education · family: the rulebook was written for a world that no longer exists
new skills, old exams
Countries reform curricula toward competencies but assessment systems still test knowledge recall
Problem statement
Many countries and jurisdictions have redesigned their national curricula to build in 21st-century competencies — student agency, creativity, collaboration, well-being, sustainability literacy — of the kind set out in the OECD Learning Compass 2030. But where exams are heavily based on mastering content, the OECD's curriculum-overload report warns, students will likely be steered toward what gets tested at the expense of other important areas of development. OECD Curriculum Content Mapping shows how unevenly the competencies are written into curricula to begin with: the share of mapped curriculum content that embeds "taking responsibility" ranges from 5% (Portugal) to 68% (Estonia), with most of the 15 countries and jurisdictions mapped below 40%, and "reconciling tensions and dilemmas" appears in just 3% of Portugal's mapped content. Teachers are caught in a bind: the reformed curriculum tells them to develop competencies, but high-stakes exams play an important role in how they set priorities, and when teachers teach to the test, the subjects and items that are tested receive disproportionately more classroom attention. What gets tested is what gets taught. Read this way, a curriculum reform that leaves assessment unchanged targets the wrong stakeholder: content designers rather than the assessment bodies whose instruments shape classroom reality.
Why this matters
Altering curriculum content is, in the OECD's words, "politically challenging and a high-stakes undertaking." Multi-stakeholder redesign processes involving educators, academics, and policymakers produce new frameworks that represent sustained effort and political capital. When these reforms are undermined by misaligned assessment — not because the reform was wrong but because the assessment system was left unchanged — the result is policy failure at national scale. PISA 2022 showed an unprecedented 15-point drop in mathematics and 10 points in reading across OECD countries. If teachers cannot teach to competencies because the exam tests facts, the reforms remain aspirational documents rather than classroom reality.
What’s been tried and why it hasn’t worked
Countries have tried several approaches. Adding competency-related content to the curriculum without removing existing content produces "curriculum overload" — teachers must cover both the traditional knowledge base and the new competencies, within teaching time that has not changed much over the last decade. The OECD documented this pattern across multiple countries: societal, technological, and economic pressures add content, but prior content is often not removed; the report points to new societal demands, pressure from lobby groups, and a general desire to retain what has always been included. Professional development programs, as in the report's Hong Kong example, keep teachers up to date on new curricula and teaching strategies, but teachers may opt to teach to the test in subjects with high-stakes exams. Some countries have tried embedding competencies within existing subjects (cross-curricular themes and competencies) as an alternative or a supplement to adding new subjects. A plausible underlying problem, which this write-up proposes and its sources do not document, is that curriculum and assessment are often governed by different bodies with different incentive structures and reform timelines. What the OECD report does say is that accountability systems and assessments signal subject priorities, "especially when accountability mechanisms do not coherently link to the curriculum," and that a well-thought alignment between curriculum goals and assessment policies and practices is critically important.
What would unlock progress
Assessment instruments that can reliably and validly measure competencies at scale — tools that test creative problem-solving, collaborative reasoning, and agency rather than factual recall. The OECD's 2025 Teaching Compass says assessments "should evolve to include formative and competency-based approaches that better capture student progress," and the 2020 report notes that international assessments like PISA are adding domains such as collaborative problem solving, creative thinking, and global competency. Governance structures that formally couple curriculum reform with assessment reform, requiring that any curriculum change include a funded, time-bound plan for corresponding assessment redesign. AI-assisted assessment approaches that can evaluate open-ended competency demonstrations (portfolios, projects, collaborative tasks) with sufficient reliability for high-stakes use.
Entry points for student teams
A student team could analyze a specific country's curriculum-assessment alignment gap by mapping the competency goals in its reformed curriculum against the actual content of its national exams, quantifying the mismatch. Alternatively, a team could design and prototype an assessment tool for one specific competency (e.g., student agency in science inquiry) that could function as a complement to existing exams, with a validity argument and feasibility analysis for classroom-scale deployment. Relevant disciplines: psychometrics, education policy, learning sciences, natural language processing (for automated assessment), comparative education.
Genes
Structural cousins: stuck the same way, in other fields
Sources
OECD Education 2030 Learning Compass, access date 2026-02-14. OECD Curriculum Overload: A Way Forward (2020), access date 2026-02-14. OECD Teaching Compass (2025), "Reimagining Teachers as Agents of Curriculum Change," access date 2026-02-14. go to source 1 ↗ go to source 2 ↗ go to source 3 ↗
Working notes
Notes made while writing and checking this brief, shown as written. They use working shorthand.
This brief connects to the health-digital-therapeutics-outcome-measurement brief — both involve a structural gap between two regulatory frameworks with incompatible evidence standards (FDA clearance vs. insurance reimbursement; curriculum competencies vs. exam knowledge). The OECD's discussion of this gap sits alongside comparative data from 37 countries and jurisdictions (its policy questionnaire on curriculum redesign) and 15 (its curriculum content mapping); that data describes curricula, not exams, and the link to assessment is drawn from the research the report reviews. The 2025 Teaching Compass publication explicitly acknowledges that "teachers face a growing list of expectations — to deliver curriculum, nurture well-being, adapt to digital and AI-assisted learning, support inclusion, and build critical 21st-century competencies. Yet they are often left without the tools, recognition or systemic support to succeed." The PISA assessment itself embodies the tension: the 2020 report says literacy and numeracy have been prioritised in part because they are consistently and reliably measured by international assessments such as PISA, while noting that recent changes to PISA are integrating new domains such as collaborative problem solving, creative thinking, and global competency. Related areas: assessment literacy, psychometric innovation, governance reform in education, AI in assessment, formative assessment theory-practice gap. Structural parallel: `education-brac-play-based-learning-transition-gap` documents the same assessment-drives-behavior claim with longitudinal outcome data from Bangladesh — BRAC pre-primary gains fully erased within 18–24 months of entering government schools where exams reward rote memorization, demonstrating that teacher training without assessment reform produces no lasting behavior change.
