For decades, policymakers, educators, and families have relied on annual state testing to glean essential information about student learning. State testing has also shined a light on racial and economic achievement gaps that were previously swept under the rug. There’s little question that summative assessments are important tools for measuring learning and forging more equitable educational outcomes.  

But for as long as there have been state tests, stakeholders have questioned whether the information gained about student learning outweighs the disadvantages of the assessments. For instance, tests promising more instructional or parental relevance have regularly delivered scores far too late in a given school year to be useful. Tests promising less administrative burden have been cumbersome and confusing in the early years of implementation. And tests promising to spur improvement in student outcomes have done so indirectly or not at all. 

With criticism only intensifying after the COVID-19 pandemic and the resulting declines in student learning, there’s growing momentum behind widespread state assessment reform with little consensus on what the changes might be between advocates who want to: 

1. Reduce the testing footprint of statewide assessments, preserving instructional time and deemphasizing tests, or 
2. Increase the instructional relevance of assessments to better support instruction throughout the school year. 

Because these two goals are at odds with each other, policymakers and advocates must embrace the necessity of trade-offs to choose a viable path forward.  

It’s difficult for tests to take up less time while offering more information: A test that provides stakeholders with real-time instructional relevance is likely to have a bigger footprint in the classroom, while a shorter test is likely to yield less data. And less data means less information on how much schools are helping their students learn. While there are several potential solutions that could give teachers and students more instructional time, many of the ideas floated for reducing time on assessments would require unacceptable losses in precision and information that make tests worthwhile — especially individual student scores. 

There is no perfect approach for either goal, but there are some potentially good paths — dependent on research, implementation, and changes to federal law — that could meet the demand for change while preserving essential sources of data for educators, policymakers, and families alike. 

Potential paths forward 

To give students and teachers some time back while producing data that everyone can use, advocates might consider a partial matrix sampling of items. 

With a full matrix sampling of items, different portions of an assessment are given to each student; this approach can produce school-level results, but no student level comparability. In the partial version, though, students have two blocks of test questions: a shared core among all students and a rotating sample of more detailed questions that vary. 

Because there are student scores to compare to each other, partial matrix sampling collects essential data and reduces testing time. This makes it a potentially strong path forward for reducing the testing footprint while retaining necessary information. 

There are caveats, however: Attention to sample design and test design would still be essential to preserve equity and ensure usable results combined with significantly less time spent testing. Changes in federal law would also be necessary. Partial matrix sampling isn’t a silver bullet to solve everything wrong with state testing, but, for many stakeholders, it could be a meaningful response to critiques about time on tests. 

In terms of increasing instructional relevance, many states are exploring through-year assessments, which offer frequent feedback to students, families, and educators. Through-year assessments could give educators and families information to use during a given school year.  

These tests are still in the early stages, though, and the states pursuing them have found both design and implementation challenges. They also take longer, cumulatively, and implementation could be a challenge. Most crucially, it’s unclear if this approach would provide comparable student data that advocates believe to be critical for identifying schools for additional resources and intervention. There’s a long way to go before through-year assessments can fulfill their promise. 

Embracing trade-offs 

There’s no perfect solution when it comes to statewide assessments. Both potential paths above face significant questions and challenges and might not be better than the status quo. State summative testing is inherently limited, and it may simply be impossible to achieve all the goals that advocates have through these assessments. While looking to make improvements in the existing state tests and preserving their role in producing comparable, reliable data, advocates might also encourage their states to offer different assessments, tools, and instructional materials to support learning and instructional reflection, without ties to accountability systems or public reporting.   

Policymakers need to consider changes with a full understanding of the trade-offs — and without risking a return to the days when there was less transparency about student learning. Making changes to state summative assessments is not a one-size-fits-all approach, but if policymakers and advocates prioritize their districts’ needs, they may get closer to a solution that improves learning outcomes for all students.