US test scores in Math in 2026?
Based on mixed but fragile data — youth rebounds vs. older-student stagnation and very weak 12th-grade results — I assess a modestly below-even chance that U.S. math scores will show a significant decline in 2026.
Analysis
**Stage 1 — Independent (blind) analysis**
The most recent objective signals show a split picture: younger cohorts (9-year-olds) posted measurable gains in the NAEP long-term trend, while adolescents (13-year-olds) were flat and 12th-grade math sits at historically low levels. That heterogeneity is the core fact guiding an independent forecast.
Key elements I weighed and how they map to probabilities:
- *Signal strength and directional consistency*: Improvements in the youngest cohort are the clearest positive signal, but they remain below pre-pandemic levels and are smaller in magnitude than the deficits observed in older grades. The 12th-grade result is a strong negative datapoint because it is an extreme low (lowest since 2005), pointing to persistent issues at the top of the K–12 pipeline. - *Scope and representativeness*: NAEP has multiple assessments (long-term trend, main NAEP). Which assessment determines a market resolution will materially change the outcome; the existing evidence varies by assessment and grade. Because the market question is general, I treat the resolution as likely to reference mainstream NAEP reporting or an aggregated view — hence I blend grade-specific signals. - *Trajectory vs. level*: Gains among younger students are trajectory-positive but insufficient to restore pre-pandemic levels quickly. Older grades show stagnation or further erosion in level. Aggregating these trends makes a large net decline less likely than a small decline, but a material decrease remains plausible. - *Policy and system response*: There is active engagement by states and districts to address math recovery (new proficiency targets, remediation programs). These interventions take time to scale and produce detectable national effects; they reduce but do not eliminate the risk of decline over short horizons.
Combining the above, I allocate probability mass as follows (independent of market prices): - **No significant difference (roughly stable / modest net change)**: 48% - **Significant decrease**: 38% (this is the probability I map to the binary "Yes" question) - **Significant increase**: 14%
Rationale for these numbers: the balance of evidence is slightly in favor of *no significant change* because positive movement in younger cohorts offsets older-grade weakness in many aggregated metrics; however the deep weakness at the high-school end and the durability of pandemic-era learning loss justify a substantial (~38%) chance of a measurable decline.
**Stage 2 — Market calibration (look at prices and interpret differences)**
Current market percentages for the three-contender structure are: No significant difference 48%, Significant decrease 44%, Significant increase 9% (market volume ~50k contracts). The market distribution is close to my independent allocation but places more weight on a significant decrease and less on a significant increase than I do.
Why the market might be pricing a higher chance of decline (market = 44% vs my 38%):
- *Salience bias toward dramatic negative headlines*: Traders may overweight the 12th-grade NAEP low (the "lowest since 2005" headline) because it is stark and newsworthy; that drives higher immediate probability for a decline across the board even though the weakness is concentrated in older grades. - *Ambiguity about resolution source*: With a general question, different traders assume different resolution criteria (NAEP main, NAEP LTT, aggregate test averages, or state assessments). Those assuming a definition that emphasizes older students will price in more decline. - *Risk premium / downside aversion*: Participants may place higher weight on possible systemic deterioration (teacher shortages, funding pressures) that could tip marginally negative cohort trajectories into a significant decline, inflating the decline side.
Why the market might be underweight increases (market = 9% vs my 14%):
- *Underappreciation of early-grade momentum*: Some market actors may not fully credit early recovery in younger cohorts or may assume that gains are too small to shift national-level results within the resolution window. - *Asymmetric attention*: Negative signals attract more trade volume; hence the increase bucket gets less liquidity and lower implied probability.
Overall calibration judgment: the market is not wildly out of line with my independent view but modestly overprices a significant decrease by ~6 percentage points and underprices a significant increase by ~5 points. That spread is plausibly explained by headline-driven positioning and resolution-source ambiguity. If one can reliably identify the specific scoring rule for resolution (main NAEP vs. LTT, grade band), there may be actionable edge; absent that, the market reasonably reflects the real uncertainty.
