Source-linked AI summary
Author Impact Factor: tracking the dynamics of individual scientific impact
Raj Kumar Pan, Santo Fortunato
TL;DR
Journal impact factor is widely used to evaluate scholars, but skewed citation distributions and changing career trajectories make it an incomplete measure of individuals’ current impact. This paper introduces AIF, an author-level measure of citations to recent work, and reports that it tracks career dynamics that current indicators do not adequately capture. The authors note that its fixed window can omit older papers whose citations arrive much later.
Problem
Skewed citation distributions make journal impact factor a partial proxy for individual research impact, while cumulative measures do not reflect changing career trajectories; AIF’s properties had also lacked empirical study.
Source: INTRODUCTION · p. 1Method
AIF extends journal impact factor to authors by averaging year-t citations to an author’s papers from a preceding time window.
Source: INTRODUCTION · p. 1, DISCUSSION · p. 5Results
AIF histories show career dynamics and sensitivity to major citation-flow variations that current individual-performance indicators do not capture adequately.
Source: DISCUSSION · p. 5Takeaways & Limitations
AIF offers a simple way to monitor the changing impact of scholars’ recent work over their careers.
Source: DISCUSSION · p. 5A fixed window can omit older papers with delayed citations; a time-window-free alternative needs longer citation histories and may make a year’s AIF change later.
Source: DISCUSSION · p. 6
Abstract
from arXiv · showhide
The impact factor (IF) of scientific journals has acquired a major role in the evaluations of the output of scholars, departments and whole institutions. Typically papers appearing in journals with large values of the IF receive a high weight in such evaluations. However, at the end of the day one is interested in assessing the impact of individuals, rather than papers. Here we introduce Author Impact Factor (AIF), which is the extension of the IF to authors. The AIF of an author A in year $t$ is the average number of citations given by papers published in year $t$ to papers published by A in a period of $\Delta t$ years before year $t$. Due to its intrinsic dynamic character, AIF is capable to capture trends and variations of the impact of the scientific output of scholars in time, unlike the $h$-index, which is a growing measure taking into account the whole career path.
INTRODUCTION
Journal impact factors can give an incomplete and potentially unfair proxy for an individual scholar’s work because citation counts across papers are skewed. The paper introduces AIF as a dynamic author-level measure, addressing the limits of cumulative indicators and the lack of empirical study of AIF’s properties.
INTRODUCTION
- Skewed citation distributions make journal impact factor a partial, potentially unfair proxy for an individual scholar’s research impact.The paper motivates AIF as an author-level measure of recent work, computed analogously to journal impact factor.Source: INTRODUCTION · p. 1
- Although AIF had appeared in earlier literature, the authors report that its empirical properties had not been studied.They present their paper as filling that gap.Source: INTRODUCTION · p. 1
- Cumulative measures such as the h-index combine a scientist’s whole career, while productivity and impact vary over time and career trajectories differ.The authors argue that past performance may not proxy current or future achievements, motivating dynamic measures.Source: INTRODUCTION · p. 1
- The authors link h-index-focused publication incentives to growth in low-quality practices, since cumulative measures do not penalize low-quality work.They suggest that rewarding high-quality work and penalizing negligible or incremental work might discourage such practices.Source: INTRODUCTION · pp. 1–2
- AIF tracks whether a scholar’s impact is rising or declining, which the authors say may matter in hiring, especially for young scientists.Because AIF averages citations to all papers published in a given window, low-quality work can keep the score down; the authors suggest this might encourage quality-focused research.Source: INTRODUCTION · p. 2
RESULTS
AIF’s history shows structure that reflects scientists’ career evolution, unlike the comparatively plain profiles of other measures. The authors use a five-year window to balance fluctuations in publication counts against the decline in citations to older papers.