Reconciliation 2026-09-30: The whole write-up was read against its sources, all fetched 2026-09-30. Curriculum Overload: A Way Forward (OECD, 2020) was read in full from the OECD's own PDF (https://www.oecd.org/content/dam/oecd/en/publications/reports/2020/11/curriculum-overload_167e276d/3081ceca-en.pdf); the landing page on the Source line refuses automated readers, so its text was read from the Internet Archive copy, which gave the PDF address. The OECD Teaching Compass (2025) was read in full from the PDF on the Source line. The Learning Compass page on the Source line was read from its Internet Archive copy; it is a short description of the framework with no figures. For the PISA 2022 sentence, PISA 2022 Results Volume I was read from the OECD's PDF (https://www.oecd.org/content/dam/oecd/en/publications/reports/2023/12/pisa-2022-results-volume-i_76772a36/53f23881-en.pdf) and Volume II was searched. Changes: (1) The text said curriculum content mapping shows "alignment between national curricula and the Learning Compass ranges from 3% (Portugal) to 68% (Estonia), with most countries below 40%." The report has no overall alignment figure. It gives, for each of about a dozen competencies, the share of each country's mapped curriculum content that embeds that one competency, for 15 countries and jurisdictions. The 3% is Portugal's figure for "reconciling tensions and dilemmas" (Figure 22); the 68% is Estonia's figure for "taking responsibility" (Figure 20) and also for ICT/digital literacy (Figure 14). The sentence now gives the range for one competency, taking responsibility, 5% (Portugal) to 68% (Estonia), where 11 of the 15 are below 40%, and gives Portugal's 3% with the competency it belongs to. (2) "Their high-stakes assessment systems still predominantly test factual recall and procedural knowledge" and "what gets tested hasn't changed" are not in the sources. The report says that when exams are heavily based on mastering content, students will likely be steered toward what gets tested, that high-stakes exams play an important role in how teachers set priorities, and that when teachers teach to the test, tested subjects and items receive disproportionately more classroom attention; the text now says that. (3) That exams are "controlled by separate assessment authorities," that curriculum and assessment are "governed by different bodies," and that assessment reform "takes 5–10 years" are not in any source. The number was removed; the different-bodies idea is kept but marked as this write-up's own proposal, followed by what the report does say about accountability and alignment. (4) "Dozens of countries ... following frameworks like the OECD Learning Compass 2030" now reads "many countries and jurisdictions ... of the kind set out in" the Learning Compass; the sources do not count reforming countries or say the reforms followed the Compass. (5) "The most resource-intensive education policy lever a country can pull" and "thousands of educators" are not in the sources; the sentence now quotes the report's own description of curriculum change. (6) "The competency-oriented reforms were supposed to address exactly these outcomes" (the PISA score drops) is not in the sources and was removed. (7) Curriculum overload was said to arise because exams were not modified and "exam syllabi don't shrink." The report gives other reasons (new societal demands, lobby-group pressure, a desire to keep what has always been included) and says teaching time has not changed much; the text now follows the report. The claim that cross-curricular themes "become optional enrichment" without assessment reform is not in the sources and was removed; embedding competencies in existing subjects is now described, as the report does, as an alternative or a supplement to adding new subjects. (8) "Finland's matriculation exam reform and Singapore's gradual reduction of high-stakes testing" are not in the sources (the report describes Singapore reducing curriculum content, not testing). The sentence was replaced with what the two OECD documents say about where assessment is heading. (9) In Source Notes, "comparative data across 40+ countries" now reads 37 countries and jurisdictions for the policy questionnaire and 15 for the content mapping, and says that this data describes curricula, not exams. (10) The Teaching Compass quotation had an extra comma; it now matches the document. (11) The statement that PISA "tests academic knowledge rather than the competencies the Learning Compass emphasizes" is at odds with the 2020 report, which says PISA is adding collaborative problem solving, creative thinking and global competency; the sentence now reports both halves: literacy and numeracy have been prioritised in part because international assessments such as PISA measure them consistently and reliably, and PISA is integrating new domains. Confirmed in the sources and left as written: the three entries on the Source line; curriculum overload as the report's term; societal, technological and economic pressures adding content; the Teaching Compass passage quoted in Source Notes; and the PISA 2022 drop of 15 points in mathematics and 10 in reading across OECD countries, which Volume I calls unprecedented. (12) "Alongside declining growth mindset and student well-being" (PISA 2022) was not found in any of the five PISA 2022 Results volumes: no volume reports a change in growth mindset since 2018, and Volume II says trends in students' sense of belonging at school between 2018 and 2022 were mixed, with equal proportions of countries and economies showing stable, improving or deteriorating trends; the clause was removed. The long title states as fact that assessment systems still test knowledge recall, which the sources support only conditionally; it carries no number and was left unchanged. The closing comparison with the BRAC write-up rests on that write-up, not on these sources. A second reader went over every repair on the same day against the same sources (all five documents fetched again independently, plus PISA 2022 Results Volumes III, IV and V), found no original claim that a named source supports, tightened five of the new sentences to the source's own hedges, and removed the growth-mindset clause (12). Genome tags untouched.