Question

Should schools do standardized testing? Should schools instead be able to all test their students differently

Answer 1

Short answer: not only one or the other. Standardized testing has important uses, but relying on it alone causes problems. A better policy is a balanced approach: keep some standardized measures for system-level monitoring and comparability, while allowing (and investing in) high-quality local or performance-based assessments as part of a multiple-measures system.

Why neither extreme is ideal
- Only standardized testing — pros: gives comparable, objective data across schools and student groups; supports accountability, resource targeting, and research. Cons: narrows curriculum, encourages “teaching to the test,” increases stress, can entrench inequities if tests aren’t fair, and misses many important skills (creativity, collaboration, critical thinking).
- Only wholly local, different tests per school — pros: can be more authentic, tailored to local priorities, and better at assessing complex skills. Cons: loses comparability (hard to know whether a school is doing better or worse than others), risks low psychometric quality (poor validity/reliability), and makes system-level accountability and funding decisions difficult.

What a practical compromise looks like
- Core standardized assessments retained for system monitoring: use a limited number of well-designed, low-to-moderate-stakes standardized tests to track literacy, numeracy, and other anchor skills across regions and demographic groups.
- Multiple measures for student outcomes and school evaluation: combine standardized tests with teacher assessments, portfolios, performance tasks, classroom-based formative assessments, attendance, graduation rates, and surveys (students, parents, teachers).
- Local innovation permitted (and supported): allow schools/districts to design supplemental assessments that reflect local curricula, project work, and competencies; but require those local assessments to meet quality standards and to be reported in ways that can be interpreted alongside the standardized data.
- Calibration and moderation: use linking/equating methods, sample-based common tasks, or external moderation panels so local scores can be meaningfully compared or interpreted across schools.
- Reduce high-stakes pressure: avoid using a single test result for major decisions (teacher pay/firing, school closure, student promotion) and use aggregated, multi-year data where possible.
- Invest in capacity: train teachers in assessment literacy, scoring rubrics, moderation, and use of assessment results for instruction. Fund the development and administration of richer assessments.
- Protect equity and access: ensure accommodations, culturally responsive items, language supports, and measures that are validated for diverse populations.

Examples and models
- Finland: minimal national testing; strong emphasis on teacher assessment and trust, high teacher education and autonomy — works because of high baseline equity and consistent teacher quality.
- Many U.S. states: mandatory annual standardized tests for system accountability (ESSA-era tests) plus district/school-level assessments; problematic when those tests become the sole basis for high-stakes decisions.
- Hybrid systems: some countries use national exams at key transition points (end of lower secondary, tertiary entrance) and local assessments for ongoing evaluation.

Practical criteria if schools are allowed to design their own assessments
- Demonstrable validity: tests measure what they claim to measure and align with the curriculum/standards.
- Adequate reliability: consistent results across raters and occasions.
- Transparency: clear reporting formats and publication of what scores mean.
- External moderation or periodic benchmarking against common tasks or sample-based standardized assessments.
- Data privacy and security safeguards.

Recommendations for policymakers
1. Keep periodic standardized assessments for core skills to provide comparable, system-level data and to identify gaps.
2. Require multiple measures for accountability and high-stakes decisions; prohibit single-test determinism.
3. Support and fund high-quality local/performance assessments, teacher training, and moderation systems so local results are credible and useful.
4. Pilot and evaluate alternatives (portfolios, performance tasks, adaptive testing) and scale what works.
5. Ensure assessments are fair, accessible, and used to improve instruction rather than only to punish.

Conclusion
Standardized tests are a useful tool but not a complete solution. Combining well-designed standardized measures with high-quality local assessments and strong safeguards produces more valid, fair, and actionable information about student learning than either approach alone. If you want, I can outline a sample policy framework or a phased rollout plan for a hybrid system tailored to a particular district or country.