Andrew Herman
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Andrew Herman has a total of 11 co-authors. Here are the top co-authors by number of shared publications:
- Jens Peter Andersen
The Danish Centre for Studies in Research and Research Policy, Aarhus University
4 shared publications - Mathias Wullum Nielsen
Department of Sociology, University of Copenhagen, Copenhagen, Denmark
4 shared publications - Bas Hofstra
Department of Sociology, Radboud University, Nijmegen, The Netherlands
2 shared publications - Andrew Caines
University of Cambridge
1 shared publication - Charles Jonathan Gomez
School of Sociology, The University of Arizona, Tucson, USA
1 shared publication - Diogo L. Pires
Copenhagen Center for Social Data Science (SODAS), University of Copenhagen, Copenhagen, Denmark
1 shared publication
Latest Publications
- Open Access
- 24 September 2026
- Published by Cambridge University Press (CUP) in Cambridge Open Engage
Multi-Label Active Learning for Doxing Detection: Enabling Large-Scale Analysis of the Dynamics and Network Organization of Harmful Behavior
Doxing is often studied as a harmful disclosure of personal information, but fine-grained detection requires distinguishing how such information is used against a target. Studying these distinctions at scale requires labels difficult to produce when relevant behaviors are rare, unevenly distributed, and overlapping. We address this challenge with a multi-label scheme anchored in the established distinction among deanonymizing, targeting, and delegitimizing doxing, and we extend that scheme to related processes through which personal information is requested, morally framed, or connected to calls for action. Using multi-label active learning with RoBERTa, we apply this framework to 1.3 million comments from three target-centered boards in a hostile forum environment from 2014 to 2025. After each annotation cycle, RoBERTa is retrained on the expanded set, while uncertainty and diversity guide selection of informative cases. This makes fine-grained labeling feasible at corpus scale while keeping reliability visible. Demonstrating this value, key findings include that deanonymizing and targeting doxing are rare, textually elaborate, strongly connected, and locally concentrated, whereas delegitimizing doxing is far more common in our setting. Doxing follows no uniform life cycle here but varies across target contexts and can continue beyond early identification across years of hostile discussion. Requests for information are consistent with a preparatory or validating role, hostile moral framing operates as a broad background condition, and action-oriented hostility is more context-specific. Timestamped reply links and reactions show that doxing-related output is concentrated among highly active and interactionally prominent users and that prior same-type interaction is associated with later doxing.
- 24 September 2026
- Published by Cambridge University Press (CUP) in Cambridge Open Engage
Open AccessMulti-Label Active Learning for Doxing Detection: Enabling Large-Scale Analysis of the Dynamics and Network Organization of Harmful Behavior
Doxing is often studied as a harmful disclosure of personal information, but fine-grained detection requires distinguishing how such information is used against a target. Studying these distinctions at scale requires labels difficult to produce when relevant behaviors are rare, unevenly distributed, and overlapping. We address this challenge with a multi-label scheme anchored in the established distinction among deanonymizing, targeting, and delegitimizing doxing, and we extend that scheme to related processes through which personal information is requested, morally framed, or connected to calls for action. Using multi-label active learning with RoBERTa, we apply this framework to 1.3 million comments from three target-centered boards in a hostile forum environment from 2014 to 2025. After each annotation cycle, RoBERTa is retrained on the expanded set, while uncertainty and diversity guide selection of informative cases. This makes fine-grained labeling feasible at corpus scale while keeping reliability visible. Demonstrating this value, key findings include that deanonymizing and targeting doxing are rare, textually elaborate, strongly connected, and locally concentrated, whereas delegitimizing doxing is far more common in our setting. Doxing follows no uniform life cycle here but varies across target contexts and can continue beyond early identification across years of hostile discussion. Requests for information are consistent with a preparatory or validating role, hostile moral framing operates as a broad background condition, and action-oriented hostility is more context-specific. Timestamped reply links and reactions show that doxing-related output is concentrated among highly active and interactionally prominent users and that prior same-type interaction is associated with later doxing.
