achievement (Chingos, 2012). Looked at in isolation, these results might justify policies to reduce class size.
But reducing class size is very expensive—often prohibitively so (Brewer et al., 1999). To get the expected improvements in achievement would require a sizable number of new teachers—presumably trained to the same standard as the evaluated cohort—and it would involve much sorting and job changing as junior teachers moved across schools. New classrooms would have to be built and furnished, and changes to the school organization would also be substantial. When these resource issues were examined, the case for class size reduction diminished (Grissmer, 1999). Moreover, when the resource commitment was evaluated against what needed to be reallocated and lost from other school practices or potential interventions, the call for smaller class sizes further weakened.
Second, there is the peril of ignoring effectiveness. Declining enrollment in schools often forces administrators to make difficult decisions about school closure. Should numerous smaller schools be maintained, or should their enrollments be consolidated into a single larger school? On the surface, the decision might seem to be related exclusively to costs. The accountant or business manager should merely calculate the cost of each alternative and implement the least costly.
Indeed, average costs may be lower in larger schools due to economies of scale (Colegrave & Giles, 2008; Lee & Smith, 1997). Larger schools may benefit from lower prices on bulk purchases of school supplies or furnishings. Perhaps most important is that the fixed costs of operating a school can be spread over a larger number of pupils. For example, a larger school needs to maintain only one playground and cafeteria, pay one electricity bill, support one library, and so forth. Numerous smaller schools would be forced to duplicate these fixed expenditures, and per-pupil costs could rise. In rural areas of developing countries, isolated schools may not enroll enough students to fill even a standard classroom. Since even five students require a teacher, per-pupil costs may turn out to be quite high. Of course it is not a foregone conclusion that larger schools are less costly: excessively large schools may have higher management costs, there may be extra transport costs, and small schools that are physically close to their communities may be better poised to generate donated resources from local parents and citizens.
Let us presume, however, that the overall cost per student declines in larger school units. Does school size alter the effectiveness of schools? There is some evidence that larger schools have lower educational effectiveness across many different outcomes, although not in all specifications (Barrow, Schanzenbach, & Claessens, 2015; Leithwood & Jantzi, 2009). Precisely why this is so is unclear. It may be that larger schools are more depersonalized and provide both students and educational professionals with less of a feeling of individual importance and involvement. Or it might be traced to the presence of a core curriculum in smaller high schools that emphasizes academic excellence for all students, regardless of their abilities or aspirations (Lee & Smith, 1997). In areas of low population density, students may be forced to travel long distances to reach a larger school. Their resulting fatigue could reduce their learning once they arrive or even deter attendance in the first place (Bray, 1987).
Despite the cost savings of larger, consolidated schools, they may entail sacrifices in effectiveness, though this obviously needs to be analyzed on a case-by-case basis. Thus, larger schools that are less costly may ultimately prove to be less cost-effective. Administrators might consider other means of cutting educational costs that will not undermine effectiveness. For example, smaller schools can reduce costs by sharing teachers and administrators and, in the case of secondary schools, by drawing on such community resources as courses offered by community colleges. They can also lease unused classroom space for child care, senior citizen centers, and private educational endeavors such as computer schools or tutoring centers. The key point is that each of these alternatives should be clearly established, so that the relative costs and effects of each alternative can be evaluated.
The third peril arises when we ignore benefits. Over the period from 2003 to 2013, the amount of public subsidies for community college fell in real terms by 6% (Desrochers & Hurlburt, 2016). Despite rapidly growing enrollments at community colleges, the inflation-adjusted amount of public funding per student has fallen. As public funding has fallen, student tuition and fees have increased to make up the difference. Some might see this as a smart efficiency gain: The students who directly gain from college have to pay for it.
But this policy change neglects the fact that college produces a substantial benefit to the local taxpayer—for example, in terms of higher tax revenues (because of higher income and spending) and lower government spending (on welfare and the criminal justice system). In fact, according to estimates from Trostel (2010), each new community college graduate contributes more than $140,000 more in fiscal benefits (higher taxes and lower public spending on other services). In other words, the public sector, by deliberately reducing the resources available to community college students, is directly reducing its own economic well-being.
Each example illustrates the pitfalls of concentrating exclusively on either outcomes or costs. If evaluators intend the results of their studies to be used for decisionmaking, the information on effects of alternatives (such as class size reduction) is not adequate
