The process by which most education policies are developed is ad hoc and only minimally driven by empirical evidence. Typically, education policy ideas are the products of theoretical musings or analogies to other fields, and they are cobbled together based on reformers’ preferences and the dictates of political expediency. Put together, in other words, like a plate of luncheon meats at a smorgasbord. The practice of rigid age-based student grouping popularized in the early 1900s, the Life Adjustment[3] curriculum of the 1940s, the New Math movement of the 1960s, and the NCTM mathematics standards of 1989, for instance, were not the products of field tests showing them to be either effective or popular, but rather were the brainchildren of education theorists. Today, charter school, voucher, and tax credit programs are most often advocated by analogy (“markets work in other human enterprises, so…”) or by reference to isolated and small-scale experimental programs that often differ substantially from the policies being proposed. Many market-inspired education reforms, moreover, tend to include or exclude particular market characteristics without explaining why some are deemed essential and others dispensable.

Obviously there are many problems with this ad hoc approach to policy design. Theories not vetted against some substantial relevant body of evidence have often been terribly misguided (viz., the medieval theory of bodily humors[4]). Some analogies simply do not hold (what if education, as some have argued, is categorically different from other human enterprises?), and some systems (e.g., ecosystems) can collapse when even a single one of their essential features is impaired or removed (e.g., sunlight[5]). Given these potential pitfalls, it is not surprising that ad hoc education reformers have tinkered feverishly and continuously with public schooling for more than a century without solving most of the problems they have set out to fix.

So how can we minimize the likelihood that we will follow in the often unsuccessful footsteps of earlier reformers? The best way of doing so is to: 1) identify a model education system capable of fulfilling our goals, 2) determine how and why that system works, and 3) develop a policy that instantiates that model system. The model should be derived as much as possible from actual examples of large-scale effective education systems, rather than being based exclusively on analogies or references to small-scale programs. In taking that policy from drawing board to reality, we must also be aware of the model’s necessary and sufficient conditions for success, and thereby distinguish compromises that might cause limited harm from those that would hobble the entire system.