Economic Evaluation in Education. Henry M. Levin
directly for their own decisionmaking—might include the school board, the district administrators, the state education agency, and the pertinent curriculum specialists and teachers. Given the many audiences, it may be useful to interview representatives to ascertain the types of information that they would like to obtain from the evaluation. The secondary audiences—that is, those with an interest in the results—might include non-English-speaking local residents, employers and residents of the district, and other school districts. These audiences may have different needs from the primary audiences. Again, setting priorities is important, since no evaluation endeavor is likely to meet all the needs of every potential primary and secondary audience.
Once the basic outlines of a cost study are agreed upon, the essential next step is to decide whether to perform CE or benefit-cost (BC) analysis. Either analysis may be appropriate. It depends on what decisions will be made as a result of the analysis. For example, imagine a nonprofit agency whose mandate is to improve literacy in kindergarten. Given the agency’s objectives are already established, it would seem that CE analysis would be most appropriate; this would inform the agency which literacy programs are the best value. By contrast, BC analysis helps establish whether it is worth investing in literacy programs at all. Presumably, the agency—in deciding its mandate—had already determined that literacy programs are worth investing in. However, BC analysis may be useful: It can clarify how valuable literacy programs are and so justify—both within the agency and to its donors or funders—further investments in literacy programs. As emphasized previously, the motivating question is this: What decisions will the audience make when presented with the information from economic analysis?
Audiences may be more or less willing to accept certain kinds of analyses that nonetheless yield similar conclusions. For example, BC analysis is less palatable to some audiences because it requires placing a monetary value on certain kinds of outcomes, effectively modeling child learning in terms of its monetary benefits. Although they are different, there is scope to substitute between CE analysis and BC analysis. One could imagine a program that is aimed at reducing high school dropouts. Since the program is designed to reduce the number of high school dropouts, an obvious measure of effectiveness is the number of dropouts averted. Eventually, of course, staying in high school might yield monetary benefits in the form of additional wages. We can conduct either CE or BC analysis, depending on the outcome measure that is chosen (dropouts averted or increased wages). It is conceivable that the CE analysis will be more warmly received by some audiences, because it does not attempt to describe the outcomes of the program in pecuniary terms. If CE and BC analyses do produce similar conclusions for decisionmaking purposes, it seems prudent to choose the analytical technique that is most likely to be well received by the primary audience. However, choosing to conduct CE analysis rather than a more comprehensive BC analysis might entail sacrifices in terms of the depth of analysis. In these cases, the researcher is best advised to allow the demands of the analytical task to guide the choice of analytical technique rather than the demands of the audience.
Thus far, we have described the audience simply as if they are readers of the research evaluation. However, their role is much more influential because they can shape the perspective of the analysis. The perspective is the framework within which costs and effects or benefits should be examined. Thus, the perspective dictates which items should be included in the analysis.
For educational evaluations, we can distinguish three important perspectives. Depending on which perspective is adopted, the analyst will have to calculate the costs, effects, and benefits that correspond to that perspective.
The primary perspective is the social perspective: This requires that all costs should be counted, regardless of who pays for them or even if they are provided in kind, and all benefits should be included, regardless of who accrues them. This social perspective is therefore all-encompassing, and other perspectives can be understood in relation to the social one. Indeed, it makes sense to start with a social perspective and then specify alternative perspectives within this general one. As discussed previously, the audience for the analysis may prefer a perspective that only helps with their decisionmaking and so excludes some costs (or even some benefits). However, this preference may be narrow-minded: If the goal is to maximize social welfare, the decisionmaker should be interested in all the resources used. This holds even when the decisionmaker uses contributed resources; these contributed resources might be used in an alternative way if their current use yields low benefits. Importantly, volunteers who contribute their time should want to know if their contribution—as part of an educational investment—is socially valuable. Volunteers for mentoring programs such as Big Brothers Big Sisters of America, for example, would presumably switch to an alternative activity if they learned that this program generated only small social benefits (or invest more if they had evidence on the social value they were creating). Finally, in order to compare interventions, it is helpful if all evaluators adopt the same perspective, and the social perspective can be adopted for all interventions.
The next perspective is the private individual perspective: This requires calculation of all the costs and consequences for the student (or the student’s household) or participant in the education program. Given that many education programs are subsidized by government and other entities, private individuals may reap substantial benefits but pay little of the costs. As such, from a private perspective many educational interventions should appear to be good value. As students are likely to choose the intervention with the highest private returns, this perspective will help explain why students participate in particular programs.
Finally, the analyst may adopt a fiscal perspective—that is, looking only at the costs and consequences from the perspective of the taxpayers or a particular government agency. This perspective is important as a way of justifying public investments in education programs. Government agencies may—by statute—be bound to consider only the implications for their agency. The challenge with using this perspective is that many educational interventions are funded from several sources, have diverse effects, or convey benefits to many individuals and in many domains. Adopting a fiscal perspective would provide a very partial picture for each agency. Nevertheless, some agencies might justify educational investments purely on these narrow grounds: Given the evidence on the effect of education on health, for example, it might be reasonable to label the high dropout rate in the United States as a “public health issue” (Freudenberg & Ruglis, 2007).
2.3. Relating Economic Evaluation to the Theory of Change
Once one has established the problem, the alternatives to be considered in addressing it, and the audiences, it is necessary to select the type of analytic framework that will be used. In the previous chapter, we identified these approaches as cost and CF analysis; CE and cost-utility (CU) analysis; and BC analysis. Here, we discuss which is appropriate across the range of educational interventions.
As noted in Chapter 1, CE and BC analyses are strictly intended to answer different questions about CE and efficiency respectively. However, the analyst may need to discover which form of analysis is appropriate for each intervention.
The best way to make this discovery is to make sure that the economic evaluation corresponds to the theory of change for each intervention. For our purposes, we can think of the theory of change in terms of (a) an educational intervention that is implemented (b) within a general context or set of existing conditions and that via a (c) connecting outcome or mechanism meets or is intended to meet (d) a set of longer-term goals (Weiss, Bloom, & Brock, 2014). Each of these elements (a–d) is relevant for deciding on the appropriate economic evaluation and how that evaluation is structured (Ludwig, Kling, & Mullainathan, 2011).
The intermediate or longer-term outcomes of some interventions explicitly can be measured in monetary values. For example, a training program may be intended to increase earnings for participants or a social emotional learning intervention may have savings to the school system from reduced conduct disorder as its goal. Given that BC analysis requires all amounts to be measured in monetary terms, this would be the most appropriate evaluative method. While the value of additional earnings and employment from an educational investment might be measured