Multidimensional evaluation of large language models on the AAP in-service examination: Assessing accuracy, calibration, and citation reliability.

PLOS Digit Health · 2026 · PMC13220994 · PMID 42213673

Europe PMC DOI

hedged sentences

0.0000

closest p · 0.0× alpha

0.0680

boldest claim

The sentences

highly significantp < 1 × 10 − 50actually significant

Relationship of citation source to citation validity Pearson’s chi-square tests revealed a highly significant association between citation source type (article vs. book) and citation validity across all three models (χ2 ≥ 236, df = 3, Bonferroni-adjusted p < 1 × 10 − 50).

also in 60,174 other papers

approaching significancep = 0.057so close (0.05 < p ≤ 0.1)

McNemar’s paired analysis corroborated this pattern, with GPT-4.0 approaching significance (p = 0.057) and GPT-5.0 showing similar directional effects (p = 0.215).

also in 1,188 other papers

marginal significancep = 0.068so close (0.05 < p ≤ 0.1)

Confidence showed marginal significance (OR = 1.013, p = 0.068).

also in 2,343 other papers

Quoted from the open-access full text in Europe PMC under the licence the publisher applied. The sentence is reproduced exactly as published; the emphasis is ours.