Tennessee high schools’ test scores didn’t fully bounce back—and funding patterns shifted

A study of 322 schools links higher pre-pandemic scores and more economically disadvantaged students to larger learning loss and slower return to baseline after COVID-19.

Abstract editorial illustration accompanying the article: Tennessee high schools’ test scores didn’t fully bounce back—and funding patterns shifted

Editorial illustration — not from the study.

In Tennessee public high schools, COVID-19–era disruption led to substantial and uneven declines in learning, measured using ACT composite scores. The study reports that the statewide average dropped from 20.2 in 2019 to 19.1 in 2021, then recovered only to 19.3 by 2024–2025.

The research also examines how recovery related to school characteristics while Tennessee received additional federal and state funding during the same period. Using statistical models, the authors find that schools with higher pre-pandemic ACT scores and higher shares of economically disadvantaged students were more likely to experience greater learning loss and less complete recovery.

Learning loss and recovery links

- Statewide ACT composite averages fell from 20.2 (2019) to 19.1 (2021) and recovered to only 19.3 by 2024–2025. - In a regression analysis, a school’s prepandemic ACT composite score and the percentage of economically disadvantaged students were statistically significant positive predictors of learning loss. - For recovery, one model indicated recovery magnitude was predicted mainly by school size. - Another model (net change from baseline) found prepandemic ACT composite score, school size, and percentage of economically disadvantaged students predicted recovery patterns. - That same net-change model indicated higher-performing schools were less likely to return fully to their prepandemic baseline. - A simple check for cost efficiency did not find evidence that schools receiving the most funding had the highest costs per outcome unit, but the authors note that 51 schools with no recovery were excluded, which may affect this evidence. - The study reports partial support for its “compound disadvantage” idea: economic disadvantage predicted greater vulnerability to learning loss and less complete recovery. - Evidence for “cost disease” (that achievement gains would be limited regardless of funding) was not confirmed at the school level; excluding the 51 no-recovery schools removed some evidence of this pattern.

Target support by risk level

The findings suggest that after COVID-19 disruption, how much learning loss a school experienced—and how completely it recovered—was related to both its starting point and the share of economically disadvantaged students. The authors conclude schools serving more economically disadvantaged students should be considered for differentiated support to improve recovery trajectories.

Data and limits

This is an ex post facto quantitative study that uses a census of 322 Tennessee public high schools and examines ACT composite score changes from 2018–2019 through 2024–2025. The study relies on statistical regression models to identify which school-level factors were associated with learning loss and recovery patterns. Limitations noted in the abstract include: it is not framed as a randomized study, and the analysis for spending/cost efficiency excludes 51 schools that showed no recovery, which the authors say removes some evidence relevant to the funding-efficiency question.

// Source

Digital Commons - East Tennessee State University (East Tennessee State University) · 2026

Open access versionPublished 2026-08-01

Authors: Wilson Boyd

Institutions: East Tennessee State University

AI food images matched real food on health and calories — but not appeal

People judged AI-generated food images similarly to real food for healthiness and calorie content, but found them less realistic and were less willing to eat them. The results suggest AI images may be useful for controlled food-perception studies, but they are not interchangeable with images of real food.

20 hours ago·3 views

Society & Economics

Preprint finds measurable patterns in how AI skills emerge and fade

A preprint reports measurable patterns for how capabilities emerge and are forgotten during language-model training. The study says emergence depends on several model parts becoming useful together, while forgetting depends on how widely and strongly a capability is stored.

1 day ago·5 views

Society & Economics

In nursing mice, brain cells link bone, fat, energy use and pain sensitivity

A study in nursing mice identified hypothalamic nerve cells that linked prolactin signaling with bone and fat metabolism, food demand and pain sensitivity. Reducing parts of this pathway lowered food demand and pain sensitivity, but the findings have not been established in humans.

3 days ago·12 views