- Open Access
- 28 February 2026
- Published by Springer Nature in Higher Education
- p. 1-18
The geographical diversity of the citation elite in STEMM is decreasing, even as the overall scientific workforce has diversified
The expansion of research systems in low- and middle-income countries has shifted science to a more diverse geographical structure. Using a global dataset of 40 million authors in Science, Technology, Engineering, Mathematics and Medical fields (STEMM), we investigate whether this shift has made it easier or more difficult for newcomers residing in the Global South to rise to prominence in their disciplines. We tracked yearly cohorts of authors who began their publication careers between 2000 and 2014, finding that the South-North gap in the likelihood of joining the ‘citation elite’ (the top 5% of most cited authors per cohort) has increased by approximately 23%. Southern Asia and South-Eastern Asia outperform Eastern Asia and Latin America and the Caribbean in terms of changing regional shares in the citation elite. South-North gaps in citation elite membership also appear to be larger and increasing in the life sciences, and smaller and shrinking in some engineering-related subfields. Using decomposition approaches, we attribute most of the widening South-North gap in citation elite membership to diverging trends in attrition rates, publication outputs and journal selection, with changes in team sizes and regional migration playing only a minor role. These findings highlight the enduring dominance of wealthier nations in the production and distribution of scientific knowledge. While the scientific workforce has diversified globally and opportunities for Global South individuals to pursue successful science careers in the Global North have expanded, regional disparities persist with scientists based in the Global South now less likely to achieve scientific prominence relative to their Global North counterparts than previously.
- 28 February 2026
- Published by Springer Nature in Higher Education
- p. 1-18
Open AccessThe geographical diversity of the citation elite in STEMM is decreasing, even as the overall scientific workforce has diversified
The expansion of research systems in low- and middle-income countries has shifted science to a more diverse geographical structure. Using a global dataset of 40 million authors in Science, Technology, Engineering, Mathematics and Medical fields (STEMM), we investigate whether this shift has made it easier or more difficult for newcomers residing in the Global South to rise to prominence in their disciplines. We tracked yearly cohorts of authors who began their publication careers between 2000 and 2014, finding that the South-North gap in the likelihood of joining the ‘citation elite’ (the top 5% of most cited authors per cohort) has increased by approximately 23%. Southern Asia and South-Eastern Asia outperform Eastern Asia and Latin America and the Caribbean in terms of changing regional shares in the citation elite. South-North gaps in citation elite membership also appear to be larger and increasing in the life sciences, and smaller and shrinking in some engineering-related subfields. Using decomposition approaches, we attribute most of the widening South-North gap in citation elite membership to diverging trends in attrition rates, publication outputs and journal selection, with changes in team sizes and regional migration playing only a minor role. These findings highlight the enduring dominance of wealthier nations in the production and distribution of scientific knowledge. While the scientific workforce has diversified globally and opportunities for Global South individuals to pursue successful science careers in the Global North have expanded, regional disparities persist with scientists based in the Global South now less likely to achieve scientific prominence relative to their Global North counterparts than previously.
- Open Access
- 20 February 2026
- Published by Center for Open Science in OSF Preprints
The geographical diversity of the citation elite in STEMM is decreasing, even as the overall scientific workforce has diversified
The expansion of research systems in low- and middle-income countries has shifted science to a more diverse geographical structure. Using a global dataset of 40 million authors in Science, Technology, Engineering, Mathematics and Medical fields (STEMM), we investigate whether this shift has made it easier or more difficult for newcomers residing in the Global South to rise to prominence in their disciplines. We tracked yearly cohorts of authors who began their publication careers between 2000 and 2014, finding that the South-North gap in the likelihood of joining the ‘citation elite’ (the top 5% of most cited authors per cohort) has increased by approximately 23%. Southern Asia and South-Eastern Asia outperform Eastern Asia and Latin America and the Caribbean in terms of changing regional shares in the citation elite. South-North gaps in citation elite membership also appear to be larger and increasing in the life sciences, and smaller and shrinking in some engineering-related subfields. Using decomposition approaches, we attribute most of the widening South-North gap in citation elite membership to diverging trends in attrition rates, publication outputs and journal selection, with changes in team sizes and regional migration playing only a minor role. These findings highlight the enduring dominance of wealthier nations in the production and distribution of scientific knowledge. While the scientific workforce has diversified globally and opportunities for Global South individuals to pursue successful science careers in the Global North have expanded, regional disparities persist with scientists based in the Global South now less likely to achieve scientific prominence relative to their Global North counterparts than previously.
