BERT, RoBERTa, or DeBERTa? Comparing Performance Across Transformers Models in Political Science Text
The Journal of Politics · 2025-01-01 · 87 (1) · 347-364
DOI: 10.1086/730737 ↗At a glance
Journal article in The Journal of Politics, published on 2025-01-01. Written by Joan C. Timoneda and 1 co-authors. Cited 45 times so far (Crossref); 51 references.
Abstract (original language)
Transformer models such as BERT, RoBERTa, and DeBERTa have revolutionized the field of Natural Language Processing in recent years with substantial improvements in the contextual understanding of text. While political scientists have begun adopting these models, their performance differences are not well understood, especially in cross-lingual applications. This article introduces Transformer models, compares their performance using three different text-as-data political science projects
Cite
APA
Timoneda, J. C., & Vera, S. V. (2025). BERT, RoBERTa, or DeBERTa? Comparing Performance Across Transformers Models in Political Science Text. The Journal of Politics, 87(1), 347-364. https://doi.org/10.1086/730737
Vancouver
Timoneda JC, Vera SV. BERT, RoBERTa, or DeBERTa? Comparing Performance Across Transformers Models in Political Science Text. The Journal of Politics. 2025;87(1):347-364. doi:10.1086/730737
BibTeX
@article{Timoneda2025,
author = {Timoneda, Joan C. and Vera, Sebastián Vallejo},
title = {BERT, RoBERTa, or DeBERTa? Comparing Performance Across Transformers Models in Political Science Text},
journal = {The Journal of Politics},
year = {2025},
volume = {87},
number = {1},
pages = {347--364},
doi = {10.1086/730737}
}
EUVIMED