PaperlayerView exact revision on arXiv

Source-linked AI summary

Quantifying Long-Term Scientific Impact

Dashun Wang, Chaoming Song, Albert-László Barabási

arXiv:1306.3293v2cs.DLcs.SIphysics.soc-ph

TL;DR

Whether early citations predict long-term impact remains a central question because early citations can underestimate future impact. The approach measures aging functions for papers with the same number of previous citations; Eq. 3 can account for a wide range of citation patterns. Citation histories across journals and disciplines collapse to a single curve, indicating a shared temporal pattern.

Abstract

from arXiv · show

The lack of predictability of citation-based measures frequently used to gauge impact, from impact factors to short-term citations, raises a fundamental question: Is there long-term predictability in citation patterns? Here, we derive a mechanistic model for the citation dynamics of individual papers, allowing us to collapse the citation histories of papers from different journals and disciplines into a single curve, indicating that all papers tend to follow the same universal temporal pattern. The observed patterns not only help us uncover basic mechanisms that govern scientific impact but also offer reliable measures of influence that may have potential policy implications.

Agreement Number W911NF-09-2-0053, DARPA under Agreement No. … D E F

The citation model predicts long-term citation histories, with 93.5% of tested papers within two standard deviations of predicted citations at year 30 after ten years of training. Early citations underestimate long-term impact for high-impact papers.

Agreement Number W911NF-09-2-0053, DARPA under Agreement No.

  • For high-impact papers with c30 ≥ 400, citations acquired two years after publication underestimate long-term impact.Source: Agreement Number W911NF-09-2-0053,… · p. 12, Figure 1
  • Figure 1 reports a data collapse for 7,775 PR-corpus papers published between 1950 and 1980 with more than 30 citations within 30 years; λ, µ, and σ parameters let Eq. 3 account for diverse citation patterns.An inset shows a data collapse for the 20-year citation histories of all papers published by Science in 1990.Source: Agreement Number W911NF-09-2-0053,… · p. 12
  • For papers selected by fitness λ≈1 or early citations c2∈[5,9], the study examines citation distributions and time-dependent relative variance across years 2, 4, 10, and 20.Source: Agreement Number W911NF-09-2-0053,… · p. 12, Figure 2
  • For Cell and NEJM, the analysis tracks ultimate impact, impact time, longevity, fitness, and immediacy alongside impact factor.Source: Agreement Number W911NF-09-2-0053,… · pp. 12–13, Figure 3
  • 93.5% of tested papers had z30 ≤ 2 under the model’s 30-year predictions using ten years of training.The test set comprised 4,492 papers published in the 1960s that received at least 10 citations in five years.Source: Agreement Number W911NF-09-2-0053,… · p. 13, Figure 4
  • Eq. (3) offered the best fit to the testing base according to the weighted Kolmogorov-Smirnov test.Source: Agreement Number W911NF-09-2-0053,… · p. 13, Figure 4
  • The prediction envelope narrows when training increases from five to ten years, although one example falls outside the envelope’s predicted range.Source: Agreement Number W911NF-09-2-0053,… · p. 13, Figure 4
Loading 1306.3293v2…