RESULTS
- AIF averages citations in year t to an author’s papers published in the preceding Δt years.The definition compares citations to the author’s recent papers with the total number of papers the author published in that window.Source: RESULTS · p. 2
- The authors choose a five-year window because longer aggregation reduces publication-count fluctuations, while citation gaps longer than five years are suppressed.For 1980–2010, the share of citations to papers from the preceding five years was 44% in Physics, 47% in Medicine, 41% in Chemistry, and 30% in Mathematics; the authors also report more fluctuation with a two-year window.Source: RESULTS · pp. 2–3, Figure 2 · p. 3
- For most scientists, yearly h-index increments fluctuate between 0 and 3, and neither yearly increments nor the five-year h-index necessarily captures current impact.The five-year h-index can count papers published at any point in a scientist’s career, so it may not reveal impact from papers published in the preceding five years.Source: RESULTS · p. 3
- AIF histories show rich structure reflecting scientists’ career evolution, while other metrics have comparatively plain profiles that reveal little beyond possible career takeoff.Years when AIF differs significantly from its preceding five-year average are also marked.Source: RESULTS · p. 3
Nobel Laureates in Physics
AIF profiles align with notable phases of Nobel laureates’ research, revealing multiple periods of impact across their careers.
Nobel Laureates in Physics
- Anderson’s AIF peaks follow his work on localization, symmetry breaking and spin glasses, and high-temperature superconductivity, while other reported indicators miss these highlights.The passage links the later peaks to work in the 1970s and 1980s.Source: Nobel Laureates in Physics · pp. 3–4
- Weinberg’s profile has a pronounced peak around 1970 and another around 1980, while Gross’s peaks follow work on asymptotic freedom and heterotic string theory.The early-1970s bump for Gross is attributed to citations to his asymptotic-freedom paper.Source: Nobel Laureates in Physics · p. 4
- Ketterle’s AIF shows a large peak after his 1995 paper establishing Bose-Einstein condensation in trapped, cooled atoms.The passage identifies the peak as indicating the impact of that work.Source: Nobel Laureates in Physics · p. 4
Nobel Laureates in Chemistry
The Chemistry laureates’ AIF profiles show peaks and bumps associated with distinct influential works across their careers.
Nobel Laureates in Chemistry
- Sharpless’s AIF profile shows a mid-1980s bump associated with asymmetric oxidation and a prominent mid-2000s spike associated with his 2001 click-chemistry paper.Source: Nobel Laureates in Chemistry · p. 4
- Cech’s 1982 work on RNA catalytic properties is associated with the most prominent bump in his AIF profile.Tsien’s 1985 Nobel paper on green fluorescent protein also feeds a major early peak in his profile.Source: Nobel Laureates in Chemistry · p. 4
- Ciechanover’s AIF profile has one major bump after his 1980 seminal paper and another from work on the ubiquitin-proteasome pathway in the 1990s.The latter work concerns the pathway and pathogenesis of human diseases.Source: Nobel Laureates in Chemistry · p. 4
Nobel Laureates in Physiology or Medicine
For laureates in Physiology or Medicine, AIF peaks align with citation activity around influential papers and later research contributions.
Nobel Laureates in Physiology or Medicine
- Baltimore’s profile has an early peak tied to 1962–63 work, a second early peak fed by his 1970 reverse-transcriptase paper, and its largest bump in the late 1980s.The late-1980s bump is linked to his 1986 co-discovery of NF-κ.Source: Nobel Laureates in Physiology or… · p. 4
- Blobel’s 1975 signal-peptide paper is associated with the most important structure in his AIF profile, while Tonegawa’s profile peaks in the years after his 1976 paper.Source: Nobel Laureates in Physiology or… · p. 5
- Horvitz’s profile has a second-highest peak after his 1986 work on cell-death genes, a more pronounced peak around 1995, and a rising trend in the latest years.The later patterns are linked respectively to a 1993 work on amyotrophic lateral sclerosis and recent work on micro-RNA expression profiles classifying human cancers.Source: Nobel Laureates in Physiology or… · p. 5
DISCUSSION
The discussion presents AIF as a dynamic indicator of recent scientific impact, while noting the limits of its citation window.
