Dataset for: Ethnic Identity Contestation on Social Media: A Bibliometric Mapping of 172 Studies and the "Karo Bukan Batak" Gap

# Open Science Repository: Ethnic Identity Contestation on Social Media (172 Studies Corpus) [](https://creativecommons.org/licenses/by/4.0/)[](http://prisma-statement.org/)[]() ## 1. Overview and Project Description This repository contains the complete, transparent, and reproducible dataset, screening documentation, keyword co-occurrence network, and replication script associated with the scientific paper: > **ETHNIC IDENTITY CONTESTATION ON SOCIAL MEDIA: A BIBLIOMETRIC MAPPING OF 172 STUDIES AND THE “KARO BUKAN BATAK” GAP** > **Authors:** Loli Gunali Ginting, Humaizi, Iskandar Zulkarnain, Dewi Kurniawati > **Affiliation:** Department of Communication Science, Faculty of Social and Political Sciences, Universitas Sumatera Utara, Medan, Indonesia > **Target Conference:** *ISAMBUDDHA 2026 (International Conference on Social and Communication Studies)* > **Ethical Approval:** Komisi Etik Penelitian Sosial Humaniora, Universitas Sumatera Utara, No. 34/UN5/D/FISIP/PT/2026 (dated 15 October 2026). The study maps 172 peer-reviewed journal articles published between 2013 and 2026, selected from an initial pool of 2,981 Scopus records through a staged literature screening protocol. The analysis addresses publication trends, leading sources, citation patterns, author keyword networks (39 nodes, 157 edges, 6 Louvain communities), manual thematic clusters, and literature gaps concerning the *"Karo bukan Batak"* identity contestation on platform X. --- ## 2. Repository File Inventory ```textrepository_open_science_conference/├── 01_Corpus_172_Included_Studies.csv # Complete bibliographic metadata of the 172 included studies (UTF-8 CSV)├── 01_Corpus_172_Included_Studies.xlsx # Formatted Excel edition of the 172 included studies├── 01_Corpus_172_Scopus_Format.ris # Standard RIS citation file directly importable into VOSviewer, Bibliometrix, Zotero├── 02_Staged_Screening_756_Records.csv # Full 4-stage screening decisions across 756 candidates (UTF-8 CSV)├── 02_Staged_Screening_756_Records.xlsx # Multi-tab workbook (PRISMA Summary, Priority A [198], Priority B [558], Data Dictionary)├── 03_Keyword_Network_39Nodes_157Edges.csv # Edge list (157 weighted undirected edges) for Gephi, VOSviewer, NetworkX├── 03_Keyword_Network_Nodes.csv # Node table (39 keywords, frequencies, Louvain community assignments 1-6)├── 04_Keyword_Normalization_Rules.csv # Mapping table documenting synonym merging (e.g. political polarization -> polarization)├── 05_Reproduce_Bibliometric_Analysis.py # Standalone Python script reproducing Tables 1-6 and descriptive metrics├── README.md # Comprehensive documentation, codebook, and reproduction guide (this file)├── CITATION.cff # Machine-readable Citation File Format for GitHub and Zenodo└── LICENSE # Creative Commons Attribution 4.0 International (CC-BY 4.0) License``` --- ## 3. PRISMA Literature Selection Architecture (Table 1) The selection followed a transparent four-stage pipeline: | Stage | Description | Records | Notes / Breakdown || :--- | :--- | :---: | :--- || **Identification** | Raw records exported from Scopus | **2,981** | Search query combining ethnic identity, social media, and polarization/SNA || **Initial filtering** | Article type, English language, duplicates removed | **2,881** | 100 non-English, non-article, or duplicate records excluded || **Title & keyword screening** | Staged screening based on thematic relevance | **756** | Priority A (198; 26.2%) and Priority B (558; 73.8%) || **Screening decision** | Categorization of 756 candidate records | **756** | YA: 415 (54.9%), MUNGKIN: 330 (43.7%), TIDAK: 11 (1.5%) || **Abstract reading** | Comprehensive abstract-level evaluation | **198** | Priority A records only: 179 YA, 19 MUNGKIN (Priority B deferred per protocol) || **Final decision** | Inclusion in final bibliometric corpus | **172** | **172 Included**, 22 CEK (under review), 3 TIDAK, 1 Paywall Excluded | *Note on Exclusion:* Exactly one study from Priority A (`10.65476/v3gra102`, *Climate Migration Discourse...*) was excluded because its full text was accessible only through a commercial subscription without institutional access, leaving exactly 172 included studies. --- ## 4. Key Bibliometric Findings & Replication Targets Executing `05_Reproduce_Bibliometric_Analysis.py` verifies the following empirical indicators reported in the paper: 1. **Publication Profile (Section 4.1):** - **Time span:** 2013–2026 (14 years). - **Recent growth:** 76 studies (44.2%) published between 2024 and 2026. - **Sources:** 143 unique academic journals.2. **Citation Profile (Section 4.2):** - **Total citations:** 2,225 citations across the corpus. - **Median citations:** 5 citations per study. - **Zero-citation rate:** 24 studies (14.0%) have zero citations (predominantly 2025–2026 publications). - **Top cited article:** Jackson, Bailey, & Foucault Welles (2018), *#GirlsLikeUs: Trans