- 20 February 2026
- Published by Center for Open Science in OSF Preprints
Open AccessThe geographical diversity of the citation elite in STEMM is decreasing, even as the overall scientific workforce has diversified
The expansion of research systems in low- and middle-income countries has shifted science to a more diverse geographical structure. Using a global dataset of 40 million authors in Science, Technology, Engineering, Mathematics and Medical fields (STEMM), we investigate whether this shift has made it easier or more difficult for newcomers residing in the Global South to rise to prominence in their disciplines. We tracked yearly cohorts of authors who began their publication careers between 2000 and 2014, finding that the South-North gap in the likelihood of joining the ‘citation elite’ (the top 5% of most cited authors per cohort) has increased by approximately 23%. Southern Asia and South-Eastern Asia outperform Eastern Asia and Latin America and the Caribbean in terms of changing regional shares in the citation elite. South-North gaps in citation elite membership also appear to be larger and increasing in the life sciences, and smaller and shrinking in some engineering-related subfields. Using decomposition approaches, we attribute most of the widening South-North gap in citation elite membership to diverging trends in attrition rates, publication outputs and journal selection, with changes in team sizes and regional migration playing only a minor role. These findings highlight the enduring dominance of wealthier nations in the production and distribution of scientific knowledge. While the scientific workforce has diversified globally and opportunities for Global South individuals to pursue successful science careers in the Global North have expanded, regional disparities persist with scientists based in the Global South now less likely to achieve scientific prominence relative to their Global North counterparts than previously.
- Open Access
- 1 January 2026
- Published by Society for Sociological Science in Sociological Science
- Vol. 13, 1132-1158
Developing Countries Face Falling Circulation Rates and Rising Migration Losses in Science
As universities and doctoral training expand in low- and middle-income (LMI) countries, transnational scientific mobility is rising. This trend is often framed as “brain circulation,” implying a mutual exchange of talent and knowledge to the benefit of all countries involved. We track 12 million scientists over 15 years, stratified by performance, to weigh this view against the prediction from structural dependency theory that expanding scientific capacity in lower-income countries may intensify rather than attenuate the loss of top-performing talent to wealthier countries. Our findings temper the brain circulation account. Circulatory migration rates have declined over time, even as the net losses from migration for LMI countries have risen. Not one of the countries that together produce 90 percent of LMI countries’ top-cited early-career scientists achieves a net migration gain relative to high-income countries. It has become more difficult for developing countries to offset migratory losses with the spillovers from circulation.
- 1 January 2026
- Published by Society for Sociological Science in Sociological Science
- Vol. 13, 1132-1158
Open AccessDeveloping Countries Face Falling Circulation Rates and Rising Migration Losses in Science
As universities and doctoral training expand in low- and middle-income (LMI) countries, transnational scientific mobility is rising. This trend is often framed as “brain circulation,” implying a mutual exchange of talent and knowledge to the benefit of all countries involved. We track 12 million scientists over 15 years, stratified by performance, to weigh this view against the prediction from structural dependency theory that expanding scientific capacity in lower-income countries may intensify rather than attenuate the loss of top-performing talent to wealthier countries. Our findings temper the brain circulation account. Circulatory migration rates have declined over time, even as the net losses from migration for LMI countries have risen. Not one of the countries that together produce 90 percent of LMI countries’ top-cited early-career scientists achieves a net migration gain relative to high-income countries. It has become more difficult for developing countries to offset migratory losses with the spillovers from circulation.
- Open Access
- 12 November 2024
- Published by Center for Open Science in OSF Preprints
Brain drain outweighs brain circulation between the Global North and South
International research mobility is increasing, as is the expectation that high- and low-income countries will mutually benefit. The once-dominant view of research mobility as “brain drain” has been replaced by a more optimistic narrative of “brain circulation”, built on the idea of cyclical talent flows that advantage all countries involved. But does this brain circulation model reflect the full picture for low- and middle-income countries, whose development depends on producing and retaining scientific talent? Our analysis of 16 years of international mobility flows in science suggests otherwise. While brain circulation increasingly occurs through temporary positions, joint affiliations, and return migration, a persistent pattern of brain drain remains, with Global South countries losing a growing share of their best-performing scientists to the Global North. Moreover, a differentiated hierarchy of talent flows is visible within the Global South. Some of the countries that lose the most talent to the Global North also attract the most talent from other Global South nations, offsetting their losses. But many others see substantial outflows of local talent to both the Global North and other Global South regions, compounding the issue. This differentiated pattern is also visible in the multiplex network structure of national inflows and outflows, where the relationship between sending talent to the Global North and receiving talent from the Global South is weaker than one would expect in random networks with a similar structure. Overall, our findings suggest a mixed reality. While visiting and return migration likely produce mutual benefits for Global North and Global South countries alike, this talent circulation is unlikely to offset the Global South's direct and indirect loss of its best-performing scientists.
