Source-linked AI summary
The golden cage of science: first-last author prestige gap buys impact over disruption
Chi Zhang, Yilin Bi, Tao Zhou
TL;DR
How prestige differences between first and last authors relate to research impact and disruption is an important question about team research outputs. This study compares those outcomes across the prestige gap and finds that larger gaps are associated with higher impact but lower disruption.
Problem
The relationship between first–last author prestige gaps and research outputs, particularly impact and disruption, is an important dimension of team research to characterize.
Source: Discussion · p. 6Method
The study compares article impact, measured by citations through 2020, and disruption, measured by the disruption index, across first–last author prestige gaps.
Source: Prestige gap buys impact over disruption · p. 2Results
Articles with larger first–last author prestige gaps tend to have higher impact and lower disruption, a trade-off that persists across team sizes, prior-collaboration intensities, and academic-age gaps.
Source: Discussion · pp. 6–7Takeaways & Limitations
The findings frame research-team organization as a balance between senior guidance and exploratory autonomy, considered through team composition and division of work.
Source: Discussion · p. 7Because the observational data cannot separate prestige-structure effects from team formation, the study cannot determine how much of the trade-off is attributable to prestige structure itself.
Source: Discussion · p. 7
Abstract
from arXiv · showhide
First and last authors make critical but distinct contributions in the majority of scientific papers. First authors tend to conduct research and analysis, whereas last authors are usually responsible for setting research directions and providing resources and supervision. The first-last author prestige gap (FL gap, $\Delta H$) is defined as the H-index difference between the last and first authors prior to the focal paper's publication. The FL gap serves as a measure of the prestige asymmetry between the two dominant collaborators. We find that larger FL gaps in either direction are associated with higher impact and lower disruption. This trade-off is robust across team sizes and intensities of prior collaboration and can not be explained by first and last authors' separate prestige effects or their academic age gap. Further analysis shows that variation in last author prestige is associated with larger differences in both the article's impact and disruption than variation in first author prestige. We refer to this coexistence of higher impact and lower disruption related to the FL gap as the ``golden cage of science'', which reflects a common collaboration structure that statistically produces highly cited but less disruptive work. These findings suggest the first-last author prestige gap as an important indicator for examining how research teams balance senior guidance with exploratory autonomy.
Prestige gap buys impact over disruption
Impact is measured by cumulative citations through 2020 and disruption by index D. A wider first–last author prestige gap is associated with higher impact and lower disruption, a trade-off examined across author prestige combinations and collaboration contexts.
Prestige gap buys impact over disruption
- Impact is V-shaped and disruption inverted V-shaped across the FL gap: both outcomes are least favorable when first- and last-author prestige is similar.For |∆H| > 5, observed impact is higher and disruption lower than predictions from a null model combining the authors’ separate prestige effects.Source: Prestige gap buys impact over disruption · pp. 2–3, Figure 2 · p. 4
Last author has higher influence
- Across different-tier pairings, placing the higher-prestige author last corresponds to worse disruption ranks but better citation ranks than placing that author first.In the Tier 5 & Tier 1 pairing, Tier 1 first rather than last is associated with a mean citation rank 1.74% worse and mean disruption rank 3.92% better.Source: Last author has higher influence · p. 4
Robustness and sensitivity analysis
- The FL-gap trade-off persists across team sizes and prior-collaboration levels; academic age gaps add little explanatory information once the FL gap is considered.It also holds across outcome percentiles, publication years, disciplines, article types, and alternative disruption-index variants.Source: Robustness and sensitivity analysis · pp. 4–6
Last author has higher influence
- With the other author’s tier fixed, differences in impact and disruption are larger across last-author than first-author prestige tiers.The Tier 5-5 pairing has the highest mean disruption and lowest median citation, while Tier 1-1 shows the opposite pattern.Source: Last author has higher influence · pp. 3–4, Figure 3 · p. 5
Discussion
- The authors propose that prestige-based endorsement, team hierarchy, and resource dependence may help explain how high impact coexists with low disruption.They suggest balancing senior guidance with exploratory autonomy; in the Tier 5 & Tier 1 pairing, disruption rank is about 4% better with the Tier-1 researcher first rather than last.Source: Discussion · p. 7
- Because the observational data cannot isolate prestige structure’s causal effects from team formation or task selection, the findings are correlational and the proposed mechanisms remain untested.The study lacks direct measures of resource support, direction-setting, and exploratory autonomy.Source: Discussion · p. 7
Methods
The methods test whether separate first- and last-author prestige effects can reproduce the observed patterns, and assess robustness across team, collaboration, career-stage, and article-type controls.
Null Model
- The null model predicts outcomes as weighted combinations of first- and last-author H-index-based predictions, without an interaction term between their prestige.It uses conditional means for disruption and conditional medians for impact; optimal weights are estimated by minimizing mean squared error.Source: Null Model · pp. 7–8
- C = 0.6463 for impact (R2 = 0.7177), yet optimized null-model predictions still deviate systematically from observed curves, suggesting author relationships may matter beyond separate prestige effects.The null model captures much of the overall pattern, but does not fully reproduce the observed trade-off.Source: Null Model · p. 8
Control Variables
- Robustness checks examine publication year, discipline, team size, prior collaboration intensity, and the academic-age gap between first and last authors.Team sizes are grouped as 2, 3–5, 6–9, and ≥10 authors; collaboration intensity is grouped into four ranges from no prior coauthorship to 0.53–1, and academic-age gaps into <−5, [−5, 5], (5, 20], and >20 years.Source: Control Variables · p. 8
- A SciBERT model [41] identifies reviews and original research from titles and abstracts, achieving 90.45% precision and a 92.79% F1-score.Applied to 5,305,920 articles, it classifies 3,989,519 (75.2%) as original research and 1,316,401 (24.8%) as reviews.Source: Control Variables · p. 8