Arguments
For
- Older-grade weakness is strong and recent: 12th-grade NAEP at its lowest since 2005 is a powerful negative datapoint that increases the chance of a measurable decline if aggregated across grades.
- 13-year-old scores are flat against the most recent comparator, indicating stalled recovery in an important cohort that influences middle-school and future high-school averages.
- Systemic, long-term downward pressure on math performance predates the pandemic and may continue to exert influence despite small short-term rebounds.
- Policy and resource constraints mean that even when interventions exist, scaling them quickly enough to produce a national uptick by 2026 is uncertain.
Against
- Clear, measurable gains among 9-year-olds show recovery pathways that, if sustained and expanding, reduce the probability of an aggregate decline.
- Many states and districts have adopted explicit math-proficiency targets and remediation strategies, which should blunt deterioration and improve odds of stability.
- Short-term improvements in early grades can offset older-grade weakness in population-weighted aggregates over time, making a large net decline less likely in the near term.
- The greatest negative headlines focus on specific grades; if the market resolution references a broader or different measure, those headlines may not translate into a decline signal.
Key drivers
- Grade-specific trajectories (improvement among younger students vs. stagnation/decline among older students)
- Which assessment or metric is used to resolve the market (NAEP main vs. NAEP long-term trend vs. other aggregates)
- Pace and effectiveness of state/district interventions and remediation programs rolled out post-pandemic
- Sampling, participation rates, and testing administration differences that can shift reported averages
Risk factors
- Resolution ambiguity — different stakeholders may assume different NAEP instruments/grade bands
- Sampling and participation swings (e.g., lower participation in lower-performing subgroups can bias averages)
- Policy shocks (e.g., sudden funding cuts, teacher strikes, or major curriculum shifts) that could depress outcomes
- Measurement noise and year-to-year volatility in national assessment scores
Scenarios
Best case
A broad-but-gradual recovery accelerates: gains among younger cohorts are larger than currently observed, targeted interventions scale effectively, and participation remains stable; the result is either small national gains or no significant downward change — the 'No significant difference' or 'Significant increase' buckets materialize.
Most likely
Mixed outcomes by grade persist: younger students continue modest recovery while older students remain flat or slightly worse. Aggregated national measures thus stay roughly stable, producing no statistically significant difference on the most commonly used national metrics.
Worst case
Stagnation in middle school and collapse in high school deepen: compounding effects of weak instruction, staffing disruptions, and stagnant remediation lead to a measurable, statistically significant drop in national math averages for the resolution metric — the 'Significant decrease' outcome prevails.
Outcomes
| Outcome | AI | MKT |
|---|---|---|
| No significant difference | 48% | 48% |
| Significant decrease | 38% | 44% |
| Significant increase | 14% | 9% |
More from this day
- FinancialsKalshi13y
Will OpenAI or Anthropic IPO first?
AI25%MKT87%Edge-62HypedI assess a 25% chance that OpenAI will IPO before Anthropic. Anthropic's earlier confidential filing and high probability of a late‑2026 announcement make it the clear favorite to go public first.
- PoliticsKalshi2y
Who will Trump pardon?
AI3%MKT46%Edge-43HypedExtremely unlikely — I assess a low single-digit chance (~3%) that Barron Trump will receive a presidential pardon before Jan 21, 2029; pardons are legally possible absent conviction, but there is no factual or political path that makes this plausible at scale.
- EconomicsKalshi2y
Will there be a Trump economic boom?
AI8%MKT51%Edge-43HypedI assess an 8% chance that U.S. real quarterly GDP (annualized) will exceed 5% in any quarter between Q1 2025 and Q4 2028 — a low-probability tail event driven only by extreme fiscal/monetary shocks, big inventory swings, or large data revisions.