Curriculum coherence: an examination of US mathematics and science content standards from an international perspective

US: 2 national standards documents (NCTM 1989, NRC NSES 1996), 21 state standards documents (in force 1999-2000), ~42-50 districts sampled PPS within a single state. International: 6 'A+' countries for mathematics (Singapore, Korea, Japan, Hong Kong, Belgium-Flemish, Czech Republic), 4 for science (Singapore, Czech Republic, Japan, Korea). Grades 1-8.

US state and national K-8 mathematics and science content standards, compared against the intended curricula of high-TIMSS-scoring countries.

Pure document/content analysis. No students, no achievement data collected or modelled at any level. The international benchmark is TIMSS-1995 General Topic Trace Mapping — a survey in which EDUCATION MINISTRY OFFICIALS ticked, grade by grade, whether a topic was in the national intended curriculum; the 'A+ composite' is then built by averaging topic counts and taking the topics endorsed by most A+ countries. 'A+' status is defined purely by mean middle-school TIMSS score, so achievement enters ONLY as the country-selection filter — it is never an outcome variable. No coherence index is computed, no correlation or regression against TIMSS is run. Their definition (p.528): content standards are coherent if they are 'articulated over time as a sequence of topics and performances consistent with the logical and, if appropriate, hierarchical nature of the disciplinary content'; they add 'rigour' (how deep, by what grade) as a second construct. The paper is the origin of the coherence language later adopted verbatim into the Common Core (via Schmidt & Houang 2002). Companion volume: Schmidt, McKnight & Raizen, A Splintered Vision (1997), the textbook/curriculum-guide analysis behind 'a mile wide and an inch deep' (grade-8 US maths textbook ~35 topics vs ~7 in Japan; US grade-4 maths textbook 530pp vs 170pp international average).

The A+ countries' intended curricula form an 'upper triangular' topic-by-grade matrix — topics enter, are taught for a few grades, and exit. US standards add topics and never drop them: in the 21 states studied, topics that A+ countries cover in grades 1-8 are covered for almost SIX years, roughly twice as long. Eleven topics had broad multi-grade coverage in two-thirds of states with no A+ counterpart, and two-thirds of states disagree on the placement of three fundamental topics. Crucially, the authors make NO causal claim and say so in print: 'Questions of causality are impossible to answer with survey data' and 'Further empirical support that coherence in content standards results in greater learning and deeper understanding must await additional study.' They then argue for acting anyway because 'coherence seems a more logical principle from a subject-matter perspective' — i.e. the case for coherence rests on disciplinary plausibility, not evidence.

Not applicable in the usual sense — no student outcomes are analysed. The latent inference (A+ countries score higher AND have coherent standards, therefore coherence causes scores) is a country-level ecological correlation with n=4-6, confounded by everything that differs between Singapore/Korea/Japan and the US, including population composition, tracking, out-of-school tutoring and selection into the tested grade.

The descriptive finding (US standards are broader and more repetitive than high-scoring countries') is uncontested and has been reproduced repeatedly. The causal reading has never been tested. Sikorski & Hammer (2017, Science Education, 10.1002/sce.21299) argue the designed-in-coherence construct conflates curriculum coherence with students' own coherence-seeking; a 2025 Journal of Curriculum Studies secondary analysis of the TIMSS video study states directly that 'though the TIMSS results were correlational, they reinforced a widespread assumption among the science education community of a causal relationship.'