Syntes: Evaluating the replicability of social science experiments in Nature and Science between 2010 and 2015 (2018)

Källa: https://doi.org/10.1038/s41562-018-0399-z (Nature Human Behaviour 2)

Verified content

• [abstract] Preregistered, high-powered replications (on average about five times the original sample size) of 21 experimental social-science studies published in Nature and Science 2010–2015.
• [abstract] 13 of 21 (62%) found a significant effect in the same direction as the original; replication effect sizes averaged about 50% of the originals.
• [full text, supplement] Sparrow, Liu & Wegner 2011 was one of the studies that did not replicate on any of the three indicators. The replicated finding was Study 1 (computer terms more accessible after hard questions, Stroop task), not the memory results. The original authors gave no feedback on the replication materials.

Limits

• [assessment] One study per original paper was replicated, so a failed replication speaks to that study's finding, not the whole paper.

Added 2026-10-16 by the project (broadening, Jonatan approved 2026-10-16). Verified 2026-10-16 by the project against the sources named in each line. [abstract] = from the published abstract; [full text] = read in the full text; [assessment] = the project's own judgement, not a fact from the paper.