What researchers studied

Multi-method study combining nationwide Italian administrative data, randomized vignette experiments, teacher belief elicitation, and a field experiment that provided personalized feedback about prior recommendations and objective information on disadvantaged students' realized success. Nationwide Italian administrative data on school-track recommendations and outcomes, supplemented by participating teachers in randomized vignette and field experiments; the public abstract does not provide a single consolidated sample size

What they found

  • Holding students' academic performance and interests constant, teachers were significantly less likely to recommend demanding academic tracks for students from disadvantaged socioeconomic backgrounds.
  • Teachers substantially overestimated the probability that high-achieving disadvantaged students would fail in academically demanding schools.
  • Providing teachers with personalized feedback on their prior recommendations and objective information on disadvantaged students' realized success substantially reduced socioeconomic gaps in recommendations for high-achieving students.
  • The strongest downstream effects were concentrated among high-achieving disadvantaged boys assigned to teachers with the largest prior recommendation gaps and most inaccurate beliefs; for that group, demanding-track enrollment increased without detectable adverse effects on short-run academic performance.

What the study does not prove

  • This is a 2026 NBER/CEPR working paper and should not be described as a peer-reviewed journal article unless a later publication is independently verified.
  • The education system studied is Italy's tracked secondary-school system, so findings should not be generalized mechanically to U.S. course placement, gifted identification, college admissions, or private tutoring.
  • The field-experiment effects were concentrated in a specific subgroup, so Noor Lyra should not imply that feedback changes every educator's recommendations or every student's placement outcome.
  • The study identifies inaccurate beliefs and differential recommendations in this setting; it does not establish that all lower-track recommendations reflect bias or that challenging placement is always the correct choice for an individual student.

Evidence strength: Strong multi-method evidence; nationwide administrative data plus randomized vignettes and a field experiment; NBER and CEPR working paper.

Why this matters for families

A recommendation from an adult can shape opportunity, but it should not be treated as a complete measure of a child's ceiling. This study suggests that even experienced educators can hold inaccurate forecasts about which high-achieving students will succeed in demanding settings, and that better outcome information can improve those judgments.

Noor interpretation

How Noor translates the evidence into practice

The Noor-relevant lesson is that adult expectations should be checked against evidence rather than treated as a substitute for evidence. A student's current performance, demonstrated understanding, interests, and response to support should guide academic recommendations; assumptions about what a student is likely to handle can become a ceiling if they are never tested against actual outcomes.

Noor Lyra Educators should use diagnostics, student work, error patterns, progress evidence, and direct observation to calibrate expectations. When deciding whether a student needs remediation, acceleration, or a more demanding course pathway, distinguish demonstrated skill gaps from assumptions about long-term potential and revisit recommendations as evidence changes.

Read the original source

Noor links to the original or authoritative source so families can distinguish the evidence itself from our interpretation.

Open original source →

DOI: 10.3386/w35701

Research notes

teacher expectationsacademic trackingstudent potentialassessmenteducation inequalityhigh school pathwayseducator judgment

No single study determines a student's plan. Noor uses research as one input alongside the learner's goals, observed performance, academic context, and response to instruction.