TY - JOUR

T1 - New methodologies for the digital age?

T2 - How methods (re-)organize research using social media data

AU - Fan, Yangliu

AU - Lehmann, Sune

AU - Blok, Anders

N1 - Publisher Copyright: © 2023 Yangliu Fan, Sune Lehmann, and Anders Blok.

PY - 2023

Y1 - 2023

N2 - As “big and broad” social media data continues to expand and become a more prevalent source for research, much remains to be understood about its epistemological and methodological implications. Drawing on an original data set of 12,732 research articles using social media data, we employ a novel dictionary-based approach to map the use of methods. Specifically, our approach draws on a combination of manual coding and embedding-enhanced query expansion. We cluster journals in groups of densely connected research communities to investigate how heterogeneous these groups are in terms of the methods used. First, our results indicate that research in this domain is largely organized by methods. Some communities tend to have a monomethod culture, and others combine methods in novel ways. Comparing practices across communities, we observe that computational methods have penetrated many research areas but not the research space surrounding ethnography. Second, we identify two core axes of variation—social sciences vs. computer science and methodological individualism vs. relationalism—that organize the domain as a whole, suggesting new methodological divisions and debates.

AB - As “big and broad” social media data continues to expand and become a more prevalent source for research, much remains to be understood about its epistemological and methodological implications. Drawing on an original data set of 12,732 research articles using social media data, we employ a novel dictionary-based approach to map the use of methods. Specifically, our approach draws on a combination of manual coding and embedding-enhanced query expansion. We cluster journals in groups of densely connected research communities to investigate how heterogeneous these groups are in terms of the methods used. First, our results indicate that research in this domain is largely organized by methods. Some communities tend to have a monomethod culture, and others combine methods in novel ways. Comparing practices across communities, we observe that computational methods have penetrated many research areas but not the research space surrounding ethnography. Second, we identify two core axes of variation—social sciences vs. computer science and methodological individualism vs. relationalism—that organize the domain as a whole, suggesting new methodological divisions and debates.

KW - big social data

KW - research methods

KW - science studies

KW - social media data

KW - word embedding

U2 - 10.1162/qss_a_00271

DO - 10.1162/qss_a_00271

M3 - Journal article

AN - SCOPUS:85186568726

VL - 4

SP - 976

EP - 996

JO - Quantitative Science Studies

JF - Quantitative Science Studies

SN - 2641-3337

IS - 4

ER -