advocacy and community building online*, *New Media & Society*, with 127 citations.3. **Keyword Co-Occurrence Network (Section 4.4, Tables 3 & 4):** - **Nodes:** 39 author keywords (threshold $\ge$ 3 occurrences). - **Edges:** 157 weighted co-occurrence ties (total edge weight = 280). - **Louvain Communities:** 6 weighted communities (Modularity = 0.200, random seed 42): - *Community 1 (11 nodes):* diaspora, digital media, echo chambers, ethnicity, immigration, Iran, nationalism, populism, public sphere, race, religion (*Interpretation: Ethnicity, diaspora, and populism*). - *Community 2 (10 nodes):* digital activism, hate speech, identity politics, Indonesia, polarization, sentiment analysis, social capital, social media, social network analysis, youth (*Interpretation: Polarization, hate speech, and social media*). - *Community 3 (8 nodes):* China, counterpublics, discourse analysis, gender, identity, network analysis, political communication, Twitter (*Interpretation: Discourse and network analysis on Twitter*). - *Community 4 (6 nodes):* content analysis, ethnic identity, Facebook, semantic network analysis, Sri Lanka, TikTok (*Interpretation: Ethnic identity on Facebook and TikTok*). - *Community 5 (3 nodes):* COVID-19, Finland, misinformation (*Interpretation: Misinformation*). - *Community 6 (1 node):* identity construction (*Interpretation: Singleton*).4. **Thematic Clusters (Section 4.5, Table 5):** - **Cluster A:** Ethnic and religious identity in Indonesia and Southeast Asia ($n = 17$, 9.9%). - **Cluster B:** Networks, polarization, and online interaction ($n = 64$, 37.2%). - **Cluster C:** Hate speech and its detection ($n = 18$, 10.5%). - **Cluster D:** Language, discourse, and online identity ($n = 42$, 24.4%). - **Cluster E:** Diaspora, minorities, and networked publics ($n = 31$, 18.0%). - **Total:** 172 studies (100.0%).5. **Term-Presence Gap Analysis (Section 4.6, Table 6):** - “Karo” or “Batak”: **0 studies** (no study addresses the dispute). - Indonesia: **16 studies** (mostly ethnic Chinese identity and hate speech). - Ethnicity and network terms: **16 studies** (mostly follower, mention, or media-based networks). - Reply or balasan: **6 studies** (rarely treating the reply structure as an analytic unit). - Sociolinguistics: **2 studies** (rarely the primary framework). - Graph neural network or GCN: **0 studies** (no application to ethnic contestation). --- ## 5. Reproduction Instructions ### Option A: Direct Execution via Python (Zero Dependencies) The reproduction script uses only standard Python libraries (`csv`, `collections`, `itertools`, `os`). Any machine with Python 3.8+ can execute it immediately: ```bash# Clone or extract repository filescd repository_open_science_conference # Run the replication scriptpython 05_Reproduce_Bibliometric_Analysis.py``` ### Option B: Visualization in VOSviewer To reproduce the keyword co-occurrence network in [VOSviewer](https://www.vosviewer.com/): 1. Launch VOSviewer.2. Select **Create** $ightarrow$ **Create a map based on bibliographic data**.3. Choose **Read data from reference manager files** $ightarrow$ select `RIS`.4. Browse and select `01_Corpus_172_Scopus_Format.ris`.5. Under **Type of analysis**, choose **Co-occurrence**; under **Unit of analysis**, choose **Author keywords**.6. Under **Counting method**, choose **Full counting**.7. Set the minimum threshold to **3** occurrences (yields the 39 nodes). --- ## 6. Data Availability Statement for Conference Paper Authors can insert the following persistent statement into Section *Data availability* of `ISAMBUDDHA_2026_Paper.docx`: > *"The Scopus-format corpus and staged screening dataset supporting the bibliometric findings of this study are openly available in the Zenodo repository at https://doi.org/10.5281/zenodo.[ID]. The repository includes complete bibliographic metadata for the 172 studies, screening decisions across all 756 candidate records, the 39-node keyword network, normalization rules, and a zero-dependency Python script replicating all tables. In accordance with platform terms and research ethics approval, no social media text from X is included in the bibliographic corpus."* --- ## 7. Citation and License ### How to Cite If you use or build upon this dataset, please cite the conference paper and dataset as follows: ```bibtex@inproceedings{ginting2026ethnic, title={Ethnic Identity Contestation on Social Media: A Bibliometric Mapping of 172 Studies and the "Karo Bukan Batak" Gap}, author={Ginting, Loli Gunali and Humaizi and Zulkarnain, Iskandar and Kurniawati, Dewi}, booktitle={Proceedings of the International Conference on Social and Communication Studies (ISAMBUDDHA 2026)}, year={2026}, address={Medan, Indonesia}, publisher={Universitas Sumatera Utara}}``` ### License This data package is licensed under the **Creative Commons Attribution 4.0 International (CC BY 4.0)** license. You are free to share, copy, and adapt the material for any purpose, provided appropriate credit is given.