Unlocking Replicability: The Seven-Year Journey in Social Sciences

A comprehensive study in the US analysed 3,900 social sciences research papers for reproducibility, revealing only half produced replicable results. The study, highlighting the 'reproducibility crisis', recommends improved practices for scientific credibility. Researchers emphasize the importance of examining replicability conditions to enhance robustness in social and behavioural sciences.

A remarkable seven-year project conducted in the US has scrutinized 3,900 research papers in social sciences to assess their reproducibility. The findings reveal that only about half of these studies were precisely reproducible, sparking discussions on scientific credibility within the social and behavioral sciences.

The initiative, part of the 'Systematizing Confidence in Open Research and Evidence (SCORE)' project run by the Center for Open Science based in Washington, D.C., shed light on this 'reproducibility crisis'. It found that 60-70 per cent of the scientific community struggles to replicate studies, particularly in fields such as economics, political science, and psychology.

With more than 850 researchers involved, the analysis also examined 'analytical robustness' and found many studies' common analytical paths lacking. By highlighting the importance of testing replicability conditions, the researchers suggest that enhanced practices could strengthen scientific evidence in the social and behavioral sciences.

(With inputs from agencies.)