Improving education for the neediest of our children is America's priority in education reform. Unlike several of our international competitors (e.g. China), we deliberately focus on the low end of performance, committing to "leave no child behind" rather than investing in the high achievers. Historically, education funders have used income as the primary criteria for deciding relative (dis)advantage in our schools, relying heavily on poverty as a proxy for student need. Though the income-based achievement gap is striking, our commitment to such a poverty-centric perspective has led us to overlook other variables that are highly correlated with educational outcomes. In our quest to do right by poor students, we have failed to take into account equally meaningful meaningful risk factors that disadvantage students.
Our current funding regime allocates dollars to school districts based on formulas that approximate the relative need of each district's student population. Almost all funding models rely on variants of the same three demographic criteria: income level, special needs and English proficiency. Federal education grants center around programs to address these three risk factors, while state funding formulas incorporate weights for these risk factors when allocating money across districts.
In terms of dollars spent, the model is highly poverty-centric, directing most of the funds allocated for high-needs students to low income students. For example, last year Illinois' directed more than two-thirds of its general state aid, greater than $2 billion, to districts with lower than average household incomes. In addition to general aid, Illinois spent an additional $1.6 billion to compensate high poverty districts with "poverty grants." In comparison, $0.5 billion was spent on special education, and $0.1 billion on bilingual programs. The lion's share of resources earmarked for high-needs children is awarded based on financial means(1).
Although it is true that each state's funding model takes a unique form, the variety seen across states is in some ways superficial. States largely share a single philosophy about what constitutes student need, focusing on the same three "needs" variables, with income at the center. States vary only in how they weight these variables, and how they weight the combination of risk factors(2). New Jersey, for instance, has the highest number of "need" categories, with the most complicated funding formula nationwide. The state has defined 10 levels of student "need," even though the formula relies on the same three demographic criteria(3). The seeming complexity emerges because the state applies weights for these factors in different combinations. The degrees of freedom to redirect funding today lay largely in the weights you apply to established risk factors, and how you quantify the additional risk of these variables in combination.
The catalogue of risk-factors that indicate disadvantage in education transcends the short list of variables applied in these formulas. Maternal education level and immigrant status are two examples of meaningful risk factors excluded. A mother's education level is highly correlated with student performance, and has been shown to have an effect independent of other risk factors, including income(4). Immigrant status (first or second generation) is likewise predictive of student outcomes, over and above considerations of English language proficiency. The Netherlands, which boasts one of the strongest education systems worldwide, has chosen both parental education and immigrant status as central variables in allocating funding(5). Most interestingly, the Dutch funding model does not include income as a neediness factor at all. The income achievement gap is so entrenched in the American psyche that our funding models look no further to find a proxy of "neediness", overlooking demographic features recognized as indicators of student need elsewhere.
Why are we so fixated on income? Why not include maternal education or immigrant status, with its well established predictive power? The answer, in part, emerges from our discomfort with labeling certain features as indicators of need. When it is less obvious to us why a particular demographic feature gives rise to inferior outcomes, we tend to dismiss the variable as a legitimate "neediness" criterion. Because the relationship between a mother's degrees and underperformance is less intuitive than the relationship between poverty and underperformance, it is more difficult to justify allocating dollars based on this criterion.
Funders should reexamine our traditional definitions of neediness. School funding in the United States relies on a limited set of "neediness" variables that largely focuses on poverty, while variables such as maternal education level or immigrant status are excluded from the assessment of need, despite robustly predicting student outcomes. Educational disadvantage transcends relative income level. Revised funding formulas that better reflect predictors of student outcomes will direct dollars to the truly neediest students.
References
2 Verstegen and Jordan, "A Fifty State Survey of School Finance Policies," Journal of Education Finance, 200
3 "A Formula for Success: All Children, All Communities," New Jersey DOE, Dec 2007