The Problem With "Proficiency": Limitations of Statistics and Policy Under No Child Left Behind

large-scale state and national test score distributions

US K-12 state assessment and NAEP score distributions

A statistical demonstration on real large-scale score distributions rather than a conceptual essay, which is why it sits at C rather than D. It is not a causal study and makes no causal claim; what it establishes is that a particular way of summarising scores destroys information in ways that are not recoverable.

Collapsing a score distribution to "percent proficient" - or to any single threshold - gives "limited and unrepresentative depictions" of trends, gaps and gap trends, and the distortions are "unpredictable, dramatic, and difficult to correct in the absence of other data." The mechanism generalises well past NCLB and straight onto an individual score report: a threshold statistic responds only to movement near the cut, so it is maximally sensitive to children sitting near the line and blind to everyone else. Ho's prescription is the one an intake tool should adopt - a distribution-wide view is required for any serious analysis of test data, including growth.

Not applicable - a claim about score summarisation.

A methodological demonstration rather than an effect; the arithmetic is not disputed.