- 12 November 2024
- Published by Center for Open Science in OSF Preprints
Open AccessBrain drain outweighs brain circulation between the Global North and South
International research mobility is increasing, as is the expectation that high- and low-income countries will mutually benefit. The once-dominant view of research mobility as “brain drain” has been replaced by a more optimistic narrative of “brain circulation”, built on the idea of cyclical talent flows that advantage all countries involved. But does this brain circulation model reflect the full picture for low- and middle-income countries, whose development depends on producing and retaining scientific talent? Our analysis of 16 years of international mobility flows in science suggests otherwise. While brain circulation increasingly occurs through temporary positions, joint affiliations, and return migration, a persistent pattern of brain drain remains, with Global South countries losing a growing share of their best-performing scientists to the Global North. Moreover, a differentiated hierarchy of talent flows is visible within the Global South. Some of the countries that lose the most talent to the Global North also attract the most talent from other Global South nations, offsetting their losses. But many others see substantial outflows of local talent to both the Global North and other Global South regions, compounding the issue. This differentiated pattern is also visible in the multiplex network structure of national inflows and outflows, where the relationship between sending talent to the Global North and receiving talent from the Global South is weaker than one would expect in random networks with a similar structure. Overall, our findings suggest a mixed reality. While visiting and return migration likely produce mutual benefits for Global North and Global South countries alike, this talent circulation is unlikely to offset the Global South's direct and indirect loss of its best-performing scientists.
- Open Access
- 18 October 2024
- Published by Center for Open Science in OSF Preprints
Reconstructing bibliometric methods for studying mobility
Understanding how the international or regional mobility of researchers affects the production of knowledge has become a fundamental goal in research on science and innovation. Bibliometric methods that use the affiliations on publications to infer when researchers move marked a large step forward in this respect. While the field now has multiple ways of going about this work, it has not paid close attention to the impact that differences in research assumptions have on reported results. This paper introduces a new framework for classifying researcher mobility in bibliometric data, the tenure events method, that makes full use of the advantages of existing approaches while avoiding their main pitfalls, improving the overall quality of resulting classifications. It also explicitly parameterizes three key research decisions, rendering the method more transparent than what already exists, and making it easier to run sensitivity checks in the spirit of researcher degrees of freedom and multiversal analysis. The new method is implemented with biblio.migration, an R package introduced here in the hopes of facilitating more robust research in the area.
- 18 October 2024
- Published by Center for Open Science in OSF Preprints
Open AccessReconstructing bibliometric methods for studying mobility
Understanding how the international or regional mobility of researchers affects the production of knowledge has become a fundamental goal in research on science and innovation. Bibliometric methods that use the affiliations on publications to infer when researchers move marked a large step forward in this respect. While the field now has multiple ways of going about this work, it has not paid close attention to the impact that differences in research assumptions have on reported results. This paper introduces a new framework for classifying researcher mobility in bibliometric data, the tenure events method, that makes full use of the advantages of existing approaches while avoiding their main pitfalls, improving the overall quality of resulting classifications. It also explicitly parameterizes three key research decisions, rendering the method more transparent than what already exists, and making it easier to run sensitivity checks in the spirit of researcher degrees of freedom and multiversal analysis. The new method is implemented with biblio.migration, an R package introduced here in the hopes of facilitating more robust research in the area.
- Open Access
- 15 September 2023
- Published by Center for Open Science in SocArXiv
What is powerful about the state’s discursive power? Road navigation in the United States during the technology transition to automobiles
States are known to hold power not only through coercion, but through social discourse. Nonetheless, the author here argues that this claim is built on shaky ground. Using a counter-factual case study with an implicit comparative design, this paper asks what happens when a state chooses not to engage with social discourse, when it otherwise would be expected to do so. Are there noticeable differences in social discourse when the state does not get involved? Considering the technology transition to the automobile in the early 20th C, historical analysis shows that the outcomes of social discourse around road navigation were extremely similar in Europe, where states were heavily involved in shaping the discourse, and in the United States, where the state was absent. This strongly suggests that neither the states in Europe, nor the state in the United States, were powerful through discourse, as what could “power” possibly mean if it has no consequences under the counter-factual? It also indicates that researchers may overstate their ability to distinguish between situations where actors are powerful through social discourse, what the author calls attributable power, and situations where actors are merely involved in social discourse, but without causing discernible changes to the outcomes with their involvement, what the author calls unattributable power.
