Over the past few decades, a handful of states have chosen to provide state financing of special education programs through a method referred to as "Census-Based" funding--an approach which involves allocated block-grant funding on an assumed basis of uniform distribution of children with disabilities across school districts. The approach has been argued to eliminate financial incentives for classification of marginal--low severity, higher incidence--disabilities. We explain herein that despite some evidence linking headcount-based financing schemes to increased classification rates (a) no evidence exists whether the incentivized rates are more or less indicative of true prevalence of disabilities, and (b) where attempts have been made to discern whether certain populations of children with disabilities are in fact uniformly distributed, researchers have found that they are not. We use U.S. Census data on families of children with disabilities to evaluate the geographic and demographic distribution of those families in Pennsylvania and New Jersey, finding high degrees of geographic clustering, relationships between census disability rates, census poverty rates, geographic locations and school district classification rates. In short, we find families of children with disabilities to be non-randomly and non-uniformly distributed across geographic spaces in Pennsylvania and New Jersey. We conclude by evaluating the equity consequences of assuming falsely that these children are distributed uniformly. (Contains 8 footnotes, 3 tables, and 16 figures.)