Source-linked AI summary
Career on the Move: Geography, Stratification, and Scientific Impact
Pierre Deville, Dashun Wang, Roberta Sinatra, Chaoming Song, Vincent D. Blondel, Albert-Laszlo Barabasi
TL;DR
The paper addresses limited knowledge about scientists’ institutional mobility and its relationship to scientific outcomes. Using affiliation data from over 420,000 papers, it reconstructs career trajectories and finds localized, highly stratified movements: elite-to-lower-rank moves associate with modestly lower impact, while moves into elite institutions show no subsequent gain.
Problem
Limited evidence exists on scientists’ institutional mobility patterns and how career movements relate to scientific outcomes.
Method
The study tracks scientists’ affiliations across over 420,000 papers to reconstruct career trajectories and approximates institutional rank using total citations.
Results
Career movements are usually early, local, and stratified by institutional rank; elite-to-lower-rank moves associate with modestly decreased impact, whereas moves into elite institutions show no subsequent impact gain.
Takeaways & Limitations
Institutional career movements are empirically patterned rather than random, and moving into an elite institution does not on average produce a subsequent performance gain.
Takeaways & Limitations
Institutional ranking is approximated by total citations, although the authors report strong correlations with institutional h-index and publication count.
Abstract
from arXiv · showhide
Changing institutions is an integral part of an academic life. Yet little is known about the mobility patterns of scientists at an institutional level and how these career choices affect scientific outcomes. Here, we examine over 420,000 papers, to track the affiliation information of individual scientists, allowing us to reconstruct their career trajectories over decades. We find that career movements are not only temporally and spatially localized, but also characterized by a high degree of stratification in institutional ranking. When cross-group movement occurs, we find that while going from elite to lower-rank institutions on average associates with modest decrease in scientific performance, transitioning into elite institutions does not result in subsequent performance gain. These results offer empirical evidence on institutional level career choices and movements and have potential implications for science policy.
RESULTS
Scientists’ careers show localized mobility in time and space, strong stratification by institutional rank, and asymmetric performance outcomes across rank changes. The analysis uses publication affiliations and five-year citation impact to reconstruct trajectories and compare movements across institutions.
- Institutional characteristics: Institution size positively correlates with publication impact (R2 = 0.85) but has little influence on productivity (R2 = 0.43).Institution size and citation totals are highly heterogeneous, with a small number of large or highly cited institutions.
- Mobility patterns: Career moves are common but infrequent: only 14% of scientists never moved, while movers typically changed institutions once or twice.Most movements occurred early in careers, likely during the postdoctoral period.
- Mobility patterns: Career-move distances follow a power law with exponent γ = 0.65±0.053, indicating that short-distance moves are more likely than distant moves.The null model predicts an approximately flat distance distribution.
- Stratification by institutional rank: Most movements involve elite institutions, while transitions between lower-ranked institutions are rare; after accounting for institution size, moves cluster within elite and lower-rank clubs.The size-adjusted transition matrix compares observed movements with a shuffled null model, where M(i, j) > 1 denotes overrepresentation.
- Performance and rank changes: Moving to a lower-ranked institution is associated with negative average performance change, whereas moving into an elite institution produces no average performance gain.The stratification pattern remains similar for movements followed by either positive or negative individual performance changes.
DISCUSSION
Career movements are stratified by institutional rank and geography, with cross-group moves showing asymmetric performance associations. The study’s conclusions are bounded by a sample biased toward physicists from the 1960s–1980s with high career longevity.
- Career movement: Career movements are stratified: scientists from elite institutions tend to move to other elite institutions, while lower-rank scientists tend to move within similar ranks.This stratification remains robust across changes in individual performance before and after moves.
- Institutional stratification: Most movements involve elite institutions, while transitions between bottom institutions are rare because elite institutions have larger populations and therefore more observed events.The likelihood measure M(i, j) accounts for institution size when assessing whether transitions are over- or underrepresented.
- Career movement and impact: Elite-to-lower-rank moves are associated with a modest decrease in scientific impact, whereas moves into elite institutions show no average impact gain.The performance comparison is based on average changes associated with cross-group transitions.
- Limitations: The dataset restricts the study to a sample biased toward physicists from the 1960s–1980s with high career longevity.The authors identify broader datasets and cross-disciplinary analyses as avenues for testing generalization.
METHOD
The study reconstructs scientists’ career trajectories from publication affiliations, using disambiguated authors and institutions and filtering affiliations to reduce spurious movements. Institutions are ranked using publication, citation, and h-index measures, with transitions grouped by rank.
- Dataset: The APS dataset contains over 450,000 publications from nine journals spanning 117 years, with publication dates, author names, and affiliations.These records provide the publication-level basis for reconstructing institutional careers.
- Author disambiguation: Author disambiguation uses author information and paper metadata, including coauthors and citations, to identify 237,038 distinct scientists.The procedure addresses ambiguities and homonymies in author names.
- Affiliation disambiguation: Affiliation disambiguation combines geocoded information, affiliation-name similarity, and disambiguated authors to identify 4,052 institutions.The source contains 319,829 distinct affiliation names before disambiguation.
- Career trajectories: Career trajectories retain only institutions reported in at least two consecutive papers to remove short-term stays and affiliation errors.This filtering rule is used to detect institutional changes as career movements.
- Institutional ranking and transitions: Institutions are ranked using total papers, cumulative citations, and h-index, and about 6,000 transitions among 1,000 institutions are binned logarithmically into five rank groups.The bins support statistically significant estimates of transition probabilities and likelihoods.
- Impact measurement: Paper impact is measured with citations accumulated within five years after publication.The citation window is applied to papers in the APS data.