- 15 September 2023
- Published by Center for Open Science in SocArXiv
Open AccessWhat is powerful about the state’s discursive power? Road navigation in the United States during the technology transition to automobiles
States are known to hold power not only through coercion, but through social discourse. Nonetheless, the author here argues that this claim is built on shaky ground. Using a counter-factual case study with an implicit comparative design, this paper asks what happens when a state chooses not to engage with social discourse, when it otherwise would be expected to do so. Are there noticeable differences in social discourse when the state does not get involved? Considering the technology transition to the automobile in the early 20th C, historical analysis shows that the outcomes of social discourse around road navigation were extremely similar in Europe, where states were heavily involved in shaping the discourse, and in the United States, where the state was absent. This strongly suggests that neither the states in Europe, nor the state in the United States, were powerful through discourse, as what could “power” possibly mean if it has no consequences under the counter-factual? It also indicates that researchers may overstate their ability to distinguish between situations where actors are powerful through social discourse, what the author calls attributable power, and situations where actors are merely involved in social discourse, but without causing discernible changes to the outcomes with their involvement, what the author calls unattributable power.
- 19 May 2023
- Published by Orvium
Editorial gatekeeping up and down the journal hierarchy
How do journal editors affect what papers and which authors get published? The research presented in the proposed presentation builds on a novel dataset that provides coverage of all journal editors across the full ecosystem of social science journals. This enables the authors to investigate the variation in how editors affect the publication process across the journal hierarchy, and across types of journals. In addition to the unique data, the authors take advantages in the computational social sciences to develop a state-of-the-art methodology for evaluating the role of editors in shaping publications and the evolution of science.
- 19 May 2023
- Published by Orvium
Editorial gatekeeping up and down the journal hierarchy
How do journal editors affect what papers and which authors get published? The research presented in the proposed presentation builds on a novel dataset that provides coverage of all journal editors across the full ecosystem of social science journals. This enables the authors to investigate the variation in how editors affect the publication process across the journal hierarchy, and across types of journals. In addition to the unique data, the authors take advantages in the computational social sciences to develop a state-of-the-art methodology for evaluating the role of editors in shaping publications and the evolution of science.
- Top Cited PapersOpen Access
- 30 May 2022
- Published by Springer Nature in Nature Human Behaviour
- Vol. 6 (7) , 919-929
Leading countries in global science increasingly receive more citations than other countries doing similar research
Citations and text analysis are both used to study the distribution and flow of ideas between researchers, fields and countries, but the resulting flows are rarely equal. We argue that the differences in these two flows capture a growing global inequality in the production of scientific knowledge. We offer a framework called ‘citational lensing’ to identify where citations should appear between countries but are absent given that what is embedded in their published abstract texts is highly similar. This framework also identifies where citations are overabundant given lower similarity. Our data come from nearly 20 million papers across nearly 35 years and 150 fields from the Microsoft Academic Graph. We find that scientific communities increasingly centre research from highly active countries while overlooking work from peripheral countries. This inequality is likely to pose substantial challenges to the growth of novel ideas.
- 30 May 2022
- Published by Springer Nature in Nature Human Behaviour
- Vol. 6 (7) , 919-929
Top Cited PapersOpen AccessLeading countries in global science increasingly receive more citations than other countries doing similar research
Citations and text analysis are both used to study the distribution and flow of ideas between researchers, fields and countries, but the resulting flows are rarely equal. We argue that the differences in these two flows capture a growing global inequality in the production of scientific knowledge. We offer a framework called ‘citational lensing’ to identify where citations should appear between countries but are absent given that what is embedded in their published abstract texts is highly similar. This framework also identifies where citations are overabundant given lower similarity. Our data come from nearly 20 million papers across nearly 35 years and 150 fields from the Microsoft Academic Graph. We find that scientific communities increasingly centre research from highly active countries while overlooking work from peripheral countries. This inequality is likely to pose substantial challenges to the growth of novel ideas.
- 31 December 2016
- Published by University of Toronto Press Inc. (UTPress)
- p. 108-112
7. Conclusion