Chhandak Bagchi
@cbagchi.bsky.social
CS PhD student at UMass Amherst.
ML and computational social science.
people.umass.edu/cbagchi
ML and computational social science.
people.umass.edu/cbagchi
📢 @przemyslslaw.bsky.social will present this work Tuesday at #ICWSM2025 at 5 PM in Room CSS 35.01.05
📄 Paper: ojs.aaai.org/index.php/IC...
Come say hi & let’s talk about the future of publishing 🚀
📄 Paper: ojs.aaai.org/index.php/IC...
Come say hi & let’s talk about the future of publishing 🚀
Effects of Research Paper Promotion via ArXiv and X
| Proceedings of the International AAAI Conference on Web and Social Media
ojs.aaai.org
June 23, 2025 at 5:25 PM
📢 @przemyslslaw.bsky.social will present this work Tuesday at #ICWSM2025 at 5 PM in Room CSS 35.01.05
📄 Paper: ojs.aaai.org/index.php/IC...
Come say hi & let’s talk about the future of publishing 🚀
📄 Paper: ojs.aaai.org/index.php/IC...
Come say hi & let’s talk about the future of publishing 🚀
This raises questions about merit vs. social momentum:
If social exposure skews citations, should we rethink how we measure impact?
We argue for:
— Better metrics
— Awareness of these dynamics
If social exposure skews citations, should we rethink how we measure impact?
We argue for:
— Better metrics
— Awareness of these dynamics
June 23, 2025 at 5:25 PM
This raises questions about merit vs. social momentum:
If social exposure skews citations, should we rethink how we measure impact?
We argue for:
— Better metrics
— Awareness of these dynamics
If social exposure skews citations, should we rethink how we measure impact?
We argue for:
— Better metrics
— Awareness of these dynamics
We find stronger social influence in CS than physics, supported by:
3. Greater variance in relative effect sizes for CS — prior works suggest that social influence leads to more randomness in rankings of items in digital platforms due to snowballing effects.
3. Greater variance in relative effect sizes for CS — prior works suggest that social influence leads to more randomness in rankings of items in digital platforms due to snowballing effects.
June 23, 2025 at 5:25 PM
We find stronger social influence in CS than physics, supported by:
3. Greater variance in relative effect sizes for CS — prior works suggest that social influence leads to more randomness in rankings of items in digital platforms due to snowballing effects.
3. Greater variance in relative effect sizes for CS — prior works suggest that social influence leads to more randomness in rankings of items in digital platforms due to snowballing effects.
We find stronger social influence in CS than physics, supported by:
2. Increasing treatment effects over time for CS (esp. tweeting & early ArXiv)
2. Increasing treatment effects over time for CS (esp. tweeting & early ArXiv)
June 23, 2025 at 5:25 PM
We find stronger social influence in CS than physics, supported by:
2. Increasing treatment effects over time for CS (esp. tweeting & early ArXiv)
2. Increasing treatment effects over time for CS (esp. tweeting & early ArXiv)
Interestingly, we find evidence of stronger social influence in CS than physics, supported by:
1. Higher relative ATE in CS across all interventions
1. Higher relative ATE in CS across all interventions
June 23, 2025 at 5:25 PM
Interestingly, we find evidence of stronger social influence in CS than physics, supported by:
1. Higher relative ATE in CS across all interventions
1. Higher relative ATE in CS across all interventions
We also looked at highly influential (HI) citations (via Semantic Scholar):
They’re affected similarly too.
They’re affected similarly too.
June 23, 2025 at 5:25 PM
We also looked at highly influential (HI) citations (via Semantic Scholar):
They’re affected similarly too.
They’re affected similarly too.
However, for physics, the effects were more modest:
📌 ArXiv-first → +4 citations
📌 Revisions → +4
📌 Tweeting → +7
📌 ArXiv-first → +4 citations
📌 Revisions → +4
📌 Tweeting → +7
June 23, 2025 at 5:25 PM
However, for physics, the effects were more modest:
📌 ArXiv-first → +4 citations
📌 Revisions → +4
📌 Tweeting → +7
📌 ArXiv-first → +4 citations
📌 Revisions → +4
📌 Tweeting → +7
Key findings for CS papers (5-year citation gain):
📌 ArXiv-first → +21 citations
📌 ArXiv revisions → +18
📌 Tweeting → +44
Statistically significant effects across the board.
📌 ArXiv-first → +21 citations
📌 ArXiv revisions → +18
📌 Tweeting → +44
Statistically significant effects across the board.
June 23, 2025 at 5:25 PM
Key findings for CS papers (5-year citation gain):
📌 ArXiv-first → +21 citations
📌 ArXiv revisions → +18
📌 Tweeting → +44
Statistically significant effects across the board.
📌 ArXiv-first → +21 citations
📌 ArXiv revisions → +18
📌 Tweeting → +44
Statistically significant effects across the board.
We analyzed 500k+ papers in CS and physics to ask:
— Does publishing on ArXiv before peer-reviewed publication boost citations?
— Does subsequent revising on ArXiv matter?
— Does tweeting your paper help?
— Does publishing on ArXiv before peer-reviewed publication boost citations?
— Does subsequent revising on ArXiv matter?
— Does tweeting your paper help?
June 23, 2025 at 5:25 PM
We analyzed 500k+ papers in CS and physics to ask:
— Does publishing on ArXiv before peer-reviewed publication boost citations?
— Does subsequent revising on ArXiv matter?
— Does tweeting your paper help?
— Does publishing on ArXiv before peer-reviewed publication boost citations?
— Does subsequent revising on ArXiv matter?
— Does tweeting your paper help?
Hi, thanks for creating this! Can you add me here as well if it's not full?
November 26, 2024 at 4:54 PM
Hi, thanks for creating this! Can you add me here as well if it's not full?