A retrospective analysis of 400 publications reveals patterns of irreplicability across an entire life sciences research field

In Drosophila immunity papers from 1959–2011, 61% of 1006 claims were verified, but 38/45 unchallenged claims tested experimentally proved non-replicable.

• Why it matters: The field-wide gap is that published findings may not hold up when rechecked, yet most claims are never independently tested.
• What they did: ReproSci retrospectively assessed conceptual replicability of 1006 claims from 400 papers, then experimentally validated 45 unchallenged claims 14 years later.
• The result: Irreplicability was linked to field dynamics and authorship patterns: challenged and unchallenged rates rose over time, top institutions were more likely to be challenged, and early-career training correlated with lower irreplicability.

AI summary of the abstract · may contain errors · not medical advice · read the paper.