DISCUSSION
- AIF tracks the evolution of scholars’ performance over their careers, especially trends, and the laureate analyses illustrate sensitivity to sharp citation-flow changes after groundbreaking papers.Source: DISCUSSION · p. 5
- The authors describe AIF as a simple author analogue of journal Impact Factor that bibliographic portals could compute.Source: DISCUSSION · p. 5
- AIF can show whether a major paper is an isolated success or part of a consistent record of significant work.The passage notes that checking each paper’s citation count can also reveal this distinction.Source: DISCUSSION · p. 5
- The authors argue that AIF penalizes low-quality work and could discourage negligible papers and honorary authorship.Source: DISCUSSION · p. 5
- AIF can compare who has been more impactful in recent years regardless of career age, unlike cumulative measures that increase with career age.The discussion frames this as relevant to identifying promising recruits or grant holders.Source: DISCUSSION · p. 5
- A five-year AIF counts citations received within five years of publication, so it can miss papers whose citations peak later.The discussion calls these papers “sleeping beauties.”Source: DISCUSSION · pp. 5–6
- An immediacy-based, time-window-free AIF would require much longer citation histories and could change later as delayed citation boosts arrive.Source: DISCUSSION · p. 6
- The authors distinguish AIF from h-index extensions, which have low discriminative power, and average-citation measures, which reuse papers across consecutive windows.They also note that i10 and total recent citations do not penalize low-quality work.Source: DISCUSSION · p. 6
METHODS
The analyzed dataset includes careers of Nobel laureates and publication profiles of scientists grouped into four broad categories.
METHODS
- The study analyzed the careers of 12 Nobel Laureates.The passage identifies 12 laureates in total.Source: METHODS · p. 6
- The study also analyzed publication profiles of 350 scientists divided into four broad categories.Source: METHODS · p. 6
- The study analyzed careers of 12 Nobel Laureates—four each in Physics, Chemistry, and Physiology or Medicine—and publication profiles of 350 scientists across four broad categories, including 100 prominent physicists.Source: METHODS · p. 6
Appendix A: Methods
The authors used disambiguated author data and publication profiles to examine Nobel laureates and selected scientist groups. Supplementary figures show AIF alongside other measures for Nobel laureates across three fields.
Appendix A: Methods
- The dataset covers 12 Nobel Laureates, with four each in Physics, Chemistry, and Physiology or Medicine.Source: Appendix A · p. 7
- The authors used disambiguated Thomson Reuters Web of Knowledge author data and ResearcherID publication profiles.Source: Appendix A · p. 7
- The 550-scientist selection includes prolific authors from Physical Review Letters, CELL, and Annals of Mathematics, plus 100 young physics assistant professors.The passage specifies 100 physicists, 100 cell biologists, 50 mathematicians, and 100 assistant professors; the mathematician criterion is truncated.Source: Appendix A · pp. 7–8
- The supplementary figures show AIF for laureates in Physics, Chemistry, and Physiology or Medicine, with the Physics figure using a 2-year aggregation period.Source: Figure S1 · p. 7, Figure S2 · p. 7, Figure S3 · p. 8
- Papers were categorized using Thomson Reuters subject categories, which were aggregated into broader scientific fields.Source: Appendix A · p. 8
Appendix B: Results
For Nobel laureates, the AIF calculated over two years fluctuates more than the five-year AIF but varies qualitatively similarly. Table S1 groups Thomson Reuters subject categories into broader fields.
Appendix B: Results
- The 2-year AIF fluctuates more than the 5-year AIF, but their variation is qualitatively similar for the prominent scientists shown.The passage cautions that a 2-year window is short for most scientists, especially younger ones.Source: Appendix B · p. 8
- Table S1 aggregates Thomson Reuters subject categories into broader fields, including categories for physics, mathematics, and chemistry.Source: Appendix B · p. 8, Table S1 · p. 9