{"ID":2871871,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2509.10389","arxiv_id":"2509.10389","title":"Beginner's Charm: Beginner-Heavy Teams Are Associated With High Scientific Disruption","abstract":"Teams now drive most scientific advances, yet the impact of absolute beginners -- authors with no prior publications -- remains understudied. Analyzing over 29 million articles published between 1941 and 2020 across disciplines and team sizes, we uncover a near-universal and previously undocumented pattern: teams with a higher fraction of beginners are systematically more disruptive and innovative. Their contributions are linked to distinct knowledge-integration behaviors, including drawing on broader and less canonical prior work and producing more atypical recombinations. Collaboration structure further shapes outcomes: disruption is high when beginners work with early-career colleagues or with co-authors who have disruptive track records. Although disruption and citations are negatively correlated overall, highly disruptive papers from beginner-heavy teams are highly cited. These findings reveal a ``beginner's charm'' in science, highlighting the underrecognized yet powerful value of beginner fractions in teams and suggesting actionable strategies for fostering a thriving ecosystem of innovation in science and technology.","short_abstract":"Teams now drive most scientific advances, yet the impact of absolute beginners -- authors with no prior publications -- remains understudied. Analyzing over 29 million articles published between 1941 and 2020 across disciplines and team sizes, we uncover a near-universal and previously undocumented pattern: teams with...","url_abs":"https://arxiv.org/abs/2509.10389","url_pdf":"https://arxiv.org/pdf/2509.10389v2","authors":"[\"Mahdee Mushfique Kamal\",\"Raiyan Abdul Baten\"]","published":"2025-09-12T16:29:39Z","proceeding":"cs.DL","tasks":"[\"cs.DL\",\"cs.SI\"]","methods":"[]","has_code":false}
