Standardised reading tests that purport to measure a general comprehension skill are invalid instruments for guiding instruction or judging schools; valid and fair assessment must be curriculum-based, testing what was actually taught.

The proposition is narrower than it sounds: Hirsch repeatedly concedes that the better standardised reading tests are valid measures of general reading ability and defends them against performance assessment; his real claim is that they are unfair and unproductive as guides to instruction and early-grade progress because they are knowledge tests in disguise. The knowledge-loading is adjudicated in the dossier as plausible with a circularity risk and no stable estimate of its size; the causal 'consequential invalidity' story rests on journalism and anecdote; and the prescription that tests must be curriculum-based is never evaluated anywhere in the packet.

Why it matters: If reading tests cannot separate school-taught from out-of-school knowledge, then attributing flat or contrary test results to a curriculum or a teacher requires care, and the case for a specified curriculum acquires a measurement rationale (thesis T5).

Standardised reading tests built on the premise that they measure a general, content-independent comprehension skill are, whatever their reliability as measures of large-group reading ability, unsuitable for guiding instruction, for measuring yearly progress in the early grades, or for attributing value to individual teachers, because what they mainly register on Hirsch's account is passage-relevant knowledge, and they cannot separate the knowledge a school taught from knowledge acquired elsewhere. The fair and productive test, at least in the early grades, is one whose passages draw on knowledge the curriculum actually assigned.

Scope: A methodological claim about the validity of a test for a purpose (instruction, accountability, teacher evaluation, early-grade progress), not about whether reading tests measure anything at all. Hirsch explicitly defends objective standardised testing against performance assessment and concedes that the better reading tests are reliable and valid measures of average reading ability; the attack is on skills-framed reading tests used as guides to schooling. The early-grade scope is present from 2006 (yearly progress in early schooling), made concrete as grades one to four in 2010, and restated in 2024.

Only 17% of asserting occurrences carry direct evidence, and 14 of the 15 'same' occurrences carry none. The registry lists 13 entries as primary-independent, and their summaries are mostly adjacent to the proposition: Koretz on portfolio unreliability and test familiarity, Morsy and Keenan on test heterogeneity, Sireci on content-test reliability, Perlstein and Walker on curriculum narrowing, NAEP and SAT trend data; underlying identities and methods were not verified from the packet, and the two Perlstein records (S0254, S1246) carry different dates and titles for what may be one book. The core 'knowledge test in disguise' claim is carried by Willingham's commentary, Hirsch's own PARCC item reading and a Norwegian parliamentary statement; no Core Knowledge Foundation report appears. The independence flags look mostly right, with the caveat that PARCC items (S0350) and Willingham (S0020) are independent sources on which the inferential work is Hirsch's.

Absent from the corpus: Absent from the packet are the three kinds of evidence the drafter named: a study partitioning reading-test score variance by passage topic familiarity, the curriculum-based-measurement literature on what such tests predict, and NCLB-era studies that identify curriculum narrowing causally. Also absent is any evaluation of a curriculum-based reading test actually administered in grades one to four.

First stated: 2006 · KD · kd:ch6_C21

The drafter dates the first statement to The Schools We Need (1996) but the first mapped 'same' occurrence is The Knowledge Deficit (2006); the packet supports the later date, with 1987 (educational formalism in writing assessment) and 1996 (defending standardised tests against performance assessment while conceding their validity) supplying precursors rather than the claim itself. The 2006 statement ('unwittingly unfair', empty criteria, shortcomings for early-grade yearly progress) is the claim proper; 2010 adds the curriculum-narrowing framing and the concrete proposal of curriculum-based tests for grades one to four. 2016 recasts it in validity-theory vocabulary (consequential validity), extends it to teacher value-added, and adds the PARCC item analysis; 2024 restates the early-grade prescription, and the selected passages show no new restriction or retraction.

• A study partitioning standardised reading-test score variance by passage-topic familiarity versus general skill (the drafter's own criterion) would either establish the knowledge-loading claim with a magnitude or bound it; D1:A5 already flags that no stable dominance estimate exists.
• An evaluation of a curriculum-based reading test administered in grades one to four, reporting reliability, predictive validity for later general reading, passage diversity, and effects on the advantaged-disadvantaged gap, would move the prescriptive half from stated necessity to tested proposal.
• A study that separates test-design effects from school-implementation effects on curriculum narrowing would settle Hirsch's own unresolved attribution (1996 and 2006 blame application; 2016 blames the tests).
• Evidence on whether gains registered by curriculum-aligned tests transfer to unfamiliar passages would show whether the Okkinga gap (d = .186 versus d = .431) reflects alignment, inflation or limited transfer, and whether a curriculum-based test can be fair and valid at once.